Target-oriented modularization - Addressing sustainability design goals in product modularization [original]

Procedia CIRP 29 ( 2015 ) 603 – 608

Available online at www.sciencedirect.com

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the scientifi c committee of The 22nd CIRP conference on Life Cycle Engineering

doi: 10.1016/j.procir.2015.02.166

ScienceDirect

The 22nd CIRP conference on Life Cycle Engineering

Target-oriented Modularization–

Addressing Sustainability Design Goals in Product Modularization

Friedrich A. Halstenberga*, Tom Bucherta, Jérémy Bonvoisina, Kai Lindowa, b Rainer Starka, b *

aTechnische Universität Berlin, Office PTZ4, Pascalstr. 8-9, 10707 Berlin, Germany

bFraunhofer Institute Production Systems and Design Technology, Pascalstr. 8-9, 10707 Berlin, Germany

* Corresponding author. Tel.: 49 (0)30 39006 358; fax: +49 (0)30 3930246. E-mail address: friedrich.halstenberg@campus.tu-berlin.de

Abstract

Through modularization, a large range of sustainability goals can be addressed in design, e.g. environmentally friendly end-of-life or improved

MRO (maintenance, repair and overhaul) processes. The development of methods for product modularization raised increasing interest in

recent years. However, published methods for product modularization still lack of flexibility and standardization. Numerous methods have been

developed that are defined for one or a given list of design goals. As a result, it is still difficult for engineers to find and apply the right method

for a defined set of design goals. In this paper, the field of modular product design methods has been analyzed with the aim to develop a Target-

oriented Modularization Method that allows defining modular product structure according to user-defined design goals. The introduced method

is demonstrated on the example of a Garrett GT2860R turbocharger.

Peer-review under responsibility of the International Scientific Committee of the Conference “22nd CIRP conference on Life Cycle

Engineering.

Keywords: product modularity; modular product design; sustainable design; product design; design methods

1. Introduction

In order to cope with the challenge of creating sustainable

value without comprising traditional success factors such as

time to market, cost and quality, new solutions for virtual

product creation are needed [1]. In broadest terms, product

modularization represents an approach for organizing

complex products and processes efficiently, by decomposing

complex tasks into simpler ones [2]. The scheme according to

which product components and functions are arranged into

chunks or modules and by which they interact with each other

are defined in the product architecture [3]. The choice of

product architecture has significant effect on the further steps

of the product development process and on the whole product

lifecycle. For complex products like automobiles or airplanes,

several alternatively/equally relevant product architectures

may compete, which makes the definition of the product

architecture a complex yet critical task in product

development.

The efficiency of a product architecture varies depending on

the goals pursued in the product development process.

Consequently, a challenge for product design teams consists

in disposing of the relevant criteria for clustering product

components and functions into modules according to a given

set of design goals. Various measures have been identified

that allow defining these design goals concretely and support

the process of grouping elements into modules and defining

interfaces [4]. Also, numerous modularization methods have

been developed which use these measures in step-by-step

procedures. However, each one of these methods has been

developed for one or a defined list of design goals, such as

mass customization or reduction of development time. This

constitutes a limitation in the support these methods can offer

to the use of product modularization in the development of

sustainable products. In reaction to this, the present article

introduces a generic approach where methodological aspects

(how to implement modularization?) are separated to

motivational aspects (what are we implementing

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the scientifi c committee of The 22nd CIRP conference on Life Cycle Engineering

604 Friedrich A. Halstenberg et al. / Procedia CIRP 29 ( 2015 ) 603 – 608

modularization for?). It introduces the concept of a Target-

oriented Modularization Method (TOMM) which can be

implemented regardless of user-defined design goals. It

allows product design teams to pursue their own goals already

in product modularization, enabling the consideration of life-

cycle and sustainability issues.

2. State of the art in modular product design

Three essential aspects of product modularization are

covered in this section: Goals, which can be achieved through

product modularization, measures which have been used to

translate goals into practice and methods that use these

measures and assist the product modularization process.

2.1. Achieving design goals through product modularization

Modularization has been found to support a broad range of

design goals. Amongst academic literature, some authors

introduced the concept of module drivers, defined as design

goals product modularization can contribute to (e.g. [5, 3]).

