AKADEMIA GÓRNICZO - HUTNICZA

im. Stanisława Staszica w

Krakowie

Wydział Inżynierii Mechanicznej i Robotyki

Katedra Systemów Wytwarzania

PRACA DOKTORSKA
mgr inż. Paweł Bałon

Analiza zjawiska powrotnego odkształcenia

w procesach tłoczenia

An analysis of springback phenomenon in

forming processes

Promotor pracy:

Prof. dr hab. inż. Andrzej Świątoniowski

KRAKÓW 2016
I thank my doctoral advisor, prof. Andrzej Świątoniowski, PhD eng.,
my wife, my both mothers and my brother-in-law for immeasurable
kindness, understanding, devoted time and invaluable discussions
which helped me in realizing the work.
What is more, I want to thank the board manager of SZEL-TECH –
Grzegorz Szeliga and the board manager of Kirchhoff Polska –
Mr. Janusz Soboń.

to my grandfather,
AKADEMIA GÓRNICZO - HUTNICZA 1
KRAKÓW 2016 1
LIST OF SYMBOLS: 5
1. THE GENESIS OF THE WORK 7
2. ANALYSIS OF THE STATE OF KNOWLEDGE IN THE RANGE OF
MANUFACTURING TECHNOLOGY FOR CAR BODIES 12
2.1.1. Formability and Blank Sheet Metal Forming 21
2.1.2. Evolution of Computer-Aided Engineering for Blank Sheet Metal
Forming 22
2.1.3. Simplistic Approaches 24
2.2. Material 26
2.2.1. Advanced High Strength Steel 30
2.2.2. DP - CP steels 31
2.2.3. Low-alloyed TRIP steels 32
2.2.4. High-manganese TWIP and TRIP steels 34
2.3. Material parameters for analysis 36
2.3.1. Anisotropy 38
2.3.2. Stress and strain curve 40
2.3.3. Forming Limit Diagram 43
2.4. Springback 46
2.4.1. Springback definition 54
2.4.2. Pure Bending – Classical Results of Springback 56
2.4.3. Rigid, Perfectly Plastic Result of Springback 59
2.4.4. Elastic, Perfectly Plastic Result 60
2.4.5. Rigid, Strain-Hardening Results of Springback 61
2.4.6. Elastic-Plastic Result of Springback 62
2.4.7. Residual stress and springback 63
2.4.8. Springback Prediction 64
2.4.9. Springback reduction methods: 65
3. AIM OF THE WORK 68
4. MODELING OF THE STATE OF STRESS AND STRAIN IN THE
STAMPED ELEMENT USING MES 70
4.1. Methods and tools in solving the problem 80
4.1.1 Defining the calculating model 84
4.1.2 Description of the model 85
4.2. Results and conclusions of the simulation 89
5. SPRINGBACK COMPENSATION 95
5.1. Trimming tool compensation 102
5.2. Realistic Shape Compensation 104
5.3. The surface compensation with scanner 107
6. TWISTING SPRINGBACK 110
6.1. Forming method in case twisting springback 112
6.2. Application of various calculation models 113
6.3. The influence of a stamping conception on dimension of a springback
deformation 117

3
 6.4. The experimental verification of FEM results 121
6.5. The part formability 122
6.6. Forming conceptions 124
7. VERIFICATION OF THE RESULT 128
7.1. Description of the equipment with technical characteristics 131
7.2. Tolerances and clients requirements 138
7.3. Description of research and work-place 139
8. SUMMARY AND CONCLUSION 140
8.1. Scientific aim of the work 140
8.2. The aim considering an industry 143
9. DIRECTION OF FUTURE RESEARCH 146
9.1. Development view 147
9.1.1. Current automotive trend hot forming production 149
9.1.2. Development of technologies for an aluminum alloy 155
REFERENCES 157
SUMMARY 162
STRESZCZENIE 164

4
List of symbols:
– ultimate strength [N/mm2]
- proof stress [N/mm2]
A – elongation [%]
– intercrystalline phases
F – ferrite
M – martensite
δ - controlled displacement [mm]
- cross-sectional initial area [mm2]
A – current area [mm2]
– initial length [mm]
σeng – engineering stress [N/mm2]
σtrue – true stress [N/mm2]
eng – engineering strain [-]
F, P – forces [N]
ΔL – elongation [mm]
- initial volume [mm3]
- volume [mm3]
OP-20, OP-30 etc. – operation number 20, operation number 30 etc.
R, r – bend radiuses [mm]
YS – field stress [N/mm2]
UTS – ultimate strength [N]
E – elastic module [N/m2]
σ – yield stress [N/mm2]
M – bending moment [Nm]
x - circumferential strain [-]
σx - circumferential stress [N/mm2]
t- sheet thickness [mm]
w – sheet width [mm]
E` - Young's modulus (plane-stress case) [N/m2]
– Poisson ratio [-]
I – inertia moment for the cross section [Nm]
σ(0) - yield stress [N/mm2]

5
K - parameter [-]
K` - effective strength parameter [-]
n – strain-hardening index [-]
- radius of curvature [mm]
- radius of curvature [mm]
- coefficient of anisotropy in 0 deg rolling direction [-]
- coefficient of anisotropy in 45 deg rolling direction [-]
- coefficient of anisotropy in 90 deg rolling direction [-]
- coefficient of anisotropy in the sample direction [-]
b - width of the sample [mm]
t - thickness of the sample [mm]
- strains in main coordinate directions [-]
G,H,F,N - Lankford coefficients [-]
- maximum fiber strain at the outer radii [-]
- maximum fiber strain at the inner radii [-]
- Coulomb's coefficient [-]
- springback value [-]
f - real compensation value [-]
t - analysed compensation value [-]
- precision of compensation [-]
- force [N]
p – pressure [N/mm2]
, - areas of blank holder [mm2]
HRC – hardness scale by Rockwell

6
1. The genesis of the work
Process of metal forming in automotive parts construction becomes more and more
demanding due to tightened up tolerance and trials to realize very complex and in many
cases unworkable design in mass production. Moreover, it is required to cut and limit
costs of die production and simultaneously keep high quality. Furthermore, construction
elements are more often produced from materials which belong to High Strength Steel or
Ultra High Strength Steel (Fig. 1.2). Nevertheless, it results in appearance of springback
effect. Springback value depends mainly the material as well as part geometry, and in
extreme cases, the deviation value from the target part might reach a high level in some
areas. Reduction of implementation time, development of metal components, and greater
restrictions about designing and producing stamping tools generate extra costs. The
process of designing dies requires the use of appropriate Finite Element Method
software to make them more economic and less time-consuming. Therefore, analysis of
the forming process alone is not enough to take into account. During the design process
it is necessary to include the die compensation to achieve an optimized blank sheet. The
prediction of springback effect by trial and error method followed by correction of
deviation is difficult, arduous, and painstaking. Virtual compensation methods make it
possible to achieve precise results in a short time. This way provides a huge economic
advantage, eliminating excessive milling and allows for just-in-time production.

Fig. 1.1. Body structure of the vehicle.

7
The optimization process can refer to individual operations, as well as take into
consideration intermediate stages in the final result, at the same time increasing the
accuracy. Die compensation using software application was experimentally verified by
prototype die. Quality requirements for products made by the sheet stamping process are
very high due to the technologies surrounding automatic assembly of formed components.
Springback, as the main source of drawpieces inaccuracy, is the function of material data,
shape of tools, and process parameters. Therefore, springback deformation becomes
a critical problem - especially for AHSS steel - when the geometry is complex. Hence, it is
necessary not only to find the springback effect value but also to include and consider it
during the design stages by tooling designers.

Fig. 1.2. Implemented materials in body structure (with SSAB cooperation).

Elements of car construction are commonly produced from materials which belong to
Ultra High Strength Steel (UHSS) or High Strength Steel (HSS) (Fig. 1.3). Application of
these kinds of materials considerably reduces construction mass. Prediction of springback
effect by trial and error and subsequent adjustment of deviation is burdensome. Numerical
compensation methods make it possible to achieve a precise result in a short time. This
method results in a huge economic advantage, eliminating wasteful milling during die
production and allows for just-in-time production according to customers’ expectations and
forecasts of vehicle demand.

8
 An innovative compensation method of both forming and trimming die for the
construction parts of vehicle parts manufactured on a transfer press (Fig. 1.4). This method
allows for correction in the accuracy of compensation, consequently reducing springback in
a more exact way than current methods. The previous methods took into account only the
influence of trimming on springback but without generating compensated surfaces for the
trimming die or the next forming operation. Moreover, there is the possibility to include
positioning effect during multi operation forming because of the huge impact of springback
results on separate operations. If positioning between operations is not taken into
consideration there are problems with proper fitting, even if the final part has correct shape.
These problems generate additional costs during die production which could be avoided by
using multi compensation.

Fig. 1.3. Designed elements of vehicle body structure (brown color – DP600, purple
color – HC380, light blue color – DP1000, green color – HX340, yellow color – DC04).

More and more stringent ecological requirements, economic competition on the global
market, as well as growing safety and comfort standards continue to impose permanent
technological changes in the productio modern cars. The rigorous norms aimed at
protecting the Earth’s atmosphere influence the construction of engines and power

9
transmission systems in order to improve fuel consumption and the minimization of
forming resistances. It also has an impact on the refinement of self-supporting car bodies in
terms of reducing their weight and improving their aerodynamics (Fig. 1.1). A 100
kilograms lighter car emits an average of 4,7 grams of CO2 less every kilometer travelled.
The issue of vehicle passengers’ safety is also a highly important factor. It requires the
durability of passenger compartment to be as high as possible. This high durability is
related to high crush and torsion resistances and also to accurately defined strain
characteristic curve of body framework.

Fig. 1.4. Transfer press with servo movement and a maximum force of 1600.000 kg.

The state of technical progress in car body structures is the proper development of material
engineering, which involves providing materials to car producers. The materials need to be
characterized by required mechanical strength, operating life, simple forming, market
availability, relatively low costs, and potential for subsequent recycling or utilization.
Despite common expectations, steel still remains the main material used for building car
bodies. However, more and more often, new sorts and technologies of steel treatment are

10
being used. In the material structure of standard a car body, the portion derived from deep
drawn steel sheets (the name derives from its high plasticity) decreased to barely 30%,
whereas they were dominant on the market at the start of the 21st century. In present
maximally stiff and light constructions, steels are absolutely more useful.
Parts made by this method have a high application in the car industry. Using elements
made by hot-forming allows reduction of car mass up to 12- % and ensures higher collision
safety due to its strength.
The author established wide scientific cooperation with the steel forming processing
department at the SSAB company. It resulted in common benefits through the development
of further research of AHSS steel springback (Dogal 600, Dogal 800, Dogal 1000).
The SSAB company is a world-class producer of the best AHSS and UHSS steels
having its own R&D laboratory. This allowed for the performance of numerous
experiments that checked the relationships between dual-phase steels in an analytical way.
What is more, the springback issue was investigated with the cooperation of the SSAB steel
producer who has many specialists in the areas of materials, steel manufacturing, and
defect-free-technology. This work has been conducted over 4 years mainly through
research and processing the research in the best possible way. A lot of material tests,
designs, and numerical analyses have been done. Furthermore, many design tests and
prototype experiments of forming tools have been conducted.

11
 2. Analysis of the state of knowledge in the range of manufacturing
technology for car bodies
Forming processes are widely used in the manufacturing of vehicle construction parts.
Their main attribute is high efficiency, repeatability, and most of all economy,
simultaneously maintaining their proper surface condition in mass production. Efforts are
being made to replace construction elements with aluminium alloy, magnesium alloy and
composites. In 2017 constructions contained about 0.5% magnesium and 8% aluminium
but were respectively, 2% and 12% inn 2020 (Fig. 2.1) (Vogt, 2011). There are some mass
production trials of construction elements as integral composite parts. Moreover, their
durability is similar to the high mechanical parameters of Advanced High Strength Steel
(AHSS). Their main advantage is low mass, being even 30% less than in the same
elements made from conventional HSS steel. However, the composite is not able to
provide all properties in constructions that have stamped parts. This is caused by a lack of
repeatability with a divergence up to 30%.

Fig. 2.1. Comparison of implemented materials in body structure in 2016 and 2020.

Technologies of composite materials are said to be transferred from construction of high-

performance vehicles to standard automotive applications. However, a problem occurs
concerning the practical possibility of servicing cars after a collision. In fact, composite
material structures are typically beyond standard repair methods after being damaged,
especially compared to steel elements.

12
 D 60 8 1000 12 14
C 0 0 DP 00 00
0 DP 0 M M
6 D
P
Fig. 2.2. Comparison of formability for steel [with SSAB cooperation].

Main qualitative expectations of customers are: exterior appearance, safety, comfort, high
engine performance and low fuel consumption. Higher standards of comfort and safety
have caused a continuous rise in average vehicle weight (Ford, 2011;Seki, 2011).
Therefore, new solutions are still being sought in order to ensure an optimal mass
supporting structure and positive crash test results. Creating some brand new steel grades
with strong mechanical properties made a ground-breaking contribution to steel research
(Fig. 2.2) (Thyssen Krupp, 2009).

5% Hotforming
17%
steel
12% cold forming
rollforming
51% 3%
alu sheet
alu extrusion
10% hybrid
2%

Fig. 2.3. Overview Current Bumper Front (% of analyzed designs).

AHSS materials are the result of much research done by joint international projects over
the last decade. They facilitate the attainment of high endurance structures, while, at the
13
same time, preserving desired plasticity. Although these two parameters do not appear
simultaneously, and an increase in one of them is often at the expense of the other, some
material groups do posses the beneficial values of both.
These competitive properties and economic considerations are the main reasons that
steel is the dominant material in the automotive industry. According to the near-term
forecast, for the next few years, the portion of HSS and AHSS materials in car bodies is
expected to be about 90%. These trends are about to coincide with requirements regarding
a decrease in overall construction mass, as well as simultaneous enhancement of safety and
reduction in fuel consumption (Ford, 2011).
Traditional design and production methods were mainly based on designers’ experience.
However, an effective method of stamping processes design must use CAD/CAM (Fig.
2.4) capabilities as well as join the abilities of numerical analysis and one based on
optimization methods. Currently, for the analysis of the forming process, as well as the
design of the forming tool, Finite Element Method (FEM) software is widely used.

Fig. 2.4. Die process design.

In the automotive, the numerical analysis is used to explore and predict forming
possibilities in practice, in order to avoid some drawpiece mistakes. The potential problems
in the stamping process during the design of the tool can be eliminated by virtual software
rather than byprocess of trial and error (Fig. 2.5). The advanced and complex software
enables fast and certain verification if the assumed method is correct, which imparts an
advantage over traditional trial methods.

14
 Fig. 2.5. Crack in FEM model and in real forming process.

The most up-to-date methods of designing, manufacturing, and production planning must
be used to meet market demands. Planning is one of the most important stages of designing
the tooling technology to shorten the time of manufacturing process start. For the process
of sheet stamping this stage is not only extremely crucial but also difficult. It is not easy to
predict the number and type of forming stages, the accuracy of springback effect and to
take into account inhibit factors, such as wrinkling and cracking (Fig. 2.5) (Kang et al.,
2002; Choil et al., 2000).
Since the second half of the 1990s intense research was carried out regarding the
linking of designing surfaces and die stampings in parametric way (Smith, 1990; Suchy,
1997). The first objective was development of the parametric models using extra surface
which strictly corresponds with the surfaces involved in the process. The next step was the
creation of a general solution for various designs by linking known features. Die softwares
such as: METHOPLAN (iCapp, Zurich), Viking (Inpro, Berlin), DIEDESIGNER
(Autoform, Zurich) were being developed which tended to design complex and parametric
assemblies of bodywork construction. Designing in a parametric way ensures the linking of
software supporting die stampings construction and FEA method in order to design
a proper surface (Fig. 2.6) (Danzberg et al., 2004).
Generally, it is required to use parametric methods of car design which enable fast
modification and implementation of essential constructional changes.

15
 Fig. 2.6. Describing forming part creation.

The optimization of the geometry of all parts is found through the application. The parts
are produced in many forming processes, creating the assembly which is built by
components manufactured in different ways. One example of such is the combination of
parts whereby one is made by cold forming and the other made by hot forming. The entire
assembly, joined by a binder must fulfil the assumptions made at the beginning of the
design process. Therefore, the optimization corresponds with a complex chain of forming
stages rather than a single operation.
The manufactured parts are very often made of AHSS steel, which are less deformable
and show a greater tendency to springback effect than mild steel. Springback compensation
by means of tooling geometry change is necessary for the part to be situated in narrower
tolerance intervals. The quality of stamped parts has greater and greater importance. It
influences the assembly stage and the final product quality. One of the factors affecting the
quality of drawpiece is the shape-dimensional accuracy connected with material
springback (Fig. 2.7). Research done during recent years caused the narrowing constraint
of the tolerance range in constituent components, as well as in assemblies. Increasingly,
some elements are designed in order to obtain high geometry stiffness. Their production
reaches limiting values for possible for mass production. In the computer age, tool
compensating is evaluated via the application of an analysis based on FEM methods.

16
 2.7. Part design process.

The accuracy problem in springback modelling is treated in literature as highly complex.

The main reason for this view has been computers’ limited design power, forcing the
application of a lower amount of nodes and number of elements at high time intervals, as
well as a lower refinement level that leads to a large computer springback error. Limits on
computer hardware possibilities vastly reduced the application of the full capabilities of
numerical analysis software before 2001 (Fig. 2.8). This resulted in low accuracy of
elements with complex geometry in relation to thickness of parts and a long calculating
time. The progress of the 64-bit software version was a huge breakthrough that led to
higher output from the Central Processing Unit (CPU) and simultaneously reduced
computing time. Numerical calculations of forming are reduced to describe a surface with
elements that are based on nodes. The number of nodes is given to ensure high calculation
precision. A good way to determine the size of elements is the functional dependence
between the smallest radius and the thickness of a sheet. Computation of springback
requires extreme accuracy. Therefore, elements with small dimensions, little time
increments, and accurate material data are used in zones where some contact with a tool
appears (Fig. 2.9). A very important factor is the use of proper elements during an analysis
of the forming process. Usually, FEM programmes dedicated for stamping use surface
elements that have more developed counting algorithms and, most significantly, they
reduce their time. Volume elements take into account 3D plasticity due to double contact.

17
Application of TTS (Through Thickness Stress) elements is well-founded for flanging on
account of process specification. Use of computer supported design and an analysis of
stamping process tools helps to achieve a target in the recommended way and most of all,
in attainable time. The analysis itself is not enough because of the necessity to take tool
compensation. into account.

Fig. 2.8. Comparison of a stamped blank with a reference part.

Nowadays, compensation calculations are carried out automatically, or at least semi-

automatically, with the use of FEM software (e.g. Pam Stamp, Autoform, Ls-Dyna, Outifo,
Mashal). Development of the compensation method for Pam-Stamp 2G software took
place in 2002-2004. The accuracy and correctness of this method was affirmed for parts
design in the years 2004-2005 for: Arcelor B-pillar and Renault bodywork elements
(similar part (Fig. 2.8).
The Pam-Stamp 2G software module is based on a method from 2005. In 2011,
research that was done on the progress of compensation methods led to enlargement of the
basic optimization process tool. Cooperation between the ESI Group and the Atlas Tool
company led to verification of the experimental method. Trials were conducted on a few
parts, for example, on a deep stamped A-pillar made from UHSS steel. In result, it was
found that using the complex process optimization allows for compensation to be achieved
with at least 80% compatibility with the part (Atlas Tool, 2011).
At present, it is possible to compensate the whole operation chain, which enables the
springback effect to be considered at every process stage. It also gives an advantage in

18
solving some problems by finding an accurate surface for the cutting operation and
matching the sheet to the tool during every individual operation. Skipping of the multi-
operational compensation tool requires taking into account unstressed surfaces at particular
trimming stations, which can change the state of stress in the part causing larger
deformation.

2.1. Metal Forming Processes

The steel forming process in the automotive industry is one of the significant
processes. Conventional stamping is a very common method because of its rapidity and
low per-piece cost. CAE technologies need an analysis of simulations for each forming
method. In order to reduce considerations, the electromagnetic and explosive forming
methods are assumed as separate forming processes that are not taken into account. The
initial condition ensuring proper results of the CAE simulation is precise mapping of
physical processes taking place during forming. The numeric analysis needs to contain
details of the physical process in order to conform to reality in a proper way. In the
conventional forming process the metal part is created by the result of a few operations
such as:
Stage 10: (drawing) forming
Stage 20: trimming
Stage 30: flanging
Stage 40: final trimming and flanging (Fig. 2.9)

Fig. 2.9. Forming process [Autoform].

19
The first operation is usually the most important because one is mapping the main
geometric shape. The first tool usually consists of a die, a punch and a blankholder.
However, there may occur a tool with two blankholders or any at all. Other stages depend

Fig. 2.10. Changing of the radius value to reduce thinning.

20
on the geometry and the degree of tool making. For example, flanging can replace the
restriking operation. The main deformation appears in the first stage of drawing on the
first tool. Generally, two types of forming may be classified: blankholder forming and
forming from the air (without the blankholder). Both methods are the opposite of each
other due to the deformation caused by the contact between die and the punch. There are
two separate forming concepts that give two different parts as a result. Each method has
some pros and cons which are used alternatively, depending on the type of forming
structure. Blankholder forming requires adouble-action press or using nitrogen cylinders
where the punch and the blankholder can be controlled independently. The bottom tool
section is stable but the top section moves towards the bottom, holding the blank in proper
position. Pad force increases with the growth of press feed. Changing the value of the
radius is highly important to reduce the thinning phenomenon (Fig. 2.10).

2.1.1. Formability and Blank Sheet Metal Forming

Before using CAE methods, the forming analysis aim, which is solving formability,
problems have to be known. Formability (Tab. 2.1.) is defined as the capability to form the
blank for designed shape geometry without any faults for particular process conditions. It
can be divided into the most important groups:
- Product design (shape of the part)
- Die-face development
- Blank shape and location
- Boundary conditions
- Material properties
- Process development
The most disappointing part is the one having the greatest interaction at one of the
factors above, for example boundary conditions and blank shape that may be more critical
factors than changing steel for aluminium alloy. In the last decade, stamping simulations
have been refined and effectively implemented. More precise and more detailed models
highly improved the capabilities of forming analysis as methods of approving, leading and
controlling the solutions.

21
Formability Formability parameteres Factors influencing
Test method
atribute characterized formability
Forming limits Full dome test Forming-limit diagram n-value, thickness
Sheared-edge Ultimate tensile strength
Hole extrusion test % hole expansion
stretching limits (UTS), r-bar value
Bending under Angular stretch- Height at failure, stretch-
(UTS), r-bar value
tension limits bend test bendability index
Yield strength, tool radii,
Springback and Springback opening angle, radius
Channel draw test draw bead restraining force,
curl of sidewall curl
tool gap
Stretch Pan forming (fully
Height at failure n-value, thickness
formability locked conditions)
1. Height at failure Uniform elongation, r-bar,
Stretch
Square draw test 2. Binder control blank size, coating,
drawability
3. Strain measurement lubrication

Tab. 2.1. Formability attribute.

However, during recent years, new requirements have suggested the benefit of getting
information like: parts quality, springback, dimensional tolerance, effectiveness of
production and assembly, fatigue and crashworthiness. These demands have become more
permanent in order to implement materials such as: aluminium alloys, ultra high strength
steels or dual phase steels into mass production with current access to stamping analyses
that are supported by formability testing reaching intended limits. New generation software
is an indisposable and required tool for industrial use. As a result, a solver considering the
algorithm elements, contact and parametric modeling of the material is compatible with
stress and deformation.

2.1.2. Evolution of Computer-Aided Engineering for Blank Sheet Metal

Forming
Using finite element method in blank sheet forming processes made scientists and
industrial users concerned, giving new approach view. At the turn of the 1970s and 1980s
most of the research was strongly limited because of the three-dimensional (3D) strain
problems concerning simple geometries. Unfortunately, the research provided overall
information for variable problems for the formed material without the possibility of using

22
it in industrial way. This technology was being continued and developed for 10 years
focusing on its industrial use. Dominant material models like Barlat’s 1989 and Hill’s 1990
yield were used for numeric analysis in practical usage. A lot of automotive applications
such as: hoods, mudguards, and bumpers were analysed in terms of using them in the FEM
analysis relying on systems and professional knowledge. In the next period of time, there
were two shifts at the beginning of 1999.

80

70

60
Simulation time

50 Springback
Forming
40
Holding
30
Gravity
20
Mesh
10

0
OPTRIS/PAM-STAMP PAM-STAMP 2G PAM-STAMP
2000 V2002 2G/DMP V2004

Fig. 2.11. Simulation time in Pam-Stamp versions.

The first one concerned improving springback predictions as a result of a few

international institutes and universities. The most important aspect in predicting the
springback effect were designs innovated by National Institute of Standards and
Technology Advanced Technology Program (NIST-ATP). Thanks to their theses, the
industry received the code provided for predicting the springback of a particular part.
Another main development was creating a more friendly and integrated system of
calculating material forming with all computer design supporting functions. The simulation
time was reduced a lot from the 1990s when the front panel analysis was limited from 50
days in 1990 to 10 days in 2002. Overall time reduction for all analyses of forming
possibilities containing prediction of springback was cut to one week in 2004. Extreme
reduction of overall simulation time approached in 2010 with the implementation of 64 bit
operating systems. Besides, time reduction influenced on standardization and high
experience of CAE analysis in producing demanded parts very fast. Proper simulation of
forming process may effectively reduce physical testing and solve some problems during
23
the early stage of design. Over the years, the simulation analysis helped reduce costs and
time in the automotive industry, partly via technological growth. In recent years, using the
simulating blank forming in industry resulted in eliminating cracking and wrinkling
problems at the design stage. The last decade has clearly changed cost reduction by
optimization of the process and implementing present possibilities of CAE programs.

