machines

Article
A Methodology for Product Development in Mobile
Machinery: Case Example of an Excavator
Nima Alaei *, Emil Kurvinen and Aki Mikkola
Department of Machine Design, LUT University, 53850 Lappeenranta, Finland; emil.kurvinen@lut.fi (E.K.);
aki.mikkola@lut.fi (A.M.)
* Correspondence: nima.alaei@lut.fi




Received: 13 October 2019; Accepted: 5 November 2019; Published: 7 November 2019


Abstract: Digital tools have become indispensable for the testing and modification of prototypes in
mobile and industrial machine manufacturing. Data that are extracted from virtual experimentation
and analysis are both affordable and valuable, due to their repeatability and because they are close
to real-world observations. Expert knowledge is a prerequisite for full deployment of computer
aided engineering tools in the design phase and concomitant stages of product development.
Currently, such knowledge, for the most part, is provided by the product development team and
the manufacturer. Yet, it is important that manufacturers and designers receive end-user feedback
throughout the product development process. However, end-users often lack sufficient know-how
about the technical and engineering background of the product development, and this lack of
understanding can become a barrier to user-designer communication. The aim of this article is
to present an alternative to traditional design approaches that is based on customized real-time
multibody simulation. This simulation-based approach can be seen as a platform that has the
potential to improve knowledge management systems for product development. End-user feedback
to the designer is given in a systematic manner throughout the design process using a multipurpose
XML-based multibody environment.

Keywords: multibody dynamics; product development; real-time simulation; product-service system;

modular design

1. Introduction
Effective simulation methods are important when the development of a system reflects
multidisciplinary design concerns [1] and when realistic behavior of the model is beneficial for
further investigation. The outcome of any simulation has to be credible as an estimation of real
experiment results. In machine manufacturing, modern simulation tools can provide interaction
between the machine (virtually presented) and the human in real time. In many cases, the machine
modeling comprises investigation of the dynamic behavior of a model made of interconnected bodies,
namely multibody simulation, combined with simulation models of electrical, hydraulic, and contact
forces [2,3].
Early efforts to improve the efficiency of real-time simulation of multibody dynamics focused on
kinematics and dynamics [4]. Nowadays, as a result of improvements to simulation design, together
with a rapid increase in computing power, modern simulator designs are able to handle complex
process models with high fidelity using groups of connected simulation elements [5]. These improved
capabilities have resulted in effective real-time simulation. In real-time simulation, a user has online
interaction with the model, so the user can utilize outcome response to investigate the behavior
of the system to their input instantly. Such systems are commonly used for operator training in
aviation, marine, mobile machinery, and other industries. In training scenarios, the system user (who

Machines 2019, 7, 70; doi:10.3390/machines7040070 www.mdpi.com/journal/machines

Machines 2019, 7, x FOR PEER REVIEW 2 of 15
Machines 2019, 7, 70 2 of 15
(who is considered the operator of the machine) is given tasks to complete, and the supervisor can
observe the reaction of the trainee to different training scenarios [6,7].
is considered
In designthe operatoreven
processes, of the machine)
though is given
the user, tasksthe
unlike to designer,
complete,might
and the notsupervisor can observe
have comprehensive
the reaction of the trainee to different training scenarios [6,7].
theoretical understanding of the machine components, providing the user with choices for actions
In design
can generate processes,
valuable even though
information the user, of
for analysis unlike
the bestthe designer, might data
practice. These not have
can be comprehensive
used for the
theoretical understanding
assessment and improvement of the machine components,
of performance providing
and productivity, the user
because with
they arechoices for actions
representative of
can generate valuable information for analysis of the best practice.
real-life reactions and behavior. Users spend considerably more time with the machine than any These data can be used for the
assessment
designer whenandtesting
improvement
machines. of performance and productivity, because they are representative of
real-life
Figurereactions
1 shows andthebehavior.
underlyingUsersstructure
spend considerably
of traditional more time methods,
design with the machine than any
as presented by
designer when
Tomiyama [8,9]testing
(on themachines.
right) and its suggested improvement (on the left). In the figure, VDI 2221
(VereinFigure 1 shows
Deutscher the underlying
Ingenieure structure
or Association of of traditional
German designSystematic
Engineers) methods, Approach
as presented to theby
Tomiyama [8,9] (on the right) and its suggested improvement
Design of Technical Systems and Products [8] represents conventional design processes for(on the left). In the figure, VDI 2221
(Verein Deutscher
mechanical Ingenieure
parts. The design orstepsAssociation
start withofdefinition
German Engineers) Systematicand
of the requirements Approach to the Design
initial searching for
aofsolution.
TechnicalThe Systems and Products
requirements [8] represents
are the needs of the conventional
customer, designwhich,processes
followingfor mechanical often
clarification, parts.
The design
specify howsteps start with
the product is definition
planned to of be
theoperated
requirements and whatand initial searching
expectations for atosolution.
need The
be fulfilled,
requirements
including are the needs
prioritization of the customer,
of performance aspects which,
of thefollowing
product. clarification,
In the secondoften step,specify how the
the conceptual
productphase,
design is planned to beare
solutions operated
soughtand to what
provideexpectations
a basis forneed to be fulfilled,
addressing including
the problem, prioritization
which requires
of performance
further analysis.aspects of the product.
More complete In including
details, the secondlayoutstep, the andconceptual
placement design
of the phase, solutionsare
components, are
sought to provide
generated a basis
in the third step.for addressing
Extras the problem,
and additional which requires
components further
(e.g., for analysis.
decoration More complete
purposes) are also
details,inincluding
added this phase.layout
Thisandstepplacement of the into
can be divided components,
modules are generatedwhich
or elements in thearethird step.on
based Extras and
solution
additional(step
principles components
2). Here,(e.g., for decoration
the modules purposes)
are parts that areare also added
difficult in but
to alter this are
phase.
taken This
as step can to
a whole be
divided
serve into modules
a purpose (e.g.,orengines
elements in which
vehiclesare are
basednotonproduct-specific).
solution principlesConcurrently,
(step 2). Here,elements
the modules are
are parts that
specifically are difficult
designed to fitto alter but
product are taken as aStep
specifications. whole to serve athe
4 considers purpose (e.g., engines
compatibility of the in vehicles
system to
are not product-specific).
different components and Concurrently,
manufacturing elements
methods areandspecifically designed
includes cost to fitThe
analysis. product specifications.
knowledge gained
Step 4 considers
enables the designerthe compatibility
to produce the of the system
final to different
design. Then, incomponents
the last step, and manufacturing
step 5, documentation methods is
and includes
prepared beforecost analysis. The [8].
manufacturing knowledge gained enables the designer to produce the final design.
Then, in the last step, step 5, documentation is prepared before manufacturing [8].

