Compatibility issues for mechanical system modelling with standard components
Transcript of Compatibility issues for mechanical system modelling with standard components
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
1/16
Compatibility issues for mechanical system modellingwith standard components
B J Hicks* and S J Culley
Department of Mechanical Engineering, Faculty of Engineering and Design, University of Bath, UK
Abstract: The success of a product and ultimately the associated company are largely dependent on the
ability of the company to generate eVective, high-quality and economical design solutions within
shortened development times. To achieve increased quality and considerable time savings, standard
components can be utilized. In order to establish a suitable solution the designer must evaluate
various component types, sizes and combinations until the desired performance capabilities for the
design are achieved. However, further improvements in economy and performance can be obtained
by the eVective utilization of various congurations of standard components. During this phase ofevaluation, the designer must consider not only system performance but also component issues such
as reliability, suppliers, cost, maintenance and internal practices. The pressure for reduced time to
market does not leave room for time-consuming trial-and-error approaches. Today, experimentally
validated computer modelling has become the preferred choice for rapidly carrying out the
assessment of performance and the evaluation of design alternatives. In using such techniques the
designer also demands the capability to assess whether the connected components are suitable both
quantitatively and qualitatively. The quantitative issues relate to whether the components are
physically connectible and matched in terms of their performance capabilities, while the qualitative
issues relate to the utilization of the component or a sequence of components. These may include
preferences for suppliers, cost considerations, reliability issues, recommended component
combinations and internal practices. These two aspects of the embodiment process are dened as
compatibility analysis in this work.
This paper outlines the process for the embodiment of mechanical systems with standard
components and identies the key aspects of compatibility analysis. Following this, the importance
of and the requirements for compatibility analysis in modelling mechanical systems are discussed
and a strategy to support compatibility analysis within a modelling environment is proposed. This
strategy is implemented into an existing system modelling tool and an example study is included to
illustrate the benets to the designer.
Keywords: standard components, component selection, compatibility analysis, system modelling
1 INTRODUCTION
In todays aggressive global markets, competitive
advantage accrues to those companies that can produce
competitive products. For a product to be truly
competitive it must be of similar quality, must deliver
comparable performance and functionality and must be
of favourable cost. This demands that designers produce
high-quality well-performing design solutions, with both
low production and development costs.
One approach to achieve high-quality cost eVectivengineering is to utilize standard components. Standar
components provide increased quality and performanc
at decreased unit cost. The reasons for this have bee
discussed by Ulrich and Eppinger [1], who state tha
the third-party producer can manufacture high volume
thus reducing unit cost, and invest in learning an
improvement of the components design and productio
process, thereby improving quality and performance
The use of standard components also yields som
indirect benets such as improved availability, glob
acceptance, established maintenance procedures as we
as assured long-term concurrency of components [2The eVective identication and utilization of standar
components can therefore signicantly improve th
2
The MS was received on 20 November 2000 and was accepted afterrevision for publication on 14 August 2001.*Correspondin g author: De partment of Mechanical Engineering, Facult y
f E i i d D i U i i f B h B h BA2 7AY UK
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
2/16
quality and performance of a design solution. In fact,
studies undertaken in the eld of machine systems
design and in particular power transmission have
shown that standard components can provide between
50 and 80 per cent of the components in an assembly
[3]. As global competition continues to rise, the utiliza-
tion of standard components in machine systems is
likely to increase, because of the many advantagespreviously discussed. The development of methods that
deal with the incorporation of standard components
during systems design is therefore an important area,
which is overlooked by many researchers. The majority
of current research deals with the selection of individual
components in isolation [46], rather than considering
the system and each of its elements as a whole. The
incorporation of selection issues into a systems approach
is necessary because of the highly coupled nature of
mechanical components, which must be considered
together w ith qualitative selection issues, in order to
develop an optimum solution [7, 8].
1.1 Standard components and embodiment design
During the transformation of a concept to an embodied
solution the designer must take decisions regarding the
type and combination of mechanical elements that are
necessary in order to perform the desired function of
the system. The relative selection of these components
and their corresponding sizes will determine the overall
performance capabilities and physical characteristics ofthe system. These physical characteristics include
measures such as spatial occupancy, reliability, working
range capabilities, transmission stiVness and importantly
cost, all of which determine the commercial success of the
machine. As a consequence, the designer must consider
all these aspects while embodying a solution and must
select various component types, sizes and congurations
to best achieve all the desired characteristics.
Pressures for reduced time to market do not leave
room for time-consuming trial-and-error approaches.
Today, experimentally validated computer modelling
has become the preferred choice for rapidly carryingout the assessment of performance and the evaluation
of design alternatives. The application of these modelling
techniques are essential where standard components are
to be considered. This is because of the large number of
manufacturers and suppliers, each providing many
diVerent types of component which are available over a
large number of discrete sizes. This makes for an
exponential number of possible congurations for
component types, sizes and combinations, which must
be investigated by the designer in order to determine
the best-performing solution.
