APIEMS2009 Dec. 14-16, Kitakyushu

Development of a framework for Disassemblability Design

Optimization for Sustainable ManufacturingRaja Ariffin Raja Ghazilla, Zahari Taha and Novita Sakundarini

Centre for Product Design and Manufacturing,University of Malay.a, Kuala Lumpur 50603, MALAYSIA,

Email: rariffin@um.edu.myzahari-taha@um.edu.mynovitas73@yahoo.com

Abstract.Sustainable Product development is seen as an inevitable solution in protecting theenvironment. Among the important strategy in sustainable product development is through productrecovery by closing the loop. Product recovery enables the reuse of depleting source of virgin materialas well as avoiding the accumulation of solid waste. Maintenance can be also considered as a strategytowards sustainable product by prolonging product life time. The success of product recovery andmaintenance is highly dependant on the time and cost efficiencies of the disassembly process. Thispaper reviews the various methods used to evaluate disassemblability. The development of the methodsand its application are reviewed to provide an understanding of any missing links that have hamperedthe wider implementation of disassemblability design evaluations. The paper then proposes adisassemblability optimization model that can provide the missing link between design anddisassembly

Keywords: designfor disassembly, design evaluation, design optimization

1. INTRODUCTION

Sustainable Product development has been a keyissue in recent years especially with a greaterawareness in the depletion of natural resources andenvironmental degradation. The heighten awareness isintensified with the introduction of various regulatorymeasures to ensure sustainable development such asEnd of Life on Vehicle (ELV) Directive, Energy UsingProducts Directive and Waste of ElectricalElectronic Equipment (WEEE) Directive would makecompliance as necessary requirements for putting. aproduct into the market. In ~rder to remaincompetitive, manufacturers must actively make effortsto ensure product compliance. The growing concern ofsolid waste generation throughout the world is also ofgreat impact whereby many countries have s~me formof legislations to control the generation andmanagement of solid waste. Production waste has hadover burden effect on the landfill cost that leads tomunicipalities in some countries to introduce wastetax for producers and consumers. Among the mainstrategies in sustainable product development isthrough product recovery which is relevant in almost

t :Corresponding Author: r_ariffin@um.edu.my

all sectors. It does not only benefit by reducingecological impact but it also have positive economicimpact to industries [I]. Product recovery is definedas the activities that lead to the salvaging of materialand energy of products at its end of life . Another formof strategy in ensuring waste generation at itsminimum is through prolonging the life of productsthrough proper maintenance. By having propermaintenance, it would not only prolong the lifetime ofa product, but also reduces emission caused byproduct wear and reduce inefficiency. Both of thesestrategies depend on the products ability to bedisassembled, disassembly for maintenance anddisassembly for end of life [2][3]. The introduction ofproduct service supply (PSS) have led to the need forbetter components salvaging for remanufacturingwhich is the critical process in the success of PSS.Similar to assembly planning and production planning,the disassembly process also highly depends on thedesign of the product.

Disassembly is a process in which a product isseparated into its components and/or subassemblies bynon-destructive or semi destructive operations [4].The term disassembly can be viewed from two

204

APIEMS2009

approaches, which are disassembly for maintenanceactivities and disassembly for end of life activitieswhere both are desirable in terms of the ecologicalimpact. The reasons are largely grouped intodisassembly for maintenance and disassembly for endof life.

Disassembly formaintenance is the removal andsubsequent reassemb.y of parts in order to prolong thelife of the product, h nee reducing the need to processnew materials and 'maintaining optimum efficientwork condition of th 'machine. Disassembly for EOLon the other hand is' the removal of parts to extractparts for reuse/remai ufacture or material extractionfor recycling. Extraction for recyclable material ismost often by destructive methods, while extraction ofparts for reuse and remanufacture must be non-destructive. This is the main reason for the limitedsuccess in reusable and remanufacturable products asremoving parts without damaging it is often a tedious,time consuming and ost inducing process with lowvalue. Disassembly for remanufacturing would not becost effective and efficient if parts are difficult to beremoved. Recycling disassembly is sometimesnecessary in order to separate parts of differentmaterial and to reduce the size of a product orsubassembly to ease the shredding and separation.Assessing and designing for product disassemblabilityis thus an important issue.