2.1.3. Simplistic Approaches

Length-of-line analysis is one of the spectrums of 3D FEA analyses for current
standards. This method simply calculates the ratio between the final length and the section
of original length. Before popularising this analysis, 2D analyses were developed.
There are two main reasons for using the 2D way. The first one concerns the impractical
using of 3D in the mid-1990s considering rapidity and better consistency of solutions
having weak computing stations. The earliest form of FEM calculations started in 1940
was calculating the parts stiffness. However, Wang and Budiansky’s work was one of the
initial completed ones for the 3D model. Despite the fact that most of the initial theses
were developed by the academy and the code was implemented in commercial software of
companies as worldwide applications, currently the market is dominated by producers such
as: AutoForm of AutoForm Engineering, LSDYNA of Livermore Software Technology
Corporation and PAM-STAMP of ESI Group (Fig. 2.11). Their well prepared codes are
based on the academy’s code. Over the years, the industry has been using all three codes.
One of the most important technologies in the 3D FEM code is the schedule of time
integration. Overally, it was divided into two main codes: explicit time integration code
and implicit time integration code. Both algorithms can be used for static and dynamic
governing equations. Moreover, there are four fundamental possibilities of forming the
FEM code:

 Static Implicit (SI)

 Static Explicit (SE)
 Dynamic Implicit (DI)
 Dynamic Explicit (DE)
Widely used implicit method is dedicated to static or quasi-static problems. Dynamic
explicit is widely used to analyse crash simulations. Main static implicit reaserches of were
initially more rigorous and adjusted for precise results of the blank forming analysis. There

24
is not a perfect method for all forming types. Everything depends on a proper choice of
calculation and for critical problems it depends on using application form for a particular
case. The implicit simulation method for a small scaling static implicit element causes an
increase of elements amount in the third power. There is also another problem in reducing
the application of static implicit codes in the production. The forming blank sheet is not
a “pseudostatic” problem indeed. In the standard stamping operation, total displacement of
the press is 150 mm in 2 seconds, with the speed peak close to 500 mm/s, There occurs
large displacement, metal rotation and large deformation. Another challenge is the highly
nonlinear contact between metal and the tool (Fig. 2.12).

Fig. 2.12. Simulation of linear and non-linear loadcases. Investigation of stress, strain
and deflection. Creation of stressmap.

The fundamental application of static implicit algorithm is inversing stiffness matrix.

Different codes may use different formulations, but the concept of the momentum equation
is the same. In the dynamic explicit equilibrium code there is a dynamic balance. The static
explicit code is presented as the one having benefits of both static implicit and dynamic
explicit codes. Autoform Engineering has its own unique code that is often classified as
implicit code. However, a lot of specialists claim that the Autoform code generates good
results.

25
 Defining and choosing the finite elements is always very important for the FEM
simulation. However, it is not practical to use precise mesh for an initially unknown
analysed element. Usually, accurate calculations are related to the high CPU loading of
computing stations and also a long calculation time. The users cannot always properly
define the deformation gradient at the beginning of the forming process analysis. This
problem particularly affects the drawing operation, where large displacements appear.
Adaption was mainly common in mid-1990s because of the limits of the used algorithms.
Soon afterwards, the development of intelligent algorithms for blank forming took place.
The algorithms solved the problems of the contact and refine mesh. The used algorithms
worked very well for each forming process. However, various parameters and needs
require particulars for special applications (Chung-Yeh, 2006).

2.2. Material

In spite of much information about elaborating and implementing cars made of

aluminium alloys, magnesium alloys, composites and plastic into the production, steels are
the dominant material in the automotive industry. This is caused not only by economic
considerations but mainly by competitive properties and simple recycling. Realising a few
international programs in the last 15 years in Europe, Japan and the USA caused appearing
attractive novelties of the used materials. They meet the requirements concerning the
reduction of body mass, which causes reduction in fuel consumption and also ensure
greater passive safety of passengers. Currently, thin sheets with small sections made by
ultra-high strength steels or aluminum/magnesium alloys are used for structural elements.

26
 Fig. 2.13. Comparison of material mechanical parameters (with SSAB cooperation).

In terms of the automotive industry, the following parameters should be taken into
consideration:
a) mechanical properties, especially high strength (the ratio of immediate strength of
the material to its density), which allows to reduce weight of vehicle (Fig. 2.13.);
b) high capability to absorb energy in case of collision;
c) properties that minimize technological problems during production and which
ensure high productivity, especially: vulnerability to plastic forming (stamping of
panels, bending, hydroforming and others), simplicity of using coverages (Zn and
Al coatings, lacquers) and also good weldability;
d) good behaviour during exploitation (fatigue strength of the material and welds,
corrosion resistant, simple exchange of the elements);
e) economic considerations;
f) recycling.
Presently, mostly conventional hot-rolled and cold-rolled low carbon steels are used in
car body building. However, new generation steels require special processing, which
causes the output of more complex structures. The dominant method of assembling car
bodies remains resistant spot welding, in spite of the spread welding and laser soldering,
gluing, riveting and clamping (Gould et al., 1998; Chuko et al., 2002).
Generally, there is not an accepted ordination of construction steels for the automotive
industry so far. According to the actual body of knowledge, they can be classified in the
three basic groups (Fig. 2.14):
27
 Fig. 2.14. Yield strength and formability (in terms of tensile ducitility) of conventional
high-strength steels and advanced high-strength steels (Arcelor Mittal).

I. mild, plastic low-carbon steels (DQSK, IF steel) with immediate tensile strength Rm -
below 300 MPa and total extension A - 30÷60%;
II. conventional High Strength Steels (Bake Hardenable, Carbon Manganese, IF with micro
additions, HSLA) with 300< Rm<700 MPa and lowered A comparing to the previous
group;
III. Advanced High Strength Steel (Rm above 700 MPa reaching even 2000 MPa) and
extension included in quite wide limits - 5÷30%, but strength increase accompanies with
lowered plasticity.

28
 Minimum bend
Forming radius -sheet
temperature thickness Springback, degrees
Test
Metal or alloy °C °F data Preferred
C-103, C-129Y Room <1r 1r 2-6
Tantalum alloys (annealed)
Tantalum Room <1r 1r …
Ta-10W Room <1r 2r 1-5
Molybdenum alloys (stress relieved)
Mo-0.5 Ti, TZM titanium-zirconium-
molybdenum 150 300 2r-5r 5r 3-8
Tungsten (stress relieved)
Tungsten 315 600 2r-5r 5r 2-8

Tab. 2.2. Formed to a 120° bend angle in a 60° V-die at a ram speed of 254 to 3050
mm/min.

The two first groups can be considered classic. These materials are widely used in building
self-supporting car bodies on a massive scale. The steels of group III are progressively
being implemented into production and their contribution is systematically rising.
However, there have been recently elaborated high-manganese TRIP/TWIP steels that are
not included in the classification. They are undoubtedly a new generation of materials
having very high plasticity and strength of 1200 MPa (Lehman et al., 1996).
Drawing Quality Special Killed and IF are ultra-low carbon steels. They have ferrite
structure and high vulnerability to forming, whereas IF steels do not contain roots forming
interstitial solutions including also carbon (C<0,005%, N<0,005%, S<0,005%).

Test t Rp0,2 Rm A50mm A80mm r n

direction mm N/mm2 N/mm2 % % 10% 5-10%
0° 1.22 433 634 29 24 0.73 0.16
45° 1.24 446 626 31 25 1.09 0.16
90° 1.23 454 636 29 24 0.83 0.16

Tab. 2.3. Chemical parameters for DP-600.

29
 The High Strength Steels from the second group (IF-HS and BH) contain micro
additions (Ti, Nb, V) bonding leavings of interstitial roots, forming carbides, carbon
nitrides and also compounds of carbon and sulfur in ferritic matrix. BH steels are hardened
in the process of lacquer firing on the finished body. High Strength Low Alloy steels also
belong to the group of carbon content below 0,1%. They are hardened by Mn and various
microscopic additive such as: Cr, Nb, Al, Si in hundredths of % (Tab. 2.3).
CMn steels owe their properties to the ferritic structure hardened by solutions. All these
steels have low carbon equivalents and very good welding as a result (Radakovic et al.,
2008; AWS, 2002; Tumuluru, 2006).

2.2.1. Advanced High Strength Steel

AHSS group contains steels hardened by phase transformation, not by hardening by
solution or hardening by dispersion as it took place in the two previous groups. However,
Complex Phase, Dual Phase and Martensitic steels hardening is done at the stage of
preparing the material, whereas TRIP hardening takes place during vehicle collision (Fig.
2.15). The part presented in the Fig. 2.15 is a cross car beam element and it illustrates
stress pattern for knees pressure during collision.

Fig. 2.15. Crash simulation knee-impact (blue color – area of high stresses, yellow color
– neutral area, brown color – area of increased stresses, green color – area of low
stresses)

30
Common technological operation for CP, DP and TRIP steels is treatment in intercritical
temperature (the area of coexisting of α and γ phases) after cold-rolling followed by low-
temperature austenite change. The appropriate choice of annealing temperature is formed
by the relation between austenite and ferrite. The final structure depends on controlling the
temperature course during cooling and might be a combination of ferrite, bainite,
martensite and retained austenite. MS steels are made during fast cooling in the field of
existing of the γ phase to obtain the 100% martensite structure. Their immediate strength
gains even 2000 MPa. In order to improve ductility, a controlled tempering has to be done.

2.2.2. DP - CP steels
The structure of DP steels is made of a mix of fine-grained polygonal ferrite or non-
carbide acicular ferrite with martensite islands in an amount of 5÷40% or even more. What
is more, there may occur small amounts of retained austenite. As was mentioned, these
steels are made by appropriate cooling after annealing in the temperature field of
coexisting of α and γ phases (Fig. 2.16). The sample presented in the picture below was
made by microsection. Then, it was examined through the scanning microscope.

Fig. 2.16. UHSS material micro structure (ferrite – light color, martensite – dark color).

During fast cooling, the austenite transforms into the martensite (the ferrite does not
transform, obviously) and the diphase F+M structure is made. Another method is thermo-
mechanical treatment that is based on heat treatment and plastic forming (hot-rolling) with
subsequent cooling (Fig. 2.17). It is possible to obtain a bit more fine-grained structure and
therefore, better mechanical properties in the second method. Essentially, the variety of the
DP steels are CP steels. Their cooling is proceeded in the way enabling to obtain the third
31
micro-constituent – bainite. Impurities in DP-CP steels are contained in order to control
austenite content in annealing temperature. The carbon content may vary widely between
0,05 and 0,2% (Dogal, 2010; Rathbun et al., 2003).

Fig. 2.17. High rolling milling (with SSAB cooperation).

The properties of DP steel (that is a sort of specific composite) are resultant of

fractions of hard and tough martensite and ductile ferrite. They depend mainly on both
phases ratio and ferrite grains size. The immediate tensile strength gains 1000 MPa with
the elongation of over a dozen %. The DP steels have also large initial difference of
ultimate strength and yield stress that rapidly decreases during cold working.
Consequently, it causes less springback effect than in low-alloyed steels (e.g. HSLA) with
the same strength, which is highly beneficial in stamping of panels (Tab. 2.2) (Dogal,
2010). Other advantageous properties are: lack of Lüders effect during deformation, crack
resistance in low temperature and low anisotropy of plastic properties (Dogal, 2010;
Rathbun et al., 2003).

2.2.3. Low-alloyed TRIP steels

During a properly processed (slower than for DP steel) cooling cycle with a stop, there
is created a mixed structure composed of ferrite, bainite and much amount of retained
austenite in low-alloyed TRIP steels (earlier processed by annealing in the field of
coexisting of α and γ phases). Similarly, as for composite materials, the properties of low-

32
alloyed TRIP steels are the resultant of particular phase properties, their fraction and
morphology.

Fig. 2.18. Car simulation testing (deformation model).

The bainite is the toughest phase and the ductile ferrite and austenite simplify the
plastic forming. Therefore, this kind of steel is beneficial because of strength and from a
technological point of view. However, its impressive properties are most saliently
revealed during violent deformation (e.g., in the case of a car crash (Fig. 2.18)). Then,
the retained austenite transforms into martensite absorbing energy and additionally
hardening the material. This last feature makes it very attractive for the automotive
industry. Moreover, during the A → M transition, there appears a particular plasticity
called the transition plasticity. Current low-alloyed TRIP steels have a diversified
chemical composition with the total fraction of alloying components equaled several
percentages by weight. Carbon content is crucial and usually oscillates in the field of
0,10÷0,25%. Higher contents, to 0,6% C, despite beneficial impact on the structure
(bainite and austenite are created and strength reaches over 1200 MPa with good
plasticity) have not come out of experimental phase because of impairing weldability
(DeCooman, 2004). Another alloying components are Mn (0,4÷2,5%), Si (0,4÷1,8%),
Al (about 1%). There are also used little additions of P to harden the ferrite by solution

33
 (unfortunately, it impaires weldability) and microadditions of Nb, Ti and V to refine the
structure and create carbon nitrides and dispersion carbides (DeCooman, 2004).

Fig. 2.19. Low speed bumper test.

2.2.4. High-manganese TWIP and TRIP steels

TWIP steels appeared on the market in 2004. They have a unusually wide grade of
deformation retaining high strength at the same time. It is a unique feature, especially
important in the event of a vehicle collision. Under these conditions, the material should
exhibit two opposed properties:
– high ductility, in order to absorb the maximal amount of striking energy during plastic
deformation;
– maximal stability of the element that protects the passenger cabin.
The already used steels were either very tough (e.g. for car frame) and of low ductilily,
which caused cracking during dynamic overloading or inversely: they were plastic and low
strength. The TWIP steels may revolutionise this segment. A number of such unique
materials have been obtained so far inter alia: in Germany (Frommeyer et al., 2003), in
Korea (Kim et al., 2006) and in Japan (Ueji et al., 2008). The steel researched at the Max
Planck Institute for Iron Research shows 1100 MPa with 90% elongation (Frommeyer et
al., 2003; Mehta, 2007).

34
 Fig. 2.20. Automotive part production

The chemical composition of the TWIP steel contains a high concentration of Mn

(15÷35%), which stabilizes the austenite to room temperature. The other alloy additions
are 2÷4% Al and/or Si. The 100% austenite structure, without martensite and other phases
(potential cracking nuclei) is characterised by high ductility and plastic forming
treatability. At high deformation speeds, there occurs a twinning. The steel deforms
locally, hardens (Re increases with the deformation speed) and transfers the remaining
energy to another region. That way, the deformation by twinning moves as a characteristic
wave, splitting and effectively absorbing the striking energy (Ueji et al., 2008; Mehta,
2007). The deformation by twinning occurs for alloys having a low stacking fault energy,
which characterises the TWIP steels (Kim et al., 2006; Ueji et al., 2008; Mehta et al., 2007;
Kliber et al., 2008).
Decreasing Mn content to 15÷20% in the presence of the other alloy additions causes
appearing the TRIP effect in the discussed steel (Frommeyer et al., 2003). The stacking
fault energy is changing and the mixed TWIP/TRIP steels are also possible. As a result of
the A → M transformation for the high-manganese TRIP steel, the energy absorption is
more effective because it runs in two steps. Initially, the hexagonal martensite is created
that transforms into hexagonal body-centred martensite afterwards. HSS and AHSS steel
have found their stable place in automotive industry (Fig. 2.20). The decreasing

35
contribution of mild low carbon steels is observed. Currently, they have lost their primacy
in favour of the HSS steel and especially Interstitial Free-HS and HSLA. The contribution
of the AHSS steels is increasing and they are predicted to become a dominant alternative in
building safe constructions in 2016. It is worth mentioning that in the USA, the DP-CP
steels from the AHSS group are preffered, whereas in Europe, the low-alloyed steels with
the Transformation Induced Plasticity effect are more desired by producers (DeCooman,
2004). Presently, the high-manganese TWIP/TRIP steels enters the commercial market and
it is probable that they will get some significant part in it in the future. Comparing these
steels with the other steels makes them quite expensive because of more difficult
technology of melting, casting and plastic forming (tendency to form hard oxides during
hot-rolling and necessity of using high pressures for cold-rolling process Fig. 2.21).

Fig. 2.21. Rolling AHSS steel process.

In the car body construction, ultra high strength steels (MS, DP, TRIP) will dominate the
strengthenings of passenger cabins, whereas Twinning Induced Plasticity steels will create
the outer zone absorbing the energy during collision, both in the vehicle axis and especially
during side impacts. The development activities concerning increasing mechanical
properties are being continued. Both for TRIP steels and for high-manganese TWIP/TRIP
steels, it appears that the most promising is commercial implementing of alloy micro-
additions (especially B, N, Ti, V, V and Zr) (DeCooman, 2004; Frommeyer et al. , 2003;
Kim et al., 2006; Ueji et al., 2008).

2.3. Material parameters for analysis

DP-1000 material is steel that belongs to the UHSS group, diphase, cold rolled
(Tab.2.4). Steel structure is formed by the compound in die form of martensite precipitates
on fine-grained ferrite warp. There also may appear small amounts of retained austenite.

36
These steels are received by suitably leaded cooling after annealing from temperature
range conforming coexisting of α and γ phases.

Tab. 2.4. Mechanical property of DP-1000 (SSAB, 2009).

0,6

0,5

0,4
σ(ε) [GPa]

ES_DC04_P-S
0,3
DC04_1-5
St14_P-S

0,2

0,1

0
0 0,2 0,4 0,6 0,8 1 1,2

Fig. 2.22. Hardening curve extrapolation for material DC04 and St14 for the roof
hardening part.

37
 12.00

10.00

8.00
DP-34
σ(ε) [GPa]

SSAB_Dogal600DP_1mm
6.00
SSAB_Dogal600DP_1,2mm
Kirchhoff_testing
Krupowsky_extr
4.00

2.00

0.00
0.00 0.00 0.00 0.01 0.01 0.01 0.01

Fig. 2.23. Hardening curve extrapolation for material DP-34, SSAB Dogal 600,
Krupkowsky extrapolation, Kirchhoff testing for the B-pillar part.

The properties of this steel which are a kind of specific composite, are the result of
a portion of hard and resistant martensite and ductile ferrite. There is also high initial yield
point and ultimate strength which quickly decline during the plastic working. Hill 48
material using isotropic hardening curve was approved despite of the fact that the software
uses the criterion of isotropic-kinematic consolidation hypothesis (Fig. 2.22 and Fig. 2.23
show hardening curves for various materials). Chosen material requires following basic
parameters describing material.
Young Modulus E, Poisson coefficient υ; density ρ, rolling direction along edge of blank
sheet and blank thickness g0.

2.3.1. Anisotropy
Reference parameters functions (G,H,F,N) are expressed by anisotropy coefficients for
characteristic rolling directions 0°, 45°, 90° (called also as Lankford coefficients- see Eq.
5.2.1) , , :

38
 (2.1)

The coefficient describing normal anisotropy r (Eq. 2.2) is treated as the average value
which can be determined from the formula:

r = 0.25(r0+2 r45+ r90)= 0.25(0.86+2*0.89+0.91)= 0.89 (2.2)

A low r- value gives higher thinning effect in opposite way a high r- value gives higher
strain levels in width direction.
Normal anisotropy can be expressed by in direction:

(2.3)

Where: ⍺ - the direction of the sample; b – the width of the sample; t – the thickness of the
sample; - accordingly initial values.

Fig. 2.24. Anisotropy properties in all directions (with SSAB cooperation).

Coefficient describing normal anisotropy r is regarded as the average value which can be
determined from the formula:
r = 0.25(r0+2 r45+ r90) (2.4)

39
Definitions describing material properties in all directions (Fig. 2.24):
 Isotropy: Equal properties in all directions
 Anisotropy: Not equal properties in all directions
 Planar anisotropy: Different properties within the plane of the sheet
 Normal anisotropy: The properties in the thickness direction differs from the
properties in the plane (Lundh et al., 1998; Bergstrom et al., 1996).

2.3.2. Stress and strain curve

Stress-strain curves are an extremely important graphical measure of a material's
mechanical properties (Fig. 2.25, Fig. 2.27). Perhaps the most important test of
a material's mechanical response is the tensile test, in which one end of a rod or wire
specimen is clamped to a loading frame and the other subjected to a controlled
displacement δ.
The engineering measures of stress and strain are determined from the measured load
and deflection using the original specimen cross-sectional area A0 and length L0.
In the early (low strain) portion of the curve, many materials obey Hooke's law to
a reasonable approximation, so that stress is proportional to strain with the constant of
proportionality being the modulus of elasticity or Young's modulus.
Engineering stress = force / initial area:
σeng = F/A0 (2.5)
True stress = force / current area:
σtrue = F/A (2.6)
Engineering strain = elongation / initial length
σeng = ΔL/L0 (2.7)
True strain = In (current length / initial length)
dε = dL / L → σtrue = ∫1/L dL = ln (L/L0) (2.8)

P P A0 PL L
 true      eng    eng (1   eng )
A A0 A A0 L0 L0 (2.9)
 true   eng
Volume  Const
A0 L (2.10)
V0  V  A0  L0  A  L  
A L0

40
As strain is increased, many materials eventually deviate from this linear
proportionality, the point of departure being termed the proportional limit. This
nonlinearity is usually associated with stress-induced “plastic” flow in the specimen.

1400

1200
True stress [MPa]

1000 Docol 800 DP

Ludwig
800
Swift

600 Hockett-Sherby

Combined S-H
400
0 0,2 0,4 0,6 0,8 1
True strain [-]

Fig. 2.25. Hardening curve by formulas for the Docol 800 DP steel for the B-pillar part.

Elasticity is the property of complete and immediate recovery from an imposed

displacement up on release of the load, and the elastic limit is the value of stress at
which the material experiences a permanent residual strain that is not lost up on
unloading. This is done because the material unloads elastically, there being no force
driving the molecular structure back to its original position. Since it is often difficult to
pinpoint the exact stress at which plastic deformation begins, the yield stress is often
taken to be the stress needed to induce a specified amount of permanent strain, typically
0.2%. The stress at the point of intersection with the σeng curve is the offset yield stress.

L  L  L 
 true  ln    ln  0   ln(1   eng )
 L0   L0  (2.11)
 true   eng

41
Here it appears that the rate of strain hardening diminishes up to a point labeled UTS,
for Ultimate Tensile Strength. Beyond that point, the material appears to soften to
strain, so that each increment of additional strain requires a smaller stress.

1300

1250

1200
True stress [MPa]

1150

1100

1050

1000

950

900
0 0,2 0,4 0,6 0,8 1 1,2
True strain [-]

Fig. 2.26. Hardening curve by Krupowsky for the Docol 1000 DP steel for the bumper part.
The apparent change from strain hardening to strain softening is an artifact of the
plotting procedure, however, as is the maximum observed in the curve at the UTS
(Roylance, 2011).

Fig. 2.27. Material parameters testing.

42
Hardening curve determined by static test of bumping for 0° direction after calculating
stress value and actual strain (Fig. 2.25). Next, the curve was extrapolated in order to
increase the range of actual strain (Firat et al., 2008). To reach that, the Krupowsky
formula was used (Fig. 2.26):
n
σw =K( 0
) (2.12)

Work hardening could be also determinate by Ludwik - Hollomons equation:

  K  pn m (2.13)

where: n determines the ability for the material to distribute the strain
n = 1; a straight line, n = 0; ideally plastic.

2.3.3. Forming Limit Diagram

During most stamping operations, the sheet metal is strained in a total of three
different modes. In the stretch mode, biaxial strain occurs on a positive basis along both
the major and minor axes. In the draw mode, strain is positive along the major axis and
negative along the minor axis. While in the plane strain mode, strain is positive along the
major axis, but at or near zero along the minor axis.
The hardening curve obtained from the forming analysis is not enough. Forming Limit
Diagram (FLD) must be used additionally, in order to precisely verify the nodes which are
critical in terms of maximal stress (Fig. 2.28). It is a diagram of steel drawability shown
with the use of main straining (Wagoner et al., 2011). The curve can be determined by the
Nakajima Test or by other experimental method. (Lundh et al., 1998; Leppin et al., 2008).
A Forming Limit Diagram, such as the one shown in (Fig. 2.31), can subsequently be used
to match up the strain modes present in a stamped part with a sheet metal that can be relied
up on to form the part without failure (Fig. 2.29 and Fig. 2.30 show comparing of FLD
charts for various materials).

43
 Fig. 2.28. FLD diagram by different type of steel (with SSAB cooperation).

A forming failure is typically defined for automotive applications as a split in the part,
a reduction in part thickness below acceptable limits or an unacceptable surface wrinkle.
To determine the limit of formability of a sheet metal, the Forming Limit Curve (FLC) is
a generally accepted tool.

Fig. 2.30. Comparing FLD chart parameters for: DP-600, HCT600 and DP-K-34.

44
 Fig. 2.29. Comparing FLD chart parameters for: DP-600, HCT600 and DP-K-34.

The FLC is determined by proportionally straining the material, in other words upon
increasing the strain during testing, the ratio between strain components should be
maintained. Experimentally, FLCs are obtained via either Nakazima or Marciniak
methods, of which the Nakazima is probably the most widely used since the Marciniak
method is more complex in operation. The initial strain is always bi-axial. (Eisso, 2002;
Leppin et al., 2008).
0,35

0,3

0,25

0,2
Major strain

0,15

0,1

0,05

0
-0,2 -0,1 0 0,1 0,2 0,3
Minor strain

Fig. 2.31. FLD diagram for t=1.2 mm.

45
 2.4. Springback

Requirements regarding reduction of cars’ fuel consumption play a key role in the
automotive as they make cars lighter by application of more and more resistant materials.
Usage of steel of increased endurance results in a higher tendency to springback. A
problem occurs not only during control of elements but mostly during process of joining of
formed parts into subassemblies. Maintenance of a narrow tolerance range taking into
consideration deviation of springback effect for Ultra High Strength Steel (UHSS) and
Advanced High Strength Steel (AHSS) is not easy, especially that newer materials tend to
release the residual stress as a result of deformation during the trimming process (Fig. 2.33)
shows or punch unloading (Vogt, 2011). Application of these kinds of materials
considerably reduces construction mass thanks to high durability. However, after removal
of forming loads, results include springback effect (Fig. 2.32).

Fig. 2.32. Springback effect for different type steel bending (with SSAB cooperation).

Springback value depends mainly on used material as well as part’s geometry and in
extreme cases deviation value may reach high level in some areas. Springback is a process
of elastic recovery (Fig. 2.36) with a local plastic deformation appearing at the end (Fig.
2.34, Fig. 2.35) show the springback value depending on various operations.