For the purpose of this work a component type
represents a specic range of mechanical elements from
ti l f t hil t i l t
to the range of discrete element sizes available for
particular component type. In order to embody fully
conceptual schema the designer must iteratively sele
component types and/or sizes in order to full th
requirements for the design. In addition to fulllin
performance and physical requirements, the designe
must also consider other aspects of the design such a
preferred suppliers, recommended practices, reliabilitcomponent standardization, scalability and rationaliza
tion of designs within the company. These aspects ar
not meant to be exhaustive but merely to illustra
some of the other qualitative considerations that
designer must evaluate. The common approach of th
designer, when embodying a concept, is to select
conguration of components and then to eVect change
in individual component types and sizes, evaluating th
system performance after each change. This i s performe
iteratively in order to achieve an optimum solution no
just in terms of performance capabilities but also again
aspects of the design previously described. Because of thinherently time-consuming nature of these tasks and th
desire to reduce development time a designer may elect t
disregard changing component type in preference of th
pursuit of a single design solution. As a consequence a
optimized solution may only be optimum for the speci
conguration. In addition to this, the designer team ma
also attempt to reduce time and eVort by undertakin
this activity in cooperation with a particular supplie
with whom they have a relationship, although under
standable this may further reduce the likelihood o
developing an optimum solution.
1.2 Computer-based support for selecting systems of
standard components
A computer-based support tool to aid the embodimen
of systems, however dened, needs to address
number of issues. In particular the ability to conside
many alternatives within shortened activity times a
well as the ability to evaluate achievable componen
congurations. To address these issues, provision fo
the identication of incompatible component types anthe suggestion of alternative component types must b
incorporated. In addition to this, the recognition o
acceptable or unacceptable component combination
or chains is demanded, in order to reduce search spac
and search time. For example a novice designer ma
attempt to connect incompatible components togethe
while a more experienced designer may utilize chains o
components that are compatible but not practised fo
reasons such as reliability, stiVness or cost. Therefor
designers need to know whether elements are compatib
and whether the combination of elements that they hav
chosen are best-practised or acceptable alternatives. Fo
the purpose of this work, the term best practised ma
l t t ith i t l t ti ll
236 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
3/16
accepted or documented design practices. For the task of
identifying compatible elements, various levels of
compatibility analysis have been practised for some
time in both electrical and uid power design tools.
These techniques are discussed in later sections with
respect to the generation of a similar support tool for
mechanical systems.
The work described in this paper deals with theembodiment of mechanical systems with standard
components. The paper develops the key aspects of the
system embodiment process and in particular discusses
the areas of compatibility analysis and component
matching. A review of supportive methods for each
phase of the process highlights the lack of support tools
and approaches in this important area. To address this,
the requirements of compatibility analysis for mechanical
systems design with standard components are established
and an approach is created and subsequently incorpo-
rated into an existing system modelling tool [9].
2 SELECTION ISSUES FOR SYSTEMS OF
STANDARD MECHANICAL COMPONENTS
For todays engineer the selection of standard compo-
nents involves the consideration of many more factors
than solely does this component deliver the required
performance? . Because of global markets and global
distribution networks, combined with the vast number
of diVerent suppliers and component types, the design
team must now consider issues covering not only perfor-mance but also aspects such as suppliers, costs, delivery,
reliability, compatibility and managed services to name
but a few. Figure 1 illustrates the key activities identied
by the research work and depicts the individual tasks for
the design of a mechanical system with standard compo-
nents. These activities are discussed in the following
sections and are categorized from studies by the present
authors into machine systems design. The activities can
be separated into four phases: identication and selection
of components, component matching and compatibility
analysis, optimization of sizes and conguration and
detailed design. These phases are separated in the processoverview in Fig. 1; however, there is a high level of
dependence and recursion between each phase and
their included activities. The core tasks or activities in
each of the phases are now discussed and research
methods which support each phase are described.
2.1 Identication and selection of components
During phase 1 of the proposed process the designer is
concerned with identifying suitable component types
and associated manufacturers. To undertake this, a
number of workers have discussed methods for the
i f f ti t t f h i l t
[10, 11], termed functional decomposition; this approac
is also a key feature of the TRIZ (theory of solvin
problems inventively) methodology [12]. For the identi
cation of preferred suppliers or particular manufacturer
Ellram [13] discussed methods for the evaluation o
relationships with suppliers, while Culley and Allen [1
described an approach to capture informal informatio
regarding the eVectiveness and performance of supplieand particular component types.
2.2 Component matching and compatibility analysis
Phase 2 of the process covers many diVerent tasks, whic
are highly dependent on each other and must be consid
ered concurrently in order for the successful completio
of this phase. This dependence is represented in Fig. 1 b
the interconnections between each task. These individu
tasks can be further grouped into four activities: syste
topology or layout, system performance, connectivity an performance matching and complementary assessment o
coupled components, depicted in Fig. 1. For the rst o
these activities there are many tools which conside
geometric relations between parts, these include th
computer aided design (CAD) packages WAVE (pa
of Unigraphics) [15] and Pro Engineer [15, 16], both o
which possess parametric modelling capabilities fo
assemblies.