To ensure product r\I1 be disassembled effectivelyand economically, product must be designed to havesuch features. Several studies have noted that designdecision contributes 70,-80 percent of total productcost [5]. This is also true for disassembly cost asdecision made during designing could causedifficulties in disassembling thus cost more todisassemble. Most studies concur that adoption ofdesign for disassembly must be at the early stage ofthe development process as early as conceptual stage[6]. Thus, it is important for engineering designers todesign with cost efficient disassembly in mind toenable the product to have a better EOL strategy.

2. PRODUCT DESIGN FORDISASSEMBLABILITY METHODS

Several studies were carried out to developmethods to evaluate product design fordisassemblability such as by [7][8] which' aredescribed in the review below. Several works that dealwith optimization of design disassemblability areincluded as well. As the term disassemblability isoften associated with the removal of fastener, soseveral studies on fastener selection for disassemblyare also included in the review below.

Through the MOST empirical data [9] had found

t : Corresponding Author: r_ariffin@um,edu.my

Dec. 14-16, Kitakyushu

that the disassemblability of a product depend on theaccessibility of the parts, precision required forpositioning a tool, amount of effort for a task and thebase time for basic task movements withoutdifficulties. Based on their findings [9] developed adisassembly task difficulty score based on the MOSTpredetermined time system that broke down all of thetasks into their basic elements. The method providedfor 16 standard operations such as unscrew, cut wedge,etc, that would cater for the disassembly of mediumsize products. The types of tool were also consideredin determining the part removal difficulty.

Research by [10] used the MTM system for itsbasis of determining disassembly difficulty. Eachdisassembly task was empirically studied in greaterdetails and scores were assigned based on the MTMvalues. The method included force requirement, parthandling in relation to part geometry requirements,tools, accessibility to joints, positioning of tIs,postural requirements as the factors for determiningdisassembly task difficulties. The method is developedfor the application in the disassembly of medium sizeconsumer products. It added on to the method bycombining assemblability and disassemblability withmaintenance tasks.

In [11] initial work developed a quantitativedisassembly evaluation based on two parameterswhich were disassembly energy and disassemblyentropy. Disassembly difficulty was determinedthrough the physical energy required to release ordisconnect a fastener or joint in which its behaviorwas determined by specific laws of mechanics such astorque for unscrewing against friction. Disassemblywas also characterized by the number ofinterconnection and the direction of the disassemblyoperation. Their findings also indicated a directcorrelation with the task time. The method onlycatered the unscrewing and unsnapping task with theattention towards electronic products.

The Hitachi Disassemblability Evaluation Method[12] is a patented method developed by HitachiCorporation to evaluate the disassemblability of theirelectronic range of products. The method wasdeveloped based on experimental study of disassemblytime and cost for various disassembly tasks. Based onthe experiments, a disassemblability index wasdeveloped for the various standard operations. Theindex was based on three basic elements ofmovements which were upward movement, lateralmovement and screw wise rotation. For each basicelements in a disassembly operation, a penalty scorewas given. Basic elements were multiplied by asupplementary element score if additional difficultiessuch as accessibility were found. Based on the totaldifficulty score, comparison of different design could

205

APIEMS2009

Dec. 14-16, Kitakyushu

be done.[13] developed a model that det~rmi~ed

disassembly difficulty based on assembly dIrectIOnand type of joint used. The score for both 0: .theassembly directions and joint were based on empirical

analysis of motion time studies.DFD method is among the commonly used

method for disassembly evaluation. [~4]Development for this method was based mformatlOnfrom various sources such as the WF method as wellas experimental work. B~sic .elements. of t~edisassembly operation were Id.entlfied as time Unl~.Due to the prior work on design for assembly, th~s

th d sed the reverse of the assembly as the basisme 0 u ddi . Ifor disassembly operation time, therefore a monatime was added based on the following factors:

• Restriction of view• Obstructed access

• Weight• Physical obstruction• Requirements for more than one person• Requirements of mechanical assistance• Requirements for special tools

• Distance to storage

[15] developed a disassemblability evaluationethod for disassembling automotive parts. Based on