46
 Fig. 2.33. The springback value for tailage part after the trimming operation.

Fig. 2.34. The springback value depending on cutting and perforating amount.

47
 Fig. 2.35. The springback value depending on trimming and bending operations.

Moreover, by currently available approaches it is not still exactly predicted. Standard

springback is expected as an angle or distance deviation to the reference model but in some
cases it would be more complicated due to the torsion. It is required to eliminate this type
of springback in the initial design phase because of problematic die compensation.
Additionally, the influence of geometric shape, springback pattern and magnitude is
determined mainly by stress level and its distribution in a stamped part. The main
influences on springback effect are:
 Coefficients: friction coefficient, anisotropy coefficient, mechanical properties of
material;
 Tool geometry: radius of a die and blank sheet geometry;
 Process parameters: mesh quality, punch displacement velocity, blank holder
strength.

48
 Fig. 2.36. Elastic recovery occurs when stress is removed from a
specimen that has already undergone plastic deformation.

The accuracy of compensation effectiveness prediction is one of the main challenges in

design of a tool construction with the aim to minimize its trials. Designing of dies demands
the use of appropriate Finite Element Method software to make them more economic and
less time-consuming. Therefore, it is not enough to take into account only analysis of
forming process. That method results in a considerable economic benefit eliminating
useless milling and allowing production of a die just in time. Geometric compensation of a
tooling surface is a direct way to eliminate springback effect of a tool in an iterative way by
system of solving nonlinear equations. Currently, the springback analysis is becoming the
routine process to form steel in CAE engineering analysis in automotive (Fig. 2.40).
However, for a big springback deviation and twisting springback the conventional methods
of compensation fail (Ford, 2011; Seki, 2011). The necessity to find another way of
forming arises for that which ensures stability of the process with the simultaneous
reduction of twisting springback to the standard case with a deviation towards the height of
part. Even though it is mostly feasible, it is time consuming and costly.

49
There is shown what is necessary to do to find the best resolution. In some cases changing
of forming conception is the only way to get a good result with controlled springback.
However, in the next step compensation is needed to reduce standard deviation in direction
towards material thickness (Fig. 2.37, Fig. 2.38).

Fig. 2.37. The springback compared to the part thickness.

Fig. 2.38. The springback compared to the part thickness.

Springback effect is a deviation occurring after the stamping process. Its size depends
on used materials and their properties, such as: elastic modulus, yield stress, strength.
Springback of designed parts may be compensated by overforming. Material overbending
mostly appears at the final surface of whole compensation process (That surface can be
analytically estimated for some simple geometry of the part. For constitutive models, the
σ(z,t) stresses are predicted in specified point and section (Fig. 2.39).

50
 Fig. 2.39. Section of overbending process.

Mostly, the springback effect occurs during low-carbon steel bending for 0,5-1,5
degree values. It can be reduced by overbending or restriking. Coefficients affecting the
springback are: ratio of bending angle to blank thickness stock; bending angle (bending
degree, flatness); bending method U-bending, V-bending (Fig. 2.41). For the bend radius
that is several times higher than blank thickness, the stress is much larger. Therefore, press
power will have to be higher than for two times lower radius. Overbending size which is
necessary to compensate the springback for low angle will be different than for high

Fig. 2.40. Springback deviation computed by FEM.

51
bending angle. This value is not directly proportional to the springback angle.
Curvilinear surfaces will be exposed to tension or compression during forming
process. As a result, it will cause returning springback depending on referent geometry of
bending flange. The springback effect is changeable during production because blank
thickness (depending on heat), blank hardness, tool temperature, tool matching and its
wear are changeable as well. Multioperational bending gives ordinary cumulative
springback (when there is more than one bending in the process) and it has to be more
controlled between particular operations. Greater amount of bending and trimming
operations make the compensation process more difficult and more complicated. Thanks to
high ratio of strength to plasticity for HSS and AHSS, those steels have high springback
value comparing to mild steels. The springback highly influences on the formed blank
because it can cause serious problems during assembling parts into set. It may also
implicate high costs of modifications for TRIP and DP steels. High strength materials such
as DP and TRIP show stress increase after forming process, which results in high
springback (the higher the stress level is, the higher the springback severity is). For
constant Young’s modulus the springback effect and steel strength increase. In many cases,
HSS and AHSS steels may save even to 40-60% of material mass by reducing blank
thickness and improving its strength. It is not possible for used Mill Steel. As it was
mentioned, springback is a function of yield-strength to modulus ratio. In result, higher
stress of hardening coefficient for DP and TRIP steels gives higher springback than for
HSLA steel despite the fact that initial yield strength of unformed blank can be similar.

Fig. 2.41. R=10 mm and R=3mm (with SSAB cooperation).

52
 Computer simulations should take into account drawability of formed elements and
also possibility of creating compensated surfaces reducing springback. The forming
process ought to minimize springback at the design stage rather than compensate it after
tryout stage. The simple way of reducing springback at the design stage is using low punch
radiuses (Fig. 2.42). However, low radiuses can also cause some changes during forming
and press-loading.
High radiuses may escalate thinning by increasing blank sheet stresses. It happens as a
result of bending of sharp edges, which may be difficult for high field strength. Springback
is a complicated physical issue and its value depends on the natural deformation of an
element. Final bending radius can be lower than reference radius of a part.

Springback ( ) (2.14)

12

10
Springback angle [degrees]

0%
6
5%
10%
4 15%

0
1 2 3 4
Bend radius (r) [mm]

Fig. 2.42. Springback angle if relations in bending radius.

53
 For superelastic materials the hardening law does not work. However, it is really
important to understand what properly defined springback depends on. The springback
does not increase for R/t ratio whereas, it is defined as variable of geometry curvature.
Steel forming tools are often covered with coatings such as PVD in order to improve life of
the coating. The blank used to stamp is mostly covered with zinc to protect from corrosion.
However, on the basis of basic steel properties it does not influence springback but it
highly changes contact friction and forming process at the same time. Springback size is
proportional to (0,2 YS+UTS)/2, where YS – field strength and UTS – ultimate strength.
The springback can be under control by reducing punch radius and bending lane (Semiatin,
2006). Hot formed parts may be done in narrower dimensional tolerance than cold formed
parts because of lower springback effect. Material properties have significant impact on
appearing springback: Elastic module E, Yield stress , Slape of the true stress/strain
curve or tangent σt/dε. Σ=Eε.

2.4.1. Springback definition

The development of currently implemented thin metal bumper structures for AHSS
steels reveal the springback effect (Fig. 2.44). The basic large-strain, elastic-plastic
deformation mechanism is simplified to the numeric and analytic methods. Decreasing
bend radius causes reduction of the springback effect. For the bend radius R> Ri (Fig. 2.43)
which is larger than the initial critical value of construction curvature, the energy is stored
in the elastic form.

Fig. 2.43. Springback for pure bending (with SSAB cooperation).

54
The recovered elastic energy concerns the springback as a definition of the geometry of
loaded and unloaded forming part. The word springback is a single word revealed in
technical use since 1940. The technical meaning of this word was initiated in formal
definitions as describing the phenomenon of changing the surface after the forming process
(Wagoner et al., 2006) . It concerns the geometry that was created as a result of forming of
the tool surface that carries loads during stamping process. If the surface change is not
predicted and exactly compensated for the designed tool, the part will not be compatible
with client’s requirements (the referential geometry). The springback effect has a big
impact on the ratio of initial and final dimensions. The changes caused by the springback
effect may be very important, despite the fact that elastic strains cause low springback.
Basically, considering a cross section schedule, three following stages can be distinguished:
a. the initial stage (flat)
b. stage during forming (with a tool)
c. unloaded stage, after springback (Fig. 2.45).

Fig. 2.44. The springback value of the bumper made from MSW-1200 steel.

Fig. 2.45. Unloaded stage, after springback.

55
Whereas the tool is being opened and the blank is being deleted from the tool, the radiuses
are being increased. This is a typical operation initiating bending with extending. The
forming areas with the U-shaped or L-shaped wall are drawn considering the variable
radius so when the press is unloaded (the press slider returns) the radius or flanging do not
change. The stress levels for each of the areas is related to the holding force, the friction
between the tool and the requirements of bending and drawing (Fig. 2.46).

Fig. 2.46. Springback degree in function tensile strength.

2.4.2. Pure Bending – Classical Results of Springback

In order to understand the springback effect it is important to initially simplify the
example.
Assumptions for the simplest state:
1. The cross section lies on the plane.
2. Lack of thicknesses changes.
3. Biaxial geometry for strains and stresses.
4. Constant curvature.
5. Lack of strains in the radiuses or in the thickness direction.
6. The neutral axis is located in the mid layer.
7. Lack of difference between real and engineering values of stress.
8. The homogeneous, isotropic material.
9. The elastic stress occurs only during the springback.

56
The main difference during the treatment that is presented in literature and is related to
predicted constitutive behaviour of the material are three basic material models:
Purely elastic or elastic-plastic model with strain-stress law for plastic range.

∫ (2.15)

The result is an almost perfectly plastic model (without hardening) or a linearly hardening
model or a power-law hardening model. The bending part is the M moment for the R
radius. The M value will get integrated form for sheet or beam:

∫ ∫ (2.16)

where: - circumferential strain, depends on z; - circumferential stress; t - sheet

thickness; w(z) – sheet width (vary with the z-coordinate).
The stress linearly related to the distance from the middle of the sheet (z) and to the
inverse of the bend radius:

(2.17)

The real and the engineering strains ( ) are less than the used ones.
The bending moment can be calculated by using constitutive relations between stress and
strain. The springback is being calculated after the bending operation, where the moment is
expressed in the units of sheet thickness M/w and the material is elastic. The isotropic
linear elastic sheet has a constitutive answer:
σx =E’εx, where E`- Young’s modulus (plane-stress case) or:

(2.18)

where – Poisson ratio.

For the elastic answer with the initially curvilinear configuration (the R radius) to the final
configuration (radius=r):

(2.19)

57
where – inertia moment for the cross section.

16
Springback 14 Pure bending
12
10 210 YS (MPa)
8 245 YS (MPa)
6 280 YS (MPa)
4 350 YS (MPa)
2
555 YS (MPa)
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
R/t

Fig. 2.47. Springback for pure bending in function ratio R to t.

Assuming the bending moment proportional to isotropic strain of the material (not
hardened) and also assuming H-M-H hypothesis and Poisson ratio 0.3, the elastic modulus
equals 1.12 times uniaxial tension. The elastic and plastic anisotropy can change this value.
The internal moments might be calculated for surface integration and be contained in the
table for variables of the standard shapes. The internal moment:

(2.20)

where t – sheet thickness.

The internal moment written for rectangular cross section and divided by width:

(2.21)

Or written in the alternative form:

[ ] [ ] (2.22)

The expression is called the springback (Fig. 2.47), whereas is the springback ratio.

Generally, the springback ( ) has positive value. The springback is measured as the
springback ratio=1/(1-R*Springback). The curvature change is proportional to the

58
springback effect. Fractional failure related to springback estimation can depend on

quantity of the second condition. If it is low, the will also be low but it may still be

important.

2.4.3. Rigid, Perfectly Plastic Result of Springback

The simplest springback result occurs for pure stiff bending (not elastic) of perfectly
plastic (not hardened) material model. Without these assumptions, the bending moment and
the springback do not depend on the referential bend radius:

∫ (2.23)

where - yield stress (flow stress).

The springback will be expressed by the equation:

(2.24)

Most frequently, the result is enough to predict the springback which is mostly dependent
on material properties (Siguang et al., 2000; Ninshu et al., 2008). The springback is
proportional to strength and stiffness and inversely proportional to sheet thickness.

A material having higher elastic modulus is more sensitive to the larger springback than
thinner materials (Fig. 2.48).

59
 7

6
Springback angle [degrees]

4 AlMg0.4Si1.2-ka-isomatt
AlMg5Mn-w-isomill
3
AlMg0.4Si1.2-ka-EDT
2
AlMg5Mn-w-mill finish
1
Sheet thickness: 1.25
0 mm without
1 2 3 4 prestreching. Position
of bending axis to
Bend radius (r) [mm]

Fig. 2.48. Springback for aluminium alloy.

Therefore, aluminium alloys have three times better properties than steel alloys considering
elastic modulus which is 1/3 higher than that for steel.

2.4.4. Elastic, Perfectly Plastic Result

Perfectly elastic bending with elastic core and the neutral axis will equal:

{ (2.25)

The external elastic core is proportional to the bend radius, while yield stress is inversely
proportional to the elastic modulus.
For typical forming steels, the bend radius is lower than the elastic region and can be
ignored without serious consequences of losing accuracy.
Estimating the values requires obtaining the moment for the standard elastic-plastic case or
the perfectly plastic case with the external elastic core:

∫ ∫ (2.26)

The elastic-plastic moment is obtained:

60
 (2.27)

And the springback result :

[ ] (2.28)

The springback equation may be described in another way by Gardiner:

(2.29)

where .

The left side of the equation is an inverse of the springback ratio. The failure caused by
ignoring the elastic core in the springback calculations may be estimated and shown tabular
R/t. The moment failure is limited in percentage terms: 1, 2, 5, 10%. For demanded data,
the accuracy level of R/r ratio is higher than the considered one.
The typical R/r ratio for forming processes in automotive industry fits in 5-25 range, even
though it is possible to find many examples beyond this range. The surfaces are not always
simple to analyse. The elastic core causes low failure in the standard materials area
(aluminium alloys with yield stress of 500 MPa, rarely possible for complex forming).

2.4.5. Rigid, Strain-Hardening Results of Springback

For the hardened material, the bending moment and the springback have:
-Empirical forms;
-Rigid, arbitrary hardening;
-Rigid, power-law hardening;
-Rigid, linear hardening.
The hardening model Power-law is changeable and used to analyse sheet forming.
Sometimes, the hardening is described by the Hollomon law for uniaxial stress and other
fixed stress- or strain-ratio forms:

or (2.30)
(2.31)

61
where: K - parameter, K` -effective strength parameter, n – strain-hardening index.
The strain and the stress have to be considered due to strain and stress state and also yield
function (anisotropic). The elasticity range is ignored.

( ) ( ) (2.32)

( ) ( ) (2.33)

2.4.6. Elastic-Plastic Result of Springback

In elastic-plastic relation of the material hardening model, the bending moment and the
springback can be a changing form containing the elastic power-law part and the stiff linear
hardening. The hardening for pure bending moment and the springback is based on the
formula:
(2.34)
(2.35)

For the surfaces strains it is important to approximate the plastic strain:

(2.36)

The moment contains the same conditions as for perfectly plastic model and for the model
caused by the hardening:

∫ ∫ (2.37)

Where z* can be estimated as explicite incremental moment:

( ) ( ) (2.38)

In this case, the complete elastic moment (plastically hardened)can be estimated:

62
 ( ) ( ) (2.39)

And the springback:

(2.40)

The springback based on the above equations is compared to the analytic elastic and
perfectly plastic material and also to elastic-plastic finite element (FE) simulations of four-
point bending (Semiatin, 2006).

2.4.7. Residual stress and springback

Whereas the springback has to be compensated by overbending, for tools with closed
surface (undercut) there needs to occur a slight tension by ironing between the punch and
the die. Overally, sharp bending cannot be done in one operation. For the settled process,
elimination of springback ought to occur without disposition to material cracking, because
the die or sharp edges will cut the blank instead of bending it. The main drawback of air
bending (without pressure) is the possibility of geometric incompatibility of bending. Air
bending requires from an operator the preset pressure on hydraulic press. The operator
needs to check every part using gauges and restrike is also necessary.
When a specimen that has been bent beyond the elastic limit is unloaded, the applied
moment M becomes zero, and the radius of curvature increases from to . For a fiber
at distance y from the neutral axis, this produces a strain difference:

(2.41)

where: - radiuses of curvature

( ) (2.42)

Elastic modulus In plane-strain :

(2.43)

63
Maximum fiber strain (Wagoner et al., 1996):

√ (2.44)

where: – maximum fiber strain at the outer and inner radii

The change in bending moment for complete unloading is dM=M, where M is the applied
bending moment prior to unloading:

∫ [ ( ) ] ∫ (2.45)

∫
(2.46)

2.4.8. Springback Prediction

In designing of cold forming tools springback should be taken into account. In

addition to the internal stresses which arise as a result of bending, there is a residual stress
distribution in the cross-section of the bent material. This residual stress distribution
depends on the extent to which the material underwent plastic deformation. Even though
the springback effect is sometimes more or less visible depending on the sort of material, it
is a result of bending the material. For parts with no complex geometry, the analytic
methods of solution are advisable in order to estimate the size in a quick way. However,
complex constructions with complicated shapes require numerical solution of the issue.

64
 Fig. 2.49. Springback prediction depend on α angle (with SSAB cooperation).

The accuracy of predicting of the springback also depends on the influence of coefficients
which are responsible for forming analysis quality (Fig. 2.49). There is a possibility to
estimate the scale of the springback effect by calculating the value in an analytic way:

D⍺ (Lundh et al., 1998) (2.47)

2.4.9. Springback reduction methods:

1. Rubber forming.
2. Increasing friction.
3. Comparing HVmF forming and conventional forming – the first one reduces the
springback effect (HVmF – High velocity metal forming; a process with high-strain rate
impulsive loading).
4. Springback compensation should be done by changing the surface geometry for multi-
operational forming process so it may follow according to geometric deviation.
5. Other method assuming compensating the designed surface after the springback, for
example spring forward and displacement adjustment techniques.
6. Computer aided engenering (CAE).
7. Using stiffeners (Fig. 2.50).

65
 Fig. 2.50. Springback for forming without and with stiffeners.

The stress results of elastic-plastic deformation appeared during the final bending,
having a bigger bend angle than the initially one. During bending (in fact, during the return
of the slide and the die) the plastic deformation called springback effect is formed. The
largest stresses appear on the outer surface of the bending material. Stress in the middle
area equals zero. In remaining areas (other than the neutral one) the material will have the
stress and the plastic deformation after exceeding elasticity. When the tool is open elastic
bending causes the return of the original flat surface part. However, some areas cannot
return to the initial shape, which is caused by plastic deformation area. Some returns of
elastic deformation to initial conditions show elastic and plastic areas reaching an
equilibrium, which is called the springback effect.
Variables that effect springback:
- hard materials have a bigger springback angle caused by a larger elastic range
- surface with lower bending radius provides higher stresses and lower springback
- during bending the plastic area is elongated, which is caused by overbending. The
springback is reduced in each degree of bending.
- overbanding
- bottoming or setting
- stretch bending
Mostly, a blank is overbended in an amount refered to the compensated bending ray after
the springback.

66
 Overbending can be done by cams, by reducing die polishning or also by setting lower
angle of die and punch that is required to V-shaped bending (Fig. 2.51) .
Bottoming or setting is about hitting the radius area by the blank (a few times). This area
places the material over the stress compensation in the bending area. Bottoming is located
on the punch in U-shaped surfaces bending. Blankholder has to set the blank position again
so its positioning to the die is appropriate. A flat blank surface usually generates buckling
because stresses do not appear in that area and the springback effect does not occur as well.
Bottoming causes the loading according to the rule: the bigger pressure is, the more it
should be under control. Bumping during bending causes that the blank has stress
depending on punch force. Prestressing that takes place before the bending operations,
causes small springback effect. In this method, only relatively small bending rays give
prestressed metal, which is caused by the final stress.

Fig. 2.51. V-shaped bending (with SSAB cooperation).

67
3. Aim of the work
The aim of the work is exploring the knowledge in the range of physical
phenomena determining the springback in the stamped car body elements. Particular
attention will be devoted to twisting springback. As was mentioned in Chapter 6, the
knowledge in this range is still highly limited because of the unique case of the
springback. Developing springback matter will create the basis for evolving and moving
into higher level the currently used compensation processes of tools. The essence is
using numerical tools of FEM software, especially the Pam Stamp or Autoform
softwares. Those softwares have the possibility of modeling the springback and its
compensation in a proper way. What is more, the commercial Pam Stamp software
contains some useful tools that describe the multioperational compensation in details.
The aim of this compensation is eliminating, or at least significantly reducing, the
negative impact of the springback on the quality of stamping parts by exceeding the
intervals of shape and dimension tolerances. As research has shown, the level of
ensuring the accuracy of the FEM numerical analysis varied between 80 and 90% of the
expected accuracy. The analysis of the state of stress and strain will be done by using
the FEM in further stamping operations. On the basis of the previous own researches,
the Pam Stamp software was chosen. However, the Autoform software was used for the
case of torsional springback in order to confirm that the unexpected form of the
springback is right. The earlier industrial use of this software has determined this.
Moreover, as our own research and experiments have shown, the Pam-Stamp
software is much more accurate than other commercial numerical programs such as:
Autoform, Dynaform and even Abaqus. This last program frequently precluded
calculating the springback for the parts over 1000 mm of length.
According to general conduct rules, the results of such an analysis have been
experimentally verified based on the measurements of the experimental prototype parts.
Those parts have been verified by checking tools, measuring CMM machines and optic
GOM scanners. The results of the numerical and experimental researches have been put
into the work and attachments in tables form.
A lot of valuable information that was used to eliminate the springback effect was
obtained through years of rigorous research in order to understand, predict and control
the springback. For the springback analysis there were chosen adequate examples that
show high springback tendency. The car structure elements such as: pillars, hardening

68
of the roof, and bumpers are important parts that need to keep the highest requirements.
Therefore, those parts have not initially met the construction assumptions of the client
so they were numerically analysed by the FEM method and verified in an experimental
way. This order allowed for accurate investigation of the problem and its elimination in
a further step obtaining the parts with appropriate quality and repeatability. The
example of the part with the springback effect did not have any other solution besides
the one presented in the work. It is caused by the fact that the phenomenon is highly
unusual and practically irresolvable in workshop conditions of the tool shop. The
research-scientific actions performed by the special group which included constructors,
process engineers, and FEM specialists, managed the set task.
It is worth adding that every examined part (contained in the tables) is a prototype
element requiring a special approach to eliminate the springback effect. In order to solve
the problem, a few dozen FEM analyses were done and over a dozen changes in the real
stamping tool were executed. However, it allowed for obtaining substantial experience
and practice in the scientific and research fields relating to forming analysis.

69
 4. Modeling of the state of stress and strain in the stamped element
using MES
For the analysis of sheet forming, there are two basic types of elements: solid
elements and thin-shell elements. Plastic deformation is adopted for two- or three-node
elements and is used for variable structures. Its main advantage is simplicity because
many elements require taking into account the sheet thickness for the springback analysis.
These elements are very stiff during bending, which makes them inadequate for springback
analysis. One of the improvements is using four-node or higher-order elements that behave
better during the springback but also append an additional degree of freedom. A lot of
variable standard elements were implemented in order to mesh without blocking the
stiffness of the proper deformation. Unfortunately, those numerical corrections often result
in some change in the phenomenon. Despite the mesh optimization, the calculation time of
forming analysis can be still very high because of the use of small elements but it has been
highly reduced for current workstations. Using typical thin-walled elements for the thin-
shell theory of forming analysis has dominated present computing.

Fig. 4.1. Meshed B-pillar die with thin-walled elements (Fig. 2.23, Fig. 2.25). The OP-
20 contained 53125 elements and 215225 nodes for computing time of 2h 54min 3s and
compensation time of 4h 0min 23s.

70
They may also generate atypical one-dimensional thin-walled solid elements with the
added function of spatial elements, e.g. for the purpose of bending thick blank sheets.
Three- or four-node elements are commonly used. Unfortunately, they are not solid but
only have virtual thickness for time saving simulation (middle surface), which is very
important in the numerical FE simulation. Usually, the deformation is described in a very
linear way by the thickness of the thin-walled structure of the evaluated steel strain.
The typical procedure utilizes additional degrees of freedom for corresponding local
rotations of material planes. The advantage of the thin-walled sheet element forming is
speed of computing. Normally, this kind of element is not used for research or testings.
These conditions refer to critically pure contact between the tool and the sheet blank,
which is important for through-thickness stresses and the R/t ratio lower than the range of
5-6. The thin-walled elements are a few times more productive than the solid elements for
the steel forming process. Therefore, they were dominant in most cases.
The blank sheet is treated by default as the middle layer without thickness. Contact
between the die and blank was adopted as automatic. It means that the solver adjusts the
type of contact depending on the corresponding conditions and type of calculation, at the
same time giving the user the chance to verify the correctness of that choice at anytime. In
OP-20 and OP-50 the blank-holder was used in order to decrease wrinkling (Fig. 4.1, Fig.
4.2).

Fig. 4.2 FE model in OP-20 with die, blank-holder, punch surface for symmetrical parts
of the roof hardening.

Defined forces equal relatively 300 kN (OP-20) and 700 kN (OP-50). The speed of
the blank-holder has been defined as 2 m/s, while the speed of forming as VBC
(Velocity Balanced Curve) is at maximal value of 5 m/s. Modeling of contact which is

71
 based on the method of penalty friction coefficient between cooperating surfaces is
described as a Coulomb bilinear model. It uses static and dynamic coefficient friction or
function, depending on the variables of the other parameters e.g., stamping velocity,
press force, etc. The friction model as Coulomb’s coefficient has been defined as
constant value μ= 0.12 (ESI Group, 2011). The friction coefficient was experimentally
determined by the trial and error method. In order to approximate the numerical model
to real conditions, covering of the material by anticorrosive coatings was considered.
What is more, covering of the tool by coatings hardening the PVD surface was also
considered. Moreover, there were used lubricants, oils and foils to reduce the friction
coefficient. In order to increase it, high surface roughness was implemented on purpose.
It allowed for estimating the Coulomb’s µ coefficient in the range of 0,09 and 0,13
(Fig. 4.4, Fig. 4.5).

Fig. 4.3. FE model in OP-50 with die, blank-holder, punch surface for symmetrical parts
of the roof hardening.

Level of refinement of the mesh is based on criterion of minimal radius value, which
enforced refinement degree on level and the minimum element value.
The part is the fabricated car element designed to ensure safety during side and head-
on collisions, especially near a front pillar. This part is an example of an element which
causes some difficulties due to its geometric specifications and the material used (UHSS).

72
 Fig. 4.4. Thinning phenomenon for the B-pillar element depending on Coulomb’s friction coefficient
t1=0.12, t2=0.11, t3=0.09.