The provided lists of drivers, however, are not exhaustive and

may vary between authors. Other authors describe similar

advantages of product modularization and label them

differently

Several authors point out that product modularization has a

positive effect on product variety for mass customization (e.g.

[6, 7, 8]). It allows decreasing internal variety by

standardizing parts in a mass customized product [9].

Breaking down product complexity in order to facilitate

design tasks and reduce development time through enabling

parallel development, shorter time-to-market at lower cost, is

another cited approach [10]. Leveraging postponement and

delayed differentiation through product modularization is also

expected to reduce production costs (e.g. [11]). The reduction

of interface complexity may simplify communication between

development teams facilitates design changes (e.g. [12]).

Next to classical cost and time driven design goals,

researchers also point out that product modularization can

contribute to addressing sustainability design goals. Product

modularization is an important factor when it comes to

product maintenance allowing separate diagnosis of product

components and isolation of wear parts [13]. Modularity also

fosters upgrade, adaptation and modification of products or

components for an extended service life that may result in a

reduction of environmental load [7]. As modular design

influences the disassembility of a product, it indirectly

influences the treatments potentially applicable at its end-of-

life and may help reducing its environmental impact (e.g.

[14]).

2.2. Modularization Measures

Researchers have described metrics which intend to measure

to what degree components should be clustered in the same

module. Gershenson et al. suggested that the affinity of

components to be grouped together can be expressed through

the generic properties of independence and similarity – two

properties that can be measured for each pair of functional

carriers within a product. Depending on the desired goal, the

generic properties can be instantiated through more specific

measures [15].

Independence is described as the measure of relations among

components inside a module versus relation between

components outside a module [16]. In other words,

independence between modules means that changing the

design of a component in a module has a minimal effect on

other modules. Different instantiations of the concept of

independence measure have been introduced: component

position pattern [17], assembly dependency [16], accessibility

[18], cost of reusability [19], interface openness [20] and

interface design effort [21].

Similarity is used to denote resemblance in processing or the

ability to be processed in a similar way [16]. The literature

provides different metrics expressing similarity in modular

product design, including assembly process similarity [16],

maintenance frequency [13], component connection pattern

[17], post life intent [22] and production cost [18].

2.3. Methods for product modularization

Methods in product modularization can be divided into two

different groups. The first group of methods aims for the

modularization of one product. Here, a single decomposition

is conducted and a single product architecture is created. The

second group consists of methods for product family

modularization. These methods decompose multiple

individual products and aggregate the elements to a family

product architecture.

Methods for single product modularization

Pahl and Beitz integrate their modularization procedure in

their generic view of the product development process [23].

The detailed approach stretches from product planning, where

customer needs are identified and requirements derived to the

definition of the product architecture and the specification of

interfaces in embodiment design. Pimmler and Eppinger

suggest a less detailed procedure which focuses primarily on

the tasks of product modularization, integrating the usage of

the interaction matrix and a clustering algorithm [24]. Lange

and Imsdahl build on the previous work by Erixon [25]. Their

approach ranges from the clarification of customer

requirements to Design for X, utilizing a range of different

matrices [5, 25]. Lanner and Malmquist leave out activities

for requirements definition and start with the establishment of

an organ structure. Alternative proposals for the product

architecture are generated through the Interaction and Lanner

Matrix [26]. Kusiak and Huang present a method for the

design of modular products for testability in the presence of

testing modules [27]. Similarity and dependency are the

central concepts for clustering modules according to

Gershenson et al. Relative modularity is calculated via a

Modularity Evaluation Matrix . Lai and Gershenson’s method

ranges from the creation of the respective similarity and

dependency matrices to checking the assembly feasibility

605

Friedrich A. Halstenberg et al. / Procedia CIRP 29 ( 2015 ) 603 – 608

[16]. The concept of different drivers is addressed by Voß and

Birkhofer. Starting out from the generation of a Driver

Selection Matrix, the component vector is calculated in order

to receive the component combination impact [4]. Qian and

Zhang’s approach has a simple structure, divided into three

steps: similarity analysis, independency analysis and

evaluation via algorithms [28]. Schmidt introduces a

procedure for effective clustering of car bodies. The author

proposes the conception of basic and variant modules [29].

As described in this paragraph, the current field of methods

for product modularization is limited to one or a defined list

of design goals. The majority of these goals address the

economic dimension of sustainability, e.g. mass customization

or reduction of the product development time. Methods

addressing the environmental dimension of sustainability are

less represented in the academic literature, while the social

dimension has not been found represented at all.