;e shapes of the product, a qualitative score wasiven for each process of disassembly such as fixing,

~nding joints, approach, disjoining, handling, etc. The

Oring was done qualitatively by the user by

sc . fl fassigning +ve, none or -ve on the m uences. 0

material type, work fixture, weight, contact point,gripping point, disassembly direction, surf~ceroughness and product structure to the specific

disassembly process.In the Matsushita Assembly-Reverse Assembly

method [16], the assembly and disassemblyassessment are done simultaneously. The factors usedfor assessing disassemblability are material type,requirements for pre-processing, di~ass~mblY direction,omponent reusability and combmatlOn of type of

~oint. For several factors which were in direct relationto the assembly process, the scores were carriedforward to the disassemblability evaluation such. asunscrewing process, postural requirements, handhng

d fixture requirements which constitute reversal of~~e assembly process. The method is very limited in itstype of disassembly operation tasks.

Using MTM model as a basis, [17].developed adisassemblability evaluation model by using the workfactor approach. The disassembly task was brokendown into work tasks. The work tasks were furtherbroken down into task elements in which disassemblytime was derived using the MTM model. The tasks

t : Corresponding Author: r_ ariffin@um.edu.my

covered preparation to disassembly processes such asidentification and search elements to the finalprocessing tasks.

[18] developed a different approach indisassemblability evaluation whereby componentsdisassemblability viability depended on the value ofthe retrieve parts as either recycled material or reusecomponents. [19] also developed a model based oncost based on a multi-factorial approach. Thedifficulty of a disassembly process was attributed tothe cost of running a disassembly facility. The methodused expert opinions on determining the cost fractions.

[20] developed disassemblability evaluationmodel based on the experimental data. The main focuswas to determine the removal time from variousfasteners. Kondo's model attributed total disassemblytime as the total amount of time to remove the fastener.Kondo also looked into the reassembly potential andthe effect fastener age but did not include theseparameters as part of the model. [21] developed adisassemblability model by using experimental data.The model known as UFI effort model lookedspecifically into the difficulties of unfasteningfasteners. The model's fastener unfastening time wascorrelated to the fastener specific design parameters.The model's main intention was to look at the cost ofdisassembly for maintenance. [22] developed a modelto optimize the selection of fastener based on artificialneural process. The model is a multi factorial modelbased on assembly, disassembly and product-in-useparameters. The information was developed based onthe expert opinions data. [23] also develop a model forselection of fasteners. But their approach used thematrix method to determine the relationship betweeneach part. The relationship depended on the productstructure. This model concurrently looks at theassembly and disassemblability of the components anddetermines the appropriate fastener. [24] developedthe disassembly model that was based on cost modelto determine the fastener selection for maintenanceand recycling. Apart from disassembly, reassemblyand recycling cost the model also introduced failureprobability of part and fastener cost due to fastener

failure.

3. DISCUSSIONBased on the review of various design for

disassemblability researches, several observations canbe made. In general, the various developmentalworks in the design for disassemblability is often torelate disassembly task to cost. By determining thecost, the value of design for disassemblability changes

206

APIEMS2009 Dec. 14-16, Kitakyushu

will be tangible for the decision making process.Generally, three approaches can be observed, whichare the process centric approach, geometry centricapproach and fastener centric approach. The processcentric approach mainly focuses on analyzing discretetasks or activities in the separation process. While thegeometry centric approach analyzes geometric shapesof parts that influence the separation process. Thefastener centric appr ach on the other hand onlyfocuses on the fastene in which views of removingthe fastener are the. primary contributors to theseparation problems.

Current research trends are also seeing moreconcurrent evaluation . proach being pursued, such asassembly-disassembly concurrent evaluation anddisassembly-recylability concurrent evaluation. Thedesign for disassemblability models can also bedivided into three types of applications, which are theevaluative model, design guide model andoptimization model. The evaluative model onlyanalyses the impact 0 design on the disassemblydifficulty. The design guide model is usually anextension of the evaluative model whereby hotspotsare determined and gen ralized design guidelines areoffered to reduce impact ef hotspots. The optimizationmodel on the other hand, optimizes the designparameter or assists in decision making in improvingthe design. Several design evaluation models havebeen e tablished and some have great industrialsuccess such as the Boo hroyd Dewhurst, but it isobserved that very few optimization models have beencarried out for the design for disassemblability. Table1 summarizes the classification of disassemblabilitymodels.