To support the undertaking of the second activity o
the process, a number of approaches for performanc
modelling and simulation are becoming availablwhich allow the modelling and analysis of system
based on standard components [17, 18]. The
approaches support the conceptual and embodimen
phases of the traditional design process and assist th
designer in rapidly embodying a design solution. Th
embodiment is frequently computer based and attemp
to populate the conceptual conguration with a set o
real mechanical components that meet the desired ove
all performance characteristics. The automation of th
embodiment process enables many more componen
congurations and combinations to be evaluated by th
designer in a relatively short time period.For the third and fourth tasks in this phase of th
proposed process, namely connectivity and performanc
matching and compatibility analysis, there is little docu
mented work, where systems of standard components ar
considered. To summarize these tasks, their primar
function is to ensure that geometric interfaces are com
patible (e.g. this may include evaluation of ts and abilit
to accommodate deections), that the magnitudes o
energy transfer are acceptable and that, where possibl
preferred component types and combination are used
Current practice requires the designer to perform suc
evaluations by hand, which is inherently time consumin
and frustrates the ability to determine an optimum solu
ti f i l di d Th d l
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 2
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
4/16
of methods which deal with this aspect of the embodi-
ment process are therefore essential for the generation
of the most eVective overall solution.
2.3 Optimization of sizes and conguration
Phase 3 of the process involves the optimization of a
particular conguration of components where the
conguration represents the types of component, the
arrangement and their preliminary sizes. The optimiza
tion strategies applied to such problems typically var
component attributes, evaluating system propertie
after each iteration, in order to converge on a solutio
which best achieves the desired attributes for thconsidered system. The examples of optimization goa
detailed in the lower portion of Fig 1 merely represen
Fig. 1 Design process for the incorporation of s tandard components
238 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
5/16
some typical objectives [7, 19, 20]. The actual goal
functions must be generated by the designer and will
depend on the original design specication. Extensive
research has been undertaken into developing optimiza-
tion strategies for searching both continuous and discrete
solution spaces [2123] and, although these have not
been used extensively in these applications, they are
considered suitable for problems involving standardcomponents which produce discrete solution spaces.
2.4 Detailed design
The nal phase of the process, phase 4, includes many of
the typical tasks for detailed design [24]. These may
include the specication of the intricacies for, say,
component housings, mountings and xtures. This
phase may also include design for X-type activities [25]
although, if diVerent component sizes are considered,
the designer may need to re-enter phases 2 and 3 of theprocess. Traditional CAD systems [2628] have been
developed for this phase of the design process. The
majority of these systems enable the geometry of
standard components in the form of two-dimensional
and three-dimensional models to be imported from
libraries, often supplied by manufacturer s or included
in electronic catalogues [29]. These libraries also include
mountings and standard xings, removing much of the
detailed work.
2.5 Overview
The previous sections discuss the individual phases of the
proposed design process for embodying mechanical
systems with standard components. These sections also
provide a brief overview of current technologies and soft-
ware tools which support the various activities. This
review highlights the fact that there is a lack of suppor-
tive methods for phase 2 of the process and in particular
what the present authors have identied as compatibility
analysis. Furthermore, the overall process depicted in
Fig. 1 shows that there is a high dependence between
each of the phases. This demands that the various activ-ities are considered concurrently in order to undertake
the process eVectively. To address these important
issues the following sections discuss and develop the
requirements for compatibility analysis in mechanical
systems and review similar approaches adopted in
other engineering domains, namely uid power and
electrical circuit design. The developed approach is
then integrated into an existing system modelling envir-
onment for the design of mechanical systems from
standard components. The incorporation of compatibil-
ity analysis into a modelling approach enables all the
four phases of the embodiment process to be undertaken
in a single environment. The collective consideration of
t f t d t l ti
issues are essential for the eVective and successfu
selection and specication of a system of standar
components.
3 REVIEW OF COMPATIBILITY ANALYSIS IN
ENGINEERING
As has been stated, the desire to reduce developmen
times and the commitment of nal cost obligated at th
conceptual phase of the design process and the abilit
to test for achievable designs as early as possib
becomes increasingly more important. Consequently
the integration of compatibility analysis routines int
software based design environments is beginning t
evolve. Such environments range from the conguratio
of personal computer hardware [30] to the compatibilit
of information systems [31]. For all these variou
applications, compatible elements are dened as thos
elements which can be used together without modication or adaptation [32] and do not produce any physic
eVects that may adversely aVect the performance of th
system considered. In the domains of uid power system
design and electrical circuit design, compatibility analy
sis techniques have been successfully incorporated int
a number of support tools. In the electrical domai
[33], compatibility analysis deals with electromagnet
compatibility utilizing complex numerical methods t
model the behaviour of elds. In contrast with thi
work in the design of uid power systems focuses o
the compatibility of component models, with respect tthe matching of input and output parameters as well a
the continuity of units of measurement [34].