Fig. 4.5. Thinning phenomenon for the B-pillar element depending on rolling directions: 90 deg, 0 deg for
Dogal 600DP material and for approximation of Krupkowsky hardening curve.

73
 During designing of the forming tool, many problems were encountered ,which had an
influence on the quality of die stamping, such as: excessive local folding, cracking (Fig.
4.7, Fig. 4.8., Fig. 4.9) and excessive springback. PAM-STAMP 2G software versions
2011.1, 2012, 2015, 2015.1 were used for the computing. It enabled numerical analysis of
the forming process, and complex, full-chain optimization of the forming die operation
OP-20 (Fig. 4.6), as well as die operation OP-30/ OP-40. In operation OP-20 (Fig. 4.10,
Fig. 4.11, Fig. 4.12) there are two blank holders, one for standard holding and a second
pad to keep the blank in place during stamping (Fig. 4.10). Operation OP-30/ OP-40 is
a standard trimming but in reality this process is divided into two operations: OP-30 and
OP-40, even though one trimming is accurate enough to achieve a correct simulation. The
next process is second forming operation OP-50 (Fig. 4.3), which is not included in FEM
analysis due to its small impact on springback effect (Fig. 4.14 shows the springback
deviation after the OP-50).
The numerical analysis Finite Elements made for material model - Hill 48 - explicitly
and implicitly uses computing methods according to the guidelines given by the software
suppliers with the aim to minimize computing time and simultaneously maintain
satisfactory accuracy of the server’s calculations. Hill 48’s law is used for orthotropic
materials because it adapts well to standard HSS steel. The Hill 48 criterion used an
isotropic Krupkowsky hardening law.
As a result, a complex state of stress is received by the blank which is not always easy
for analytical notation as the reduced strength hypothesis. The material used in the forming
process has a nominal thickness of 1.5 mm and the dimensions of this blank are 539 x 470
mm. For the sake of a simulation producing the left and right side parts, a single, blank
sheet symmetry can be used, which has reduced the computing time. Springback can also
be computed taking into consideration the plane symmetry resulting from the cutting
operation before the calibration of the tool. Use of symmetry makes it necessary to include
the boundary conditions. This meant that six translations had to be locked in a statically
determined way (Fig. 4.15).

74
Fig. 4.6. Sheet thickness after the OP-20 for the FEM model and the real part.

75
Fig. 4.7. Cracking in the FEM model.

Fig. 4.8. Cracking in the FEM model.

76
 Fig. 4.9. Cracking in the real part.

Fig. 4.10. The springback value for the crashbox element.

Fig. 4.11. The analysis of OP-20 in the FEM software and after tests.

77
 Fig. 4.12. Comparison between the FEM model and the real part.

Fig. 4.13. FEM B-pillar model with visible areas: die, punch, blank holder, pad and
blank (Fig. 2.23, Fig. 2.25).

The elements of triangle and quadrangle surface types of nonlinear stress distribution
towards thickness were used. This is based on the Belytschenko-Tsay model which relies
on uniform, coherent and reduced integration (Atlas Tool, 2011). The thickness integration
rule was set to the Gauss rule by default number integration points. Five integration points
are needed for good quality results. The number of elements was optimized relative to the
computing time, so that the accuracy reflects the actual conditions. Forming die operation
OP-20 contained 53125 elements and 49842 nodes, whereas the blank sheet initially
contained 5766 elements and 5922 nodes in such a way that the final stage of the process
attained about 20000 elements and it 215255 nodes.

78
 Fig. 4.14. Springback deviation computer by the FEM after OP-50.

The number of triangles for the blank sheet is 186 at the end of the process. However
the number of quadrangles is variable: at the beginning the amount was 5766 whereas at
the end it was 205815.

Fig. 4.15. Springback boundary condition for B-pillar Part (Fig. 2.23, Fig. 2.25).

79
 4.1. Methods and tools in solving the problem

An analysis of forming operation considers variable springback effect that is caused by

blank sheet bumping in order to obtain a required part shape, its internal properties (yield
strength) and also external loadings and strains following the springback. Geometric
complexity of structure for the three-axial curve of an automotive part requires a process of
discretization. For pure bending case, where the elastic-plastic springback may occur (over
the strain close to the yield stress), material behaviour depends on time. An increase of
blank sheet bumping (in order to reduce the springback) is relatively variable for industrial
use and for adequate sheet elasticity. Finite Element Method offers many advantages. Most
FE programs accept the material behaviour law containing anisotropy of elastic-plastic
behaviour and level of hardening sensitivity. More compound programs operate on large
deformations and appropriate rotations. Choice of elements for unloading process and for
integration schedule have to reflect the overall strains and configuration of the parts.

Fig. 4.16. Distance to reference part after compensation without full chain for the roof
hardening part (Fig. 2.22).

80
 Fig. 4.17. Distance to reference part after compensation with full chain for the roof
hardening part (Fig. 2.22). The OP-20 contained 698267 elements and 173554 nodes
and the OP-50 contained 937747 elements and 232379 nodes for the computing time
of 2h 49min 21s and compensation time of 4h 30min 36 s.

The FE method uses forming analysis containing the following elements:

1. Assuming a controlling equation for dynamic examples.
2. Discretization of continuous structure by choosing a mesh and elements type.
3. Partial replacement of various equations standing for motion continuum for
linear displacements of nodes.
4. Solving linear equations, sequential sending and repeating.
Assumptions 1 and 2 are particularly important for the springback analysis. In the first
one, there exists many materials that control the statically determinate equations (in the
nodes) or the F=MA momentum equation for each commercial codes. The implicit method
solves an equilibrium in every iteration step starting from the experimental solution. Each
of the commercial programs have calculations in the static implicit range. The programs
that simulate the forming process solve an equation by the explicit method concentrating
unbalanced force in an increase of time and acceleration but without iterating in order to
find the solution.
Appropriate mesh defining requires choosing elements connected to the amount of
nodes responding to the particular elements. It also requires taking into account the number
of degrees of freedom for each element and the internal connection between the elements
as well. Deformation is connected to a small amount of variables. Distribution of the

81
elements can be completely different for solution continuity and it is often reduced by
remeshing and choosing more adequate elements. Similarly, work done by deformations of
elements equals the work done by displacements of nodes related to the internal forces of
those nodes. An amount of integral locations of points balances between efficiency and
accuracy for the required stability.

Fig. 4.18. Distance to reference part after OP-50 compensation with multi
compensation for the roof hardening part (Fig. 2.22).

Fig. 4.19. Measured finish part without multi compensation by Coordinate Measuring
Machine for the roof hardening part (Fig. 2.22).

The forming analysis of the springback procedure contains the utilized boundary
conditions (punch or die movement, drawbeads using, friction, constraints), holding at the
end of the forming process, replacing the variable’s contact forces by fixing the external
forces (without changing the parts shape) and then, unstressing external forces until their

82
dissapearence. The last step brings out the springback of part surface caused by unstressing
the die. In the analysis of forming and springback, there may be chosen two important
types of solving the alghorith and the equation: the explicit and the implicit methods. Both
methods are used for the springback calculations.

Fig. 4.20. Measured finish part with multi compensation by Coordinate Measuring
Machine for the roof hardening part (Fig. 2.22).

The explicit method is used for dynamic calculations and the implicit method is used
for static calculations. Frequently, both methods are used in numerical analyses, which
allows for the computation of complicated simulations. Most of the currently used analyses
of sheet forming in the industry use the explicit method. Compound shape and contact
conditions in an industrial application allow for very small steps of the dynamic explicit
method. Usually, simulations inaccurately and inadequately represent the reality.
Therefore, the experience of manufacturers can greatly improve the software by
implementing necessary changes. On the other hand, the accurate setting of program
variables needs to generate the proper result for the explicit method.
Otherwise, swapping the method for the implicit one cannot lead to the overall
improvement of the process. It is important to realize that simple simulation of the bending
springback phenomenon can be calculated in an analytical way and then, it may be tested
and compared to the FEM method. Inverse forming analysis is focused on blank sheet
bumping in order to reduce springback. For a perfectly plastic case, the springback
appears, whereas bumping moves in an appropriate yield stress. For deep drawing, the
springback mainly appears during forming of sidewalls and drawn over. The shape of the
final part surface is highly determined by the shape of the tool surface because of the large
friction contact with the stamped part, which often shows the springback effect. The
springback simulation is better matched with the implicit method model because the

83
operation is dominated by elastic quasi-static deformation which is ineffectively calculated
by the dynamic explicit method. Therefore, the springback analysis using the implicit
method is preferable to the explicit method. Especially, the dynamic explicit analysis
should not be used for more precise springback analysis.
Fig. 4.21. The OP-50 operation without multi compensation.

4.1.1 Defining the calculating model

The numerical calculations have been done using commercial ESI Pam-Stamp
software in the 2011.1, 2012.1, 2015, 2015.1 versions that are owned by New Tool Shop of
Kirchhoff Automotive Poland. The preprocessor and postprocessor Autostamp Solvers and
Optimizer 64-Bit 2011.1, 2012.1, 2015, 2015.1 prepare numerical models, then mesh them
(DeltaMesh) and process the results. There are highly specialized applications allowing for
solving linear and nonlinear problems (geometric and physical) and also contact issues.
The Pam-Stamp version 2015.1 is much more accurate than other commercial numerical
programs such as: Autoform, Dynaform and even Abaqus due its use of commercial
computing code. The code is created on the basis of MS NASTRAN and PASTRAN
programs that are widely used for nonlinear studies of construction stability.

84
The calculations have been done on a 64 bit workstation with an Intel Xeon CPU E5-1620
3,6 GHz processor and 32 GB RAM memory. During the computing, about 50% of the
CPU and 23,2 MB of physical memory were used. This was caused by reduced cores
amount of the software license.

4.1.2 Description of the model

In all presented calculations, the three- and four-node elements were used. Their
amount depended on the overall dimensions of the element and the value of the minimum
radius that was formed. In this work, there are considered three different elements:
B-pillar, roof hardening part and front bumper. Each part has its optimized and appropriate
amount of FEM elements. What is more, each element is made from different material
(UHSS steel) with different thickness and geometry. As a result, every part has different
minimal bending radius value, which has a great impact on refinement level and, as
a result, on size and type of the elements. In the work, an amount and type of the used
elements for particular car parts are shown. For all the presented elements the amount
equaled:
1. B- pillar: Forming die operation OP-20 contained 53125 elements and 215255 nodes
for computing time of 14min 43s and compensation time of 5h 37 min 17 s.
2. Front bumper: Forming die OP-20 contained 474205 elements and 126742 nodes for
computing time of 2h 54min 3s and compensation time of 4h 0min 23s.
3. Roof hardening part: Forming die OP-20 contained 698267 elements and 173554 nodes
and Forming die OP-50 contained 937747 elements and 232379 nodes for the computing
time of 2h 49min 21s and compensation time of 4h 30min 36 s.

Remeshing was used as complementary action allows for improvement of mesh quality
whenever necessary. The remeshing operation will automatically detect faces with a bad
quality mesh due to the presence of some anomalous surfaces in the CAD model. The face
detection is done by applying three different elements of quality criteria. DeltaMESH will
select these bad mesh quality faces and delete them.

Remeshing is based on the criterion of minimal radius value r = 2 mm, which determines
the minimal size of the element – 0,64 mm. The refinement level is usually matched to this
criterion so an amount of the elements cannot increase too much. It should be noted that

85
using an inappropriate amount of the elements causes far-reaching inaccuracies and faults
at the stage of tool design. Using few of the elements causes inaccuracies in calculations
and as a result, larger springback, wrinkling, and cracking in limiting areas (Fig. 4.22 and
Fig. 4.25 show the wrinkling aspect for the FEM model, the actual part, and a comparison
between various materials). However, too many of the elements increases the computing
time in exponential function and raises false alarms e.g. pointing out cracking areas that
will not appear in reality (Fig. 4.22 and Fig. 4.24 compare cracking between the FEM
model and various steel materials).

Fig. 4.22. Comparing wrinkling for the FEM model and the tested part.

86
 Fig. 4.23. Comparing cracking for the FEM model and the tested part.

Fig. 4.24. Comparing cracking for the FEM model for: DP-600, HCT600 and DP-K-34.

87
Fig. 4.25. Comparing wrinkling for the FEM model for: DP-600, HCT600 and DP-K-34.

88
 4.2. Results and conclusions of the simulation

The natural limitation of using the methods of tool shape correction are the stamping
parts that have vertical walls. After the correction, the tools having wall angles of less than
90° are obtained, which requires special drives for shaping the geometry of the parts.
Similar restrictions occur for stamping dies with large flat surface areas (Fig. 4.28) where
other buckling forms appear (Fig. 4.27). It is extremely hard to compensate the shape
where the corrected angles require considering minus values or have a prebuckling form
with an excessive amount of material.

Fig. 4.26. Comparison of the surfaces in the B-pillar part.

Fig. 4.27. The analysed part of the roof hardening with visible buckling form
(simulation) and implemented sttiffening (tested part).

89
 Fig. 4.28. The analysed FEM part and the result after tests.

Fig. 4.29. Distance to reference part after compensation with full

chain for B-pillar (Fig. 2.23, Fig .2.25).

As it was numerically and experimentally prove, springback values reached the

assumed level of about 8 mm for OP-20. The numerical analysis showed that the initial
tool compensation without full chain compensation brought an unsatisfactory result.
The deviation value after compensating is above 2 mm, which vastly exceeded complex

90
tolerance range. It is caused by optimizing only of OP-20 (forming) without taking into
account OP-30 and OP-40 (cutting). During the trimming operation little deformation of
an unstressed part occurred, which resulted in parts which could not meet the
requirements of imposed tolerance +/- 1.25 mm. Surface histograms and the FEM
simulations were done for the B-pillar part (Fig. 4.33, Fig. 4.26, Fig. 4.34). The
diagrams of average and maximal distance to reference B-pillar part presented in the
charts (Fig. 4.32, Fig. 4.31).

Fig. 4.30. Distance to reference part after compensation without full chain for the
B-pillar (Fig. 2.23, Fig. 2.25).

91
 Fig. 4.31. Diagram of maximal distance to reference B-pillar part (horizontal axis –
iterations amount, vertical axis – maximal value of the springback after compensation.

Fig. 4.32. Average distance to reference B-pillar part (horizonstal axis – iterations amount,
vertical axis – average value of the springback after compensation).

92
 However, the application of the complex springback compensation enabled obtaining
good parts which reached the tolerance required by the customer (Fig. 4.16, Fig. 4.17, Fig.
4.18, Fig. 4.19, Fig. 4.20, Fig. 4.29, Fig. 4.30). It must be emphasized that some parts’
zones are less important and may fit into general tolerance which is usually wider than for
mating zones and amounts +/- 1.5 mm. Sometimes, bigger springback tolerance is also
allowed in the spot of subsequent calibration operation and in that way it is used in the
presented part. The zone with the deepest overpress is calibration in another OP-60
operation and therefore, compensation with a lower tolerance is acceptable.

Fig. 4.33. Comparison between the standard compensation of the B-pillar, compensated trimming die
and the histogram of % surface deviation during compensation.

93
Fig. 4.34. Distance to reference model of the B-pillar part and the histogram of %
surface deviation during compensation.

94
 5. Springback compensation
The main idea to correct the die consists of application of the classic springback
compensation used in technological conditions, which means extra overbending or sheet
overpress during forming, in order to reduce the springback effect (Ken-ichiro, 2001). For
simple forming, the problem comes down to the additional sheet overbending (Marciniak
et al., 2005). However, if the geometry is more complicated, it is not easy to predict the
behaviour of material. Nonlinear behaviour during stamping process depends on a state of
stress, a state of strain and on the material’s deformation history. According to the material
forming analysis containing springback, its description of tool loading changeability has
a distinct effect on the final analysis. The definition of tool and blank sheet interaction

Fig. 5.1. Reverse engineering process.

determined as a contact friction of constitutive model is also used to model the used
density of mesh for a charge and a tool.
Development of compensation methods based on successive approximation methods
with the possibility to define the measured direction and quality of compensated elements
enables accurate determination of the unstressed surface. The algorithm of the die
correction, with the use of dislocation in an elastic range, assumes the iteration procedure
of springback effect compensation.
Springbacks come out after lightening of the press parts by the strength or after release
of the residual stresses during trimming operations. The run of forming process and
planning of hardening process also has great influence (Wie et al., 2011).
The important issue is springback value which can be calculated by the formula:

δE= |Δf – Δt| (5.1)

95
The precision while forming can be expressed as:

–
(5.2)

Where δE- springback value, Δf- real compensation value, Δt- analyzed compensation
value, ηδΔ- precision of compensation (Bekar, 2011).
The main compensation targets are:
- identification of tooling geometry which adjusts the surface to minimize the distance
between assumed geometry, in such a way that the initial tooling geometry will be
modified (overbent),
- compensation should be studied in a virtual, numerical way to reduce the classic ways of
modification to minimum (e.g.- trial and errors method) or eliminate them completely:
- presentation of compensation range with the use of tool mesh,
- description of surface in CAD format (e.g. IGS),
- description of surface including the history of all surfaces during consecutive iterations,
- transformation of compensation results by the surface description:
- mesh transfer with the modified die surface,
- possible formats of transfer (Ansys, Asci, Ideas, Nastran).
It is possible to obtain CAD surface as clean mesh based on reverse engineering (Fig. 5.1)

Fig. 5.2 Compensation process.

96
method but most often it leads to low surface quality.
There are die and part compensation which are closely connected. However, in
industrial settings the compensated surface of the tool is searched, which is necessary
during the process of stamping die design. It means the modification of tool surface
geometry during design of the forming process and the necessity to optimize the effect.
The algorithm of die correction assumes iteration procedure of springback effect
compensation (Fig. 5.2). It often happens that optimization can end up successfully on the
condition that the entire forming process linked in the correlation chain (forming,
trimming, bending and separating operations) will be analyzed. Springback most often
occurs after cutting and separating operations. The natural restriction to apply methods of
die shape correction are drawpieces which have vertical walls because after correction we
get tools of wall angles smaller than 90 degrees, which require special tools for shaping.
Similar restrictions appear for die stampings which have a large flat surface area
(Gronostajski et al., 2004).
FEM software enabling optimization of the full process chain is not enough. It happens due
to the fact that such programs do not generate surface in CAD but only as the collection of
nodes. It is adequate in theory, but not under industrial conditions.
Creation of the die stamping requires milling of each of its parts. Therefore, in the next
step it is necessary to convert the mesh in the form of different formats (e.g. Nastran,
Ansys files) into more general CAD formats (e.g. IGS). The most common transition
between two softwares generates imprecisions in the form of non-tangency discontinuity
and small surface smoothness. The die optimization requires the input data which include
the original compensated mesh and CAD data of the die. Thanks to this, 2 codes of the net
of heterogeneous CAD surface zone dislocation are given. It means, that every surface is
tightened to the new shape which is not usually bigger than the same surface before the
compensation (e.g. cylindrical surface will not be a cylindrical surface, flat surface will not
be the a surface etc.).
Obviously, NURBS surface model data base on following the increase of surface
flexibility in nodes will conform to compensated elements. Therefore, the main loop for
every surface is checking the accuracy, local nodes till the assumed tolerance will be
reached everywhere other than surface definition. This process generates the path with
maximum continuity determined by double tolerance. Searching the correct, cut surface we
need to make a global dislocating area definition. We solve for the unknown vector valued
parameters λ:

97
 ∑ ‖ ‖ (5.3)

belongs to the set [1,n]: interpolate - harmonic radial basis

function.
Where n is a number of movement vectors.
The adjustment of nodes to the 3D model of the part takes place so that the surface is
stretched on the mesh. The process is implemented by PanelShop software (iCapp)
(Danzberg et al., 2007; iCapp, 2011).

Fig. 5.3. Comparison compensation surfaces to reference.

The innovative compensation method that additionally contains compensation of

trimming stations was taken into account in the process. A standard compensation method
contains only the cutting process during a springback calculation but does not include
compensation of shape of the trimming stations (changes in the trimming operations
surfaces). The compensated surface of the trimming station should be geometrically
positioned between a nominal surface of the forming operation and the compensated
surface of the same forming operation (Fig. 5.3). Mostly, when the springback
compensation process was predicted, the surface of the trimming operation is the nominal
surface or the compensated one. The parts divided after the trimming operation were
analysed by the FEM method (Fig. 5.4). The simulations of joining the parts by external
and internal bridges were done (Fig. 5.5, Fig. 5.6, Fig. 5.7, Fig. 5.8).

98
 Fig. 5.4. The springback value for divided parts.

Fig. 5.5. The springback value for parts joined by external bridge.

99
 Fig. 5.6. The springback value for parts joined by internal bridge.

Fig. 5.7. The springback value by vektor depending on bridge joining type.

100
 Fig. 5.8. The springback value depending on bridge joining type.

Real process Simulated process

Gravity Gravity
OP-20 Holding OP-20 Holding
OP-20 Stamping OP-20 Stamping
OP-30/ OP-40 Triming OP-30/ OP-40 Triming
Springback after OP-40 Springback after OP-40
OP-50 Holding OP-50 Holding
OP-50 Stamping OP-50 Stamping
OP-60 Triming OP-50 Triming
Springback after OP-50 Springback after OP-50

Tab. 5.1 Real standard process and simulated operations.

101
 Using the compensated surface on the trimming and the forming stations increases an
accuracy of the whole process. It allows for getting a better compensation result, even
a few percentage points better than the standard compensation process (Tab. 5.1). If the
compensation is not included at the trimming station it causes growth in the reference
surface deviation. What is more, it also influences on stamped blank positioning errors that
appear at the trimming station because of the springback difference before and after the
trimming process. Therefore, it is recommended to apply this method in order to enhance
dimensional tolerance and surface quality. In some particular tools analysed by the authors,
situations occurred where the springback effect was eliminated by standard compensation.
However, the positioning problem was so significant that the blank sheet between stations
distorted itself, which was visible in the finished element. As a result, it led to a search for
a compensated surface that was usually separated between the operations, which reduced
the effect of irregular positioning. Obviously, this situation ought to be predicted during
design of the tool in order to avoid extra costs related to rework of the existing tool and
time delays in delivering prototype elements.

5.1. Trimming tool compensation

Trimming operations usually do not occur individually, but are preceded by a forming
operation. This function enables semi-automatic calculation of stress-relieving in the

Fig. 5.9 Visualisation of simulation process from forming OP-20 to

springback OP-50 for the B-pillar part (Fig. 2.23, Fig. 2.25).
102
optimization process of the tool.

On the basis of the calculation results, the surface of the trimming tool gets overbend
and the positioning of the drawing parts stays effective. In fact, the surface of the trimming
tool has a shape between a calibrated surface and a stress relieved forming surface, which
may lead to a positioning problem with the blank in the trimming tool. It is necessary to
take into consideration springback transfer from operation OP-20 (new surface of the
forming die) to the tool in operation OP-30/ OP-40 (with the overbend for springback in
operation OP-30/ OP-40). Taking into consideration optimization of the trimming tool
allows one to achieve a true forming process. Optimization of the trimming operation in
order to improve efficiency of standard stamping tool compensation and solving the
positioning problem of the blank (Fig. 5.9).
Because of the variation in springback of formed parts on the trimming operation,
there may appear to be a problem with positioning of the blank during transfer from
operation OP-20 (forming operation) to operation OP-30/ OP-40 (trimming operation).

Fig. 5.10. Compensation process taking into consideration optimization of trimming

tool allows one to achieve a real forming process.

The innovative compensation method that additionally includes compensation for

trimming stations was taken into account in the process (Fig.5.10). A standard
compensation method includes only the cutting process during springback calculations but
does not include compensation for shape of the trimming stations (changes in the trimming
operations surfaces). The compensated surface of the trimming station should be

103
geometrically fit between a nominal surface of the forming operation and the compensated
surface of the same forming operation. Mostly, when the springback compensation process
is predicted, the surface of the trimming operation is the nominal surface or the
compensated one. Using the compensated surface on the trimming and the forming stations
increases the accuracy of the whole process. It allows for getting a better compensation
result, even more precise than the standard compensation process. If the compensation is
not included at the trimming station it causes an increase in the reference surface deviation
(Fig. 5.12). Further, it also has an influence in stamped blank positioning errors that appear
at trimming station because of the springback difference before and after the trimming
process. Therefore, it is recommended that this method be applied to enhance dimensional
tolerance and surface quality. In some particular tools analysed by the authors, situations
occurred where the springback effect was eliminated by standard compensation. However,
the positioning problem was so substantial that the blank sheet distorted itself between
stations, with a distortion which was visible in the finished element. As a result, this led to
a search for a compensated surface that was normally separated between operations, which
reduced the effect of irregular positioning. Obviously, this situation should be predicted
when designing the tool in order to avoid extra costs related to reworking of the existing
tool and time delays in delivering prototype elements.

5.2. Realistic Shape Compensation

CATIA Realistic Shape Optimizer – allows for optimization of the 3D model by the
realistic shapes resulting from the stamping process simulation or measurement of real part
(Fig. 5.11, Fig. 5.13). RSO also allows for the execution of geometric morphing of the
nominal surface or with analysis of the finite element method for the discrete model. The
surface, created for load conditions, is defined on the basis of the comparison between the
measurement and the initial surface as defined by CAD. This technology ensures access to
the process of designing and constructing the tool in order to save some time. Therefore, it
is possible to include particular displacement to create a realistic shape, which is required
by the construction or the designed tool. For the basic blank elements, really simple text
files containing points and coordinates vectors, simulations or measurements are required.

104
 Fig. 5.11. Deviation analysis of the B-pillar part in Catia (Fig. 2.23, Fig. 2.25).

Realistic Shape Optimizer (RSO) tool and Generative Shape Optimizer (GSO) allow
for simplistic deformation of the tool using referent geometry of tools’ surfaces as
a compensation aim because it accepts the nominal geometry of the tool. Realistic Shape
Optimizer allows for the use of point cloud received from numerical analysis FEM which
can be saved as a Simple .txt file so it conforms to the whole part. It is also possible to
import surfaces from numerical calculations programs in the form of nodal points. In order
to get the proper geometry of the referent part it is necessary to use springback
compensation from CMM measurement export in the form of point cloud or from the
numerical analysis as a format e.g. Nastran.