The tasks of a designer have been identified as themain cost contributor towards the productdevelopment process. The ..decision made during theearly stage of product development or design processcontributes towards 70-80 percent of the design cost.Making a wrong decision early in the design processcould determine the success of a product later in itslifecycle. The tasks of balancing the various designissues rest on the designer's shoulders; anddetermining disassemblability for maintenance andend of life is are some issues that need decisive trade-off decision among other issues such as functional,assemblability and manufacturability issues. Designersare also faced with the difficult task of making adecision with such limited and often ambiguousinformation.

This is especially true during conceptualizationwhere information is mostly estimates of actualcondition. Current design for disassemblabilitymethod most often only provide means of assessing adesign but the design changes are often left entirely to

t :Corresponding Author: r_ariffin@um.edu.my

the designer's experience with limited design fordisassembly guidelines. As most designer are notfamiliar with the area, then the task will be dauntingespecially due to the contradictive nature of thedisassembly process towards a functional design. Anoptimization model that balances the disassemblabilityissues against functional and assemblability couldreduce the decision making process of a designer inmaking design changes.

It could also avoid or overcome the designcontradiction. Currently, only Gungor's and Jyh-Cheng &Yi-Ming's model attempts to tackle theproblems above. But several short comings areobserved. Gungor's model is based on qualitativeinformation while Jyh-Cheng &Yi-Ming's model onlycovers a part of the relationship which is insufficientto make the trade-off decision. Both of the modelsonly focus on the fastener selection. Unfortunatelydisassembly cost is not only attributed to just thefastener removal but also to the other tasks whiChcould consume more time than the fastener removal.

It is also noticed that most of the designs fordisassembly model require substantive amount ofinformation in order to appropriately evaluate oroptimize the design. Information such as a detailedunderstanding of the task required for disassembly ormotions required to achieve tasks may not be availablein the conceptual design. Most often, the currentmodels concentrate their application during theembodiment stage where substantial definitiveinformation is available. Changes made during theembodiment stage reduce the innovative opportunitiesthat can be introduced in a product.

207

APIEMS2009

Dec. 14-16, Kitakyushu

Table 1: Approach of Different Models

Process Centric Geometry Centric Fastener Centric

Evaluative Desai et al Mok et al Sodhi et al

Suga et al Boothroyd Dewhurst

Kondo et al

Hwa Cho Yi et al

Villalba et al

Kroll&Hanft

Gungor-Gupta

.Design Guide Desai & Mital

Shu & Flowers

Desai

Hitachi DEM

I Matsuhita

OptimizeGungor et al

Jyh-Cheng &Yi-Ming

Furthermore, the parameters and language beingused in most models are not reflective of engineeringdesigner's terms. For example, in some modelsdesigners are required to define product'sdisassembly motion, disengage force or handlingrequirements which may not be easily identifiableduring the conceptual design by designers. Designersare more familiar with parameters such as dimension,weight, surface friction or surface properties, jointstrength etc which are related to design parameterthat are identifiable during the conceptual design.

In summary, numerous studies have beenconducted in developing the design fordisassemblability models. Some have been proven tobe effective and considered to be industry standards

such asBoothroyd Dewhurst method while others are

still at the lab simulation phase as seen through thevarious simple lab case studies. None the less therestill exist opportunities for further research.Limitations such as information/data requirementtrade-off, designer based language, approach designfor disassemblability at the concept design level

design and parameter optimization issues are some ofthe noticeable elements that can be improved.

4. FRAMEWORK FOR METHODOLOGYDEVELOPMENT

From the review of the disassemblability models,two questions that will contribute to the improvementof design for disassemblability model arises.

Can a relationship be established between less aformed design specification data with a more defineddisassembly parameter?

If a relationship can be establish, how can we usethe established relationship to generate better conceptdesigns that are functional wise and disassemblywise?

