You are on page 1of 6

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261612686

Systematic optimization of the Product Development Process with PLM-ML

Article · April 2013

CITATION READS

1 1,202

4 authors, including:

Jörg W. Fischer
Karlsruhe University of Applied Sciences
20 PUBLICATIONS 110 CITATIONS

SEE PROFILE

All content following this page was uploaded by Jörg W. Fischer on 15 April 2014.

The user has requested enhancement of the downloaded file.


48 PROJECTS

Systematic optimization of the


Product Development Process
with PLM-ML
Prof. Dr. Jörg W. Fischer, Martin Rebel, Bernhard Lammel, Jürgen Gruber

Increasing productivity, driven by the pressure of costs and competition, is a core strategy element of all global
acting companies. Manufacturing companies persecute this goal, beginning from continuous improvement of
their manufacturing processes as far as optimizing global supply chains. The idea of optimization is progressively
applied to the Product Development Process (PDP) and discussed under the term Product Lifecycle Management
(PLM). Contrary to this, PLM projects basically focus on IT implementation. Hereby, the idea of optimization
often becomes lost. One cause is to be seeked in the complexity of the PDP. The PDP can hardly be captured
and optimized with existing modeling methods. Against this background PLM-ML was developed by the
authors as an instrument to easily capture and optimize the PDP. This article explains PLM-ML and exemplifies
how the PDP can be optimized utilizing PLM-ML.

The idea of PDP optimization manufacturing and production technolo- coupled to implementation of IT techno-
gies. Currently the idea of optimization is logies, PLM is prevalently conducted as
Manufacturing companies are facing the increasingly adapted to the PDP as the an IT implementation project within IT
challenge of continuously increaseing awareness of so far unutilized potentials departments. Thereby the intrinsic idea
their productivity in order to withstand within the PDP, particularly for its of optimization is oftentimes lost. Appa-
the pressure of global competition in the downstream processes, grows. rently one main cause for this is the
future. Thereby, they are achieving great distinct focus on information technolo-
success through optimization of operation These discussions are primarily bound gies. Agley a further problem, which
procedures, improvement of organizatio- under the buzzword Product Lifecycle typically gets lost in the fog of the IT
nal structure as well as the use of new Management (PLM). Since PLM is tightly projects and currently drawing insuffi-

Figure 1: A new point of view to understand the Product Development Process (PDP)

ProductDataJournal No. 1 I 2013


PROJECTS 49

cient attention, exists. The PDP is extra-


ordinarily complex. Using existing model-
ing methods, it is nearly impossible to
depict the PDP. Modeling Tools for a
transcend discussion on optimization are
missing as of today. Hence, in the Frame
of the development of Do(PLM)Con
methodology the PLM Modeling Langua-
ge (PLM-ML) was designed as a tool to
enable a discussion on optimization of
the PDP.

A new approach to perceive the


PDP

Genesis of PLM-ML was expert discussions


of PLM architects which aimed to define
“industry best practices” for PLM. In
context of these discussions conducted, Figure 3: Aspects of (Partial-)Product Models
analysis of industrial applications consi-
derable weakness in existing modeling
instrument, such as functional analysis,
process chain representation or IT system
landscape maps, were found. The result-
ing depictions were quickly getting too aspects of the future product for the schedules. The daily operating processes
complex. Beyond that essential PLM specific application area. Hence each of of the PDP are activated through a mile-
aspects could not be represented with those models is a (Partial-)Product Model stone of the management process or
existing methods at all. (shortly: Partial Model) of the future they end up by one of its milestones.
product. They can be subdivided into long-term
After some time of intensive discussions stable processes (meaning release and
a new view based on a systemic model- Based on those considerations, the change processes, see Eigner 2012, p. 21)
ing element emerged, whose significance modeling approach shown in Figure 1 and daily engineering processes. Contrary
was not recognized sufficiently. To develop was developed. It combines the process to daily engineering processes, stable
a product, engineers are imagining it and oriented view with the idea of the processes are explicitly defined and
it's production beforehand. Therefore systemic modeling element “Partial generally controlled by workflows in data
they use models. Each domain (e.g. Model”. Processes of the PDP are thereby management systems. All daily processes
Requirements Management, mechanical differentiated in three categories. within PDP produce, utilize, change or
and mechatronic design, documentation, Management processes are controlling extract information into or from Partial
etc.) is using their own specific models. the PDP (Figure 1, top). Typically they are Models. As a result, Partial Models can
Those models provide ample detailed processed with the help of milestone be seen as the seed crystals of the pro-
cesses within PDP (Figure 2).