105
 Fig. 5.12. B-pillar surface creation by Shape Optimizer (Fig. 2.23, Fig. 2.25).

Thanks to the RSO it is possible to use actual measurement of the part which contains
real deviations from the referent surface. Then they are defined as values of necessary
correction to eliminate springback effect. For this, the compenstation coefficient is given as
-1 value in order to reverse measured values and replace them with the compensated
values. It allows for the reducection of springback in engineering terms on the basis of
numerical analysis or CMM measurement. Obviously, matching a proper scale efficient is
determined by the experience of program user and his observations of previous results for
similar elements. It turns out that for different materials and complex part geometries,
proper surface overbending for compensating the springback effect is not so simple and is,
in fact, highly time-consuming. Therefore, using opportunities for engineering function is
an irreplaceable work tool for a modern tool constructor. It is worth mentioning that some
cases such as torsional springback are not possible to implement in such simple
engineering software. In this kind of situation the only possible way of solving the problem
is using FEM software with an opportunity to compute nonlinear springback.

106
 Fig. 5.13. Measured real B-pillar part in PolyWorks software Deviation in [mm] (Fig.
2.23, Fig. 2.25).

Fig. 5.14. Tortional springback of the bumper part and its boundary conditions.

5.3. The surface compensation with scanner

During the removal of the sheet from the tool used in a forming process, additional
deformation is generated. This deformation is called the springback effect. Its severity
depends on the coefficient between residual stress and Young modulus.
107
Fig. 5.15. B-pillar deviation analysis for springback compensation (Fig. 2.23, Fig. 2.25).

The springback effect is one of the most significant factors that influence the
quality of stamped metal parts. Currently, it is the largest defect in blanks. This
phenomenon is observed for high strength steels which are widely used in the
automotive industry as a material reducing vehicle weight and improving the properties
during crash tests. Estimating the springback size in the final forming process is a well-
known problem in metal forming. The compensation surface after springback is highly
difficult to parametrise in a traditional way. For simple bending of a U-shaped surface it
is possible to predict the compensated surface using an analytic method. However, the
problem appears for complicated bending-drawn parts or for some parts with a majority
percentage of flat surfaces. The complex measurement method, for example with an
optical scanning method, is dedicated for drawing surfaces (Fig. 5.14, Fig. 5.15). First
of all, the appropriate parametrisation of the outer surface received after the springback
represents the final forming surface using the radius length and the angle length.
Without extensive analysis of the issue, this method is imprecise and not appropriate for
the surface description. In the case of measurement of a surface with the springback
effect it is necessary to get the proper amount of points. Otherwise, it causes loss of
accuracy for the compensated surface indirectly by measurement. For the basic
U-shaped surface in the final forming stage it is possible to conduct the parametrisation
process of force, stamping force, velocity etc.

108
 Fig. 5.16. Scanned B-pillar surface (Fig. 2.23, Fig. 2.25).

Metal forming is defined as blank deformation in order to obtain the demanded surface
shape. It is done by plastic deformation in the forming tool. Before the 1980s, forming
parts production was treated as sometimes difficult to implement in practice. However,
many fundamental plasticity theories have been published since then. Application of theses
has been limited to theoretical speculations because of simple surface shapes that were
repeatedly limited to perfect conditions. Therefore, it has not had large implementation in
industrial production. Computer-aided CAE engineering analysis with the finite elements
method was a huge success for many applications in 1980. Unfortunately, in the mid-1990s
this technology was not sufficiently refined for mass production because of hardware
limitations of computers.

109
 6. Twisting springback

Compensation of springback effect is not the easiest method but by using commercial
programs based on the Finite Element Method (FEM) it has become more effective than
the heretofore standard method of trial and error. The problem may occur when the
springback does not result in the special case of twisting springback which deposits
partially a lot of energy unintentionally. The appearance of such deviations from
intentional geometry complicates the compensation of the tool (Li et al., 2011). The result
can be such an unstable state that we will not be able to predict the divergence of twisting.

Fig. 6.1 Twisted springback is caused by stress distribution in the sheet.

Generally speaking, springback depends mainly on engineering factors, e.g. material

properties, coefficient of friction, geometry of tooling and process parameters. Springback
effect concerns all steels, however some materials, such as AHSS, exhibit particularly
flexible- elastic properties which reduced with trial and error methods are expensive and
time consuming.
Twisting springback (Fig. 6.1) is caused by elastic redistribution of internal stress
during the unloading process, which is usually the result of longitudinal dissymmetry for
materials of an increased plasticity limit. From the mechanical perspective, springback is

110
a result of a torque moment in the cross-section of a work piece which causes the turn of
one end relative to the other.
The torsional dislocation expands through non-balanced elastic deformation and
residual stress acting. The torsional moment can come from the in-plane residual stresses
in the flange, sidewall, or both (Abdullah et al., 2012) . The influence of hardening law
type (kinetic or isotropic) on twisting springback angle and longitudinal curvature was not
found. It results in discrepancy on the final surface and problems during the design
process of a metal forming tool. Through years of thorough studies carried out with the aim
of understanding, predicting and controlling springback, much important information has
been gleaned which has been used to eliminate the effect. For springback analysis, an
adequate example has been chosen which shows the tendency to twisting springback. It is
useful to pay attention to the way in which springback can appear:
 height differences (standard springback deviation)
 transversal changing curvature
 longitudinal changing curvature
 change of section curvature
 torsion of all geometry (special case) (Mangas et al., 2006).

Fig. 6.2 Twisting springback (with SSAB cooperation).

111
Very often springback reveals itself in a few cases simultaneously, which additionally
complicates the complex deviation of a part. The authors would like to show the special
case with torsion (Fig. 6.2) of all geometry including also the changed section of curvature.
Twisting with angle deviation is not a typical case, especially when the part geometry is
close to the symmetric part in longitudinal deviation.

6.1. Forming method in case twisting springback

Due to the geometry of a double U and more precisely the flange having a curvature
of an arc it is required to apply the elements of Thorough Thickness Stress (TTS) which
take into account 3D plasticity. These are the elements of the solid type which consider 6
constituents of stresses, including axial compression resulting from contact pressure
which is observed in a flange area. It has an essential meaning for the flanging process,
especially when the bending radius is lower than 3- steel thicknesses (ESI Group, 2011).
For the stamping conception including blank holder force, the blank holder force
(Eq. 4) for the numerical model was estimated at 310 kN value and it was confirmed
experimentally that a 200- 300 kN force assured correct holding.

(6.1)

Where: p- pressure; A1, A2- the area of blank holder for case 1 and 2
For the additional simulated pad the force used was lower the than the force of a basic
blank holder so that the kinematics remained appropriate, therefore blank holder force
was F2=150 kN.
The proper modelling should allow for the gravity of a blank sheet, what makes the
model more accurate and comparable with a real forming process. In the numerical
analysis the influence of gravity (Fig. 6.3) in the middle part was 48 mm, whereas for
a real measurement it was about 45 mm.

112
 Fig. 6.3. The gravity influence in the springback compensation for the bumper part (Fig.
2.26). Forming die OP-20 contained 474205 elements and 126742 nodes for computing
time of 2h 54min 3s and compensation time of 4h 0min 23s.

6.2. Application of various calculation models

Many tests have been carried out in order to find the best method for the simulation
calculation. The same calculations were made for computing methods such as: standard
implicit, explicit, and advanced implicit (Fig. 6.4, Fig. 6.5 and Fig. 6.6 illustrate the
twisting springback depending on solver computation methods). It must be stressed that
the explicit method is usually used for a major nonlinearity, whereas the implicit one is
commonly used for the statically linear processes such as gravity and springback. The
advanced implicit algorithm contains an improved implicit algorithm using the same
base. It is fairly clear that the most reasonable approach seems to be the advanced
implicit analysis, however the experiment has shown that standard implicit is enough to
provide results comparable to reality (Fig 6.7, Fig. 6.8). The explicit method fails totally,
resulting in both sides twisting (Fig. 6.7). Taking into consideration research on
a prototype die, the next analysis includes the standard implicit model as the most well-
adjusted to reality.

113
Fig. 6.4. Twisting springback depending on solver computation method: Standard Implizit, Explicit, Advance
Implicit

Fig. 6.5. Twisting springback depending on solver computation method: Standard Implizit, Explicit,
Advance Implicit.

114
Fig. 6.6. Twisting springback depending on solver computation method: Standard Implizit, Explicit,
Advance Implicit.

Fig. 6.7. Types of calculations: Standard Implicit, Explicit, Advanced Implicit for the
bumper part (Fig. 2.26).

115
 Fig. 6.8. Distance between the object for calculation types: Implicit, Explicit, Advanced
Implicit for the bumper part (Fig. 2.26).

All analyses for different types of computing were done in the commercial Pam-
Stamp 2G software. However, prior to the production of the trial tooling, with the aim of
costs reduction, a numerical analysis was carried out using a similar type of software
Autoform, to verify meaningful springback (maximum angle deviation approx. 71 deg.).
It is important to stress that all parameters including material data were set exactly in the
same way as in the Pam-Stamp 2G software. As a result, a similar twisting type was
obtained even though it was more moved than twisted (Fig. 6.8). Therefore, in subsequent
tests it was assumed that the FEM model for the Pam-Stamp 2G software was reasonable
and such a springback type could be expected.

Fig. 6.9. The comparison of the FEM calculation for different commercial software
products for the bumper part Autoform (left) Pam-Stamp (right) (Fig. 2.26).

116
 6.3. The influence of a stamping conception on dimension of a springback
deformation

The object of research was a front part of a bumper (Fig. 6.10) which evolved during
the designing of the die tool, three times changing the geometry of the part. These
changes were necessary in regard to the strategic significance of the parts during a head-
on vehicle collision. They were caused by inadequate stiffness of the element during the
crash test.

Fig. 6.10. Autoform FEM analysis of the bumper part (Fig. 2.26).

Unfortunately, the new changed geometry triggered off formation of a twisting

springback, a high tendency for material wrinkling, and oil canning phenomenon. All of
this necessitated the performance of numerous formability analyses, putting an emphasis
on the springback problem. Hence, a few material models were included and a few dozen
virtual tests were conducted. What is more, without the inclusion of computer analyses,
even the most experienced tooling designers were not able to predict twisting springback

117
 for set parts geometry. Only the trial tool showed how serious of a defect it was. A few
stamping concepts were tried both virtually and in real trials, verifying the method of
springback deformation. Generally, stamping conceptions can be divided based on:
I. Blank shape- initially rectangular blank (Fig. 6.11) was used for calculated values: 200
x 1057 mm, however, as the trials showed, that shape caused the biggest twisting.
Therefore, the decision was made to use the blank shape of a maximum width of 160
mm, whereas the length was unaltered. It means that forming with final blank outline is
without any trimming after stamping except piercing holes.

Fig. 6.11. Type of the Blank sheet outline (shaped and rectangular).

II. Holding method:

a. Crash forming- not including any holding (Fig. 6.12). It is the least complicated
conception of stamping which reduces springback effect, but at the same time exacerbates
the wrinkling of a blank sheet and oil canning.

118
 Fig. 6.12. The process and springback result for crash forming method for the bumper
part (Fig. 2.26).

b. Classic forming with the use of a blank holder that is holding and stamping (Fig.
6.13) - enables the assurance of holding in the required area but sometimes it causes oil
canning phenomenon, when there is inadequate holding force or too little holding surface.

Fig. 6.13. The process and springback result (cross section) for holding and stamping
method for the bumper part (Fig. 2.26).

c. Classic forming with extra re-striking- for the case of twisting springback it
eliminates twisting value or leaves a slight deviation depending on the press load.
Generally speaking, the bigger the force of the press the more compatible the result will
be with the expected geometry.
d. Forming with extra pad (Fig. 6.14) - considerably reduces springback value and
oil canning phenomenon by extra holding. Taking into consideration two holding forces

119
of a standard blank holder and pad, which are facing it is required to apply one of the
alternatives if it is a transfer tool:
- press with delayed cushion back move
- gas spring with delayed system including special cooling
Each of these options involves extra costs of equipment or investment in a more
expensive press.

Fig. 6.14. The process and springback result for standard forming with an additional pad.

e. Stamping in two forming stages (Fig. 6.15) - in the first one U-shaped forming
and then flanging was done. Such a process develops the tooling significantly and
emphasizes the springback which was shown for the set geometry (Fig. 6.16).

Fig. 6.15. The process in two stages forming for the bumper part (Fig. 2.26).

120
 Fig. 6. 16. Twisting after two stages forming.

6.4. The experimental verification of FEM results

The above-mentioned concepts of forming have a significant influence on the shape

of springback deformation and the value of its deviation. All of the methods except
forming with extra pad and stamping in two stages were experimentally verified by trial
to choose the most appropriate method of forming which eliminates twisting springback,
however the aim was to replace twisting springback with a standard deviation towards the
height of a part (Fig. 6.17).

Fig. 6.17. Twisting and standard deviation of springback for the bumper part (Fig. 2.26).

121
 6.5. The part formability

Taking into consideration geometry specification and the material model profile
inappropriately chosen forming parameters (Fig. 6.18), especially the holding force, can
cause local cracks, which practically renders the quality of stamped parts as unacceptable.
The von Mises stress analysis of the bumper part was simulated (Fig. 6.19, Fig. 6.20). As
was proven in the trials, even a small change in blank holder force caused more
wrinkling. It is important to use the nominal surface first and then the compensated one to
prevent it from cracking. However, usually in the standard compensation, including
restriking it is done in reverse order.

Fig. 6.18. The FLD diagram of for the bumper part after forming (holding plus stamping)
for Dogal 1000DPX (Fig. 2.26).

122
Fig. 6.19. Comparison of FEM results for reduced Huber-Mises-Hencky (H-M-H)
hypothesis with real cracks problem for the bumper part (Fig. 2.26).

Fig. 6.20. Von Mises stress analysis of the bumper part (Fig. 2.26)

123
 6.6. Forming conceptions

Forming for the rectangular blank which gets cut in the last operation, surprisingly
resulted in a twisting springback. The FEM model showed in the most cases of
calculation about 51 degrees of twisting for 310 kN holding force, while in the
experiment the value of the twisting angle using 200 kN (4 bar) holding force was 54
degrees and for 330 kN (6 bar) it amounted to 42 degrees. As it is visible, the value of the
twisting angle depends strictly on holding force and boundary conditions.

Fig. 6.21. Numerical and Experimental twisting springback for rectangular blank
forming.

During the virtual modelling the maximum value of the twisting angle amounted to
an even 71 degrees. (Fig. 6.21), which confirms the instability of a twisting angle in the
value and twisting side. After the forming process, the part possessed a huge amount of
energy which was so unstable that simply with the use of muscle strength the twisting
side could be switched from left to right (Fig. 6.22, Fig. 6.23) or in the opposite direction
with the distinctive crack of an unbalanced elastic strain energy.

124
 Fig. 6.22. Twisting of the left and right side in the FEM bumper model (Fig. 2.26).

Fig. 6.23. Twisting of the left and right side in experiment of the bumper part (Fig. 2.26).

For the crash forming process (which does not take into consideration the holding
force) (Fig. 6.24) with the finished shaped blank of part, the springback was limited to the
standard one in the height direction. The success was that there was no, or just barely
noticeable, twisting of the part. However, the stamping process required application of
a special based blank sheet to retain its correct arrangement and to receive repeatability of
parts. Hence, additional pins were used in the middle of blank sheet (limitation of
a longitudinal movement) and slot in the later openings resulting from geometry
(limitation of transverse movement). The only defect of the stamped part which appeared
after forming using that method, was the excessive wrinkling of the material on flange in
the form of oil canning phenomenon.

125
 Fig. 6.24. Results for crash forming method of the bumper part (Fig. 2.26).

Forming using classic holding and stamping method (Fig. 6.25) for holding force
310 kN resulted in a slight twisting of the entire part (by a few degrees), which was
eliminated by the additional calibration operation with the press force of 5000 kN (Fig.
6.26). The suitable holding force reduces the oil canning, effect ensuring proper steel
holding. Moreover, the holding force during calibration is so strong that it is necessary to
take it into consideration when choosing the type of press. The bumper part was
measured by the CMM machine and the results are presented (Fig. 6.27).

Fig. 6.25. Results for holding and stamping method with soft twisting for the bumper part
(Fig. 2.26).

126
 Fig. 6.26. Experimental and FEM results for holding and stamping method with
additional re-striking operation for the bumper part (Fig. 2.26).

Fig. 6.27. The CMM measuring of the springback value of the bumper made from the
MSW-1200 steel (Fig. 2.26).

127
7. Verification of the result
The experimental verification of the stamping tool concerned the strategic parts of
car construction: the B-pillar part, the roof hardening part and the front bumper. It
showed that all the presented elements fulfilled the required tolerances to 90%.
Moreover, all the parts had to maintain their condition of repeatability in production
which ensures an appropriate assembly by welding. In automotive fabrication, the
amount of pieces produced yearly varies between a few thousands and a few million so
repeatability is highly important.
The tolerance of surface shape for all the presented elements equaled +/- 1.0 mm
and some areas of the parts had the general tolerance in the range of +/- 1.5 mm.
Therefore, the numerical analysis process allowed for the computation of springback
value and then compensate it precisely, which resulted in fitting the expected shape
tolerance to the finished elements. Those elements were measured and visually analysed
in order to eliminate anomalies that could lead to holds on the production line of
a client. The FEM numerical analysis fully confirmed the springback value and
precision of the compensation. Further, the positive influence of considering the
multioperational compensation, the compensation , was that of the trimming operation
and the positioning the blank in every operation were revealed.

128
 Fig. 7.1. Comparison of FEM numerical analysis (the springback value after
compensation for a reference part) and measurement of the experimental B-pillar part
(Fig. 2.23, Fig. 2.25).
129
 Fig. 7.2. Comparison of FEM numerical analysis analysis (the springback value
after compensation for a reference part) and measurement of the experimental bumper
part (Fig. 2.26).

130
 7.1. Description of the equipment with technical characteristics

The measurement of the formed part for an automotive industry is not the easiest
job. This is due to its complicated shape and reduced tolerance, especially for
measuring having above 2000 mm.

Fig. 7.3. Methods of the springback measurement.

The terms and conditions in determining which measuring method to use in laboratory
conditions:
- size of measured object,
- sort of material that the detail is made of,
- type of measured object,
- type of measuring characteristic,
- tolerances of the characteristics,
- accuracy of the measuing device,
- way of clamping the part,
- knowledge and experience of staff,
- waiting time of measuring,
- necessity of keeping statistical analyses,
- Accuracy of measuring systems.

131
 In the stamping process, the most widespread measuring methods are the manual
ways: standard gauge tools, control attributive tools that control both single parts and
whole assemblies. Checking tools allow one to measure the shape of a compound surface,
trimming course, and also the locations and sizes of holes, very quickly.

Fig. 7.4. Controlling attributive device, accuracy of measuring arm 0,04 mm.

Most often, the measurement takes place for certain points of Reference Point System
(RPS) that express the setting in the configuration of the whole assembly or vehicle. This
kind of measurement allows for the fast tolerance verification of the shape, location and
trimming course by using the gauges. The example of the checking tool shows the picture
(Fig. 7.4).
Obviously, the measurement using the gauges depends on many aspects: quality of the
tool +/- 0,1, the way of fixing of the part and even the order of the part setting. Therefore,
the checklists are used in order to measure the stamping part in a proper order. Very often,
the measurement made by the checking workshop tool is not enough. For that reason,
measurements by Coordinate Measuring Machine (CMM) or the methods of optic
scanning of the surface (Fig. 7.3, Fig. 7.9) are used. The boot’s roof parts with and without
multi compensation are measured by CMM (Fig. 7.7, Fig. 7.8). These measurements are
compared to the results of the FEM analysis (Fig. 7.5, Fig. 7.6).

132
 Fig. 7.5. Distance to reference part after compensation without full chain.

Fig. 7.6. Distance to reference part after compensation with full chain.

Fig. 7.7. Measured finish part with multi compensation by Coordinate

Measuring Machine.

133
Fig. 7.8. Measured finish part without multi compensation by Coordinate Measuring Machine.

Fig. 7.9. Accuracy of CMM Zeiss machine 0.01 mm.

Another important factor that eliminates the measurement by workshop attributive tool
is the actual cost of performed checking tool, which is relatively high and is used only for
measuring one specific stamping part. Therefore, any engineering changes generate an
expensive gauge modification.

134
 Fig. 7.10. GOM Scaner measurement in manipulator.

That is why, using these kinds of measuring tools is expensive and inflexible. CMM
machines with a measuring tolerance of +/- 0.01 mm the measuring lengths of 1000 mm
(Fig. 7.10, Fig. 7.13) have been in use for over a dozen years.What is more, scanning tools
have also been in use for a few years.There are intermediate tools with jointed contact arm
with the measuring tolerance of +/- 0.04 mm or tools with the scanning head (+/- 0.05 mm)
that are now commonly used for surface measuring (Fig. 7.12, Fig. 7.13).

Fig. 7.11. Failures and innacuracy of measurement (C- specification phase, D -

verification phase, specification field in the specification range; 1- out of specification;
2- compatibility field; 3- incompatibility field; 4- inaccuracy range; 5- increase of
measuring inaccuracy.

135
 Metrological aspects:
- Drawing requirements (constructional, technological);
- Tolerance of made part, subassembly and assembly;
- Referential elements of the measurement;
- Base system (RPS, in the controlling device, surfaces, holes);
- Measuring conditions (measuring laboratory, workshop conditions);
- Human factor (sensitivity of machines e.g. keeping perpendicularity of the measurement);
- Measuring method (touch, laser, optical);
- Measuring failures (Fig. 7.11);
- Innacuracy of measurement .
The optical measuring tools are more and more often used thanks to their incredibly
important feature of providing a surface measurement with relatively high accuracy of over
0.09 mm for measuring range of over 1200 mm. They are used for large and complicated
thin-walled elements. Currently, optical scanning tools are more and more sophisticated.
Using two cameras for the scanning optical tool gives an opportunity to detect movement
during the measurement process, increases the field of vision and simplifies the
measurement in places that are hard to reach. Using blue light removes any noises and
porosities of registered data, increasing flexibility in regard to the work in production line
conditions and eliminating heating of the scanner head (allows for maintaining calibration
of the scanner for a longer time). Other benefits of the blue light are the possibility of
measuring without chalking, measuring without the necessity of searching for shaded
places and what is more, the blue light manages with anti-reflex very well.

Fig. 7.12. The measuring FARO tool with a jointed contact arm.

136
 Errors Zeiss [%] GOM [%] HGV Faro [%]
[%]

software-algorithm 0 5 5 0

measures against CAD- 0 0 20 5

data

sensor head 5 5 5 5

accurancy IR 0 10 10 10

clima control 0 5 5 5

fixture and parts 5 5 5 5

Total 10 % 30 % 50 % 30 %

Measurement tolerance +/- 0,03 +/- 0,25 +/- 0,5 +/- 0,06
in tool shop [mm] [mm] [mm] [mm]

Fig. 7.13. Measurement tolerance by equipment.

An accuracy of a single scan is 0.015 mm and it depends on a few factors. For the
measuring field of 500x350x300 the measuring accuracy equals about 0.03 mm and it is
the most optimal quantity (Fig. 7.10). The narrowest field of vision has to completely
describe the measured part. The lesser the field of vision is, the higher the accuracy is. A
greater number of scans reduces the measuring accuracy but a greater number of scanned
points gives a more accurate representation of the measured object.
The advantages of the scanning arm:
 Mobile measuring system;
 Possibility of using the tool directly in the workshop in case of fast control without high
requirements.

137
 The drawbacks of the scanning arm:
 Quality and resolution of scans depends on the travelling speed of the head;
 Necessity of constant controlling of the head position during scanning;
 Failure of head accuracy increases the failure of arm encoders;
 Incidence angle of laser and the head path impacts on the measuring accuracy;
 The human factor highly influences accuracy of the tool;
 Lack of constant controlling of the measured object movement;
 Lack of checking the correctness of a single scan;
 Low work ergonomics by using the head may cause significant accuracy drop and in the
worse case, the carbon arms can lead to uncontrolled displacement of the object;
 Lack of possibility to automate another measurement of the same part;
 Necessity of high engagement of the user during scanning – simultaneous controlling
perpendicularity of the head, distance between the head and the object, arm controlling and
controlling the scanned areas at the monitor;
 An orientation of the head in regard to the object is aggravated during measuring of
large details.

7.2. Tolerances and clients requirements

Quality requirements regarding the products of the blank sheet stamping process are
significantly high because of the technology of automatic assembling of the stamping
components e.g. plasma welding or laser welding where the maximum edge deviation
cannot exceed +/- 0,2 mm. During the trimming operation, the unstressed part is subtly
deformed, which prevents fulfillment of the imposed tolerance.
The springback, as a main source of stamping parts inaccuracy, is a function of
material data, tool shape and parameters of the process. Therefore, the springback
deformation becomes a critical problem, especially for UHSS steel with compound part
geometry. This is why it is necessary to find the severity of the springback and also to take
into account this effect during the early stage of design by tool designers. However, plastic
forming processing is distinguished by very high efficiency, repeatability and especially
the economic aspect keeping good surface condition in the mass production at the same
time. Using curvilinear surfaces for components aims at obtaining smooth reflexes so the
formed surfaces do not contain any defects.

138
 7.3. Description of research and work-place

This work contains the results of experimental studies that are divided in view of used
technologies and the springback form. Therefore, some parts of the results are put directly
in the text. However, the work is based on a number of studies and tests that were done on
automatic transfer presses with a press force of 1200 tons or on the test press with a press
force of 1000 tons equipped with an additional blank holder in order to ensure an extra
press force. Size of the tools used for testing and making the prototype and serial parts is
about 6m in length and 2,5m in width. Mass of a single transfer stamping die is about 30-
40 tons. Every tool made is a prototype because in spite of many similarities, changing
even a few details causes significant problems e.g. changing a supplier of normalize
material generates excessive thinning of the material and even cracking. A material model
is so important that material properties should already be considered during the springback
calculations. Manufacture of a single stamping tool that produces a single part – takes 4-5
months from the design stage to its termination.
During the tests and experiments the environmental model of the numerical analysis
was approximated. The appropriate spring pressure was ensured in order to control the
press force. What is more, adequate press force and friction conditions were also set i.e. the
blank sheet was lubricated with a stamping Helcoted oil and drawing Profi Stanz B3 oil
etc. The produced parts were transferred to the testing laboratory where the shape tolerance
and the tolerance of positioning of surface and trimming lines were checked in details. The
way of fixing of measured parts and the way of measuring are described in chapter 7.