Methods for product family modularization

Dahmus et al. start their product development method with

developing separate function structures for each product in the

product family. These are merged into a single family

function structure in the next step. Using these functions a

modularity matrix is created, which aids in constructing

different possible product portfolio architectures [30]. Kimura

et al. propose a similar procedure, where functional

dependency modules are described in a graph structure and

then superimposed [31]. Kong et al.’s detailed procedure

contains 14 steps in an extended V-model, including product

family planning and the identification and definition of

interfaces [32].

3. Problem Statement

Three major findings have been identified through literature

research:

1. Various design goals addressing the economic and

environmental dimensions of sustainability can be

addressed through the design task of modularization;

2. For a significant amount of them modularity

measures have been developed;

3. Current methods for product modularization are

specifically focused on addressing one or a defined

list of design goals

In this paper, the Target-oriented Modularization Method is

introduced. This method allows defining modules according

to user-defined design goals and measures. It is defined

independent of any modularization measure and can therefore

be adapted to any design goals and can contribute to more

sustainable performance of the final product.

4. Research Methodology

Within the scope of this research, existing modular product

design methods have been analyzed. Their procedures have

been extracted and compared systematically. Distinctive steps

and methodological concepts have been extracted and

examined regarding the question: Can this step be utilized for

a generic method, which allows taking arbitrary

modularizations measures into account?

The useful elements which have been identified were

rearranged to Target-oriented Modularization Method

(TOMM). The method has been tested on a Garrett GT2869R

turbocharger.

5. Method Conception

In this section, the development of the proposed method

Target-oriented Modularization Method is described. The

results of the analysis of existing methods are stated in section

5.1. The procedure of TOMM is specified in section 5.2.

5.1. Identification of suitable steps for Target-oriented

Modularization Method (TOMM)

Researchers widely agree that an essential task in product

modularization is breaking down the product into elements.

Several approaches for this task, such as the one by Pahl &

Beitz, Lanner & Malmquist, Gershenson et al. and Ulrich &

Eppinger have been defined in the academic literature [3, 12,

23, 26]. For TOMM, we selected “Decomposing the system

into elements” by Pimmler & Eppinger. The advantage of

their approach consists in allowing a decomposition in either

physically or functionally described elements.

In early stages of the product development process, it can

be more useful to describe the product in functional elements,

using a function structure, since most of the physical elements

have not been defined yet. In later stages, physical elements

or components can be used. In the majority of cases, however,

a mixture of both will be applied.

Two approaches for merging structural diagrams of several

products into one structural product family diagram have been

identified [30, 31]. For TOMM we decided to use the

approach of Dahmus et al. since it is described in more detail.

The only concept described in literature which takes user-

defined measures into account is the one by Gershenson et al.

[16]. Therefore a product representation in the form of

similarity and dependency matrices was chosen for TOMM.

These matrices list physical and functional elements of the

decomposed products on both axes. Their cells contain either

similarity or dependency relationships between the elements

of the product. The goal of similarity and dependency

representation is to quantify and visualize the element-to-

element relationships in the cells of the matrices [16].

In order to cluster the elements into chunks, rows and

columns of developed similarity and dependency matrices

have to be reordered by an algorithm, so that the highest

values are located closer to the diagonal. This way, product

elements which should rather be clustered in the same module

are located next to each other after the algorithm has clustered

the matrix. The heuristic swapping algorithm used by

606 Friedrich A. Halstenberg et al. / Procedia CIRP 29 ( 2015 ) 603 – 608

Pimmler & Eppinger or the one developed by Kusiak can be

applied [24, 27].

5.2. Application of Target-oriented Modularization Method

(TOMM) to a turbocharger

In this section, we present the concept for a six-step

approach for product modularization which allows the

consideration of goals and measures by the product design

team. In order to test it, the method has been applied in the

redesign of a Garrett 2860R turbocharger.

In Step 1 of TOMM, the products which are subject to the

modularization task are decomposed into smaller elements.

The elements are represented in one scheme, i.e. a

hierarchical diagram modeling the functional or physical

elements of a product. [24]. A product scheme represents the

product design team’s understanding of the constituent

elements of the product [3].