Figure 1. Focus of the proposed approach in relation to the traditional design approach in the VDI 2221
guideline [10,11].
Machines 2019, 7, 70 3 of 15

Alongside such methods, product-service systems (PSS) have been introduced, with the aim of
serving users (customers) by associating physical commodities with services designed to help users
make optimal use of an already purchased product [9]. Improving services provision and satisfying
user needs to obtain higher loyalty can be addressed using the PSS concept. In the model presented in
this paper, these aims are examined through use-oriented and results-oriented consideration of the
product before design finalization. Additionally, better communication and project knowledge transfer
between the consumption and production parts of the PSS are expected [12,13].
The frame of reference of the methodology in this process is constructed around decision making
and systematic data acquisition in steps 2 and 3 of the design phase of product development, and the
methodology also incorporates concepts of PSS. The first step in this process follows the traditional
guideline of the design process with a statement of need from the user side, e.g., from the user
of the machine. This stage normally includes a combination of several concepts or modules that
together form an interdisciplinary task (e.g., the design of hydraulics and the sizing of the bucket).
The specification is then produced based on the user needs and the next step, i.e., concept design of the
product, commences [13]. Preliminary sketches are made, which are then assessed based on cost or
performance priorities, providing a path towards detailed design. This stage is normally constructed
with careful consideration of boundaries and limitations, for example, total acceptable weight or
overall dimensions, unless there is a very specific requirement governing the design concept (e.g., a
requirement for operability in extreme arctic conditions). Mechanical strength calculation and other
detailed design processes then complete draft revisions. Once final optimization and reviews have
been performed, engineering drawings can be produced and further steps towards manufacturing
started [10,11].
In complex systems, the machine is not a final product but plays the role of a tool. Consequently, it
has to be able to fulfil a variety of tasks and operate in different environmental conditions; furthermore,
it is influenced by the techniques and the skills of the user. Forklifts, excavators, and wheel-loaders
are examples of industrial machines for which the designer might not have all necessary input data
about machine operations [12,14] at the design clarification stage. This uncertainty regarding operating
conditions and constraints becomes even more marked when the tasks that users should carry out
are case-dependent on other variables, such as environmental conditions. The design process then
becomes highly fluid and numerous cases and scenarios need to be investigated for successful selection
of the final design. The interplay between the product and the desired outcome of product utilization
is in line with the PSS orientation, i.e., consumers are not demanding the product itself but are in
pursuit of the utility deriving from utilization of the product in their application area [2].
Several studies have been conducted around the development of PSS and design knowledge-reuse
as a knowledge management framework [15,16]. These works aim to develop an effective and
appropriate platform for knowledge flow that is suitable for different disciplines and stages of the
product life cycle. Even though knowledge-reuse was modeled and applied in the above papers,
information and communication of information resources (i.e., the data that are collected from the
user, including associated processed information) are considered to remain within the domain of the
manufacturing company side. The user is involved in the initial need and task clarification, but user
involvement is limited or non-existent in the following stages.
This paper demonstrates integration of simulator use into the concept design stage and detailed
design stage of product development. The simulator functions on the basis of dynamic computation in
a real-time multibody approach, where bodies, joints, and forces interact with each other. The objective
of the study is to introduce a design methodology in which customization is available for defined
modular parameters of machine components. In parallel, this study aims to present the potential of a
multibody dynamics approach to solve real-world problems. The example case discussed consists
of one hydraulic part and one mechanical part. The user, as the client, has the freedom to customize
the machine model without needing in-depth knowledge or special expertise in engineering and
simulation. Consequently, user contribution to product design enhancement is possible from early
Machines 2019, 7, x FOR PEER REVIEW 4 of 15
Machines 2019, 7, 70 4 of 15
possible from early stages of prototyping—as shown in step 2 and step 3 in Figure 1—and
improvements in decision making are achievable in terms of features of the multibody system
stages of prototyping—as shown in step 2 and step 3 in Figure 1—and improvements in decision
dynamic real-time simulation. To shed light on the introduced design concept, an excavator is
making are achievable in terms of features of the multibody system dynamic real-time simulation. To
considered as an example. Excavators are among the most commonly used mobile machinery
shed light on the introduced design concept, an excavator is considered as an example. Excavators are
equipment in civil engineering and their design and use include complex interconnected bodies and
among the most commonly used mobile machinery equipment in civil engineering and their design
sophisticated design [17]. The digging mechanisms and complex components of excavators mean
and use include complex interconnected bodies and sophisticated design [17]. The digging mechanisms
that they are a suitable example for examination of the potential of the proposed approach for product
and complex components of excavators mean that they are a suitable example for examination of the
development of mobile machinery.
potential of the proposed approach for product development of mobile machinery.

2.2. Materials
Materialsand
andMethods
Methods
Inthis
In thissection,
section,aaschematic
schematicof ofthe
theidea
ideaisisfirst
firstexplained
explainedandandthe themultibody
multibodyreal-time
real-timesimulation
simulation
part, which is the backbone of the method, is then described as it relates
part, which is the backbone of the method, is then described as it relates to the PSS concept. to the PSS concept. Next, the
Next,
background
the background forforthethe
hydraulic
hydraulic design,
design,the themodularity
modularityusing
using XML
XML (Extensible Markup Language)
(Extensible Markup Language)
format, and the customer interface are
format, and the customer interface are presented. presented.
Oneiteration
One iterationof ofthe
thedesign
designprocess
processisisexplained
explainedfor forthe
thecase
caseexample.
example.In Ineach
eachiteration,
iteration,the
themodel
model
wasupdated
was updatedbased
basedononthe the outcome
outcome of of
thethe previous
previous stepstep until
until the desired
the desired results
results werewere achieved.
achieved. It wasIt
was assumed that more than one part of the machine was involved in
assumed that more than one part of the machine was involved in the design case and that the partsthe design case and that the
parts interacted, and their performance was interdependent. In the illustration
interacted, and their performance was interdependent. In the illustration of the applicability of the of the applicability of
the method,
method, the goal
the goal waswas to make
to make decisions
decisions about
about proper
proper bucket
bucket size
size andandthethearm-attached
arm-attachedhydraulic
hydraulic
cylinder, which represent the goals in the mechanical and hydraulic
cylinder, which represent the goals in the mechanical and hydraulic design, respectively. design, respectively.
Themodeling
The modelingpart partof
ofaadesign
designprocess
processtraditionally
traditionallystarts
startswith
withan aninitial
initialestimation.
estimation.This
Thisinitial
initial
estimation is based on prior knowledge and previous experience of the designer.
estimation is based on prior knowledge and previous experience of the designer. For example; when For example; when
the breadth of the bucket for a 20-ton excavator has a range between 300 and
the breadth of the bucket for a 20-ton excavator has a range between 300 and 1200 mm, a value of 750 1200 mm, a value of 750
mmisistaken
mm takenas asaastarting
startingpoint.
point. The
The same
same procedure
procedure isispossible
possiblefor forhydraulic
hydrauliccylinders
cylinders with
withbore
bore
diameter ranges. Figure 2 shows the proposed concept design and detailed
diameter ranges. Figure 2 shows the proposed concept design and detailed design steps (stage 2 and design steps (stage 2 and
3
3 in Figure 1) with respect to the traditional design
in Figure 1) with respect to the traditional design approach. approach.

Figure 2. Dynamics data flow for the proposed concept design and detailed design steps.
Figure 2. Dynamics data flow for the proposed concept design and detailed design steps.
Three values, i.e., minimum, mid, and maximum, were used for mechanical dimensioning. Based
Three values, i.e., minimum, mid, and maximum, were used for mechanical dimensioning. Based
on the parts, a model could be built as a combination of every single choice (here, nine models, as
on the parts, a model could be built as a combination of every single choice (here, nine models, as a
a multiplication of three by three). The choices were then given to the target group, i.e., the users,
multiplication of three by three). The choices were then given to the target group, i.e., the users, to utilize
to utilize in real-time simulation. The designers recorded the simulation data in their area of interest
in real-time simulation. The designers recorded the simulation data in their area of interest (see the
(see the results section). On the basis of the first results, refinements for the second round of choices (i.e.,
results section). On the basis of the first results, refinements for the second round of choices (i.e.,
minimum, mid, and maximum values) could be proposed. Once the desired result met the designer
minimum, mid, and maximum values) could be proposed. Once the desired result met the designer
priorities (based on the target requirements of the designer), dimensioning was completed, and the
priorities (based on the target requirements of the designer), dimensioning was completed, and the
design process proceeded to the next steps for detailed design, e.g., material selection, tolerances.
design process proceeded to the next steps for detailed design, e.g., material selection, tolerances.
Machines 2019, 7, 70 5 of 15