The benefit of the considerations in Figure


1 and Figure 2 is the accomplished
simplification. Following this approach
the PDP can be analyzed and depicted
through its Partial Models. Since the
number of Partial Models is much less
than the number of processes within the
PDP, their level of detail can far better be
harmonized and additionally they can
easier be localized which is an essential
modeling advantage

Several fundamental characteristics of


Partial Models can be described as follows.
A Partial Model is always an instance of a
Partial Model type. Figure 3 exemplary
shows instances of various Partial Model
types for the product line “Dumper
Truck” (Figure 3, top) and “Digger”
(Figure 3, bottom). Partial Model instances
of the same type often possess commu-
Figure 2: (Partial-)Product Model as seed crystals of the processes within PDP nality, meaning shared usage of compo-
nents with, where required, various

No. 1 I 2013 ProductDataJournal


50 PROJECTS

occurrence in context of the specific individual PDP realizations. It uses the substitutes the information object used
product line (Figure 3, shared component Partial Model as the central modeling in EPC with the modeling element “Par-
“underbody”). Different Partial Model element. PLM-ML is inspired by value tial Model”. The groups “Partial Model
types (Example Figure 3, bottom, Model stream design and especially adopts the Detail” and “Partial Model Interaction”
types of product line “Digger” from left idea of alignment of the PDP to specific focus on the Partial Models itself. They
to right) often contain redundant infor- Customer Values as well as the idea of an enable for analyzing static structure
mation and similar model elements since easily understandable, less formal graphic (semantics), time behavior (progression)
they are a reproduction of the same notation technique. Also, the separation as well as integration of different Partial
entity. A usage in a usage model (e.g. between current state and a future state Models.
tire) has an equivalence in the bill of to be reached is adopted. Therefore, the
material model (the for this usage supp- PLM-ML provides several types of dia- Optimizing the PDP of a medium
liable tire with varying profile and specific grams, which can be categorized in the sized machine tool manufacturer
rubber composite) and an equivalence in following groups:
the virtual prototype. The validity and n Partial Model Landscape Following the usage of PLM-ML is shown
significance in those model types is n Partial Model to Process on the example of a typical tool manu-
however diverging (e.g. a CAD model is n Partial Model Detail facturer. The case is anonymized and
not designed for each type of rubber n Partial Model Interaction simplified. The company builds machines
compound). To consolidate diverging with far more than 1000 parts per
model types of the same product a com- With diagrams of the group “Partial machine in single or limited-lot producti-
parison of the significance of the model Model Landscape” maps of the relevant on. They offer an open variant configura-
type’s elements is necessary. This leads to Partial Models of the PDP can be depicted tion to their customers, meaning the cus-
an adaption of semantic and, therefore, including their cross linking over a time tomer can order individual variants which
to a change in mode of operation in at axis. Areas with high potential for opti- are then design and produced for the
least one of the two model types. mization can be easily located on those particular order. In the context of a dis-
landscape maps. Such areas are then cussion of weak points with the company
PLM-ML as an instrument for further detailed with the help of the design and manufacturing teams, two
modeling and optimizing the diagrams of the other groups. This essential problems were recognized as
PDP approach enables reducing the analyzing essential pains.
efforts by keeping the focus on the
To enable the usage of the previously essential key aspects. The group “Partial The company does not have any
outlined theoretical view in practice, the Model to Process” combines the Partial alignment process to keep the CAD and
PLM-ML was developed within the scope Model concept with the process view. the BOM-Data synchronized. Thus often
of Do(PLM)Con methodology (Fischer The incorporated “Partial Model/Process the wrong parts or components are
u.A. 2011). PLM-ML is a semiformal gra- Chain” diagram is based on the event- supplied and provided in prototype and
phical modeling language which enables driven process chain (EPC) (see Keller, G.; order production, leading to massive
analyzing and optimization of company Nüttgens, M.; Scheer 1992, p. 13), but wastage of time and enormous additional