139
 8. Summary and conclusion
Currently, the automotive industry is changing rapidly in terms of implementing new
restrictions and regulations that require reducing vehicle mass or fuel consumption.. In
North America, the government regulatory body has increased the fuel mileage
requirement for 2025 to 54.5 mpg. In Europe, the passenger vehicle emission target is 95
grams per kilometer for 2021. These new standards will surely promote innovation and
investment in advanced technologies. Three of the five core products we studied showed
a strong growth tendency towards lightweight design, in particular Cross-Car Beams
(CCB), Crash Management Systems (CMS), and Front End Modules (FEM).
In Europe, the IKA institute investigated 195 CCBs and it showed 64% of these
beam assemblies are steel. Similarly in North America, the majority of the beam
assemblies are steel too. Initial design concept of the beam assembly will be steel material
and as the vehicle matures and if there is weight classification concern, then alternative
material will be considered based on incremental cost over mass saved. For 2020, the
cross-car beam material choice is steel to keep costs down and develop advance
technology and/or innovation in AHSS for tubing and stamping to reduce mass. The
current trend indicates that 60% of the rear bumper markets are using steel material and
a majority of them are hot forming to keep the mass low as compared to conventional roll
formed design. Alternative materials will enter hot stamping (higher strength 1900MPa,
lower strength 500MPa) and opens new opportunities for tailored welded blanks.

8.1. Scientific aim of the work

First of all, the most important is accuracy of simulation which allows for exact data
concerning material, boundary conditions and process parameters. The material model is
so important that it should take into account material data of the specific supplier already
during springback calculation. If the exact value and type of element is allowed for, it
guarantees keeping necessary accuracy and saving time. Defining of tool correction range
influences calculation time so that taking into consideration only a partial correction may
be enough. It is also required to build the correct surface from nodes of defined coordinates
to obtain a uniform CAD surface of the tool. Compensation is determined so much by
springback calculation accuracy that if they are not exact, then in practice it will make it
impossible to achieve the required result.

140
 A breakthrough in compensation quality has occurred recently with the use of multi-
operation compensation. It enabled the better simulation of reality by correction of
individual tooling stations, which increased the precision of computation.
As the numerical analysis and physical experiment confirmed, not taking into
consideration springback during the forming process, leads to reception of parts which do
not comply with requirements.
Many factors need to be taken into consideration to get a compensated surface which
will meet the assumed tolerance:
 Accurate material data taking into account the hardening curve and FLD chart;
 Correct forming process without cracks or severe wrinkling;
 Meshing strategy as compensation (includes maximal angle criteria which must be
set as 7.5 deg);
 Defining of punch velocity as a ramp with a maximum speed of 5 m/s (including
scale);
 Use of correct boundary condition strategy (set statically or manually determined
by RPS points);
 Correct computation of solver type (implicit, explicit, advance implicit);
 Set up springback after all operations
 Correct mass scaling (CPU control)
First, the most important aspect is accuracy of simulation which allows for exact data
concerning material, boundary conditions, and process parameters. The material model is
so important that it should take into account material data of the specific supplier even
during springback calculation. Compensation is determined so much by springback
calculation accuracy that if it is not exact, achieving the required result will be, in practice,
impossible.
The breakthrough in compensation quality has occurred in step with the use of multi-
operation compensation. It has enabled better simulation of reality by correction of
individual tooling stations, which has increased the precision of computations.
Achieving compensated surfaces for forming and cutting operations at the design stage
enable time savings and, especially, cost reductions connected with the trial-and-error
method. These kinds of parts with such a complex geometry and material specification
require approximately a few months of hard work in order to achieve the assumed
geometry that is contained in the tolerance range. Drawpieces’ defects such as cracking

141
and excessive folding are easy to predict and eliminate, however doing the same for the
springback effect often remains problematic without being sure of a virtual environment.
The range of the research that was done included wide material tests (steel anisotropy,
hardening curve, graph of forming ability, chemical composition), impact of geometrical
factors on the steel springback (changing of radius and angle, additional stiffnesses and
changing of the surface shape), predicting the springback value and the way of springback
compensation. What is more, the UHSS steel producer - Swedish SSAB company has
worked in cooperation for a few years in order to complete the knowledge in the range of
mechanical and chemical properties of UHSS materials and to improve their parameters.
Moreover, there was cooperation with the Pam Stamp 2 G software producer – the ESI
Group company in order to numerically design the forming process in an appropriate way.
Thanks to that, the numerical analyses strictly imitated reality and gave correct results.
The torsional springback research was used to develop the FEM software in order to
improve possibilities for an engineering approach aimed at solving the problem and
thereby speed up the design process.
Furthermore, the covering impact on the friction coefficient was tested for PVD
bainite coverages of the forming tools. The blank coverings of anticorrosive coverages
such as Zn, Fe-Zn etc. were tested as well. What is more, there were tests done concerning
the control of the press slide speed at the servo presses with the additional electrical engine
that allows for precise speed control. Additionally, the Docol 1400 steel (high mechanical
properties, especially forming ability and the springback value) was tested. This steel is an
alternative material for the parts produced by the hot forming process that are able to reach
a strength of 2000 MPa.
All of this research allowed for the precise correction of the forming tool so the aim of
the work which was the properly reproduced shape geometry of the designed vehicle’s part
- was accomplished. All the results were presented as the comparison of the FEM analyses
with their experimental verification and subsequently put intowork.

142
8.2. The aim considering an industry

Because of the fact that there is not an industrial attitude of solving the
springback problem in the literature, the work contains broadly defined topics related to
this problem. In 2010, when the author conducted the research on steel springback, tool
shops focused on different blank defects such as: cracks, wrinkles, min. sheet thickness
or slip lines, rather than the springback phenomenon and its compensation. It was
caused by using mild steels, simpler part geometry and higher manufacturing tolerances.
Presently, the trend has changed and it is surely going to change to more restrictive
requirements for car producers. The work has become an important source of
knowledge about the springback effect that engineers need to consider (not only in the
automotive industry).
New materials such as UHSS have great structural potential but unfortunately
they require using new technologies in the rework process of these materials (Fig. 8.1).
More and more intensive economic competition and an increase in expectations of the
final product have forced increased efficiency in stamping departments. Quality
expectations of customers refer mostly to: appearance, security, comfort and low fuel
consumption. This is related to visible elements for example outer element coverage,
pillars that also have an aesthetic function which cannot accept any geometric
deviations. Using curvilinear surfaces on components is meant to get smooth reflections
so the formed surface does not contain surface defects. In regard to tightening geometric
tolerance requirements, compensation of springback effect should meet high customer
expectations. Therefore, improved methods for reducing springback are being sought.
The subject of the discussion was reduction of the springback phenomenon
during forming process of a vehicle’s construction parts in order to meet the
requirements of a client. Because of the fact that this issue especially concerns UHSS
and AHSS steels, the author presented their basic mechanical properties and main
features describing forming ability of a particular steel. Furthermore, the springback
phenomenon was described in order to properly estimate its value, character and
location. It allowed for minimization of the springback effect by changing the forming
method, steel supplier, adding additional stiffenings and reducing bending radius. All of
these actions reduced the springback effect by minimizing or highly simplifying the
process of springback compensation. The author faced many problems in the range of
springback inter alia the torsional springback that is significantly hard to compensate

143
for. It required reducing torsion to a minimum and then replacing it with a transverse
value of compensation, which was done and described in Chapter 6.

2% 2% front chassis
4%
steel sheet
2% 9%
48%
2% steel sheet + steel
tubes
11%
Alu: casting

Alu: casting +
20% extrusion
Alu: casting +
sheet

2%
2% rear chassis
2% steel sheet + steel tubes
2% 4%
2% steel sheet
14% 39%
Alu: casting
33%
Alu: casting + tubes

Alu: casting + sheet +

% of analyzed designs tube + extrusion

Fig. 8.1. Front and rear chassis in % of analyzed designs using materials.

The presented problems were numerically analysed by the FEM method and
experimentally verified afterwards on the test stand for a few complicated stamping
tools. Therefore, it can be assumed that the aim of the work was reached and the
presented results confirm the assumptions of the numerical analysis more than 85%.
The main problem that the author faced was determining and assuming an appropriate
friction coefficient, μ, and the Poisson ratio, υ, that are not easy to determine in
industrial conditions and which are highly important during springback calculations.
The aim of the work was reached by the appropriate compensation of the tool for the B-
pillar, the roof hardening and the front bumper of a car. As was experimentally verified
after the forming process, the compensated parts fulfil the assumed technical
requirements and expectations of a constructor. It is worth adding that only the accurate
determination of the material’s mechanical parameters and the appropriately designed
numerical analysis allowed for the consideration of the compensation in the forming
tool in a short time and without generating extra costs. The aim of the work was
accomplished in terms of research and experiments. In 2015, monthly level of part
144
complaints for a whole factory reached a value presented in Fig. 8.2 for particular
defects. Most amount of complained parts occurred as a result of elastic deformation,
which shows a high need of developing the springback issue and its reduction. The
author has presented a description of the springback problem and techniques of its
reduction in the way of practical use, even in extremely difficult conditions. The
presented work will be a basis for the future research.

16

14

12

10
Amount

Fig. 8.2. Amount of complaints depend on type of defects.

145
 9. Direction of future research
In the 21st century, industrial tools development in North America with major
experience in production capacity and market competition allowed for scientific attitude
towards computer-aided systems. Expensive original tooling of OEM producers and first-
tier metal parts suppliers began appearing at low-cost producers all over the world. This
negatively impacted the traditional tool industry in developing countries. These changes
affected many tools and tool shops that dropped out of the market in the last few years of
the new millennium. The producers in the tool market had to adapt new practices such as
Lean Manufacturing Process.
These practices were presented as a result of reduction of production costs. One of the
innovative practices of CAE forming was the opportunity to verify the product and its die-
face development in the early stage. Appropriate use of this technology helped to improve
cost estimations, reduce design time and increase quality and rapidity of reaction for
meeting market demands.
The CAE Technology and the numeric FEM method replaced the previous traditional
trial-and-error method. What is more, toolmakers employed as high class specialists more
and more often use CAE tools with great accuracy and effectiveness. The design of
automobile parts often requires esthetic design features, which may be especially noticed
in the last 10 years. Furthermore, modern vehicles have aerodynamic shapes of outer
coatings surfaces with big curves, small radiuses and occasionally “retro” shapes, which
determines round deep-drawing outer surfaces of vehicles (Fig. 9.1). In new designs, it is
common to emphasise the production of futuristic car concepts, which is related to
difficulty of producing forming tools.

146
 Fig. 9.1 Deep-drawing surfaces of vehicles.

This trend simultaneously brings a lot of uncertainties and progress for the metal
industry. The other aspect connected to development of the green technologies market is
reducing fuel consumption with the highest safety level. This trend imposed creating and
producing new materials starting from structural composites through new aluminium alloys
and ending with AHSS and UHSS steels for steel suppliers and OEMs. Using dual phase
steels, complex-phase steels, transformation-inducted plasticity steels and even ultrahigh-
strength steels enhanced highly over the last 10 years. These materials are considerably
stronger than conventional mild steels for crashworthiness. Using higher-strength steels
brings better possibilities but also serious problems related to formability (cracking,
wrinkling) and dimension accuracy caused by the springback effect.

9.1. Development view

More and more restrictive ecological requirements, economic competition on the

world-wide market and growing safety and comfort standards impose everlasting
technological changes in producing modern cars.

147
 Fig. 9.2 Typical hotformed component.

The rigorous norms aimed at protecting the Earth’s atmosphere influence the
construction of engines and power transmission systems in order to improve the
economics of fuel consumption and minimization of forming resistances. It also has an
impact on refining self-supporting car bodies in terms of reducing their weight and
improving aerodynamics. A 100 kilograms lighter car emits an average of 4,7 grams CO2
less at every travelled kilometer.
The issue of vehicle passenger safety is also a highly important factor. It requires the
stability of the passenger compartment to be as high as possible. The high stability means
high crush and torsion resistances and also accurately defined strain characteristic curve
of body framework in areas of controlled cold work.
The condition of technical progress in car body structures is the proper development
of material engineering, which involves providing materials to car producers. The
materials need to be characterized by required mechanical strength, operating life, simple
forming, market availability, relatively low costs and possibility of later recycling or
utilization.
Despite common expectations, steel still remains the main material used for building
car bodies. However, more and more often, new sorts and technologies of steel treatment
are being used. In the material structure of the standard car body, the contribution of deep

148
drawing steel sheets (the name derives from its high plasticity) decreased to barely 30%,
whereas they were dominating on the market in the beginning of the 21st century. In
present maximally stiff and light constructions, steels are absolutely more useful.

9.1.1. Current automotive trend hot forming production

Parts made by this method have high application in car industry. Using the elements
made by hot-forming allows for a reduction in car mass of over 12- % and ensures higher
safety during collision because of its strength.

Fig. 9.3. External heating (indirect).

The hot formed parts have very high Rm - 1000 MPa (Fig. 9.2). In hot forming,
incandescent semi-finished products are heated to about 950 and then are put into the
forming tool (Fig. 9.3, Fig. 9.4, Fig. 9.5, Fig. 9.7). They are formed during the stamping
process and subsequently it need to be cooled fast.

Fig. 9.4. Hot forming bumper part.

149
Thanks to the used technological process, it is possible to obtain following parameters:
 Reduced springback effect of steels;
 The perfect flow of grains and microstructure that improves mechanical properties
of the parts;
 Relatively high production factors with minimal defects.

Fig. 9.5. Hot formed part after laser trimming.

The required shape is obtained by relatively low stamping force and minimal level of
strains in the parts.
Hot forming technology is more and more often used by the producers of the
construction frame elements that belong to the safety frame of a vehicle. There are
construction elements of the body such as: pillars, bumpers, hardening of the roof,
crossmember floors, doors, side impact beams, door framings, inner rocker panels, front
cab walls etc. Most of the car body elements are made of steel with a strength of 250-
1400 MPa. However, hot formed steels are able to achieve a strength of 1500-2000 MPa.
Vehicle producers want to reduce the body structure weight and to improve safety of
passengers during collisions, so they use high strength steels with low thickness. It causes
the reduction of the car weight, which lowers the CO2 emissions during vehicle use. The
limits of CO2 emission for new cars are increasingly low. They are being mandated for
for all cars by EU legislation. Every car producer will be forced to pay high extra
charges for every produced car if the limits are not fulfilled.

150
 Fig. 9.6. Alloyed Al-Si-Fe vs heat treatment transfer time 10 s.

The parts after hot forming with simultaneous hardening reach the strength of 1500-
2000 MPa. The steels with such a high strength cannot always be cold formed because
they usually have a low elongation factor. What is more, these steels crack during
forming processes so reaching the required shape is not possible in that case. In order to
obtain appropriate designed geometry from the high strength materials, the cut blanks are
being heated up to a temperature of 940° and after putting them in the tool and forming
them, they are immediately being cooled to the temperature of 140-150° (Fig. 9.6). In
order to get the high strength effect, the speed of cool down has to be at least 30° per
second in every area of the part but the coolant cannot directly be in contact with the part.

Fig. 9.7. Hot forming die with upper and lower die (opening position).

151
 This may be obtained thanks to the special design of the forming tools with the web
of cooling channels that are evenly located towards the part shape. The channels also
have to be as close as possible to the forming surfaces where the coolant continuously
flows under demanded pressure.

Fig. 9.8. Thermogram of the blank before transformation.

The heating process of the tools and holding time in a furnace has to be strictly
controlled with the use of special sensors and thermal imaging cameras that register the
real temperature of the parts just before the tool is closed and after it is opened (Fig. 9.8).
It is highly necessary because only conscientious compliance with the proper conditions
of the temperatures and the particular phase times guarantees appropriate strength
parameters and obtaining the austenitic structure. Every temperature deviation registered
by the cameras results in scrapping the part. In order to protect the surfaces of the
elements from oxidation caused by the high temperature the sheets are covered with the
special Al-Si coating. There can also be used the protective atmosphere composed of
inert gases in the furnace. The tools design needs to guarantee the proper shape of the
element after cooling down. It means that it needs to predict the elongation and thermal
shrinkage factors during the forming process of the hot blank and then during its fast cool
down (min 1%). Designing these kinds of tools requires the appropriate software for the
forming process simulation and also for simulating the heat removal process during the
152
cool down. On the basis of simulations, the tools, paths and dimensions of the particular
cooling channel webs located in the tool are designed in engineering CAD programs such
as Catia V5 (Fig. 9.9).

Fig. 9.9. Designed in CATIA hot forming die.

For more effective usage of the hot forming line there are designed two-section and
even four-section tools. It means that in one stroke there is a possibility to get even four
different elements from one tool. These kinds of tools require using a special sort of steel
with the cold formed and hot formed high strength and with very high heat conductivity
and high dimensional stability at such large temperature deviations during the forming
process. High class CNC machines such as: 5-axial milling and drilling machines (with
hardness of punches and dies – 60 HRC) and machines drilling deep holes in at least 5
axes (hole diameters of 6-14 mm and depths of 350-500 mm) are necessary to produce
the hot forming tools. Before the produced tool moved mass production it requires many
tests of forming and cooling. What is more, leak tightness of the cooling system has to be
tested and quality of the parts is controlled with the use of the testing system that controls
shape, hardness and strength in various sections.
The tailored blank elements are also used in modern car bodies (Fig. 9.10). They locally
stiffen loaded constructional nodes or reduce overall vehicle mass by using low thickness
sheets in low-loaded parts e.g. in some areas of floor panels. At present, the method of
joining various steel materials is commonly used in variable constructional hollow

153
 sections and also during forming in the areas of controlled cold working. They are
composed of welded (by a laser) various sheet sorts, e.g. the deep drawing high- and
ultra-strength sheets. The deep drawing sheet absorbs the striking energy during
a collision. Then, high strength material is deformed if the deep drawing sheet can not
dissipate the entire energy strike. The stiffest element can be at most slightly deformed
because it needs to keep stiffness of the vehicle body (the space of survive for a driver
and passengers). The production of the tailored blanks and tailored tubes is handled by
plants having processing lines appropriate for methods of steel treatment that have the
highest strength and minimal plasticity. One such specialized method is hot stamping
based on hot forming elements made of sheets that cannot be cold formed.

Fig. 9.10. Tailored blank element of the B-pillar part for a small truck.

The methods make it possible to produce light and resistant parts that have reduced
springback effect caused by the production technology. It allows one to reduce time and
production costs without losing mechanical properties of construction elements at the same
time. Although, the hot forming method has been used for over a dozen years it is
a developing future for steels with high mechanical properties. Moreover, it is an
alternative for standard forming technologies.

154
 9.1.2. Development of technologies for an aluminum alloy

Currently, the main objective of the project is developing technologies for producing
B-pillars out of 7xxx series aluminum alloy, which will have mechanical properties similar
to those of currently used pillars made of boron-manganese steel. The use of lightweight
high-strength aluminum alloys in the automotive industry will enable significant reduction
of vehicle weights and consequently, reduction of fuel consumption.

Due to low deformability at ambient temperature, 7xxx series alloys are of limited use
in the automotive industry. There are ongoing very intensive works on warm forming of
that alloy in the aged condition. Unfortunately, it appears that improvement of formability
at a temperature of approx. 230-250°C results in considerable reduction of strength
properties of the product and hence there is no worldwide developed technology of making
B-pillars of that alloy. As regards the first technology, the base material is a T6 aged
material to be formed in hot tools (the temperature should not exceed 230°C), the second
technology assumes the base material to be rolled material (much cheaper), then: heated to
the temperature of over 500°C, cooled during forming, and aged.

Nowadays, vehicle bodies are made of thin-walled profiles produced by bending and
stamping, which subsequently are joined by welding, seam welding and spot welding. Due
to the costs and properties, low carbon steel and high strength steels prevailed at the end of
the previous century. Recently, in order to reduce fuel consumption by reducing the weight
of vehicles, new materials with a high strength to weight ratio are sought (reducing the
weight of car by 100kg allows for reduction of fuel consumption by 0.3-0.6 l/100 km,
depending on the vehicle size). The use of lightweight, advanced materials in the
automotive industry allows for significant reduction of the weight of vehicles. In that
regard, aluminum alloys are the most promising materials. Aluminum (2.7 g/cm³) is three
times lighter than steel (7.86 g/cm³). With the exception of elastic module, parameters
describing characteristic properties of aluminum and its alloys related to their weights have
higher values than for steel. Advantageous properties of aluminum include its high thermal
conductivity, low thermal expansion, good electromagnetic shielding and very high
potential for vibration damping. Those last two features will become increasingly
important in the future. Also, recyclability is important, easier for metals than for
competing materials which are plastics. One of the most critical parts in the body of each
passenger vehicle is the B-pillar, due to the fact there is very little space in that part of the

155
bodywork where the collision energy absorbing elements and stiffeners may be applied.
The B-pillar acts as part of the safety cage, thus should have the highest possible strength
properties. Currently, about 89 million cars are produced annually in the world, of which
about 20 million in Europe (ACEA data - European Automobile Manufacturers'
Association). Subsequent legislation imposed on manufacturers has resulted in the need to
reduce emissions of greenhouse gas and noise.

Car manufacturers, therefore, have to seek cost savings mainly in bodywork weight.
According to the Regulation (EC) no 443/2009 of the European Parliament and of the
Council, carbon dioxide emission per 1 km is 120 g for the year 2015 and is to be lowered
to 95 g in 2020. The penalty is provided for, in the amount of EUR 95 per g/km CO2.
Consequently, every car manufacturer would have to pay extra 2375 Euro. Due to low
deformability at ambient temperature, aluminum alloys of 7xxx group are of limited use in
automotive industry. There are ongoing very intensive works on high-temperature
deformation of that alloy in the aged condition. Unfortunately, improvement of formability
at about 230-250°C occurs to cause significant reduction of product strength properties. In
the literature there are few research results on formability, structure, properties of final
products depending on the parameters of deformation, so there is also a high demand for
results of such studies. The cognitive effect, including enhancement of the knowledge in
the field of plastic processing technology will be the development of the rational basis for
selecting conditions to effectively shape the structure and mechanical properties of
automotive structural elements of aluminum alloy, in sheet forming processes.