In machine systems design, the present authors con
sider compatible designs to be those that are achievabl
and consist of preferred or best practiced congura
tions of components. Achievable designs in this contex
are those that are fundamentally based on existin
technology and principles. Such designs incorporat
standard components or standard designed componen
and are congured so that components are physical
connectible and mutually complement each other, i
terms of the type(s) and magnitudes of their outpuand inputs. As the number of elements that viola
these conditions increases, so the condence in th
achievability of the design reduces. Preferred design
utilize congurations and combinations of componen
that are deemed to be suitable or better performing fo
the particular application. Reasons for this perceive
better performance range from cost concession
provided by certain suppliers, a drive to use existin
stocks and the use of well-tested and reliable componen
combinations, an important consideration in aero
nautical disciplines [35]. As the number of preferre
components are reduced, the subjective or qualitativ
attributes, such as reliability or eciency, become le
t i d i i l t ti d i t l lid ti
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 2
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
6/16
must be undertaken, especially where performance is
critical to safety.
For the purpose of machine systems design the
objectives of compatibility analysis as dened above
are concerned with the assessment of connected compo-
nents rather than system eVects, such as the elds
produced by electrical circuits. Hence an elemental
approach is all that is required. This elemental approachevaluates connected components and sequences of
coupled components throughout the system. This is
similar to the approach adopted in the uid power
domain, where attributes between coupled components
are evaluated. However, modelling environments in the
uid power domain deal only with owrate, pressure
and consistency of units. In order to develop an elemen-
tal approach for compatibility analysis of machine
system models a number of issues must be addressed.
These are discussed in the next sections.
4 COMPATIBILITY ANALYSIS IN
MECHANICAL SYSTEMS
For the purpose of this work, compatible elements are
dened as those which are capable of being used in com-
bination and are well suited. This latter aspect ensures
that coupled components are matched in terms of their
capabilities and do not impose any adverse eVects on
to each other. In order to evaluate the compatibility of
a system of mechanical elements the present authors pro-
pose three activities: connectivity analysis, performancematching and complementary assessment, illustrated in
Fig. 2.
1. Connectivity analysis ensures that the geometric
interfaces are matched and that energy interfaces
are compatible [27]. This ensures that coupled
components t together and that energy can be trans-
mitted in the desired manner across their interface.
An example of the importance of this latter condition
is for the selection of a bearing on a shaft. If the shaft
transmits an axial load and a radial load, then, while
many bearings may be geometrically compatible, on
bearings that support both loading conditions a
suitable for the particular design.
2. Performance matching ensures that the magnitudes o
energy transfer are acceptable output and input l eve
for coupled components. This matching may not b
exact; the designer may wish to include safety facto
or additional capacity for variations in performancconditions. In addition to this, performance matchin
encompasses component attributes that are no
exchanged at the interface. These may include lubr
cation type, material properties, working tempera
ture, life and reliability.
3. Complementary assessment provides for the qual
tative considerations which the designer mu
undertake. These may include cost, reliability, manu
facturer and supplier issues and recommended o
standard practices. The objective of complementar
assessment is to identify particular components o
combinations which are preferred for one or moof the reasons previously discussed. An example o
complementary components is in the case of th
attachment of a gear to a lay shaft. It may be tha
the manufacturer recommends the attachment o
the gear with a bush rather than a keyway. While
keyway is a compatible element, it would not b
considered complementary or preferred.
As previously discussed in Section 2, there is a hig
level of dependence and recursion between the variou
stages of phases 2 and 3 of the proposed design proces
(Fig. 1). Because of this dependence, supportive methodand tools for each of the respective tasks within thes
phases (geometric and performance consideration
component matching, complementary evaluation an
optimization) must be considered collectively. Th
really can only be achieved within a single modellin
environment or by linking a range of computer-base
design tools. Research into an integrated modellin
environment for mechanical systems has been unde
taken by the present authors [9, 36]. This approac
provides the representation, performance modellin
Fi 2 Th d f ti f tibilit l i i h i l t
240 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
7/16
and optimization of mechanical systems from compo-
nent-based models. These component-based models are
the electronic procedures that either select or specify a
particular mechanical component, ranging from bearing
and electric motor catalogues to parametric models for
gear pairs and for shafts. The following sections describe
an approach for incorporating the three phases of
compatibility analysis into an integrated modellingenvironment.
5 COMPATIBILITY ANALYSIS IN AN
INTEGRATED MODELLING ENVIRONMENT
In order to provide for verication of compatible ele-
ments as dened above, the approach of modelling
tools in the uid power domain is to inspect the input
and output parameter elds of connected component
models. This inspection ensures that there is an exhaus-
tive and exact match of output parameters to inputparameters for the coupled components. The introduc-
tion of such a scheme into a mechanical systems support
tool is problematic, because components may demand a
range of design parameters that can be fully met by all or
only some of the available output parameters from other
compatible components. An example of this is in the case
of a shaft and bearing. Here the design data referring to
the power and torque transmitted by the shaft are
redundant, i.e. they are not required for execution of
the bearing model, in this case an electronic catalogue.