Figure 4: Life Time & Information Flow diagram “Current State“

ProductDataJournal No. 1 I 2013


PROJECTS 51

costs. A further problem is caused by the duct-BOM. The fundamental information this available components of existing
much too long cycle times while proces- for the Pre-Product-BOM he attains from products are copied and then adapted
sing orders with high amount of customer CAD-Assemblies stored unstructured for the specific new product. The use of
specific development. The main competi- within the Team Data Management identical modules between different
tors are often able to deliver faster and (TDM) system (Figure 4, CAD Cluster). product lines, so that enhancements of
therefore important orders are frequently From the TDM-System he extracts the a component during its lifecycle can be
lost. part information which is enclosed in the shared by all products, is not possible.
CAD-Assemblies semi-automatically into Since the company suffers enormous
The company spends a lot of time trying Excel files and puts the necessary infor- difficulties, arranging and passing on
to improve this situation. However, they mation in the Pre Product BOM (Figure 4, current working data and collaboration
were trapped in the typical pitfall of iso- Pre Product BOM). Necessary customer with suppliers is only possible with tre-
lated optimization without understanding specific adaptations are then discussed mendous efforts and so is mostly avoided.
the whole PDP. To improve this the Life with product engineers responsible for Area B is related to interaction between
Time & Information Flow diagram of specific CAD-Assembly during meetings product engineers and purchasing.
PLM-ML was used to analyze the com- and on the telephone. The results of Material master and Pre-Product-BOM
plete company individual PDP on a high these meetings are also documented lists are neither linked by it technology
abstraction level (Figure 4 and 5). This by the product engineer within the Pre nor semantically associated. Hence, pur-
diagram type provides a time-related Product BOM. Subsequently, the equiva- chasing has no knowledge about the
general view of interacting Partial lence between CAD and BOM for adap- materials being relevant for development
Models in course of the PDP. Rectangles ted parts (Figure 4, material master) is of new products. Therefore the product
are representing the relevant Partial clarified with different ERP specialists by engineer is usually not informed about
Models. Arrows and lines are depicting telephone. Before completion of the changes of the material master. Further-
the crosslinking between Partial Models. design phase, the product engineer more, inapplicable components are often
The dedicated symbols at the arrows are initiates the creation of the Master Product ordered, then having to be amended and
furthermore used to characterize the BOM (Figure 4, Master Product BOM) newly ordered. Extended delivery times
crosslinks. The time axes document based on the Pre Product BOM. and gratuitous additional costs are the
essential milestones or events of the PDP. consequences. An attempt to reduce
Clouds are used to define out of scope Areas A & B (Figure 4, circuited) were costs through purchasing strategies
areas for simplification of the diagram. identified as substantial origins of the aggravates this situation. This only leads
problems during analysis of the PDP. Area to further changes of the material master
The diagram depicted in Figure 4 shows A affects mechanical design. CAD data is and in consequence incalculable effects
the analyzed “Current State” of the PDP. managed and released within the TDM on orders in development. Figure 5 illus-
Upon a receipt of an order, a product system, but there is no differentiation trates the enhanced and already success-
engineer arranges the necessary assem- between research and product compo- fully implemented “Future State”. The
blies for this specific order based on his nents. Development at this company unstructured CAD cluster was substituted
experience. For this he creates a Pre-Pro- takes place only in scope of orders. To do by CAD product structure being calibrated

Figure 5: Life Time & Information Flow Diagram “Future State“

No. 1 I 2013 ProductDataJournal


52 PROJECTS

for each order (Figure 5, CAD structure). Structure. In the ERP Material only pur- and reorganization of responsibilities of
It also replaces the Pre Product BOM. chasable standard lengths are recorded. the created information during their life-
Furthermore, an explicit partial model for Allowing a distinct assignment between cycle in the PDP.
standard part being synchronized with the parts in the Master-Product-BOM and
the material master was introduced. As a the models of the parts within the CAD- PLM-ML enables a discussion on optimi-
basis for a research and development Product-Structure was done through zation of the PDP. Diagrams of PLM-ML
(R&D) process that is decoupled from introducing the possibility to define the are independent of future applicable IT
order development, a CAD cluster for obstructed lengths and additionally the systems, so that a discussion can be con-
R&D components (Figure 5, R&D CAD resulting standard lengths of bars already ducted neutral to vendors and systems.
model) was installed and connected to within the CAD-Models in the CAD-
the CAD product structure through so Product-Structure. Only with the solutions In scope of the Do(PLM)Con consulting
called “Cross Model Based On Clone”. for those allocation cases it was possible approach, PLM-ML can show off yet
Thus modules being adopted from R&D to invert the direction of supply from another advantage. Since PLM-ML enables
model into the CAD product structure BOM to CAD driven and also to control representation not only for a functional
are copied, but remain linked to their the ERP Material as well as the Master view but also semantics and progression
origin. So they are decoupled for their Product BOM by the CAD Product Struc- of relevant information of the PDP, it is
origin and could be developed indepen- ture. Contrary, it is generally defined possible to plan implementation of
dently but are still able to access their ori- impossible that changes are initiated by appropriate PDP IT systems more precisely
gin. This enables the share of innovation the ERP Master. Beyond that, deriving a and more systematically than ever before
of the R&D model to all derived CAD logistically oriented 3D model (not (see Fischer, Brinkmeier 2012). n
product models. Organizationally this is shown by diagram) which can serve as a
conducted by a “Module Owner”, basis for installation and repair by service
responsible for a specific module across employees being on-site with the custo-
multiple orders. This concept gives the mer is made possible by the rearrange-
company the flexibility to develop modules ment of structure semantics of the CAD
earlier and independent of customer Product Structure to a logical orientation.
orders and to use them in several pro-
ducts. Further optimization