156
 References
A.B. Abdullah; S. M. Sapuan; Z. Samad; N. A. Aziz, A comprehensive Review of
Experimental Approaches Used in the Measurement of Springback, Advances in Natural
and Applied Sciences, 6(2); (2012); 195-205;

Atlas Tool inc., ESI Group, Develops a sounds springback compensation process with the
help of Pam-Stamp 2G software, France, 2011;

Bekar D., Acar E., Ozer F., Guler M. A., Robust Springback Optimization of DP600 Steels
for U-Chanel Forming, London, United Kingdom, World Congress on Engineering, Vol
III, 2011;

Bergstrom, M. Bronnestam, L. Gustafsson, L. Ingvarsson, B. Setter, Sheet Steel

Handbook- Design and fabrication in high strength steel, SSAB Tunnplant AB, Borlange,
Sweden, (1996) 5:12-5:27;

J. C. Choi1, C. Kim, Y. Choi, J. H. Kim, J. H. Park, An Integrated Design and CAPP

System for Deep Drawing or Blanking Products; Int J Adv Manuf Technol (2000) 16, 803-
813;

Chuko W.L., Gould J.E.: Development of Appropriate Resistance Spot Welding Practice
for Transformation-Hardened Steels. Welding Journal, 81 (2002);

Chung-Yeh sa, Computer-Aided Engineering in Blank sheet metal forming, General

Motors, 2006;

J. Danzberg, S. Schmaalze, Parametrization of the tool geometry as prerequesite for the

shape optimization of tool surface, Zurich, Switzerland, 2004;

DeCooman B.C.: Structure-properties relationship In TRIP steels containing carbide-free

bainite. Current Opinion in Solid State and Materials Science 8, 2004;

Dogal 600 i 800P SSAB Swedish Steel, 2010;

157
Eisso H. Atzema, Ardy Duwel, Louisa Elliott, Peter F. Neve,Henk Vegter; “Appreciation
of The Determination of The Forming Limit Curve”;pg -, Numisheet 2002; Jeju Island,
Korea;

ESI Group PAM-STAMP 2G 2011, User's Guide, 100-102 Avenue de Suffren Paris;
France; 2011;

M. Firat, B. Kaftanoglu, O. Eser, Sheet metal forming analysis with any emphasis on the
sprinback deformation, Journal of Material Processing Technology 196 (2008), 135-148;

Frommeyer G., Brux U., Neumann P.: Supra-Ductile and High-Strength Manganese-
TRIP/TWIP Steels for High Energy Absorption Purposes. ISIJ International 43 (2003);

J. Gronostajski, A. Niechajowicz, S. Polak, Springback deformation of drawpieces after

trimming and parting, Umformtechnik Tagungsband, Freiberg,Institut fuer Metallformung
TU Bergakademie (2004), 457-476;

Gould J.E., Workman D.: Fracture Morfologies of Resistance Spot Welds Exhibiting Hold
Time Sensitivity Behavior. Proc. Sheet Metal Welding Conf. VIII, 1998, Detroit (MI),
Paper No 1-1;

S.S. Kang, D.H. Park, Application of computer-aided process planning system for non-
axisymmetric deep drawing products. Journal of Materials Processing Technology 124,
(2002), 36-48;

Ken-ichiro Mori, Proceedings of the 7th International Conference UMIFORM 2001,

Toyohashi, Japan 18-21 June 2001, Simulation of Material Processing: Theory, Methods
and Application p. 3-10;

Kim S.K., Choi J., Kang S.C., Shon I.R., Chin K.G.: The Development of TWIP Steel for
Automotive Application. POSCO Technical Report 10 (2006);

158
Kliber J., Kursa T., Schindler I.: Hot Rolling of steel with TWIP effect. Hutnik 75, nr 8
(2008);

Lehman L.R., Gould J.E., Holmes S.: The Use of Design of Experiments for Investigating
the Resistance Spot Weldability of High-Strength Steel. Proc. Sheet Metal Welding Conf.
VII, 1996, Detroit (MI), Paper No C1;

Leppin, J. Li, D. Daniel, Application of a Method to Correct the Effect of Non-

Proportional Strain Paths on Nakajima Test Based Forming Limit Curves, Numisheet
2008, Interlaken, Switzerland, 217-221;

H. Li, G. S. G. Li Z. Gong, D. Liu, Q. Li; Elsevier; On twist springback in advanced high-

strength steels; Materials and Design; 32; (2011) ;

H. Lundh, P.A. Bustad, B. Carlsson, G. Engberg, L. Gustafsson, R. Lidgren, Sheet metal

forming, SSab Tunnplant, Goteborg, Sweden, (1998) 2:1, 2:2, 4:3, 4:4, 4:10, 4:17, 5:14;

Mangas, A. Arroyo, I. Fernandez, B. Gonzalez; Improvement of springback FEM

simulation by means of an optimization procedure based on an experimental trial, 2006;
Z. Marcinika, J.L. Duncan and S.J. Hu, Mechanics of Sheet Metal Forming, second
edition, 2005, p.82-106;

Materials Property Ultra High Strength Steel Dogal, SSAB Tunnplat AB, Borlange,
Sweden, 2009;

Mehta R.: Twinning Induced Plasticity (TWIP) Steel Improves Car Safety. Materials Word
Magazine 12 (2007);

M. Ninshu, U. Yasuyoshi; Springback prediction by Youshida-Uemori model and

compensation of tool surface using JStamp; Interlaken; Switzerland, (2008);

Radakovic D.J., Tumuluru M.: Predicting Resistance Spot Weld Failure Modes in Shear
Tension Tests of Advanced High-Strength Automotive Steels. Welding Journal, 87 (2008);

159
Rathbun R.W., Matlock D.K., Speer J.G.: Fatigue Behavior of Spot Welded High-Strength
Sheet Steels. Welding Journal, 82 (2003);

Recommended Practices for Test Methods for Evaluating the Resistance Welding Behavior
of Automotive Sheet Steel Materials. AWS/SAE D8.9M Document, 2002 AWS Miami
(FLA);

D. Roylance, STRESS-STRAIN CURVES, Cambridge, August 2001;

N. Seki, Successfully controls springback in thin Ultra-High Strength Steel stamping parts
using PAM-STAMP 2G, TOA Industries, 2011;

S.L Semiatin, ASM Handbook, volume 14B, Metalworking: Sheet Forming, 2006;

X. Siguang, Z. Kunmin, T. Lanker, J. Zhang, C. T. Wang; Springback prediction

compensation and correlation for automotive stamping; (2000);

D. A. Smith, Die Design, 1990, Michigan, 6-15, 6-16;

I. Suchy, Handbook of Die Design, Mc Graw- Hill, 1997, New York, 8-28, 8-33;

Thyssen Krupp Steel, High- Strength steels for shallow formed part, Duisburg, Germany,
2009;

Tumuluru M.D.: Resistance Spot Welding of Coated High- Strength Dual Phase Steels.
Welding Journal, 85 (2006);

Ueji R., Tsuchida N., Terada D., Tsuji N., Tanaka Y., Takemura A.: Tensile Properties and
Twinning Behavior of High Manganese Austenitic Steel with Fine-Grained Structure.
Scripta Materialia 59 (2008);

Ultra-High-Strength Steels, Enhancing vehicle safety and improving fuel efficiency, Ford
7/2011;

160
User’s Guide PanelShop iCapp Automated Springback Compensation version 11.0, (2011),
Zurrich, Switzerland;

H.-P. Vogt, Modern lightweight design with magnesium flat products, Magnesium
Flachprodukte GmbH, Freiberg, Germany, 2011;

R. H. Wagoner, J.-L. Chenot, Metal Forming Analysis, 2011, p. 177-198;

Robert H. Wagoner, Jean-Loup Chenot, Fundamentals of Metal Forming, first edition,

1996, p. 209;

R.H. Wagoner, J.F. Wang, and M. Li, The Ohio State University, ASM Handbook,
Volume 14B: Metalworking: Sheet Forming, 2006;

H. Wie, W. Chen, L. Gao, Sprinback Investigation on Sheet Metal Incremental Formed

Parts, 2011.

161
 Summary
Cold forming, especially for steels of increased mechanical properties, encounters
problem with a springback. Most of the tools require compensation of that effect,
however it is not always feasible in a conventional way. In some special cases such as
twisting springback, compensation of a tool remains an ineffective method of solving
the problem.
Elements of automotive production are often fabricated from materials the types
Ultra High Strength Steel (UHSS) or High Strength Steel (HSS). Applications using
these kinds of materials considerably reduce construction weight. Nevertheless, their
usage results in the appearance of a springback effect. Springback value depends mainly
on the material used as well as, in the part, the geometry of part forming, thus, in
extreme cases the deviation value from the reference part might in some cases, reach an
unacceptably. Prediction of springback effect by the trial-and-error method and
subsequent correction of deviation is difficult and labor-intensive. Numerical
compensation methods make it possible to receive precise results in a short time. This
method results in a huge economic advantage, eliminating unnecessary milling during
production of a die and facilitates its production on a just-in-time basis according to
customers’ expectations and forecasts for vehicle demand.
In this work there is presented an innovative compensation method for both forming and
trimming die for the construction of vehicle parts manufactured on a transfer press. This
method allows one to optimize the accuracy of compensation, consequently decreasing
springback in a more exact way than methods which are used currently. Previous
methods took into account only the influence of trimming on springback without
generating compensated surfaces for the trimming die and also for the next forming
operation. These problems generate additional costs during die production which could
be avoided by using multi- operation compensation.
The work is exploring the knowledge in the range of physical phenomena
determining the springback in the stamped car body elements. Particular attention will
be devoted to the twisting springback. The current trend indicates that 60% of the rear
bumper markets are using UHSS steel material and a majority of them are hot forming
to keep the mass low as compared to conventional roll formed design. For the
springback analysis there were chosen adequate examples that shows high springback
tendency. The car structure elements such as: pillars, hardening of the roof and bumpers

162
are important parts that need to keep the highest requirements. Therefore, those parts
have not initially met the construction assumptions of the client so they were
numerically analysed by the FEM method and verified in an experimental way.

163
 Streszczenie

Geneza pracy

Procesy tłoczenia są dziś powszechnie wykorzystywane w masowej produkcji

elementów wchodzących w skład konstrukcji pojazdów. Decyduje o tym zarówno
możliwość uzyskania w tych procesach bardzo wąskich przedziałów tolerancji
wymiarowo – kształtowej, powtarzalności wytłoczek, dobrego stanu ich powierzchni
jak i dużej wydajności oraz ekonomiczności. Równolegle podejmowane są też na
szeroką skalę próby zastąpienia stalowych elementów konstrukcji samochodów przez
stopy aluminium, stopy magnezu oraz kompozyty.
Można przytoczyć, iż jeszcze w 2002 roku konstrukcje samochodowe zawierały
średnio 0,5% stopów magnezu oraz 8,8% stopów aluminium, ale już w 2015 roku
wartości te wzrosły odpowiednio do 2.0 % oraz 12.0 % (Vogt 2012).
W literaturze przedmiotu często przedstawiane są też wyniki prób masowej produkcji
elementów konstrukcji jako integralnych części kompozytowych, których wytrzymałość
jest na porównywalnym poziomie jak dla stali o wysokich parametrach mechanicznych
Advanced High Strength Steel (AHSS).
Jakkolwiek kompozyty charakteryzują się niską masą własną - nawet do 30% mniejszą -
niż te same elementy stalowe, to jednak zapewnienie wszystkich korzystnych
własności, jakie posiadają elementy tłoczone z tego materiału okazuje się wciąż trudne
do realizacji także ze względów ekonomicznych. Dotyczy to zwłaszcza przejęcia
technologii materiałów kompozytowych z konstrukcji pojazdów wyczynowych do
samochodów masowej produkcji. Pojawia się też tak istotna w XXI wieku kwestia
recyklingu tych materiałów po okresie eksploatacji, co pozostaje ściśle związane
z ochroną środowiska.
Uszkodzenie struktury materiału kompozytowego praktycznie eliminuje bowiem
możliwość jego naprawy, a z pewnością nie jest to tak technologicznie jednoznaczne
i dopracowane jak w przypadku elementów stalowych.

164
Jakościowe kryteria oceny nadwozi samochodowych dotyczą dziś głównie: masy,
estetyki1, bezpieczeństwa, komfortu, oporu aerodynamicznego przy jednoczesnym
niskim zużyciu paliwa (Ford 2011).
Dążenie do osiągnięcia wyższych standardów komfortu i bezpieczeństwa powoduje
ciągły wzrost masy pojazdów (Ford 2011, Seki 2011).
Dlatego też wciąż poszukuje się nowych rozwiązań, które zapewniając optymalną – ze
względów użytkowych - strukturę nośną pojazdu spełnią też wymagania testów
zderzeniowych.
Znaczącym, wręcz przełomowym osiągnięciem było tu skomponowanie w ostatniej
dekadzie wielu nowych gatunków stali o bardzo wysokich własnościach
mechanicznych.
Materiały te umownie podzielono na grupy według wartości granicy wytrzymałości na
zerwanie Rm.
Są to:
 konwencjonalne stale o wysokiej wytrzymałości High Strength Steel 300 <Rm< 600
MPa.
Należą do nich stale: High Strength Low Alloy (HLSA), Bake Hardenable (BH),
CMn (Carbon Manganese) oraz częściowo Dual Phase (DP);
 zaawansowane stale o bardzo wysokiej wytrzymałości UHSS o Rm > 600 MPa,
a sięgające nawet do Rm = 2000 MPa.
Należą do nich stale: Dual Phase, Martensitic Steel (MS) oraz Transformation
Induced Plasticity (TRIP) (Thyssen Krupp 2009).

1
Dotyczy to widocznych elementów nadwozia jak np.: fragmenty karoserii, słupki
boczne - pełniące również funkcję estetyczną, a które nie mogą wykazywać żadnych
odchyłek geometrycznych.
Z kolei stosowanie krzywoliniowych powierzchni na częściach karoserii ma na celu
uzyskanie gładkich refleksów, tak by formowana powierzchnia karoserii nie zawierała
defektów powierzchniowych.

165
Stale te są efektem szerokich badań prowadzonych w czasie ostatniej dekady, które
były realizowane poprzez wspólne projekty ośrodków badawczych i przemysłu
samochodowego na arenie międzynarodowej.
Pozwalają one na uzyskanie wysokiej wytrzymałości materiału Rm przy jednoczesnym
zachowaniu jego dobrej plastyczności A5.

Chociaż zwykle te dwa skojarzone ze sobą parametry nie idą w parze i wzrost jednego
następuje kosztem drugiego, to jednak dla niektórych grup stali udało się zachować
wysokie wartości obu.
Te właśnie konkurencyjne właściwości oraz względy ekonomiczne powodują, że stale
stopowe są nadal materiałem dominującym w motoryzacji.
Jak pokazują prognozy na najbliższe lata udział materiałów z grupy HSS oraz UHSS
osiągnie ok. 70% udziału w konstrukcji nadwozia samochodu.
Jednocześnie z wprowadzaniem stali o wyższych własnościach wytrzymałościowych –
lecz trudniej poddających się plastycznemu kształtowaniu - w przemyśle
motoryzacyjnym obserwuje się stały trend do nadawania wytłoczkom coraz to bardziej
złożonych kształtów, tak aby poprzez ograniczenie liczby elementów tworzących
strukturę pojazdu ograniczyć czas i koszty montażu.
Łączy się to ściśle z wymaganiami dotyczącymi zawężenie przedziałów tolerancji oraz
zapewnienia dobrego stanu powierzchni. Ograniczenia kosztów poszukuje się też na
drodze bardziej efektywnego projektowania tłoczników i wykrojników oraz
wprowadzania nowych – bardziej odpornych na zużycie - materiałów narzędziowych.
Realizacja przedstawionych powyżej postulatów wymaga jednak coraz to lepszego
poznania zjawisk towarzyszących procesom tłoczenia, tak aby na tej podstawie
możliwym stało się rozwiązanie szeregu pojawiających się wciąż nowych problemów
technologicznych.
Jednym z nich jest konieczność wyeliminowania lub co najmniej znaczącego
ograniczenia zjawiska sprężynowania zwrotnego, będącego skutkiem zaniku sprężystej
składowej odkształcenia po zdjęciu nacisku wywieranego przez narzędzie.
Sprężynowanie, jako główne źródło niedokładności wytłoczek, jest funkcją ich
nominalnego kształtu, właściwości wytrzymałościowych materiału wsadowego,
parametrów procesu - w tym warunków tarcia na powierzchni kontaktu materiału
i narzędzia - oraz geometrii narzędzia (promienie matrycy itp.). W skrajnych
przypadkach może ono lokalnie osiągać wartości rzędu kilkunastu milimetrów, co

166
uniemożliwia prowadzenie procesu montażu na zautomatyzowanych
i zrobotyzowanych liniach produkcyjnych.
Warto przy tym zwrócić uwagę, na wielorakość form w jakich sprężynowanie to może
się ujawniać, a to :
- poprzez zmianę wysokości wytłoczki (standardowa dewiacja sprężynowania);
-poprzez poprzeczną zmianę krzywizny wytłoczki;
- poprzez wzdłużną zmianę krzywizny wytłoczki;
- poprzez zmianę krzywizny w przekroju wytłoczki;
- poprzez skręcenie całej części (przypadek szczególny).

Z reguły jednak sprężynowanie ujawnia się w kilku formach jednocześnie, co

dodatkowo komplikuje i tak złożoną odchyłkę wytłoczki.
Wspomniane wyżej wymagania odnoszące się do skrócenia czasu projektowania
i wytwarzania narzędzi tłoczniczych, a także ograniczenia ich kosztu sprawiają, iż
przewidywanie efektu sprężynowania dotychczasową metodą prób i błędów staje się
dziś nie do zaakceptowania.
W tej sytuacji kompensacja sprężynowania wymaga predykcyjnej zmiany geometrii
narzędzia w kolejnych operacjach. Zagadnienia te – niezwykle dziś ważne i aktualne –
stały się impulsem do podjęcia badań przedstawionych w niniejszej rozprawie.

Cel pracy

Celem pracy jest pogłębienie naszej wiedzy o stanie naprężeń i odkształceń

w poszczególnych złożonych operacjach tłoczenia, i na tej podstawie rozwinięcie
techniki kompensowania całego łańcucha operacji, co pozwali w efekcie na
uwzględnienie efektu sprężynowania na wszystkich etapach procesu.
Można zasadnie przewidywać, iż przyniesie to istotne korzyści w procesie
rozwiązywania problemów znajdowania dokładnej powierzchni dla operacji cięcia oraz
dopasowania blachy do narzędzia na poszczególnych operacjach. Zaniedbanie problemu
kompensacji wielooperacyjnej prowadzi tu bowiem z reguły do konieczności
uwzględniania odprężonych powierzchni na poszczególnych stacjach okrawania, co
zawsze jest obarczone pewnym błędem i wpływa ostatecznie na dokładność
wymiarowo – kształtową wytłoczki.

167
Tak więc, wyniki niniejszej pracy mają służyć podniesieniu jakości wytłoczek
stanowiących elementy konstrukcji nadwozia pojazdu.

Analiza stanu wiedzy w zakresie problemów technologicznych występujących

podczas wytwarzania nadwozi pojazdów

W konwencjonalnych metodach projektowania i wytwarzania bazowano

głównie na wiedzy oraz doświadczeniu inżynierów projektantów. Dziś jednak
efektywna metoda projektowania procesów tłoczenia może wykorzystywać coraz
szersze możliwości stwarzane przez oprogramowanie CAD/CAM, a także wyniki
numerycznych symulacji i analiz opartych na metodach optymalizacji (Lundh, Bustad,
Carlsson, Engberg, Gustafsson, Lingren 1998 ).
Obecnie do analizy procesu tłoczenia, a zatem i projektowania narzędzi tłoczniczych,
powszechnie używa się oprogramowania Metody Elementów Skończonych (MES).
W przemyśle motoryzacyjnym, analiza numeryczna wykorzystywana jest jako
praktyczna metoda badania możliwości formowania wytłoczek oraz przewidywania ich
finalnego kształtu, a to w celu wyeliminowania błędów i niedokładności.
Potencjalne niedokładności występujące podczas projektowania narzędzi – skutkujące
wadami wytłoczek - mogą być znacznie lepiej wyeliminowane poprzez wirtualne
środowisko, niż gdyby wykorzystać do tego celu tradycyjną metodę prób i błędów.
Rozbudowane i zaawansowane oprogramowanie pozwala też na skrócenie czasu
weryfikacji słuszności podjętych korekt.
Aby spełnić aktualne żądania odbiorców powinno się stosować nie tylko najnowsze
metody projektowania, lecz także - kompleksowo - wytwarzania i planowania
produkcji.
Jednym z najważniejszych etapów projektowania technologii narzędzi, jest jego
planowanie w celu skrócenia czasu uruchomienia procesu wytwarzania.
Dla procesu tłoczenia blach etap ten jest nie tylko bardzo ważny, ale i bardzo trudny
(Kang et al.,2002). Przewidywanie liczby, typów operacji tłoczenia, uwzględnienie
czynników ograniczających takich jak: fałdowanie, pękanie oraz odchyłek związanych
ze sprężynowaniem blach, okazuje się wciąż bardzo trudne (Kang, Park 2002 , Choi1,
Kim 2000, Danzberg, Schmaalze 2007).

168
Dlatego też począwszy od drugiej połowy lat 90-tych prowadzone były intensywne
prace nad powiązaniem w sposób parametryczny projektowanych powierzchni
wytłoczek i narzędzi tłoczących (Smith,1990, Suchy,1997).
Pierwszym celem było rozwinięcie parametrycznych modeli opartych o wprowadzenie
dodatkowej powierzchni ściśle korespondującej z powierzchniami biorącymi udział
w procesie.
Kolejnym krokiem było stworzenie uniwersalnego rozwiązania dla różnych projektów
poprzez łączenie znanych cech.
Rozwijano oprogramowanie narzędzi takich jak: METHOPLAN (iCapp, Zurich),
Viking (Inpro, Berlin) i DIEDESIGNER (Autoform, Zurich) umożliwiających
kompleksowe i parametryczne projektowanie całych zespołów konstrukcji nadwozia.
Projektowanie parametryczne wymusza łączenie oprogramowania wspomagającego
konstruowanie narzędzia tłoczącego CAD i metody MES w celu zaprojektowania
poprawnej powierzchni, a w efekcie również i narzędzia tłoczącego (Danzberg,
Schmaalze 2007).
Projektowanie konstrukcji samochodów oparte na parametryczności pozwala na szybką
modyfikację i wdrożenie niezbędnych zmian konstrukcyjnych. Coraz częściej sięga się
do optymalizacji całej geometrii elementów, które odpowiadają więcej niż jednemu
cyklowi formowania, tzn. tworzą zespół, w skład, którego wchodzą komponenty
wytwarzane w różnych technologiach tłoczenia. Reprezentatywnym przykładem może
być tu połączenie części wykonanych podczas formowania na zimno oraz części
wykonanych metodą formowania na gorąco (hot forming).
Całość podzespołu, jako złożenie musi przy tym odpowiadać założeniom, które
postawiono na początku procesu projektowania.
Dlatego też optymalizacja odpowiada dziś raczej kompleksowemu łańcuchowi etapów
formowania aniżeli pojedynczej operacji.
Jak już wspomniano rozwój komputerowych technik wspomagania projektowania
sprawia, iż kompensowanie narzędzia ze względu na sprężynowanie zwrotne wytłoczki
dokonuje się dziś przy wykorzystaniu analizy bazującej na metodach MES.
Są to pakiety: Pam-Stamp, Autoform, Ls-Dyna, Outifo, Mashal i inne mniej
rozpowszechnione w praktyce przemysłowej.

169
Obliczenia te wymagają dużej dokładności, dlatego koniecznym jest uwzględnienie
ścisłych danych materiałowych 2 2, a także stosowanie mało - wymiarowych elementów
skończonych. Ograniczenia dotyczą też wielkości przyrostów czasowych.
Wielkość elementów pozostaje przy tym zależna od relacji pomiędzy najmniejszym
promieniem gięcia wytłoczki, a grubością blachy.
Numeryczne obliczenia formowania sprowadzają się do opisania powierzchni
elementami, które bazują na węzłach, a których ilość jest zadawana w taki sposób, aby
zapewnić założoną dokładność obliczeń.
Ważnym czynnikiem jest też zastosowanie odpowiednich elementów podczas analizy
procesu formowania. Zwykle programy MES dedykowane dla tłocznictwa używają
powierzchniowych elementów, które posiadają bardziej rozbudowane algorytmy
obliczeń oraz - co najważniejsze - skracają czas ich realizacji. Zastosowanie elementów
hybrydowych
TTS (Through Thickness Stress) jest natomiast uzasadnione dla gięcia ze względu na
specyfikę procesu.
Przedstawione powyżej uwarunkowania sprawiły, że we wcześniejszej literaturze
przedmiotu (do około 2001 r.) zwracano głównie uwagę na problemy związane
z niedostateczną dokładnością modelowania sprężynowania. Głównym powodem
takich opinii była ograniczona ówcześnie moc jednostek obliczeniowych, wymuszająca
stosowanie relatywnie małej – nawet w przypadku skomplikowanej geometrii wytłoczki
- liczby elementów i węzłów, co obok błędu w wartości tak obliczonego sprężynowania
skutkowało długimi czasami obliczeń. Tym samym potencjalne możliwości analizy
numerycznej były redukowane przez poziom rozwoju zarówno hardwerowej jak
i softwerowej techniki komputerowej. Przełomem stał się tu rozwój 64-bitowych wersji
oprogramowania systemu (2005 r.), który zaowocował wyższą wydajnością CPU
(Central Processing Unit), - tym samym skracając czas obliczeń.
Przedstawiając powyższe warto też wskazać, iż moduł oprogramowania Pam-Stamp
2G - bazujący na metodzie opracowanej w 2005 roku –został w kolejnych latach

2
Model materiałowy jest ważny do tego stopnia, że powinno się uwzględniać dane
materiałowe konkretnego dostawcy już podczas obliczeń sprężynowania.

170
rozwinięty tak, że w 2011 r. doprowadziło to do rozszerzenia podstawowego narzędzia
optymalizacji procesu kompensacji narzędzia.
Co nie mniej istotne, grupa ESI przy współpracy z firmą Atlas Tool dokonała w
ostatnich latach eksperymentalnej weryfikacji wspomnianych wyżej metod
obliczeniowych. Przeprowadzone próby wykonano na kilkunastu wytłoczkach w tym
między innymi na głęboko tłoczonej części słupka A karoserii wykonanej ze stali
UHSS.
W rezultacie stwierdzono, że użycie kompleksowej optymalizacji procesu pozwala na
otrzymanie kompensacji na poziomie przynajmniej 90% zgodności z częścią (ESI
2011).

Metodyka rozwiązania problemu sprężynowania poprzez kompensację narzędzia

Idea korekcji kształtu narzędzi wywodzi się z klasycznej kompensacji

sprężynowania stosowanej w warunkach technologicznych i polega na wprowadzeniu
dodatkowego przegięcia lub przetłoczenia blachy podczas jej formowania w celu
redukcji efektu sprężynowania (Ken-ichiro, 2001)
O ile dla prostego gięcia problem ten sprowadza się z reguły jedynie do dodatkowego
przegięcia blachy, to w przypadku bardziej złożonych procesów tłoczenia
i skomplikowanych kształtów – tak jak ma to miejsce w większości wytłoczek
samochodowych - działania w tym zakresie muszą być oparte o szczegółową analizę
problemu (Marciniak et all.,2005).
Użycie narzędzi komputerowego wspomagania projektowania oraz analizy procesu
tłoczenia pomaga osiągnąć ten cel w sposób przemyślany, a przede wszystkim
w realnym czasie.
Nieliniowe zachowanie materiału podczas procesu jego formowania zależne jest od
panującego w jego objętości stanu naprężeń i odkształceń oraz historii odkształcenia.
Można przy tym wykazać, iż zachodząca w procesie zmienność nacisków i sił tarcia na
powierzchni kontaktu materiału i narzędzia wywiera istotny wpływ na końcowy kształt
wytłoczki. W istocie mamy tu bowiem do czynienia z specyficzną postacią interakcji
pomiędzy materiałem, a narzędziem.
Ewoluowanie metod kompensacji bazujących na metodach kolejnych przybliżeń
z możliwością wskazania wektora zmian oraz liczby kompensowanych elementów,
pozwala na dokładne znalezienie odprężonej powierzchni. Algorytm korekcji narzędzia

171
z uwzględnieniem odkształceń w zakresie sprężystym, zakłada iteracyjną procedurę
kompensacji efektu sprężynowania.
Jej rezultaty wyrażane są zwykle przez wartości średniego sprężynowania (δE)
formowanych części. Sprężynowanie ujawnia się po odciążeniu wytłoczki przez siłę
nacisku prasy lub też po uwolnieniu naprężeń resztkowych podczas operacji okrawania.
Znaczny wpływ na sprężynowanie ma również sposób przebiegu procesu formowania
oraz planowanie przebiegu umocnienia (Wie, Chen 2011).
Najogólniej ujmując, cele kompensacji sprowadzają się do:
a) identyfikacji kształtu narzędzia, podczas której dopasowuje się jego powierzchnię
celem zminimalizowania odchyłki od geometrii referencyjnej w taki sposób, iż
początkowa geometria narzędzia będzie modyfikowana (przeginana).
b) kompensacji prowadzonej w oparciu o modelowanie numeryczne, tak aby
ograniczyć bądź wyeliminować konwencjonalne sposoby modyfikacji np. metodę prób
i błędów.
Należy tu wykonać:
 przedstawienie obszaru kompensacji za pomocą siatki narzędzia;
 zapisanie powierzchni w formatach CAD (np. IGS);
 dokonanie opisu powierzchni w taki sposób, aby zawierał on niejako historię
wszystkich powierzchni podczas kolejnych iteracji.
c) transformacji wyników kompensacji poprzez:
 eksport siatki ze zmodyfikowaną powierzchnią matrycy;
 wykorzystanie dostępnych formatów eksportu (Ansys, Asci, Ideas, Nastran).