Consequently, the design parameters or input param-eters for a bearing are fully met by only some of the
possible output parameters from the shaft. Hence,
there is not an exact and exhaustive match of output to
input design parameters between two compatible
components in machine systems design. Therefor
compatibility cannot be determined through exa
matching of design parameters between two mechanic
components, or indeed partial matching, which woul
be adequate in the case previously described. Howeve
if the partial matching approach were adopted and
bearing was considered compatible with a shaft on th
basis that all the bearings input parameters can be mby only a proportion of the shafts output parameter
then as a consequence the bearing would be unable t
distinguish between a shaft and a gear which outpu
essentially the same parameter elds [36].
The fact that the compatibility of two connected el
ments in the mechanical domain cannot be determine
by purely whether they have exact, similar or parti
correlation between parameter input and output requir
ments is one of the challenges addressed by the work. I
order to distinguish between compatible and incompat
ble elements, permissible component combinations mu
be explicitly dened within the modelling environmenThis approach is further frustrated by the inherent com
plexity of a mechanical systems representation [37]. Th
representation for a mechanical system is a comple
network of component chains linked together by cor
elements such as a shaft, shown in Fig. 3. This form o
representation has been developed for the integrate
modelling environment in order to capture system
connectivity and to propagate information throughou
the system model.
The requirements for compatibility analysis describe
in earlier sections discuss the need for support during thconguration and construction of a design solution an
explicit evaluation of the system for compatibility an
matching once constructed. In order to provide for th
latter, requirement analysis routines must be execute
Fi 3 O d f l ti d d t ti d i t l ti
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 24
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
8/16
during the system resolution process. To develop the
requirements for this, it is rst necessary to understand
the protocol for resolution, which entails the propaga-
tion of design parameters between models and the
execution of these models.
5.1 Data propagatio n in the integrated modellingenvironment
The integrated modeller used for this work adopts an
elemental approach, where each element is a constituent
of the system model. These elements can represent
mechanical components, parts such as mounts and
system inputs outputs. The data propagation cycle
implemented in the modelling environment has two
distinct phases: rstly, data are propagated inwards
from the outlying elements, referred to as unitary ele-
ments, to the core elements; following this the outward
resolution phase propagates data from the core elementsto the unitary elements. Unitary elements possess only a
single connection and are the outlying system elements
such as inputs, outputs and ground points. Core ele-
ments possess more than two connections and are the
principal elements through which all other components
or component chains are connected, such as a shaft.
These elements demand full data propagation from all
the connected elements in order that all the inputs are
available prior to the model execution. The two-stage
propagation cycle is therefore necessary in order for
component data elds to be fully populated and for thesystem model to be successfully resolved. However, the
bidirectional nature of this cycle also complicates
the application of compatibility analysis to the system
model order of resolution, as shown in Fig. 4. During
the inward phase of resolution the bearing is coupled
to the shaft, while during the outward phase the shaft
is coupled to the bearing. Permitting a bearing to
shaft coupling or vice versa would satisfy both these
conditions. However, such a condition would also b
satised in the case of the component conguratio
shown in Fig. 4c, where the bearing is connected i
between two shafts, which is incompatible. Consequent
the explicit denition or specication of permissib
component combinations for both the inward and th
outward resolution phases is necessary. To provide fo
the specication of compatible components aknowledg
base is implemented within the modelling environmen
This knowledge base can be interrogated during th
model construction phase as well as being explicit
searched during resolution. The architecture for th
knowledge base and its functionality are discussed i
detail in the next section.
6 COMPATIBILITY KNOWLEDGE BASE
The requirements for the compatibility analysis o
mechanical components within an integrated modellinenvironment were introduced in the previous section
In order to full these requirements a knowledge bas
of some form is required. A knowledge base is dene
as a formal and explicit representation of the knowledg
pertaining to a given domain [38]. In this work th
knowledge needs to describe the compatibility of compo
nent combinations. The structure developed for th
knowledge base is a multi-dimensional matrix, depicte
in Fig. 5. This matrix provides for compatibility analys
during the bidirectional phases of system resolution, a
well as providing for the distinction between compatiblincompatible and complementary component combina
tions. This distinction is achieved through the adoptio
of a weighted compatibility identier which can assum
a value of 0, 1 or 2. An identier is specied for eac
possible component combination. A value of 0 denote
an incompatible pair while a value of 1 indicates
compatible pair and 2 denotes a component pairin
that is both compatible and complimentary.
Fi 4 O d f l ti th h ibl t
242 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
9/16
6.1 A two-dimensional matrix for compatibility analysis
In order to provide for compatibility analysis during the
two phases of resolution a two-dimensional compatibil-
ity matrix can be incorporated into the architecture of
the modelling environment. The incorporation of the
knowledge base enables the customization of compatibil-
ity matrices by the designer to accommodate their indivi-
dual preferences for manufacturers or suppliers. During
each phase of resolution the matrix can be concurrentlyreferred to in order to evaluate component combina-
tions. To account for the change in order of data
propagation and resolution, the search order for the
matrix is transposed for each phase of resolution. For
the inward phase of the resolution cycle the search
order evaluates the ij dimensions respectively; this is
inverted for the outward phase to the ji dimensions,
shown in Fig. 5. This example matrix contains seven
component types and demonstrates the use of the com-
plementary weighting between various types of generic
component, such as bearings. In the case of a bearing
on a shaft, bearing 2 is the preferred choice for reasonspreviously discussed, although bearing 1 is still a
compatible alternative. The matrix system therefore
provides for the inclusion of this preference, although
the system does not capture the reasons (intent) for this
precedence. Possible reasons for this may be associated
with reliability, cost or favoured suppliers and aVords
an area for future work.