2.1. Multibody Systems

Multibody system dynamics is a straightforward computational approach that can be used to
analyze dynamic responses of bodies that are in interaction with each other, i.e., connected by joints.
The bodies can be defined as either rigid or flexible, and the multibody system dynamics stay valid for
large rotational and translational movements. When this interaction extends from machine components
to a human (as a human-in-the-loop) the real-time simulation concept moves to the foreground.
Running a real-time simulation in practice has challenged simulators, because of the complexity of
finding reliable simulation models and the considerable computational demands. Study of multibody
systems is attempting to address these limitations [18].
Relative coordinates in open chain systems where the main dynamic formulation is founded on a
topological system form the basis of the semi-recursive method. When each body (index j) coordinate
is pointed relative to the previous body (index j − 1) in the model tree, relative accelerations, velocities,
and positions, as well as Lagrange multipliers, are employed to carry out dynamic analysis. In this
approach, the position vector of body j in the global coordinate system can be written as:

¯
r j = Rcm
j−1 + A j−1 u j−1 + d j−1, j , (1)

where Rcm
j−1
is a position vector for the parent of body j − 1 in its center of mass, A j−1 is a rotation
¯
matrix, and u j−1 is the body coordinate system in the parent body reference. In Equation (1), d j−1, j is
the relative displacement vector for body j with respect to body j − 1, as shown in Figure 3. For body j,
orientation matrix A j can be expressed:

A j = A j−1 A j−1, j , (2)

where A j−1, j is the relative rotation matrix of body j referenced to the previous body in the chain.
The velocity associated with body j can be obtained by differentiating Equation (1) with respect to time,
as follows: . cm .
. ~
r j = R j−1 + ω j−1 u j−1 + d j−1, j (3)
. cm ~
where R j−1 is the velocity vector of the parent body j − 1, ω j−1 is the skew symmetric matrix of the
relative angular velocity (between body j and its previous body in the chain), u j−1 is the result of
¯ .
multiplication of rotation matrix A j and u j−1 , and eventually d j−1,j is the velocity of the body j to its
..
precedent body in the chain. Then, the vector of acceleration in the global coordinate system r j of
Equation (3) can be derived as:

.. cm . ..
.. ~ ~ ~
r j = R j−1 + ω j−1 u j−1 + ω j−1 ω j−1 u j−1 + d j−1, j , (4)

.. cm .
~
where R j−1 is acceleration of the body j − 1, in its center of mass, ω j−1 is a time derivative of the skew
..
symmetric matrix of the relative angular velocity, and d j−1, j is relative angular acceleration.
Velocity and the acceleration vector in the center of mass are written based on relative rotation
matrices and relative skew symmetric matrices of angular velocity [19,20]. Solving such a system of
equations consisting of generalized coordinate systems demands considerable computation resources,
while practical requirements for running a real-time simulation must not be neglected. To overcome
this barrier, velocity transformation can be employed to express the equation of motion in terms of the
generalized velocity and acceleration. Generalized velocity and acceleration, correspondingly, can be
written as:
. .
q = Vz, (5)
.. .. . .
q = Vz + Vz, (6)
components to a human (as a human-in-the-loop) the real-time simulation concept moves to the
foreground. Running a real-time simulation in practice has challenged simulators, because of the
complexity of finding reliable simulation models and the considerable computational demands.
Study of multibody systems is attempting to address these limitations [18].
Relative
Machines 2019, 7,coordinates
70 in open chain systems where the main dynamic formulation is founded 6 ofon
15
a topological system form the basis of the semi-recursive method. When each body (index j)
coordinate. is ..pointed relative to the previous body (index j − 1) in the model tree, relative
where q and q are generalized velocity and acceleration, respectively, V is a velocity transformation
accelerations,
. velocities,
.. and positions, as well as Lagrange multipliers,
. are employed to carry out
matrix, z and z are relative joint velocity, and acceleration V is the first derivative of the velocity
dynamic analysis. In this approach, the position vector of body j in the global coordinate system can
transformation with respect to time. In three-dimensional space, the velocity transformation matrix
be written as:
for each body—without a parent body—is a matrix of six rows by c-columns, where c is the number
of constraints of degrees of freedom. 𝐫 =As 𝐑 an + 𝐀 𝐮 a+spherical
example, 𝐝 , , joint in body j had a six by three (1)
transformation matrix, which was multiplied by all its parent bodies (Bodies j − 1, j − 2 . . . ) in the
where 𝐑 is a position vector for the parent of body 𝑗 − 1 in its center of mass, 𝐀 is a rotation
model tree to form a full velocity transformation [19]. The equation of motion, using Equation (5), can
matrix, and 𝐮 is the body coordinate system in the parent body reference. In Equation (1), 𝐝 ,
be written as: . . respect to body 𝑗 − 1, as shown in Figure 3. For
is the relative displacement vector for body.. j with
M(Vz + Vz) + Qv = Qe , (7)
body j, orientation matrix 𝐀 can be expressed:
..
where M is the mass matrix, q is the generalized accelerations, and Qe and Qv are vectors of generalized
𝐀 =𝐀 𝐀 , , (2)
forces and quadratic velocity, respectively. In the semi-recursive method, the equation of motion
where 𝐀 , ofisEquation
in the form the relative
(5) rotation matrixbyofthe
is multiplied body j referenced
transpose of thetovelocity
the previous body in the
transformation chain.
matrix to
The velocity
assure associated
invertibility with
of the mass body j canComputationally
matrix. be obtained by differentiating [20] Q∗ and
parallelizableEquation (1)M ∗ terms
with respect
can tobe
time, as follows:
introduced as: . .
Q∗ = VT (Qe − MVz − Qv ), (8)
𝐫 =𝐑 +𝛚 𝐮 +𝐝 , (3)
M∗ =VT MV, (9)
where 𝐑 is the velocity vector of the parent body j − 1, 𝛚 is the skew symmetric matrix of the
relative
and the angular
solvable velocity (between
form of the body
equation canjbe
and its previous
combined with body
penaltyin terms
the chain), 𝐮 for
to account is the result
closed of
loops.
multiplication
So, the final form of rotation matrix 𝐀
of the equation and 𝐮 is written
of motion , and eventually
as: 𝐝 , is the velocity of the body j to
its precedent body in the chain. Then, the vector of acceleration in the global . coordinate system 𝐫 of
.. .
Equation (3) can be derived (M∗ + αΦ T ∗ T
as: z Φz )z = Q − αΦz (Φzt z + Φtt + 2ξΩΦ + Ω Φ),
2
(10)

where Φz and Φzt are Jacobians 𝐫 =of𝐑the constraints

+ 𝛚 𝐮 with + 𝛚 respect
𝛚 𝐮to the + relative
𝐝 , , joint displacement vector(4)
and its 𝐑
where first derivative, and matrix
is acceleration α contains
of the body 𝑗 − 1,values of the penalty
in its center of mass,term,
𝛚 as is well as Ωderivative
a time and ξ, which are
of the
matrices
skew for corresponding
symmetric natural
matrix of the frequency
relative angular and damping
velocity, and ratios
𝐝 , of the penalty
is relative systems
angular defined for
acceleration.
each constraint condition [18,21,22].

Figure 3. Schematic of relative bodies for recursive coordinate definition.