The answers to the research questions above canincrease the designer's foresight towards disassemblyproblems during the conceptual design stage itself.This will allow more flexibility in the changes, hencemore innovation opportunities regarding to the designfor disassembly and the design for function trade-offwill be available.

208

APIEMS2009

Based on the research questions above, aframework is proposed to specifically answer thequestions. The framework attempts to add upon theexisting knowledge in the design fordisassemblability; several contributions are thefunctional design trade-off and parameteroptimization to reduce disassembly time respectively.The framewor ,model is a hybrid of geometric andfastener centric approach. Figure I below shows theproposed framework. The model considers part andfastener as complimenting objects. Each disassemblyelement consists of part disassembly and/or fastenerremoval. The u -'r provides minimal information suchas mating geometry, mating part material, partoperating conditi n, aesthetic requirement, end of lifeoption, maintenance requirement and joint strength.By using the information, the Analytical HeuristicsProcess approach will be used to determine theappropriate type of fastener for the element that hasthe minimum basic unfastening time. The unfasteningtime will be calculated based on the model by [25].Workspace allowance and motion direction, althoughmay be unnecess ry during initial definition will also

Dec. 14-16, Kitakyushu

be factored in to determine the final unfastening timewhen such information is made available. In order todetermine the part removal time information such assurface quality, material, weight and size should bedirection, although may be unnecessary during initialdefinition will also be factored in to determine thefinal unfastening time when such information is madeavailable. In order to determine the part removal timeinformation such as surface quality, material, weightand size should be defined. Based on the information,the knowledgebase search algorithm by using Desai[9]model will determine the part removal time. Thedisassembly cost will be calculated based onunfastening and part removal time. This cost willlater be an offset to generated revenues. If theelement is not cost efficient in its disassembly, partfeature hotspot can be identified by using thesensitivity analysis so that the parameter range can bebroadened to allow more flexibility in improv gdisassembly efficiency.

209

APIEMS2009Dec. 14-16. Kitakyushu

SensitivityAnanlysis

No

AsaembIy Tone~Tome

Figure I: Disassembly Design optimization framework

Mating GeomebYMating Part MalerialOperation ConditionAe8II1etic RequilementEnd 01 ute OptionMaintenanCe requiIemeI1tJoint Mating Force

Faslenlng Directionwnspace Availability

Unfa8ten1ng Direction~---I wnspace Availability

Determine PartTotal Assembly! Assembly!Disassembly 1-----1Time-Cost DisassemblyTime

Time

Offsets

Recycle maIerial Price~ __ -e- Reuse Price

ManufacturabiIiF cost

partassemblyefficiency

Disassemblyefficiency

Maintenanceefficiency

210

APIEMS2009

5. CONCLUSIONDisassembly activities are gaining greater importanceespecially due to the heighten awareness of theecological concerns. New business, manufacturingand end of life approach such as PSS andremanufacturing depend highly on the ability todisassemble. Mnintenance efficiency also depends onthe ability for a product to be disassembled efficiently.Due to these two main drivers, Design forDisassemblability is gaining acceptance that canassist designers" identifying disassembly problems

REFERENCESGehin, A., Zwolinski, P., Brissaud, D., 2008, A tool toimplement sustainable end -of-life strategies in theproduct development phase, Journal of CleanerProduction, Vol. 16.

Willems B., Dew IfW., Duflou J.R" 2008, A methodto assess the lifetime prolongation capabilities ofproducts, International Journal on SustainableManufacturing, VoU, pp.122-144.

Xanthopoulos A., Lakovou E., 2009, On the optimaldesign of the disassembly and recovery processes,Waste Management, Vol 29.

Lambert A.J.D, pta S.M., 2005, Disassemblymodeling for assembly, maintenance, reuse andrecycling, CRC Press.

Boothroyd G, Alting L., 1992, Design for Assemblyand Disassembly, Annals of the CIRP, Vol 41.

Kroll E., Carver B. S., 1999, Disassembly analysisthrough time estimation and other metrics, Roboticsand computer integrated manufacturing. Vol. 15, pp:191-200.

Kroll E., Beardsley B., Parulian A. , 1996, Amethodology to evaluated ease of disassembly forproduct recycling, lEE Transactions, Vol. 28.