6.2 A three-dimensional matrix for compatibility
analysis
The two-dimensional matrix described above addresses
tibilit i b t t i h
it does not suggest compatible structures, although it wi
evaluate structures by inspecting consecutive componen
pairings. The ability to suggest component structures is
desirable feature for any design support tool. Such
feature enables a designer who does not possess an
knowledge of preferred components or suppliers t
congure a system that consists of preferred o
recommended components. By adding a virtual thir
dimension to the matrix, permissible chains or structure
may be represented and displayed to the designer. Ththird dimension is achieved by searching consecutiv
instances of the compatibility matrix, illustrated i
Fig. 6. If the designer selects component type A from
instance (i) of the matrix then evaluation of instanc
(i 1) for component type A will generate a list of a
possible compatible components. This approach can b
extended to inspect instance (i 2) for each componen
derived form instance (i 1). In this m anner, all possib
component sequences can be displayed to the designer a
a hierarchical tree structure, shown in Fig. 6 and for th
case example shown later in Fig. 9. As the designer pro
ceeds through the design space committing to particula
components, so the number of feasible combinations
reduced. This approach can be used to identify a
compatible components or limited to complementar
(preferred) components only, thereby ensuring a
achievable design solution that takes account of th
preferences of the respective design team or company.
6.3 Dynamic manipulation of the compatibility
knowledge base
The incorporation of a two-dimensional matrix and a
extended three-dimensional matrix provide for th
l ti f tibl d f d t
Fig. 5 A two-dimensional compatibility matrix
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 24
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
10/16
well as allowing the identication and suggestion of
achievable components sequences. This approach is
adequate for compatibility decisions undertaken at a
component level and explicitly dened in the matrix;
however, as discussed previously in Section 3, decisions
on compatibility and matching may depend on the
inspection of individual component parameters. This
parameter or attribute dependence means that certain
components may only be compatible given acceptable
coupling conditions, in terms of both types and magni-
tudes of coupling parameters. An example of such acoupling parameter is between a shaft and a roller bear-
ing. In this case, all types of roller bearing from diVerent
manufacturer s and suppliers are compatible elements for
a shaft. However, some may not support axial loads or
shaft deections beyond a certain value. Consequently,
components may only be considered to be well matched
if corresponding parameters between them are within
certain ranges or exactly matched. These exact matches
may be numerical values as well as textual properties,
such as lubrication type or material.
The provision for both the specication and the
inclusion of compatibility conditions is achieved bymanipulation of the compatibility matrix. The procedure
for this manipulation is shown in Fig 7 The designers
input to the process is to construct conditions or equa
ities between attributes for related components o
between a predened value and a particular attribut
These conditions are evaluated and the matrix elemen
are updated to reect the change in condition. Th
evaluation and manipulation of the matrix occur afte
the system has been resolved, at which stage all com
ponent models have been interrogated and comple
specications for components determined. This enable
the evaluation of parameters for individual componen
or between component groups against the predeneconditions set by the designer to be made. In th
manner, any component(s) that does (do) not satisf
the matching conditions is (are) identied and th
designer is prompted to take the appropriate action.
7 AN INTEGRATED MODELLING
ENVIRONMENT WITH COMPATIBILITY
ANALYSIS
The integrated modelling environment used as a basis fo
this work has been discussed in detail elsewhere [9, 337]. The environment allows the representation o
engineering systems for their embodiment with standar
Fig. 6 A virtual three-dimensional compatibility matrix
244 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
11/16
components. The approach expedites the concept to
embodiment phase of the design process, enabling a
solution principle, termed a conceptual schema in this
work, to be entered into the system. This schema
describes the class of mechanical component, such as a
gear, a shaft or a bearing, and their arrangement in
terms of connectivity, similar to a concept sketch. Forthe example considered, the conceptual sketch and the
associated schema are depicted in Figs 8a and b respec-
tively. In addition to describing the class of mechanical
component, each icon also has an electronic representa-
tion assigned to it. This representation is assigned by
the designer and may be any one of the available
models for the particular class of component. The elec-
tronic representations range from electronic catalogues,
numerical codes, spreadsheets and databases to models
constructed in CAD systems. In the case of a bearing
this electronic representation may be an electronic
catalogue from any of the included manufacturers orsuppliers. It is at this stage that the designer commits
to a particular supplier or component type, where this
component type represents a specic range of mechanical
elements from a particular manufacturer. It is during
the construction of the schema that the rst level of
compatibility analysis can be invoked. This supports
the selection of component types, sequences and impor-
tantly component manufacturers or suppliers. To
achieve this, the virtual three-dimensional matrix is
utilized. This provides a mapping of the possible compo-
nents sequences, which can be coupled to a particular
component and suggests compatible elements and/or
preferred elements. For the example considered in
Fi 8 th tibilit t i i i k d d i
model construction for the shaft and the keyway. Th
process automatically references the compatibilit
matrix and produces a dendritic structure, whic
describes the various component sequences that a
capable of being coupled to each element. For the cas
of the shaft and the keyway, the possible componen
arrangements are shown in Figs 9a and b respectively.The second level of compatibility analysis occurs afte
the system has been resolved. This resolution consider
the performance of the system and interrogates th
various electronic representations to determine
system of components that are capable of deliverin
the required performance for the design (see Fig.