Figure 3. Schematic of relative bodies for recursive coordinate definition.
2.2. Hydraulics
Velocity and the acceleration vector in the center of mass are written based on relative rotation
Hydraulics
matrices in this
and relative study
skew were modeled
symmetric matricesusing a semi-empirical
of angular method
velocity [19,20]. that assumed
Solving that the
such a system of
pressure variation within a volume was negligible. In this approach, valves and long hose lines were
modeled as throttles. A pressure within a hydraulic volume index i (pi ) can be estimated using a first
order differential equation as:
. B .
pi = ei (Qin,i − Qout,i − V i ), (11)
Vi
. .
where pi is the derivative of pressure pi with respect to time, Vi and Vi are volume size and its derivative
with respect to time, Bei is effective bulk modulus, and Q is flowrate, with Qin,i and Qout,i the entering
Machines 2019, 7, 70 7 of 15

and outgoing flow rates of the control volume, respectively. Hydraulic valves introducing pressure
losses and flow rate through a valve can be estimated as:
p
Q = Cv U dp, (12)

where Cv is a semi-empiric flow constant, dp is the pressure difference between two sides of the flow,
and U, as the spool position symbol, is calculated by integration of the following equation with respect
.
to time U:
. Ure f − U
U= , (13)
τ
where Ure f is the reference spool position, and τ is a time constant [23].

2.3. Collision and Soil Model

The computational cost of collision detection algorithms is a key concern in real-time simulation.
This challenge increases exponentially with respect to the number of colliding elements. Contact force
is only calculated when a collision between bodies is identified in a prior time step, so contact forces
are dismissed when there is a gap between two bodies. This approach is possible through multi-stage
collision detection strategies, such as rough spherical boundaries and detailed spherical boundaries, as
explained in [24]. Spherical and cylindrical geometries were employed to approximate the outermost
boundaries of a body and elevate the computational efficiency of the real-time simulation. At the same
time, the contact forces were sensitive to the time step of the simulation, since high velocity actuation
can introduce inaccuracy in the behavior of colliding bodies, leading to dynamic instabilities in the
real-time simulation.
Contacts and collisions are one of the most complex parts of real-time simulated models. Bounding
geometries methods like oriented bounding box (OBB) can be applied to calculate the interaction forces
after intersection between objects is detected. Triangular definition of the geometry to formulate the
model average and its covariance matrix can be considered as:

1 X i
u= (h + li + mi ), (14)
3b
b
1 X i i
C jk = (h j hk + lij lik + mij mik ), j ≤ 1, k ≤ 3, (15)
3b
i=1

where u is model average, b is number of triangles, vertices of triangle i are shown by hi , li and mi , and
C jk is the correlated covariance matrix [21,25].
In a real-time environment, a realistic soil model plays an important role in the dynamics of mobile
machinery. A number of methods for soil modeling are available in the literature using algorithms such
as RAPID, V-Collide and I-Collide [25], which are based on the principles of OBB. In such models, the
best-fitting polygon breaks into triangulated geometries. Overlapping geometries are then treated using
an impulse-based or constraints-based approach. When contact is not detected, particles follow their
trajectory under the influence of gravity until they either land on each other, bounce, or slide through
each other. The stacked particles from a planar soil surface simulate shrinkage and compression by
consideration of a soil recovery factor and the time history of the normal terrain stress. In the case
example under consideration, the excavator remained stationary and the excavation maneuver was
performed by movements of the cabin (see Figure 4), boom, and other connected bodies. Since crawlers
(which are attached to the carriage) were not active, the tread pattern of the excavator grouser shoes
was not modeled and the crawlers were presented as a flat surface rather than a corrugated pattern [26],
which helped channel computational power and avoid unnecessary complications.
Machines 2019, 7, x FOR PEER REVIEW 8 of 15

minimal. For simplification of the topology, tracked parts of the excavator, which produced traction
inMachines 2019, 7,with
interaction 70 the ground, are not illustrated in the figure. 8 of 15

Machines 2019, 7, x FOR PEER REVIEW 8 of 15

minimal. For simplification of the topology, tracked parts of the excavator, which produced traction
in interaction with the ground, are not illustrated in the figure.

Figure 4. Topological schematic for the studied multibody model.

Figure 4. Topological schematic for the studied multibody model.
2.4. Excavator Model
The mathematical representation of the system, including definition of the bodies, constraints,
In this paper, the pair (as the part 1 and the part 2 which previously were mentioned in Figure 2)
and forces, was implemented in an XML database, where each component of the machine sat in one
in the design process were the bucket size and the arm-attached double-acting hydraulic cylinder,
subfolder, accompanied by its attributes. Storage of the data in this way benefited the modeling
considering that the bucket should be functional such that the user has enough force to excavate
within the context of knowledge management as the two correlated key components of the excavator
and fill the bucket. The excavator in this case example was modeled using 10 independent bodies.
served the proposed method. These components remained independent and discrete, while the
The bucket was used as a tool to transfer sand with a density of 2000 kg/m3 to a destination (a hopper).
whole model stayed integrated as a group of segmented components in XML format. First and
The carriage of the excavator was floated on the ground, which made it the only non-holonomic
foremost, bucket size—which is also representative of bucket capacity—was taken as an alternative
constraint in the model. The topological map of the excavator is presented in Figure 4. The chain of
that was determined by the operator of the excavator. Three different bucket sizes were used in this
bodies of the excavator was contiguous and the structure of the excavator was robust in exposure to
experiment so the user could select their own configuration. The other optional key component was
normal excavation maneuvers, so the flexibility of parts such as the arm and boom (rigid bodies) was
the hydraulic actuator that generated the force between the arm body (stick) and bucket connector
minimal. For simplification ofTopological
Figure 4. the topology, trackedforparts
schematic of the excavator,
the studied multibody which
model. produced traction in
link. Figure 5 presents the interface for user selection. The software replaced the modular properties
interaction with the ground, are not illustrated in the figure.
of the component based on the selection made by the operator.
The mathematical representation of the system, including definition of the bodies, constraints,
and forces, was implemented in an XML database, where each component of the machine sat in one
subfolder,accompanied
subfolder, accompaniedbybyitsits attributes.
attributes. Storage
Storage of data
of the the data in way
in this this benefited
way benefited the modeling
the modeling within
within
the the context
context of knowledge
of knowledge management
management as the
as the two two correlated
correlated key components
key components of the excavator
of the excavator served
served
the the proposed
proposed method. method. These components
These components remained independent
remained independent and discrete,and discrete,
while while
the whole the
model
whole integrated
stayed model stayed as aintegrated as a groupcomponents
group of segmented of segmented components
in XML in XML
format. First and format.
foremost, First and
bucket
foremost, bucket
size—which is alsosize—which is also
representative representative
of bucket capacity—wasof bucket
takencapacity—was
as an alternativetaken
thatas andetermined
was alternative
thatthe
by was determined
operator of the by the operator
excavator. Threeofdifferent
the excavator.
bucket Three different
sizes were usedbucket
in this sizes were used
experiment so theinuser
this
experiment so the user could select their own configuration. The other optional
could select their own configuration. The other optional key component was the hydraulic actuator key component was
the hydraulic
that generated actuator
the forcethat generated
between the armthebody
force(stick)
between the armconnector
and bucket body (stick)
link.and bucket
Figure connector
5 presents the
link. Figure
interface for 5user
presents the interface
selection. for user
The software selection.
replaced The software
the modular replaced
properties thecomponent
of the modular properties
based on
of the
the component
selection madebased
by theon the selection made by the operator.
operator.
Figure 5. Graphical user interface for bucket and hydraulic-circuit selection.

The risk of collision for larger bucket sizes was higher as handling a larger bucket is
geometrically challenging. Collisions were possible between the bucket and hopper edge, and also
between the boom/arm and an obstacle, which in the case under study was an electric pole. The
obstacle was intentionally placed near the operation area, so if the user slewed further than the
required maneuver, the boom, arm, or bucket may approach the pole. As an example, Figure 6 shows
a top view of the initial location of the excavator and a schematic of the main objects located nearby.

Figure
Figure 5. Graphical user
5. Graphical user interface
interface for
for bucket
bucket and
and hydraulic-circuit
hydraulic-circuit selection.
selection.