Table 1: Procedure of TOMM

Step

Task

Result

Decompose products into

physical or functional

elements and represent in

schemes

Product schemes

Union multiple product

schemes into a single product

family scheme

Product family scheme

Identify goals and related

modularity measures

List of goals and modularity

measures

Determine value of

modularity measures

Similarity and dependency matrices

Use algorithmic support to

reorder the matrices

Alternative proposals for

product architectures

Choose final product

architecture by comparing the

different product structures

created in step 5

Product architecture

In figure 1 the developed scheme for the Garrett 2860R is

presented. The product has been decomposed into 6 separate

elements. Since the scheme has been developed within a

redesign process, the elements are described physically. The

changes which will be performed to the product within this

process are incremental and take place on the component

level. At this point of the product development process, the

design team already knows which functions are fulfilled by

which physical element, so they can be described through

components.

In Step 2 the diagrams which have been developed for each

product in Step 1 are merged into a single diagram using the

procedure described by Dahmus et al. [30]. The product

family scheme represents a single diagram showing every

functional or physical element of the considered products in

the product family. If the TOMM is applied on a single

product, only one scheme has to be developed and Step 2 can

be skipped. In the case of the Garrett 2860R only one product

is subject to the redesign process. Thus, no family product

scheme has to be developed.

In Step 3, the product design team decides upon the goals

and related measures for product modularization. For this

step, the design team is provided with a database of standard

modularity metrics where design teams can find the right

metric corresponding to their design goals. The detailed

presentation of this database is not in the focus of this article

and will be the subject of a future publication. For means of

this example, the design goals decrease development time,

improve end-of-life treatment and improve maintenance have

been chosen.

Pimmler & Eppinger describe the energy-type interaction

as the necessity of energy transfer in between two components

and as a valid measure for decreasing development time.

Within the development of a product different design teams

may be assigned to the development of individual

components. Components sharing high energy flows are

interdependent and their respective design teams need to

interact frequently, which enhances coordination effort and

increases development time [24]. The energy-type interaction

was therefore chosen as a dependency measure for the

modularization of the Garrett GT2860R. Disassembly time is

described by Qian and Zhang as a dependency measure for

improving the end-of-life treatment of a product as well as

service and maintenance [28].

For every measure selected by the design team, a similarity

/dependency matrix is created in Step 4. Dependency and

similarity matrices line the different functional elements or

physical elements, which have been identified through the

creation of the product scheme on both axes. For every

modularity measure which has been identified, the

similarity/dependency of each element to every other

functional element or component has to be evaluated by the

design team. Measures can be either quantitative or semi-

quantitative. All measures have to be normalized into a

common scale. In every cell of a dependency matrix the value

states the dependency of an element to another element

according to a dependency measure. In similarity matrices the

value states the similarity of an element to an element

according to a similarity measure.

The dependency matrices for the measures energy-type

interaction and disassembly time can be seen in table 2 and 3.

In order to make the two measures comparable, a common

scale of 1-4 has been chosen. The scores for energy-type

interaction have been determined by evaluating the respective

flow of energy in between the components. The Garrett

GT2860R has been disassembled and the time for

Figure 1: Scheme of a Garrett 2860R turbocharger

607

Friedrich A. Halstenberg et al. / Procedia CIRP 29 ( 2015 ) 603 – 608

disassembly has been measured in order to identify the scores

for disassembly time.

Table 2: Dependency matrix for energy-type interaction

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure Can

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure

Can

Table 3: Dependency matrix for disassembly time

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure Can

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure

Can

In Step 5 rows and columns of individual dependency and

similarity matrices are reordered in order to find a suitable

cluster of the functional elements and components. Once rows

and columns are reordered, different modules can be chosen

and combined into a proposal for a product architecture for

every dependency and similarity matrix. In the clustered

matrices, elements which should be arranged in the same

module are located next to each other. This way, an efficient

product architecture according to the measure can easily be

identified. Table 4 shows the reordered dependency matrix for

disassembly time and the reordered dependency matrix for

energy-type interaction. It can be seen that the matrix for

disassembly time recommends the clustering of the turbine

housing, compressor housing and pressure can into one

module and compressor, turbine and crank into another. The

matrix for energy-type interaction factor on the other hand

shows strong dependencies between turbine housing, turbine,

crank and compressor, which indicates clustering them into

one module.