Możliwe jest też otrzymanie powierzchni CAD w postaci siatki bazującej na metodzie
inżynierii odwrotnej , ale najczęściej skutkuje to otrzymaniem niskiej jakości
powierzchni.
Rozróżnia się przy tym kompensację narzędzia oraz kompensację wytłoczki, które są
ściśle ze sobą powiązane, chodź różnią się wartościami. Jednak w warunkach
przemysłowych zwykle poszukuje się skompensowanej powierzchni narzędzia
tłoczącego, której znajomość jest konieczna podczas procesu jego projektowania.
Oznacza to modyfikację kształtu powierzchni narzędzia podczas projektowania procesu
z koniecznością optymalizacji tego efektu.

172
Optymalizacja może zakończyć się powodzeniem tylko pod warunkiem, jeśli poddany
analizie zostanie cały proces kształtowania, powiązany w łańcuch korelacji pomiędzy
następującymi po sobie kolejno operacjami formowania, cięcia, gięcia i rozkrawania.
Najczęściej jednak sprężynowanie ujawnia się po operacjach cięcia i rozkrawania.
Istotnym ograniczeniem zastosowania metod korekcji kształtu narzędzi są przypadki
wytłoczek
o pionowych ściankach, ponieważ dla takich wytłoczek po korekcji otrzymujemy
narzędzia
o kątach ścianek mniejszych niż 90°. Wymaga to zastosowania szczególnych rozwiązań
w konstrukcji narzędzia w celu zapewnienia ruchów poszczególnych jego elementów.
Podobne ograniczenia występują dla wytłoczek charakteryzujących się dużym obszarem
płaskiej powierzchni.
Trzeba przy tym zwrócić uwagę, iż oprogramowanie Metody Elementów Skończonych,
pozwalające na optymalizację pełnego łańcucha procesu tłoczenia okazuje się
w praktyce niewystarczające. Dzieje się tak dlatego, iż programy tego typu nie generują
powierzchni
w postaci CAD, lecz spójną siatkę. Jest to wystarczające dla celów analizy teoretycznej,
lecz nie sprawdza się w warunkach rzeczywistych procesów przemysłowych.
Zbudowanie tłocznika wymaga obróbki poszczególnych jego części. Dlatego też
w następnym kroku konieczne jest konwertowanie siatki w postaci różnych formatów -
najczęściej są to pliki: Nastran i Ansys - na bardziej uniwersalne formaty CAD
(IGS).
Parametryzowanie danych powierzchni opisujących kształt części narzędzia, wymaga
wprowadzenia danych początkowych zawierających oryginalną kompensowaną siatkę
oraz danych CAD narzędzia. W ten sposób zawarte są dwa kody siatki niejednorodnego
przemieszczenia obszaru powierzchni CAD. Oznacza to, iż każda powierzchnia jest
„dociągana” do nowego kształtu, który zwykle jest nie większy niż ta sama
powierzchnia przed kompensacją. Dlatego też główną pętlą dla każdej powierzchni jest
sprawdzenie dokładności lokalnych węzłów, tak aż założona tolerancja zostanie
osiągnięta wszędzie za wyjątkiem definicji powierzchni. Proces ten generuje ścieżkę
z maksymalną ciągłością określoną przez podwójną tolerancję. Następuje dopasowanie
węzłów do modelu części w przestrzeni 3D, tak by powierzchnia była rozciągnięta na
siatce. Proces ten jest realizowany przez oprogramowanie Panel Shop (i Capp)
(Danzberg, Schmaalze 2007, Capp 2011).

173
Program badań własnych i jego realizacja

Mimo swej dużej różnorodności wynikającej tak z typu struktury pojazdu jak
i pełnionej w niej przez daną wytłoczkę funkcji, po analizie udało się wytypować grupy
wytłoczek sprawiających w procesie tłoczenia zbliżone do siebie charakterem problemy
technologiczne, w tym szczególnie związane z postacią zjawiska sprężynowania
zwrotnego.
Reprezentatywnym przykładem pierwszej z wyselekcjonowanych grup była wytłoczka
fragmentu części dachowej pojazdu. Z tej racji istotnym do spełnienia kryterium jakości
było zapewnienie jej wymaganej sztywności niezbędnej do ochrony pasażerów
w przypadku kolizji. Spełnienie tego kryterium wymagało zarówno zastosowania
gatunku stali UHSS o dużej wytrzymałości jak i nadania wytłoczce skomplikowanego
kształtu. Konsekwencją tego było pojawienie się w procesie tłoczenia szeregu zjawisk
negatywnie wpływających na jakość wytłoczki. Dotyczyło to takich wad wytłoczki jak:
nadmierne lokalne fałdowanie, pękanie oraz stosunkowo duże sprężynowanie
szczególnie w kierunku grubości blachy.
Wady te należało wyeliminować przez zaprojektowanie odpowiedniego kształtu
narzędzia.
Proces formowania powierzchni części podczas odkształcenia materiału wymaga
przyłożenia obciążeń zginających oraz obciążeń rozciągających (względnie
ściskających).
W efekcie otrzymuje się złożony stan naprężenia, który skutkuje pewnymi trudnościami
przy jego analitycznym ujęciu w postaci zredukowanej hipotezy wytrzymałościowej
(ESI 2011).
W matrycach OP-20 i OP-50 użyto dociskaczy w celu zmniejszenia fałdowania.
Wywierały one siły docisku wynoszące odpowiednio 300kN oraz 700kN.
Siły te zostały dopasowane do istniejącej geometrii progów ciągowych, tak by
w odpowiednim momencie przygotówka została uwolniona od ich działania pozwalając
tym samym na zmniejszenie naprężenia. Duże wartości sił docisku - 300kN oraz 700kN
- a także specyficzny kształt wytłoczki skutkowały znaczną wartością siły nacisku
prasy wynoszącą około 10000 kN.
Prędkość dociskacza ustalono na poziomie 2 m/s, natomiast tłoczenia - jako VBC
(Velocity Balanced Curve) - na maksymalnym poziomie 5 m/s.

174
Do obliczeń użyto oprogramowania PAM-STAMP 2G wersji 2015.1. Pozwoliło to na
wykonanie analizy numerycznej procesu formowania oraz na kompleksową
optymalizację matrycy formującej OP-20 oraz matryc OP-30 i OP-40.
Analiza numeryczna MES została wykonana dla modelu materiałowego Hill 48
z wykorzystaniem metody obliczeń „implicite”. Postępowanie takie jest zgodne
z zaleceniami twórcy oprogramowania i ma na celu zminimalizowanie czasu obliczeń
przy jednoczesnym zachowaniu zadowalającej dokładności obliczeń solvera.
Blacha wsadowa użyta w procesie formowania miała grubość nominalną 1,35 mm
i wymiary 873 x 222 mm. Ze względu na duży wymiar wsadu podczas numerycznych
obliczeń modelu wykorzystano pojedynczą symetrię, co znacznie skróciło czas
obliczeń.
Sprężynowanie liczone było również z uwzględnieniem płaszczyzny symetrii ze
względu na operację cięcia tuż przed kalibracją narzędzia. Wymusiło to również
odebranie stopni swobody w warunkach symetrii, tak by sześć translacji zostało
zablokowanych w sposób jednoznaczny i statycznie wyznaczalny. Użyto elementów
typu powierzchniowego trzy i cztero węzłowych o nieliniowej dystrybucji naprężeń
w kierunku grubości. Elementy typu powierzchniowego oparte były na modelu
Belytschko-Tsay bazującego na jednorodnej, spójnej i zredukowanej integracji (ESI
2011).
Ilość elementów została dobrana względem czasu obliczeń, tak by dokładność
odzwierciedlała rzeczywisty przebieg zjawisk podczas procesu.
Matryca formująca OP-20 zawiera 698267 elementów oraz 173554 węzłów, natomiast
sama przygotówka początkowo zawierała 131245 oraz odpowiednio 33151 węzłów.
W końcowej fazie procesu wartości te osiągnęły odpowiednio 131245 elementów oraz
215255 węzłów.
Zastosowany materiał DP-800 jest stalą należącą do grupy UHSS, dwufazową, zimno
walcowaną. Strukturę stali tworzy układ w postaci matrycy drobnoziarnistego ferrytu
oraz „wyspy” martenzytu. Dodatkowo mogą wystąpić małe ilości austenitu
szczątkowego. Stale te są otrzymywane przez odpowiednio prowadzone chłodzenie po
wyżarzaniu z zakresu temperaturowego odpowiadającego współistnieniu faz α i γ.
Właściwości stali, która jest rodzajem specyficznego kompozytu są wypadkową udziału
twardego i wytrzymałego martenzytu oraz ciągliwego ferrytu. Również, co jest
charakterystyczne dla tego gatunku to duża wyjściowa granica plastyczności
i wytrzymałości, która podczas obróbki plastycznej szybko maleje.

175
Przyjęto model materiałowy Hill 48 wykorzystujący izotropową krzywą umocnienia
(Lundh, Bustad 1998, Bergstrom, Bronnestam 1996).
Wybrany model wymaga wprowadzenia następujących parametrów wejściowych
opisujących materiał:
- grubość blachy g 0 = 1,35 mm, moduł Younga E = 210 GPa, współczynnik
Poissona;
- υ= 0,3; gęstość  = 7.8 10 -6 g/cm3, oraz wskazania - dla analizy anizotropii
kierunku walcowania.
W ślad za tym, w pracy wyznaczone zostały współczynniki anizotropii dla
charakterystycznych kierunków walcowania: 0°, 45°, 90° (nazywane również, jako
współczynnikami Lankford’a) oraz referencyjne parametry funkcji (G, H, F, N).
Krzywą umocnienia - dla kierunku walcowania 0° - wyznaczono na podstawie
statycznej próby rozciągania po przeliczeniu wartości naprężenia i odkształcenia
rzeczywistego. Następnie krzywą tę ekstrapolowano w celu zwiększenia zakresu
analizy.
Sama znajomość krzywej umocnienia jest niewystarczająca dla przeprowadzenia
analizy odkształcenia materiału w procesie formowania, dodatkowo - w celu dokładnej
weryfikacji węzłów, które są krytyczne pod względem maksymalnego naprężenia –
koniecznym okazało się użycie wykresu FLD (Forming Limit Diagram).
Jest to wykres tłoczności blachy, przedstawiony za pomocą odkształceń głównych.
Krzywą tę wyznaczono przeprowadzając Nakajima Test 3
Wsad traktowany był domyślnie, jako środkowa warstwa bez grubości. Kontakt
pomiędzy narzędziem, a materiałem (przygotówką) przyjęto jako automatyczny,
w którym solver dopasowuje typ kontaktu w zależności od korespondujących
warunków i typu obliczeń.
Kontakt pomiędzy narzędziem, a materiałem (przygotówką) przyjęto jako
automatyczny, w którym solver dopasowuje typ kontaktu w zależności od
korespondujących warunków i typu obliczeń.
Przyjęto model tarcia poślizgowego Amontonsa – Coulomba, przypisując
współczynnikowi tarcia µ stałą wartość μ = 0,12 (ESI 2011).

3
Można też posłużyć się inną metodą eksperymentalną (Lundh, Bustad 1998, Leppin, Li
2008).
176
Zagęszczenie siatki oparto na kryterium minimalnej wartości promienia r = 2mm, co
determinowało minimalną wielkości elementu 0,64 mm.
Wstępnie - przez szacunkowe wyliczenie wartości w sposób analityczny -
przewidywano maksymalną wartość sprężynowania na około 9 mm w kierunku
grubości materiału. Dokładność przewidywania sprężynowania zależy od wpływu
współczynników odpowiedzialnych za jakość analizy formowania.
Przeprowadzona analiza stanu naprężeń zakładała poziom poniżej 20% pocienienia
(wymagania klienta) mimo, że początkowo stwierdzono dużą tendencję fałdowania
materiału oraz obszar propagacji pękania.
Zmiana geometrii powierzchni docisku OP-20 oraz optymalizacja metody gięcia półek
pozwoliła na dopasowanie najbardziej odpowiednich parametrów procesu.
Jak wykazano numerycznie i eksperymentalnie wartość sprężynowania osiągnęła
zakładany poziom około 8 mm dla OP-20. Analiza numeryczna wykazała też, że
początkowa kompensacja narzędzia bez uwzględnienia kompleksowego ujęcia
przyniosła rezultat niezadowalający. Wartość odchyłki po kompensacji narzędzia
wynosi ponad 1,5 mm, co znacznie wykracza poza założone pole tolerancji. Przyczyną
tego jest optymalizowanie jedynie OP-20 (formowania) bez uwzględnienia OP-30 i OP-
40 (tnące). Podczas operacji okrawania następowała nieznaczna deformacja odprężonej
części co skutkowało nie spełnieniem narzuconej tolerancji /- 1,25 mm.
Zastosowanie kompleksowej kompensacji sprężynowania pozwoliło w rezultacie na
uzyskanie części mieszczących się w tolerancji wymaganej przez klienta. Należy
zaznaczyć, że niektóre obszary części wytłoczki są mniej ważne i mogą mieścić się w
tolerancji ogólnej, która zwykle jest szersza aniżeli dla obszarów o zawężonej tolerancji
i wynosi +/- 2 mm. Czasem dopuszcza się większą odchyłkę sprężynowania w miejscu
kolejnego formowania. Tak też było dla prezentowanej części. Obszar
o najgłębszym przetłoczeniu jest formowany w kolejnej operacji OP-50 stąd też
akceptuje się kompensację z mniejszą tolerancją.

W kolejnej serii symulacji MES analizie poddano fragment przedniego zderzaka

wykonanego ze stali MSW-1200. Jest to element o podstawowym znaczeniu
w przypadku czołowego zderzenia pojazdu. Stąd też musi on spełniać rygorystyczne
wymagania co do wytrzymałości i sztywności potwierdzone w testach zderzeniowych
NCAP. W konsekwencji zastosowanej procedury optymalizacyjnej - z wartością

177
sztywności jako funkcją celu - kształt elementu podczas jego projektowania ulegał
trzykrotnym zmianom, a wraz z nim i kształt tłoczników.
Zoptymalizowanie sztywności elementu doprowadziło wszakże do znaczącego
pogłębienia się zjawiska sprężynowania skrętnego kumulującego w wytłoczce dużą
ilość energii sprężystej bez możliwości jej uwolnienia, a także do wystąpienia tendencji
do fałdowania materiału wraz z zjawiskiem tzw. oil canning’u (falowania materiału
wzdłuż osi).
W najogólniejszym pojęciu sprężynowanie skrętne jest rezultatem zmiany stanu
naprężenia w objętości wytłoczki po jej odciążeniu w chwili zakończenia procesu.
Z mechanicznego punktu widzenia sprężynowanie jest rezultatem działania momentu
skręcającego w poprzecznym przekroju wytłoczki tak, że powoduje to obrót jednego
z jej końców względem drugiego. Skrętne odkształcenie plastyczne rozwija się poprzez
niezbilansowane naprężenia resztkowe (Abdullach et all.,2012). Natomiast sam rozkład
naprężenia może prowadzić do wystąpienia różnic w postaci skręcenia pomimo stałego
kierunku momentu.
Pojawienie się tego typu odchyłek względem geometrii referencyjnej w znaczącym
stopniu skomplikowało problem kompensacji narzędzia. Efektem tego był stan na tyle
niestabilny, iż nie było można przewidzieć a’priori kierunku skręcenia, który stał się
wielkością losową.
Opisane powyżej uwarunkowania wpłynęły na konieczność przeprowadzenia w pracy
wielu analiz tłoczności ze szczególnym naciskiem właśnie na ten rodzaj sprężynowania.
W tym celu uwzględniono kilka modeli materiałowych oraz wykonano kilkadziesiąt
symulacji numerycznych w oparciu o model MES (liczba elementów matrycy OP-20
była równa 474205 zaś liczba węzłów odpowiednio 126742). Prowadzone obserwacje
potwierdziły przy tym, iż w rozpatrywanym przypadku - bez uwzględnienia wyników
symulacji komputerowych - nawet najbardziej doświadczeni konstruktorzy narzędziowi
nie byli w stanie przewidzieć ani końcowej postaci sprężynowania skrętnego ani jego
wartości. Także próby z użyciem narzędzia pilotowego pokazały z jak poważnym
efektem należało się zmierzyć.
Wypróbowano przy tym - zarówno w sposób wirtualny, jak i w trakcie rzeczywistych
prób - kilka koncepcji tłoczenia sprawdzając ich wpływ na wartość sprężynowania
elementu.
Przyjęte koncepcje tłoczenia różniły się ze względu na kształt wsadu oraz sposób
wywierania docisku. Były to:

178
 - crash forming - najprostsza, nie uwzględniająca żadnego docisku koncepcja
tłoczenia, która redukuje efekt sprężynowania, lecz wzmaga fałdowanie materiału
- konwencjonalne formowanie z użyciem dociskacza - pozwalające na zapewnienie
w docisku w żądanym obszarze, lecz mogące - przy niedostatecznej jego
wartości lub też zbyt małej powierzchni - wywoływać zjawisko oil
canning’u. Dla tej części generuje również sprężynowanie skrętne.
Przeprowadzone badania wykazały dużą zależność wpływu gatunku materiału oraz
geometrii wytłoczki na sprężynowanie skrętne materiału. Początkowa analiza MES dla
geometrii przed zmianami inżynierskimi nie wykazywała takiego efektu. Jednak
pojawiająca się niewielka asymetria ujawniła się w postaci sprężynowania skrętnego.
Dołożono wszelkich starań tak, by model numeryczny odpowiadał w jak największym
stopniu rzeczywistym warunkom tłoczenia. Po wykonaniu narzędzia do prób
stwierdzono podobne zachowanie modelu i eksperymentu. Wykorzystano to w celu
sprawdzenia innej koncepcji formowania tak, by wyeliminować sprężynowanie skrętne,
redukując je do przypadku odchyłki wzdłuż wysokości części. Pozwoliło to na
całkowitą eliminację skręcenia, czyniąc kompensację znacznie uproszczoną. Ponadto
stwierdzono, iż w zależności od siły docisku, uzyskuje się inny charakter
sprężynowania skrętnego. Ostatecznie zdecydowano o użyciu metody crash forming ze
względu na prostą konstrukcję narzędzia oraz eliminację sprężynowania skrętnego.
Przeprowadzone symulacje modelu MES nie wykazały wpływu przyjętego prawa
umocnienia materiału (kinematycznego lub izotropowego) na wyznaczoną tą drogą
wartość kąta skręcenia (sprężynowania) oraz krzywiznę wzdłużną. W rezultacie
zbliżone były także i wartości odchyłek pomiędzy obliczonymi, a rzeczywistymi
wartościami sprężynowania wytłoczki zderzaka.

Podsumowanie

W rozprawie wykazano, iż w procesie projektowania tłoczników koniecznym jest

skompensowanie kształtu narzędzi ze względu na efekt sprężynowania zwrotnego.
Dotyczy to szczególnie wytłoczek ze stali grupy HSS oraz UHSS.
Istotną innowacją w tym zakresie jest przedstawiona w pracy metoda kompensacji,
która dodatkowo uwzględnia odprężenie stacji okrawających. Standardowa metoda
kompensacji podczas obliczeń sprężynowania bierze pod uwagę jedynie proces cięcia,
natomiast nie uwzględnia kompensacji kształtu stacji tnących. Kompensowana

179
powierzchnia stacji tnącej powinna zawierać się pomiędzy powierzchnią nominalną
operacji formującej, a odprężoną powierzchnią tej samej operacji formującej.
Najczęściej powierzchnia operacji tnącej jest powierzchnią nominalna danej operacji
lub też odprężoną jeśli przewidziano proces kompensacji sprężynowania. Zastosowanie
odprężonej powierzchni na stacji tnącej i formującej zwiększa dokładność procesu,
który pozwala na uzyskanie lepszego rezultatu kompensacji nawet o kilka procent, niż
w przypadku standardowego procesu kompensacji. Stwierdzono, że zaniedbanie
kompensacji na stacji tnącej powoduje narastanie odchyłki względem powierzchni
referencyjnej oraz błędy pozycjonowania przetłoczonej blachy na stacji tnącej, a to ze
względu na różnice sprężynowania przed i po procesie cięcia.
Dlatego też zalecić należy stosowanie powyższej metody w celu poprawy tolerancji
wymiarowych oraz jakości powierzchni. W niektórych analizowanych narzędziach
zdarzały się sytuacje, gdzie poprzez standardową kompensację wyeliminowano efekt
sprężynowania, ale problem pozycjonowania był na tyle duży, że wsad pomiędzy
stacjami ulegał znacznemu odkształceniu, co było widoczne na gotowej wytłoczce.
Kompensacja jest zdeterminowana dokładnością obliczeń sprężynowania tak, że jeśli
będą one mało dokładne, to w praktyce nie pozwoli to osiągnąć zakładanego rezultatu.
Wysoką jakość w procesie tłoczenia można osiągnąć poprzez zastosowanie
wielooperacyjnej kompensacji. Pozwala to na lepsze modelowanie jego rzeczywistych
warunków poprzez korekcję poszczególnych stacji narzędzia, w efekcie zwiększając
dokładność obliczeń.
Uzyskane skompensowane powierzchnie operacji formowania i cięcia na etapie
projektowania dają oszczędność czasu i przede wszystkim redukują poniesione koszty
związane z metodą prób i błędów. Tego typu wytłoczki o tak skomplikowanej geometrii
oraz specyfice materiałowej wymagają średnio kilku miesięcy ciężkiej pracy by
osiągnąć zakładaną geometrię mieszczącą się w polu tolerancji. O ile wady wytłoczek
takie jak pękanie i nadmierne fałdowanie są stosunkowo łatwe do przewidzenia
i wyeliminowania to efekt sprężynowania bez wirtualnego środowiska często pozostaje
nie w pełni zidentyfikowany. Jak potwierdziła analiza numeryczna oraz narzędzie
prototypowe, nie uwzględnienie sprężynowania podczas procesu formowania prowadzi
do otrzymania części wadliwej i nie spełniających wymogów klienta.

180
Perspektywy dalszych badań

Jakość wytłoczek wyrażona poziomem dokładności wymiarowo – kształtowej pozostaje

zdeterminowana technologicznymi parametrami procesu tłoczenia, w tym kompensacją
efektu sprężynowania zwrotnego poprzez odpowiednią konstrukcję narzędzi.
Zagadnienia te odnoszące się do konwencjonalnych procesów prowadzonych na zimno
stały się przedmiotem niniejszej rozprawy.
Nie wyczerpują one jednak w najmniejszym stopniu całości tej tak złożonej
problematyki. Rozwiązania problemów technologicznych związanych z procesami
tłoczenia poszukuje się dziś bowiem także poprzez wprowadzanie nowych technologii.
Trzeba tu przede wszystkim wymienić stosowaną już od kilkunastu lat technologię
Tailored Blanks. Istotą jej jest łączenie blach o różnej grubości i z różnych gatunków
stali celem miejscowego usztywnienia silnie obciążonych węzłów konstrukcyjnych lub
też zmniejszenia ogólnej masy pojazdu dzięki zastosowaniu w mniej obciążonych
partiach konstrukcji blach o mniejszej grubości, np. w pewnych partiach płyty
podłogowej. Technologię tę stosuje się też w odniesieniu do konstrukcyjnych profili
zamkniętych o zmiennych przekrojach, a także podczas tworzenia tzw. stref
kontrolowanego zgniotu. Są to pakiety wielofunkcyjne, złożone z zespawanych
laserowo różnych rodzajów blach, np. głęboko tłocznych oraz o podwyższonej
i najwyższej wytrzymałości. Każdy z tych składników pełni w pakiecie inną funkcję.
Blacha głęboko tłoczna pochłania energię uderzenia podczas ewentualnej kolizji
drogowej. Materiał o podwyższonej wytrzymałości odkształca się wówczas, gdy blacha
głęboko tłoczna nie zdołała rozproszyć całej energii zderzenia. Element najsztywniejszy
może przy tym ulegać najwyżej nieznacznym odkształceniom, gdyż jego zadaniem jest
zachowanie sztywności kabiny pojazdu, czyli tzw. przestrzeni przeżycia dla kierowcy
i pasażerów.
Technologię Tailored Blanks i Tailored Tubes stosują zakłady dysponujące liniami
technologicznymi odpowiednimi do metod obróbki stali o najwyższej wytrzymałości
i minimalnej plastyczności. Do takich wybitnie specjalistycznych metod należy też Hot
Stamping, czyli tłoczenie na gorąco elementów z blach niepozwalających się formować
na zimno.
Podane metody umożliwiają produkcję lekkich, a zarazem sztywnych i wytrzymałych
wytłoczek, w których zjawisko sprężynowania - ze względu na specyfikę kształtowania

181
– zostało znacząco zredukowane. Koniecznym jest wszakże podjęcie badań nad
lepszym poznaniem zjawisk zachodzących w tej technologii.
Drugą technologią – wciąż znajdującą się w początkowej fazie rozwoju jest proces
tłoczenia na gorąco – Hot forming.
Wytłoczki wytworzone w tym procesie mają coraz większe zastosowanie w przemyśle
samochodowym, gdyż ich zastosowanie pozwala na obniżenie masy samochodu nawet
o kilkanaście procent oraz zapewnia większe bezpieczeństwo w czasie wypadku ze
względu na ich wytrzymałość.
Części formowane na gorąco mają przy tym bardzo wysoką granicę plastyczności na
rozciąganie Re około 1000 MPa.
W procesie formowania na gorąco, materiał wsadowy podgrzewa się do około 950
stopni Celsjusza, a następnie odkształca w narzędziu formującym. Po zakończeniu tej
operacji schładza się do temperatury pokojowej.
Efektem tego jest:
- spadek wymaganej siły nacisku prasy
- zmniejszenie poziomu naprężeń w wytłoczce
- ograniczony efekt sprężynowania
- korzystna mikrostruktura pozwalająca na uzyskanie wysokich parametrów
mechanicznych
- uzyskiwanie stosunkowo wysokich wskaźników produkcyjnych i mała liczba
wybraków

Również i w tym przypadku koniecznym jest podjęcie szczegółowych badań i takie też
są zamiaru autora niniejszej rozprawy.

182