The risk of collision for larger bucket sizes was higher as handling a larger bucket is
geometrically challenging. Collisions were possible between the bucket and hopper edge, and also
between the boom/arm and an obstacle, which in the case under study was an electric pole. The
obstacle was intentionally placed near the operation area, so if the user slewed further than the
required maneuver, the boom, arm, or bucket may approach the pole. As an example, Figure 6 shows
Machines 2019, 7, 70 9 of 15

The risk of collision for larger bucket sizes was higher as handling a larger bucket is geometrically
challenging. Collisions were possible between the bucket and hopper edge, and also between the
boom/arm and an obstacle, which in the case under study was an electric pole. The obstacle was
intentionally placed near the operation area, so if the user slewed further than the required maneuver,
the boom, arm, or bucket may approach the pole. As an example, Figure 6 shows a top view of the
initial
Machineslocation
2019, 7, xof thePEER
FOR excavator
REVIEWand a schematic of the main objects located nearby. 9 of 15

Machines 2019, 7, x FOR PEER REVIEW 9 of 15

Figure
Figure 6. Top view
6. Top view of
of the
the excavation
excavation environment
environment to
to introduce
introduce collision
collision risk.
risk.

There
There is is aa trade-off
trade-off for
for bucket
bucket size
size in
in an
an excavator;
excavator; aa higher
higher bucket
bucket volume
volume helps
helps operators
operators
achieve
achieve higher
higher productivity
productivity of of soil
soil excavation,
excavation, but but large,
large, bulky
bulky buckets
buckets can
can lead
lead to
to higher
higher risk
risk of
of
collision, controllability problems, difficulties in soil excavation, and more
collision, controllability problems, difficulties in soil excavation, and more demanding hydraulic demanding hydraulic
circuits
circuits toto actuate.
actuate. Figure 6. Top view of the excavation environment to introduce collision risk.
AA similartrade-off
similar trade-off waswasseen for actuation
seen for actuationof theofbucket with the
the bucket hydraulics
with of the arm,
the hydraulics ofasthe
illustrated
arm, as
schematically
illustrated in Figure
Thereschematically 7.
is a trade-offinfor The bucket
bucket
Figure excavation
sizebucket
7. The force,
in an excavator; which
excavation aforce,
higheris an
bucket
which important
volume
is an parameter
helps
important for
operators
parameter
excavation
achieve
for excavation productivity,
higher mustmust
productivity
productivity, be
of considered
soil together
excavation,
be considered butwith
together thebulky
large,
with cycle time.time.
buckets
the cycle This
canhelps
Thisleadtotodetermine
helps higher
to theof
risk
determine
most
the suitable
collision, cylinder
controllability
most suitable geometry
cylinderproblems, and to address
geometry difficulties
and to address user
in soilneeds—involving
userexcavation, soil
and more
needs—involving properties
demanding
soil and the
properties hydraulicsite
and the
field situation—accordingly.
circuits to actuate.
site field situation—accordingly.
Selections
A similarbytrade-off
the user led
wastoseenthe implementation
for actuation ofofthe changes
bucketin the
withmodel tree. Changes
the hydraulics of had effectsas
the arm,
inillustrated
different areas, for example, bucket selection would merge a subfolder in the
schematically in Figure 7. The bucket excavation force, which is an important parameter XML-based model.
Definition of
for excavation theproductivity,
mass and inertia,
must beas well as the together
considered graphics with
and geometrical
the cycle time.definition
This helps of to
thedetermine
bucket,
were
the most suitable cylinder geometry and to address user needs—involving soil properties and in
provided through the correlated subfolder. Three-dimensional geometries were introduced the
3DS
site(3D
fieldStudio) format [27]. An example of properties for hydraulic components is shown in Table 1,
situation—accordingly.
which presents data for a sample hydraulic cylinder.

Figure 7. Schematic of decision-making considerations for the arm-attached hydraulic actuator.

Selections by the user led to the implementation of changes in the model tree. Changes had
effects in different areas, for example, bucket selection would merge a subfolder in the XML-based
model. Figure
Definition of the mass
7. Schematic and inertia, as
of decision-making well as thefor
considerations graphics and geometrical
the arm-attached hydraulicdefinition
actuator. of the
bucket, Figure
were provided
7. Schematic of decision-making considerations for the arm-attached hydraulic actuator. were
through the correlated subfolder. Three-dimensional geometries
introduced in 3DS (3D Studio) format [27]. An example of properties for hydraulic components is
shownSelections
in Table 1,by the user
which led to
presents thefor
data implementation of changes
a sample hydraulic in the model tree. Changes had
cylinder.
effects in different areas, for example, bucket selection would merge a subfolder in the XML-based
model. Definition of the Table 1. Sample
mass of stored
and inertia, asproperties for graphics
well as the a hydraulic cylinder.
and geometrical definition of the
bucket, were provided through the correlated subfolder. Three-dimensional geometries were
Property Value/Units
introduced in 3DS (3D Studio) format [27]. An example of properties for hydraulic components is
Cylinder type double acting
shown in Table 1, which presents data for a sample hydraulic cylinder.
Friction properties defined as a spline
Cylinder piston diameter 190 mm
Table 1. Sample of stored properties for a hydraulic cylinder.
Machines 2019, 7, 70 10 of 15

Table 1. Sample of stored properties for a hydraulic cylinder.

Property Value/Units
Cylinder type double acting
Friction properties defined as a spline
Cylinder piston diameter 190 mm
Cylinder piston rod diameter 95 mm
Machines 2019, 7, x FOR PEER REVIEW 10 of 15
Cylinder inner pipe inner diameter not used
Piston length 150 mm
Minimum
Cylinder stroke
attachment length 100 mm
2150 mm
Minimum
Maximum stroke
stroke 1600 100mm mm
Maximum
Cylinder stroke
material bulk modulus 2101600
GPa mm
Cylinder material bulk modulus 210 GPa
Oil bulk modulus 1.3 GPa
Oil bulk modulus 1.3 GPa
Cylinder coefficient
Cylinder coefficient 0.95
0.95
Leaks between cylinder chambers
Leaks between cylinder chambers 0.01 L/min
0.01 L/min
Pressure difference for rated
Pressure difference for rated leak leak 130130MPaMPa
End damper damping coefficient 8.2 × 10 7 N/m
End damper damping coefficient 8.2 × 10 N/m
7
End damper spring coefficient 5
3 × 10 Nm/s
End damper spring coefficient 3 × 105 Nm/s
Hydraulic end damper not used
Hydraulic end damper
Damper length not used 0
Damper length
Viscous damping coefficient 0 0
Viscous damping coefficient 0
User selections for the cylinder led to the implementation of an XML subfolder, as displayed in
User selections for the cylinder led to the implementation of an XML subfolder, as displayed in
Figure 8, including graphical representation, geometrical specifications, stiffness/damping properties,
Figure 8, including graphical representation, geometrical specifications, stiffness/damping
relevant friction coefficients, and connecting hydraulic elements, as an interpretation of Table 1.
properties, relevant friction coefficients, and connecting hydraulic elements, as an interpretation of
It should be mentioned that the bulk modulus of the hydraulic oil in the system was here assumed
Table 1. It should be mentioned that the bulk modulus of the hydraulic oil in the system was here
to be constant with respect to pressure, and the bulk modulus presented in the table belongs to the
assumed to be constant with respect to pressure, and the bulk modulus presented in the table belongs
cylinder alone and does not include other connected volumes.
to the cylinder alone and does not include other connected volumes.