Table 4: Reordered dependency matrix for energy-type interaction

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure Can

Turbine

Housing

Turbine

Crank

Compressor

Housing

Pressure

Can

Table 5: Reordered dependency matrix for disassembly time

Turbine

Housing

Compressor

Housing

Pressure Can

Turbine

Crank

Compressor

Turbine

Housing

Compressor

Housing

Pressure

Can

Turbine

Crank

Compressor

In the final Step 6 of TOMM the design team critically

reviews the different product architectures created in Step 5.

Based on the judgment of the design team, a final product

architecture is selected. In the case of the redesign of a Garrett

GT2860R, the design team has several different options based

on the information the two reordered matrices provide. One

efficient choice would be to cluster turbine housing,

compressor housing and pressure can into one module and

turbine crank and compressor into another one. In the further

redesign process, changes in the construction of the elements

clustered in one module can be conducted much easier. This

will result in a shorter developing time and an improved end-

of-life treatment of the product.

For the future development of the method, a weighting of

criteria will be made possible and supported through an

integrated clustering algorithm in order to enable an objective

decision for the preferred solution.

6. Conclusion

In this article, existing methods for product modularization

have been analyzed and it has been claimed that the field of

608 Friedrich A. Halstenberg et al. / Procedia CIRP 29 ( 2015 ) 603 – 608

product modularization methods still lacks of flexibility and

standardization. In reaction to this, the concept of a Target-

oriented Modularization Method (TOMM) has been

introduced. The method assists product design teams in

defining modular product architecture according to criteria

fitting their design goals. It helps generating alternative

proposals for product architectures based on similarity and

dependency analysis and modularization measures. This new

generic method allows for the consideration of design goals

addressing all dimensions of sustainability. It has been

demonstrated on the example of a Garrett GT2860R

turbocharger and considering the design goals decrease

development time, improve end-of-life treatment and improve

maintenance.

Acknowledgements

The research reported in this publication has been funded

by the German Research Foundation DFG in the frame of the

Collaborative Research Center 1026 “Sustainable

Manufacturing – Shaping global value creation” (SFB 1026/1

2012-2015, Collaborative Research Center CRC 1026).

References

[1] Stark, R., Krause,F. L., Kind, C., Rothenburg, U., Müller, P., Stöckert,

H. 2009. Competing in Engineering Design – the Role of Virtual Product

Creation, in “Proceedings of the 19th CIRP Design Conference –

Competitive Design”, Cranfield University, 30-31 March 2009, pp1

[2] Baldwin, C. Y., and Clark, K. B., 1997. Managing in an Age of

Modularity, in: “Harvard Business Review 75”, no. 5, 84–93

[3] Ulrich, K. T.; Eppinger, S. D., 2012. Product design and development. 5.

ed., internat. ed. New York: McGraw-Hill.

[4] Voß, M.; Birkhofer, H., 2006. Identifying and assessing environmentally

benign modules, in: “Innovation in Life Cycle Engineering and

Sustainable Development”. Dordrecht: Springer Netherlands, S. 255–

270.

[5] Erixon,G., 1998. Modular Function Deployment (MFD), Support for

Good Product Structure Creation. Centre of Industrial Engineering at

Dalarna University and KTH.

[6] Ulrich, K. T.; Tung, K., 1991. Fundamentals of Product Modularity:

Sloan School of Management, Massachusetts Institute of Technology.

[7] Tseng, T., Huang, C., 2008. Design support systems: A case study of

modular design of the set-top box from design knowledge

externalization perspective. In: Decision Support Systems 44 (4), S.

909–924.

[8] ElMaraghy, Waguih; ElMaraghy, Hoda (2014): A New Engineering

Design Paradigm – The Quadruple Bottom Line. In: Procedia CIRP 21,

S. 18–26.

[9] Mikkola, J. H., 2007. Management of Product Architecture Modularity

for Mass Customization: Modeling and Theoretical Considerations. In:

IEEE Trans. Eng. Manage. 54 (1), S. 57–69.

[10] Duray, R.; Ward, P. T.; Milligan, G. W.; Berry, W. L., 2000.

Approaches to mass customization: configurations and empirical

validation. In: Journal of Operations Management 18 (6), S. 605–625.

[11] Fixson, S. K., 2005. Product architecture assessment: a tool to link

product, process, and supply chain design decisions. In: Journal of

Operations Management 23 (3-4), S. 345–369.