and the activity of performance modelling for th
system. This level of compatibility analysis involves th
dynamic manipulation of the matrix. During th
process, certain attributes between related componen
are evaluated. If conditions are not satised, the
the component pairing is deemed incompatible and thdesigner is notied. An example of this might be th
shaft deection and the maximum allowable angle o
misalignment for a bearing. The angle of misalignmen
for a given bearing tends to be range specic; i.e. th
value of the attribute is constant across the give
range. Consequently, static attributes such as these ar
evaluated after the system resolution phase, at whic
stage initial values for the various components an
their attributes are available for comparative assessmen
These conditions will have been specied in advanc
by the designer by virtue of the constraint specier
depicted in Fig. 10. Constraints between elements ar
specied as logical conditions ( and ), which ar
l t d t l ti Th diti ll
Fig. 7 Dynamic matrix manipulation by inspection of component parameters
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 24
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
12/16
expressed as numerical equalities, although stringcomparisons can be evaluated, but this is restricted to
exact matching only. This feature can be set to ensure
that attributes such as lubrication types or materials
are consistent throughout a design. If the constraints
are not satised, the designer is notied and prompted
to take the appropriate action, in order to resolve the
conict and to satisfy the constraint.
8 CONCLUSION
This paper discusses a design process for the embodiment
of systems with standard mechanical components. This
i d l d f t di f d i i t
from standard components and separates the overaprocess into three activities: selection of componen
types, performance modelling, component matching an
compatibility analysis and detailed design. The issue
associated with this embodiment process are describe
and the lack of supportive methods and tools currentl
available for the important area of performance mode
ling, compatibility analysis and matching of standar
components during the early stages of machine desig
is highlighted. The work also emphasizes the importanc
of considering all the various aspects of embodimen
design within a single modelling environment. Th
work describes an integrated modelling environmen
for the performance modelling of mechanical system
d di th i t f th i ti
Fig. 8 A connectivity model of a concept principle in the integrated modeller
246 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
13/16
Fig. 9 Compatibility analysis during model construction
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 24
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
14/16
compatibility analysis in the analysis of such systems.These requirements cover three aspects: connectivity
analysis, performance matching and complementary
assessment. A review of compatibility analysis techniques
in other engineering domains is also undertaken and the
associated issues for the generation of compatibility
analysis in mechanical systems are discussed.
A strategy for compatibility analysis within an inte-
grated modelling environment is then developed. This
strategy provides for the distinction of compatible and
incompatible components as well as for the utilization
of preferred combinations of components, such as
shafts and particular bearings or gears and keyways.The incorporation of compatibility analysis ensures
that achievable designs are congured and through
complementary evaluation makes certain that preferred
components are utilized where possible. Achievable
designs are those which are congured from components
that can be procured exactly as specied; usually these
will be standard components and are matched such
that all components are physically connectible and
mutually complement each other in terms of the type(s)
and magnitudes of their inputs and outputs. This ensures
that the design can be produced and that it will operate
within the performance capabilities for which it was
designed. The ability to congure achievable systems
ithi d lli i t i i l bl f th
early stages of system design and for rapid developmenFurthermore, for complex large systems consisting o
hundreds of components, this type of assessment
essential.
REFERENCES
1 Ulrich, K. T. and Eppinger, S. D. Product Design an
Development, 2nd edition, 2000 (Irwin McGraw-Hill, Ne
York).
2 Wallace, A. P. The m odelling of engineering assembli
based on standard catalogue components. MPhil thesiUniversity of Bath, 1995.
3 Theobald, G. Modelling standard components for enginee
ing systems. PhD thesis, University of Bath, 1995.
4 Webber, S. J. Techniques for the analysis of search resul
in computer aided selection systems. PhD thesis, Universi
of Bath, 1994.
5 Reinemuth, J. Designing with computer-based compone
catalogues using hypermedia. In Proceedings of the Inte
national Conference on Engineering Design, The Hagu
August 1993, Vol. 3, pp. 17001707.
6 Harmer, Q. J., Weaver, P. M. and Wallace, K. M. Desig
led component selection. Computer Aided Des., 1998, 30(5
391405.