Figure 8. Sample XML file structure for the arm-attached hydraulic actuator.
Figure 8. Sample XML file structure for the arm-attached hydraulic actuator.
The XML definitions allowed the user to make the model communicate with a variety of other
The XMLHere,
components. definitions allowed
the model the user
tree, which tothe
was make theofmodel
input communicate
the simulation with was
software, a variety of other
also based on
components.
XML Here,
format, and the model could
components tree, which wasbe
therefore the input of the
integrated simulation
to make software,
a complete modelwas
thatalso based
included
on XML format, and components could therefore be integrated to make a complete model that
included all bodies, constraints, forces, input definitions, environment data, and colliding geometries
in the final stage [22]. A dynamic solver then processed the XML files to make a real-time physics-
based visualization. The data required for post-processing were obtainable visually and numerically,
and produced based on user inputs. A sample of numerical results is presented in the next section.
Machines 2019, 7, 70 11 of 15

all bodies, constraints, forces, input definitions, environment data, and colliding geometries in the
final stage [22]. A dynamic solver then processed the XML files to make a real-time physics-based
visualization. The data required for post-processing were obtainable visually and numerically,
and produced based on user inputs. A sample of numerical results is presented in the next section.

3. Results
During the product development process, the designer considers different aspects of the design
with the aim of moving towards reduced energy and material consumption. It is thus beneficial to
study the effect of changes to different components on the dynamic behavior of the real-time simulation;
an example of such studies is presented in [3]. In this case study, samples were taken from the effect of
changing the arm-attached cylinders on excavation force or the effect of change in bucket size on the
fuel consumption, applicability, and speed of the process. These samples were meant to demonstrate
how the capability of coupled parts can be checked to answer user needs based on user criteria
and preferences.
Samples
Machines 2019, 7, xof such
FOR collected
PEER REVIEWdata are presented in Figures 9a and 10. Considering the actuator 11force,
of 15
the figures show change in the performance of three different arm-attached cylinders in response to an
study the effect
input signal. Theofmaximum
changes values
to different
for thecomponents on the dynamic
small, the medium, behavior
and the large of are
actuator the marked
real-time in
simulation; an example of such studies is presented in [3]. In this case study,
the figure to ease comparison of generated forces. The forces in the figures are for a sample trenchingsamples were taken
from the effect
movement of changing
filling the bucket the arm-attached
with cylinders onrepresent
sand. The fluctuations excavation theforce or the
friction andeffect of change
separation in
forces
bucket size on the fuel consumption, applicability, and
needed to excavate the ground during the operation shown in Figure 9b. speed of the process. These samples were
meantOneto demonstrate
difficulty inhow thework
design capability of coupled parts
for complicated can be checked
machine-like to answer
excavators user needs
is ambiguity based
about the
on user criteriaof
consequences and preferences.
changes that do not exhibit a direct effect or have a linear relationship with respect
Samples parameters.
to changing of such collected data are presented
This challenge in Figures
can be tackled in two9a and 10.
ways. Considering
Firstly, the designerthe actuator
is free to
force,
choosetheandfigures show change
try a different setup in the performance
through of three
initial estimation, different
which resultsarm-attached cylinders in
in a rapid trial-and-error
response
process totooutline
an input signal. The
a general planmaximum
and can helpvaluesto for the the
define small, the medium,
boundaries. and theoptimization
Secondly, large actuator in
are
certain ranges can be performed through a loop by adding a small amount of change (as aare
marked in the figure to ease comparison of generated forces. The forces in the figures for in
step) a
sample trenching
one variable, movement
followed filling of
by analysis thethebucket
result,with sand. The
especially when fluctuations
a mechanicalrepresent the friction
part (e.g., a bearing)andis
separation forces needed
chosen from a catalogue. to excavate the ground during the operation shown in Figure 9b.

(a) (b)

Figure
Figure9.9. (a)
(a)Comparison
Comparisonofofpick
pickforce
forcefor
forthree
threetypes
typesof
ofarm-attached
arm-attachedcylinder;
cylinder;(b)
(b)schematic
schematicfor
foraa
sampletrenching
sample trenchingmovement.
movement.

Fueldifficulty
One consumption is an work
in design issue that is difficult tomachine-like
for complicated investigate with simpleismathematical
excavators ambiguity aboutformula.
the
The difficulty is a consequence of the multi-lateral dependency of engine fuel consumption
consequences of changes that do not exhibit a direct effect or have a linear relationship with respect on other
variables,
to changing such as actuation
parameters. of challenge
This components caninbe
thetackled
hydraulic circuit.
in two Figure
ways. 10 shows
Firstly, the effects
the designer on fuel
is free to
consumption for three types of arm-attached cylinder when the excavator was in a parked
choose and try a different setup through initial estimation, which results in a rapid trial-and-error position
(i.e., when
process the crawlers
to outline wereplan
a general not operated for transferring
and can help theboundaries.
to define the bulk material) in response
Secondly, to a similar
optimization in
certain ranges can be performed through a loop by adding a small amount of change (as a step) in
one variable, followed by analysis of the result, especially when a mechanical part (e.g., a bearing) is
chosen from a catalogue.
Fuel consumption is an issue that is difficult to investigate with simple mathematical formula.
The difficulty is a consequence of the multi-lateral dependency of engine fuel consumption on other
Machines 2019, 7, 70 12 of 15

control input signal for the arm-attached cylinder. As can be expected, the highest value of fuel
consumption was found for the largest rod area. It should be noted that higher productivity depends
on the excavation application and comes at the expense of higher fuel consumption. Therefore, in
Machines 2019, 7, x FOR PEER REVIEW
the
12 of 15
final selection of the size, this interplay needs to be taken into account simultaneously.

Machines 2019, 7, x FOR PEER REVIEW 12 of 15

Figure 10.Comparison
Figure10. Comparisonof
offuel
fuelconsumption
consumptionfor
forthree
threearm-attached
arm-attachedcylinders.
cylinders.
4. Analysis and DiscussionFigure 10. Comparison of fuel consumption for three arm-attached cylinders.
4. Analysis and Discussion
Small
4.Small bucket
Analysis andsize and small cylinder size were taken as the first guess of the design process. The
Discussion
bucket size and small cylinder size were taken as the first guess of the design process. The
results section presented one step for the design iteration and expected outcome from each iteration.
results section presented
Small bucket onesmall
size and step cylinder
for the design
size were iteration
taken asand expected
the first guess ofoutcome
the design from each The
process. iteration.
The process was repeatable, and users were therefore trained through their choices of combinations
The results
process section presented one
was repeatable, andstep for the
users design
were iteration
therefore and expected
trained through outcome from each
their choices of iteration.
combinations
of bucket selection and hydraulic-circuit selection. This approach can be employed for other similar
The process
of bucket selectionwas and
repeatable, and users were
hydraulic-circuit therefore
selection. trained
This through
approach cantheir
be choices
employed of combinations
for other similar
mobile machinery using a bucket for loading bulk material (e.g., wheel loaders) or with a similar group
mobile machinery using a bucket for loading bulk material (e.g., wheel loaders) other
of bucket selection and hydraulic-circuit selection. This approach can be employed for or with similar
a similar
of mechanical or hydraulic parts.
mobile machinery using a bucket for loading bulk material (e.g., wheel loaders) or with a similar
group of mechanical or hydraulic parts.
One
groupelement of theordesign
of mechanical hydrauliciteration,
parts. based on Figure 1, was an integration check, which also
One element of the design iteration, based on Figure 1, was an integration check, which also
included evaluation
One elementofofthe thepracticality of usebased
design iteration, of different
on Figure bucket
1, wassizes
an and possible
integration riskswhich
check, (as introduced
also
included
includedevaluation ofofthe practicality of use
use ofof different
differentbucket
bucket sizes
andand possible risks (as
by the numberevaluation
of collisions)the practicality
associated withofsuch design alternatives. sizes
Productivity possible
was risks (as by
measured
introduced
introduced by the number
by the of
number of collisions) associated with such design alternatives. Productivity was
the amount of sand delivered to collisions)
the designatedassociated with such
destination indesign alternatives.
a limited time span. Productivity
Change in wasbucket
measured
measured by by thethe
amount
amountofofsandsand delivered
delivered to to the
thedesignated
designated destination
destination in ain a limited
limited time time
span. span.
size and attached hydraulics could be adopted based on the preferred strategy by considering budget,
Change
Changein bucket
in bucketsize and
size andattached
attachedhydraulics couldbebeadopted
hydraulics could adopted based
based on the
on the preferred
preferred strategy
strategy by by
safety, or machine performance.
considering budget, safety, or machine
considering budget, safety, or machine performance. performance.
Figure 11 depicts
Figure
thethe
11 depicts
optimization
optimization and
and integrationcheck
and integration
checkinitially
initially introduced
introduced in the
in first
the first chapter
Figure 11 depicts the optimization integration check initially introduced in thechapter
first chapter
on the basis
on the of
basis modeling, customization,
of modeling,customization, interaction,
customization, interaction, and results. The key idea in this method
was was
on the basis of modeling, interaction,and andresults. TheThe
results. keykeyideaidea
in this in method
this method was
communication
communication of information
of informationaboutaboutthethe
effects of of
effects component
componentdesign designdecisions
decisionsvia via anan accurate real-time
accurate real-
communication of information about the effects of component design decisions via an accurate real-
simulation that showed
time simulation the combined
that showed the combinedimpact of parameter
impact of parameter changes
changes(for(forthe
the example
example in inFigure
Figure 11,
time simulation that showed the combined impact of parameter changes (for the example in Figure
11, the first component and second component). Evaluating such multi-dimensional
the first component and second component). Evaluating such multi-dimensional outcomes is extremely outcomes is
11, the first component
extremely challenging andoften
and second component).
impossible using Evaluating
traditional such
design multi-dimensional
approaches alone. outcomes is
challenging and often impossible using traditional design approaches alone.
extremely challenging and often impossible using traditional design approaches alone.