McGlothlin S., Kroll E., 1995 ,Systematic Estimationof Disassembly Difficulties: Application to ComputerMonitor,. Proceedings of the 1995 IEEE InternationalSymposium on Electronics and the Environment.

Desai A., Mital A. 2003, Evaluation ofDisassemblability to enable design for disassembly inmass production, Journal of Industrial Ergonomics,Vol. 32.

Dec. 14-16, Kitakyushu

during the design stage. Throughout the years,numerous research works have been conducted on thearea. None the less, gaps are still found and they canbe improved so that acceptance and utilization of theDesign for Disassemblability can further beimprovised. This paper then introduces theframework for the development of a novel design forDisassemblability optimization approach that hopesto fill in the gap.

Desai A.,2006, An integtarted Methodology ForAssembly, Disassembly and Maintenance ForConsumer Products, Phd Dissertations, Division ofResearch and Advance Studies, University ofCincinnati.

Suga T., Saneshige K., Fujimoto J., 1996,Quantitative disassembly evaluation, Proceedings' ofthe 1996 IEEE International Symposium onelectronics and the environment, pp: 19-24.

Arimoto S. et ai, 1993, Method of Evaluatingeasiness of works and processing performed onarticles and evaluation apparatus, United StatesPatent 5586022.

Gungor A., Gupta S.M, 1997. An EvaluationMethodology for Disassembly Processes, Computerin Industrial Engineering, Vol. 33.

Dewhurst B., 1996, Design for Environment Manual,Boothroyd Dewhurst Inc.

Mok H.S, Kim H.J, Moon K.S, 1997,Disassemblability of Mechanical Parts in Automobilefor Recycling, Computer in Industrial Engineering,Vol. 33.

lin K. et. aI., 1999, Method and device for evaluatingassemblability and reverse assemblability, UnitedStates Patent 6434438 B I.

Villaba 0., Segarra M., Chimenos 1.M., Espiell F.2004, Using the recylability index of materials astools for design for disassembl", EcologicalEconomics, Vol. 50.

Das S.K, Yedlarajiah P., Narendra R., 2000, Am

211

APIEMS2009Dec. 14-16. Kitakyushu

approach for estimating the end-of-life productdisassembly effort and cost, International Journal ofProduction Research, Vol. 38.

Kondo Y., Deguchi K., Hayashi T., Obata F. , 2003,Reversibility and disassembly time of part connectio",Resources, Conservation and recycling, Vol. 38.

Sodhi R., Sonnenberg M., Das S., 2004, Evaluatingthe unfastening effort in design for disassembly andserviceability, Journal of Engineering Design, Vol. 15.

Yi H.C, Park Y.C, Lee K.S. , 2003, A study on themethod of disassembly of a product using work factormethod, IEEE Transaction.

Gungor A, 2006, Evaluation of connection types indesign for disassembly (DFD) using analytic networkprocess, Computer and Industrial Engineering, Vol.

50.

Shu L.H., Flowers W.C. , 1995, Consideringremanufacture and other End-of-life Options inselection of fastening and joining methods,Proceedings of IEEE International Symposium onElectronics and the Environment.

Jyh-Cheng Y., Yi-Ming L., 2006, Structurerepresentation for concurrent analysis of productassembly and disassembly, Expert System with

Application, Vol. 31.

AUTHOR BIOGRAPHIESRaja Ariffin Raja Ghazila is a lecturer and PhDstudent at Department of Engineering Design andManufacture, Faculty of Engineering, University ofMalaya. His email addressisrajaariffin@gmail.com

Zahari Taha is a Professor at Department ofEngineering Design and Manufacture, Faculty ofEngineering, University of Malaya. His researchinterest is in Manufacturing Design. His emailaddress is zahari_taha@um.edu.my

Novita Sakundarini is a PhD student in Departmentof Engineering Design and Manufacture, Faculty ofEngineering, University of Malaya, Malaysia. She isalso lecturer at UPN "Veteran" Yogyakarta,Indonesia.Her main research interests in ComputerBased Eco-Design. Her email address

is novitas73@yahoo.com

212