The quintessence of the new concept is Future State shown in Figure 5 is currently
the implementation of an automated serves as a basis for further optimization. References:
Fischer, Jörg W. u.a.: Implementierung domänenintegrie-
process between CAD Product Structure, Therefore, it is now the Current State render PLM-Lösungen. Do(PLM)Con – ein Ansatz
Master Product BOM and ERP Material. that shall be transformed into a new zur Konzeption und Realisierung domänenintegrie-
render PLM-Lösungen. In: Industrie Management,
The CAD Product Structure is defined Future State. Potentials still rest in a more Berlin, 27(2011)5, S. 17-21.
as the leading model. Each newly created precise clarification of coactions between Fischer, J.W., Brinkmeier, B.: Do(PLM)Con – Proposal of a
taxonomy for the development of PLM-Architectures.
CAD model automatically generates a R&D models and CAD Product Structure In: ProductDataJournal, (2012)1, S. 24-28
counterpart in the ERP Material and if (Figure 5, detail 2). Currently models are Eigner, M.: PLM versus PPS: Zukünftige Lösungen für ein
durchgängiges Produkt- und Prozessmanagement,
necessary in the Master Product BOM. functionally held together by one person In: eDM Report; electronic Data Management, Hop-
Non-synchronous BOM and CAD data and modules are copied with reference penstedt Publishing, Darmstadt, (2012) 1, S. 20-23.
Keller, G.; Nüttgens, M.; Scheer, A.-W.: Semantische
is eliminated for this reason and possi- to their origin. An improvement can be Prozeßmodellierung auf der Grundlage „Ereignis-
bility to automatically generate a conducted if utilizing the same modules gesteuerter Prozeßketten (EPK)“, in: Scheer, A.-W.
(Hrsg.): Veröffentlichungen des Instituts für Wirt-
Master Product BOM from the CAD and context-sensitively adjusting them schaftsinformatik, Heft 89, Saarbrücken 1992.
Product Structure at any time exists. succeeds. Above all, considerable possi- (http://www.iwi.uni-sb.de/iwi-hefte/heft089.pdf)
Realization of this approach demanded ble savings are identified in even more
for an adjustment in information consequent usage of modules through
management and hitherto mode of opera- products and product lines. Implementa-
tion. tion requires a more precise analysis of
the product spectrum to define which
The Models were semantically modified information has to be adjusted context-
in such an extent that elements of the sensitively in particular products and
CAD Product Structure have distinct which can be kept generic.
equivalences in each other’s Partial Contact
models (Figure 4, detail 1, Model Type Conclusion and Outlook Prof. Dr. Jörg W. Fischer.
Alignment). The original functional and Hochschule Karlsruhe
structural semantics of CAD Product Evidently the example shows that optimi- – Technik und Wirtschaft
Structure were thereby converted into zation of production is different from Fakultät für Maschinenbau
logically oriented semantics. To ensure optimizing the PDP in principle. Production und Mechatronik
unambiguousness at any time, relevant optimization generally focuses on cut Phone: +49 173 3675220
allocation case between models were down of capacities of human resources, E-Mail: joerg.fischer@hs-karlsruhe.de
carved out and specific solutions were while optimization of PDP focuses on
defined in each case. As an allocation intelligent, context- and time-sensitive Martin Rebel, M. Eng.
cases bars shall be mentioned as an crosslinking of information. Optimization Siemens Industry Software
example. Those bars are obstructed in is thereby successful when combining GmbH & Co. KG
different lengths. Only actually obstructed optimization of information flow between Phone: +49 151 42248392
lengths are modeled in the CAD Product relevant partial models with clarification martin.rebel@siemens.com

ProductDataJournal No. 1 I 2013


View publication stats

You might also like