Figure 11. Interaction

Figure among
11. Interaction components
among components in
in modeling basedonon
modeling based accurate
accurate real-time
real-time simulation.
simulation.

This work demonstrated a method for decision making about the size of the bucket and/or sizing
Figure
of the 11. Interaction
arm-attached among
hydraulic components
actuator (also in modeling
known based arm).
as dipper on accurate real-timeapproach
The described simulation.
can
benefit designers when it is used as a tool for product development and design optimization. The
This work demonstrated a method for decision making about the size of the bucket and/or sizing
of the arm-attached hydraulic actuator (also known as dipper arm). The described approach can
Machines 2019, 7, 70 13 of 15

This work demonstrated a method for decision making about the size of the bucket and/or sizing
of the arm-attached hydraulic actuator (also known as dipper arm). The described approach can benefit
Machines 2019, 7, x FOR PEER REVIEW 13 of 15
designers when it is used as a tool for product development and design optimization. The excavator
model was used
excavator as a was
model virtual
usedtest
as bench fortest
a virtual thebench
project. Easy
for the customization
project. was available
Easy customization and allowed
was available
usersand(e.g., end-users as machine operators) to choose and try different components
allowed users (e.g., end-users as machine operators) to choose and try different components of the excavator.
of
In previous methods,
the excavator. as presented
In previous by Zheng
methods, [28] orbyXu
as presented [29], [28]
Zheng evenorthough
Xu [29],the model
even thoughparameters
the model were
partlyparameters
adjustable,werethe partly
options adjustable,
were notthe options
meant were an
to make notinteractive
meant to makemodel anwith
interactive model
operator with
contribution,
operator contribution, and model components were thus fixed from
and model components were thus fixed from the operator’s point of view and the focus wasthe operator’s point of view andon the
the focus was on the compatibility of components in modules. Different aspects
compatibility of components in modules. Different aspects of inter-disciplinary engineering design of inter-disciplinary
engineering design and the data circulation of the design work were addressed in [30] and [31], but
and the data circulation of the design work were addressed in [30] and [31], but user interaction in
user interaction in these works considered user input mainly in the final stages of the design and
these works considered user input mainly in the final stages of the design and product process.
product process.
In the work in this paper, validation of the proposed method was performed using a desktop
In the work in this paper, validation of the proposed method was performed using a desktop
computer with
computer with Logitech
LogitechForce
Force3D3Dpro
projoysticks.
joysticks. A PC with
A PC withananIntel
Intel Core
Core i7-6700
i7-6700 processor
processor equipped
equipped
with with
64 GB 64 of
GBRAMof RAM andandananNVIDIA
NVIDIAQuadro
Quadro M2000
M2000 GPUGPUwas wasable
ableto to
runrun
thethe simulation
simulation with with
a1 a1
millisecond timetime
millisecond step,step,
which permitted
which useruser
permitted interaction as real-time
interaction simulation.
as real-time simulation. AnAnexample
example test screen
test
screen is given in Figure 12. Based on the on-screen images, the excavator operators
is given in Figure 12. Based on the on-screen images, the excavator operators could instantly observe could instantly
observe
the effects of the effects
their of theiron
selections selections
aspects onsuchaspects such as productivity
as productivity (here this(here
wasthis was excavation
excavation cycle and
cycle times)
times) and
fuel consumption. fuel consumption.

Figure 12. Environment for test and validation of the model.

Figure 12. Environment for test and validation of the model.

Test results are are

Test results alsoalso
available totobeberecorded
available recorded and collectedfor
and collected forfurther
further analysis,
analysis, where
where simulation
simulation
repeatability provides
repeatability a basis
provides for for
a basis decision making
decision makingabout
aboutthe
thereal
realproduct
productandand the
the simulated product can
simulated product
be seen as a digital twin. Combining parameterizations of different components,
can be seen as a digital twin. Combining parameterizations of different components, which which is discernable
is
as rapid prototyping
discernable as rapidin designers’
prototypinghands, reduceshands,
in designers’ the time required
reduces for modeling,
the time required forvalidation,
modeling, and
validation,
testing, and testing,
as modular componentas modular
modelingcomponent
is usedmodeling is used for
for the machine to the machine to Additionally,
be designed. be designed. the
Additionally,
proposed approach theisproposed
suitable approach
for use iniscrowdsourcing
suitable for use in crowdsourcing [32].
[32].

5. Conclusions
5. Conclusions
This This paper described an approach for customized modeling for machine design using state-of-
paper described an approach for customized modeling for machine design using
the-art multi-body simulation techniques that made it possible to run a complicated multi-aspect real-
state-of-the-art multi-body simulation techniques that made it possible to run a complicated multi-aspect
time simulation on a normal budget PC. Use of this methodology provides the designer with the data
real-time simulation on a normal budget PC. Use of this methodology provides the designer with the
required to meet customer needs already in the early and middle steps of the design process. To
data illustrate
required application
to meet customer needs already
of the methodology, theinpaper
the early and middle
presented stepsofofthe
an example thedesign
designof process.
an
To illustrate
excavator bucket for bulk material excavation. The proposed approach paves the way for fasterof an
application of the methodology, the paper presented an example of the design
excavator bucket
product for bulk and
development material excavation.
improved The proposed
user satisfaction. approach
User priorities paves
can the
be met inway
a moreforsystematic
faster product
development and improved user satisfaction. User priorities can be met in a more systematic manner
Machines 2019, 7, 70 14 of 15

when the product development process is handled on the basis of user-generated data that do not
require the user to consider complex engineering details.
The proposed method creates a basis for introducing user feedback through an XML-based
multibody model of an excavator, and clients who use the machine as a tool can make a meaningful
contribution to the design when searching for optimum parameters for the specific application and
environment conditions. Environment conditions here were interpreted as soil density, cohesive
modulus, friction modulus, and other terramechanical variables that may affect the excavation process,
such as risk of collision with obstacles.
When interrelated design factors are involved in product design of a complex system, the presented
approach can benefit the designer by enabling better visualization of the impact of changes—through
customized modular parameters—on other components and related processes of the system.