[12] Gershenson, J.K., Prasad, G.J., Zhang, Y., 2003. Product modularity:

Definitions and benefits. J. Eng. Des. 14, 295–313.

[13] Newcomb, P. J.; Bras, B.; Rosen, D. W., 1998. Implications of

Modularity on Product Design for the Life Cycle. In: Journal of

Mechanical Design 120 (3), S. 483–490.

[14] Seliger, G., 2007. Sustainability in manufacturing. Recovery of

resources in product and material cycles. Berlin: Springer.

[15] Gershenson, J. K.; Prasad, G. J.; Allamneni, S., 1999. Modular Product

Design. A Life-Cycle View. In: J. Integr. Des. Process Sci. 3 (4), S. 13–

26.

[16] Lai, X.; Gershenson, J. K. 2008. Representation of similarity and

dependency for assembly modularity. In: Int J Adv Manuf Technol 37

(7-8), S. 803–827.

[17] Ji, Y., Jiao, R.J., Chen, L., Wu, C., 2013. Green modular design for

material efficiency: a leader–follower joint optimization model. J. Clean.

Prod. 41, 187 – 201.

[18] Hsiao, S.-W., Ko, Y.-C., Lo, C.-H., Chen, S.-H., 2013. An ISM, DEI,

and {ANP} based approach for product family development. Adv. Eng.

Inform. 27, 131 – 148.

[19] Nepal, B., Monplaisir, L., Singh, N., 2005. Integrated fuzzy logic-based

model for product modularization during concept development phase.

Int. J. Prod. Econ. 96, 157 – 174.

[20] Chen, K., Liu, R., 2005. Interface strategies in modular product

innovation. In: Technovation 25 (7), S. 771–782.

[21]Hölttä, K.M.M., Otto, K.N., 2005. Incorporating design effort complexity

measures in product architectural design and assessment. Des. Stud. 26,

463 – 485.

[22] Smith, S.; Yen, C., 2010. Green product design through product

modularization using atomic theory. In: Robotics and Computer-

Integrated Manufacturing 26 (6), S. 790–798.

[23] Pahl, G., Beitz, W., Feldhusen, J., Grote, K-H., 2005.

Konstruktionslehre. Grundlagen erfolgreicher Produktentwicklung

Methoden und Anwendung. 6. Aufl. Berlin: Springer.

[24] Pimmler T. U.; Eppinger S. D., 1994. Integration Analysis of Product

Decompositions. Massachusetts Institute of Technology (MIT), Sloan

School of Management.

[25] Lange, M.W., Imsdahl, M., 2014. Modular Function Deployment –

Using Module Drivers to Impart Strategies to a Product , in:” Advances

in product family and product platform design. Methods &

applications”. Springer. New York.

[26] Lanner, P., Malmquist, J., 1996. An approach Towards Considering

Technical and Economic Aspects in Product Architecture Design. The

2nd workshop on Product Structuring, TU Delft, The Netherlands.

[27] Kusiak, A. 1990, Intelligent manufacturing systems. Englewood Cliffs,

N.J.: Prentice Hall, Prentice Hall international series in industrial and

systems engineering.

[28] Qian, X., Zhang, H. C., 2003. Design for environment: an environmental

analysis model for the modular design of products. In:” Conference

Record. 2003 IEEE International Symposium on Electronics and the

Environment”. Boston, MA, USA, 19-22 May 2003, S. 114–119

[29] Schmidt, H., Seliger, G., 2002. Beitrag zum Variantenmanagement und

zur Prozessoptimierung im Wagenkastenbau von Schienenfahrzeugen.

Fraunhofer IRB Verlag.

[30] Dahmus, J. B.,Gonzalez-Zugasti, J. P.; Otto, Kevin N., 2001. Modular

product architecture. In: Design Studies 22 (5), S. 409–424..

[31] Kimura, F., Kato, S., Hata, T., Masuda, T., 2001. Product

Modularization for Parts Reuse in Inverse Manufacturing. In: “CIRP

Annals - Manufacturing Technology” 50 (1), S. 89–92.

[32] Kong, F. B., Ming, X. G., Wang, L.; Wang, X. H., Wang, P. P., 2010.

On Modular Products Development. In: Concurrent Engineering 17 (4),

S. 291–300.