7 Papalambros, P. Y. and Wilde, D. J. Principles of Optim
D i 1988 ( C b id U i it P C b id )
Fig. 10 Specication of compatibility constraints between related attributes
248 B J HICKS AND S J CULLEY
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
15/16
8 Flood, R. L. and Carson, E. R. Dealing with Complexity: An
Introduction to the Theory and Application of Systems
Science, 1988 (Plenum, New York).
9 Hicks, B. J. and Culley, S. J. An integrated modelling
environment for the design and synthesis of mechanical
systems. J. Computer Aided Des., 2000 (accepted).
10 Counsell, J., Porter, I., Dawson, D. and DuVy, M. Scheme-
builder: computer aided knowledge based design of mecha-
tronic systems. Assembly Automn, 1999, 19(2), 129138.
11 Hundal, M. S. and Langholtz, L. D. Conceptual design by
computer-aided creation of function structures and the
search for solutions. J. Eng. Des., 1992, 3(2).
12 The TRIZ Journal. http://www.triz-journal.com, 2000.
13 Ellram, L. S. The supplier selection decision in strategic
partnerships. J. Purchasing Mater. Managmnt, 1990,
32(18), 814.
14 Culley, S. J. and Allen, R. D. Informal information
denitions and examples with reference to the electronic
catalogue. In Proceedings of the International Conference
on Engineering Design (ICED 99), Munich, Germany,
1999, pp. 19611964.
15 UG WAVE. http://www.ugsolutions.com/products/
unigraphics/cad/wave/wavecontrol, 2000.
16 Pro/ENGINEER solutions. http://www.ptc.com/products/
proe/index.htm, 2000.
17 AMESIM. http://www.amesim.com/amesim.htm, 2000.
18 MSC software, working model. http://www.workingmodel.
com, 2001.
19 Adeli, H. Advances in Design Optimization, 1994 (Chapman
and Hall, London).
20 Samuel, A. and Weir, J. Introduction to Engineering Design:
Modelling, Synthesis and Problem Solving Strategies, 1999
(Butterworth Heinemann, Oxford).
21 Walsh, G. R. Methods of Optimization, 1975 (John Wiley,London).
22 Sen, P. and Yang, J. Multiple Criteria Decision Support in
Engineering Design, 1998 (Springer-Verlag, New York).
23 Hajela, P. Nongradient methods in multidisciplinary design
optimization status and potential. J. Aircr. Des., January
1999, 36(1), 255265.
24 Pahl, G. and Beitz, W. Engineering Design: A Systematic
Approach (Transl. K. Wallace), 2nd edition, 1996
(Springer-Verlag, London).
25 Ullman, D. G. The Mechanical Design Process, 199
(McGraw-Hill, New York).
26 AutoCAD 14 Users Guide, 1997 (Autodesk, USA).
27 Solidworks 98 Plus Users Guide, 1998 (SolidWor
Corporation, Concord, USA).
28 Introducing I-DEAS. http://sdrc.com/ideas/index.htm
2000.
29 Culley, S. J. and Webber, S. J. Implementation requir
ments for electronic standard component catalogue
Procs Instn Mech. Engrs, Part B, Journal of Engineerin
Manufacture, 1992, 206(B4), 253260.
30 Coleman, J. M., McGuV, P. C. and Koederitz, W. L. P
compatibilityis there such a thing anymore? In Proceed
ings of the Petroleum Computer Conference, 1996, p
129134.
31 Mailoi, C., Penzo, W., Sola, S. and Vitali, F. Using
reference model for information system compatibility. I
Proceedings of the Hawaii Conference on Systems Scienc
1994, Vol. 3.
32 Collins Concise Dictionary, 1990 (Collins, Glasgow).
33 Parry, J., Bailey, C. and Aldham, C. Multiphysi
modelling for electronic design. In Proceedings of th
Seventh Intersociety Conference on Thermal and Therm
mechanical Phenomena in Electronic Systems (Ither
2000), 2000.
34 Sidders, J. A., Tilley, D. G. and Chapple, P. J. Therma
hydraulic performance prediction in uid power system
Proc. Instn Mech. Engrs, Part I, Journal of Systems an
Control Engineering, 1996, 210(I4), 231242.
35 Keynote lecture. In Aerospace Transmission Systems
Concurrent Design and Manufacture, IMechE Semin
Publication, 1998.
36 Hicks, B. J. and Culley, S. J. A protocol for communic
tion in a component based modelling infrastructure. ProInstn Mech. Engrs, Part B, Journal of Engineering Manufa
ture, 2001, 215(B4), 453464.
37 Hicks, B. J. and Culley, S. J. Understanding engineerin
assemblies and their individual elements for an integrate
modelling environment. In Proceedings of the Internation
Conference on Engineering Design (ICED 99), Munic
Germany, 1999, pp. 11131116.
38 McMahon, C. and Browne, J. CAD/CAM form principles
practice, 1993 (Addison-Wesley, Wokingham, Surrey).
COMPATIBILITY ISSUES FOR MECHANICAL SYSTEM MODELLING 24
-
8/14/2019 Compatibility issues for mechanical system modelling with standard components
16/16