Author Contributions: Draft preparation and investigation, N.A.; review, editing and methodology, E.K.; review,
editing, resources, and supervision, A.M.
Funding: This research received no external funding.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Boschert, S.; Rosen, R. Digital twin—The simulation aspect. In Mechatronic Futures; Hehenberger, P.,
Bradley, D., Eds.; Springer: Cham, Switzerland, 2016; pp. 59–74. [CrossRef]
2. Jahangirian, M.; Eldabi, T.; Naseer, A.; Stergioulas, L.K.; Young, T. Simulation In manufacturing and business:
A review. Eur. J. Oper. Res. 2010, 203, 1–13. [CrossRef]
3. Shen, R.; Zhang, X.; Zhou, C. Dynamic Simulation of the Harvester Boom Cylinder. Machines 2017, 5, 13.
[CrossRef]
4. Cuadrado, J.; Cardenal, J.; Bayo, E. Modeling and solution methods for efficient real-time simulation of
multibody dynamics. Multibody Syst. Dyn. 1997, 1, 259–280. [CrossRef]
5. Tuegel, E.J.; Ingraffea, A.R.; Eason, T.G.; Spottswood, S.M. Reengineering aircraft structural life prediction
using a digital twin. Int. J. Aerosp. Eng. 2011, 1–14. [CrossRef]
6. Becker, M.C.; Salvatore, P.; Zirpoli, F. The impact of virtual simulation tools on problem-solving and new
product development organization. Res. Policy 2005, 34, 1305–1321. [CrossRef]
7. Balakrishnan, A.; Kumara, S.R.T.; Sundaresan, S. Manufacturing In the digital age: Exploiting information
technologies for product realization. Inform. Syst. Front. 1999, 1, 25–50. [CrossRef]
8. Ullman, D.G. A taxonomy for mechanical design. Res. Eng. Des. 1992, 3, 179–189. [CrossRef]
9. Tukker, A.; Tischner, U. Product-services as a research field: Past, present and future. Reflections from a
decade of research. J. Clean. Prod. 2006, 14, 1552–1556. [CrossRef]
10. Tomiyama, T.; Gu, P.; Jin, Y.; Lutters, D.; Kind, C.; Kimura, F. Design methodologies: Industrial and
educational applications. CIRP Ann. 2009, 58, 543–565. [CrossRef]
11. Pahl, G.; Beitz, W.; Feldhusen, J.; Grote, K.H. Engineering Design. A Systematic Approach, 3rd ed.; Springer:
London, UK, 2007.
12. So, J.C.Y.; Proctor, R.W.; Dunston, P.S.; Wang, X. Better retention of skill operating a simulated hydraulic
excavator after part-task than after whole-task training. Hum. Factors 2013, 55, 449–460. [CrossRef]
13. Beuren, F.H.; Ferreira, M.G.G.; Miguel, P.A.C. Product-service systems: A literature review on integrated
products and services. J. Clean. Prod. 2013, 47, 222–231. [CrossRef]
14. Bernold, L.E. Quantitative assessment of backhoe operator skill. J. Constr. Eng. Manag. 2007, 133, 889–899.
[CrossRef]
15. Baxter, D.; Roy, R.; Doultsinou, A.; Gao, J.; Kalta, M. A knowledge management framework to support
product-service systems design. Int. J. Comput. Integr. Manuf. 2009, 22, 1073–1088. [CrossRef]
16. Nemoto, Y.; Akasaka, F.; Shimomura, Y. A framework for managing and utilizing product–service system
design knowledge. Prod. Plan. Control 2015, 26, 14–15. [CrossRef]
17. Shabana, A.A. Computational Dynamics, 3rd ed.; John Wiley & Sons, Ltd.: Chichester, UK, 2010.
18. De Jalón, J.G.; Bayo, E. Kinematic and Dynamic Simulation of Multibody Systems: The Real-Time Challenge;
Springer: New York, NY, USA, 2012.
Machines 2019, 7, 70 15 of 15

19. Jiménez, J.M.; Avello, A.N.; de Jalón, J.G.; Avello, A.L. An efficient implementation of the velocity
transformation method for real-time dynamics with illustrative examples. In Computational Dynamics
in Multibody Systems; Pereira, M.F.O.S., Ambrósio, J.A.C., Eds.; Springer: Dordrecht, The Netherlands, 1995;
pp. 15–35. [CrossRef]
20. Avello, A.; Jiménez, J.M.; Bayo, E.; De Jalón, J.G. A simple and highly parallelizable method for real-time dynamic
simulation based on velocity transformations. Comput. Methods Appl. Mech. Eng. 1993, 107, 313–339. [CrossRef]
21. Baharudin, M.E.; Rouvinen, A.; Korkealaakso, P.; Mikkola, A. Real-time multibody application for tree
harvester truck simulator. Proc. Inst. Mech. Eng. K J. Mul. 2014, 228, 182–198. [CrossRef]
22. Korkealaakso, P.M.; Rouvinen, A.J.; Moisio, S.M.; Peusaari, J.K. Development of a real-time simulation
environment. Multibody Syst. Dyn. 2007, 17, 177–194. [CrossRef]
23. Novak, P.; Guinot, V.; Jeffrey, A.; Reeve, D.E. Hydraulic Modelling: An Introduction, Principles, Methods and
Applications, 1st ed.; CRC Press: London, UK, 2010. [CrossRef]
24. Lai, K.C.; Kang, S.C. Collision detection strategies for virtual construction simulation. Automat. Constr. 2009,
18, 724–736. [CrossRef]
25. Hu, S.; Yu, L. Optimization of collision detection algorithm based on OBB. In Proceedings of the 2010
International Conference on Measuring Technology and Mechatronics Automation, Changsha City, China,
13–14 March 2010; IEEE: Piscataway, NJ, USA, 2010; Volume 2, pp. 853–855. [CrossRef]
26. Negrini, S. On the Modelling of Deformable Tyre on Deformable Soil for Tread Pattern Design Optimization.
Ph.D. Thesis, Politecnico di Milano, Milano, Italy, 26 March 2013.
27. Dai, T.; Wang, Z.; Xu, S. Research of creating and fetching 3D models of virtual reality based on OpenGL.
In Proceedings of the 2006 International Conference on Mechatronics and Automation, Luoyang, China,
25–28 June 2006; IEEE: Piscataway, NJ, USA, 2006; Volume 1, pp. 1991–1995. [CrossRef]
28. Wang, Z.; Liu, Z.; Tan, J.; Fu, Y.; Wan, C. A virtual environment simulator for mechanical system dynamics
with online interactive control. Adv. Eng. Softw. 2006, 37, 631–642. [CrossRef]
29. Xu, Z.; Xi, F.; Liu, L.; Chen, L. A Method for Design of Modular Reconfigurable Machine Tools. Machines
2017, 5, 5. [CrossRef]
30. Melville, C.; Yan, X.-T.; Gu, L. TiV-Model—An Attempt at Breaching the Industry Adoption Barrier for New
Complex System Design Methodologies. In Mechatronic Futures; Hehenberger, P., Bradley, D., Eds.; Springer:
Cham, Switzerland, 2016; pp. 41–57. [CrossRef]
31. Panarotto, M.; Wall, J.; Bertoni, M.; Larsson, T.; Jonsson, P. Value-driven simulation: Thinking together
through simulation in early engineering design. In Proceedings of the 21st International Conference on
Engineering Design ICED17, Vancouver, BC, Canada, 21–25 August 2017; Volume 4, pp. 513–522.
32. Hirth, M.; Hoßfeld, T.; Tran-Gia, P. Analyzing costs and accuracy of validation mechanisms for crowdsourcing
platforms. Math. Comput. Model 2013, 57, 2918–2932. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).