A novel comparative design procedure for reconfigurable assembly fixtures

G Model CIRPJ 427 No. of Pages 13

CIRP Journal of Manufacturing Science and Technology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

CIRP Journal of Manufacturing Science and Technology journal homepage: www.elsevier.com/locate/cirpj

A novel comparative design procedure for reconfigurable assembly fixtures Ilker Erdema,* , Christoffer Levandowskia , Cecilia Berlina , Henrik Kihlmanb , Johan Stahrea a b

Department of Product and Production Systems, Chalmers University of Technology, Gothenburg, Sweden Prodtex AB, Gothenburg, Sweden

A R T I C L E I N F O

Article history: Available online xxx

Keywords: Fixture Tooling Reconfigurable Assembly Manufacturing Production

A B S T R A C T

While market requirements demand that manufacturing systems increase their responsiveness, assembly fixtures remain limited in corresponding to the same demand. Fixture designers as practitioners are left without guidance to design reconfigurable fixtures. This study proposes a comparative design procedure for reconfigurable assembly fixtures that can adapt to manufacturing system characteristics by using efficiency metrics. In this study, a theoretical analysis based on manufacturing systems is presented to establish efficiency metrics. Later, these metrics are utilized in a design procedure that offers guidance and determines the efficiency of fixtures in conceptual and detailed design stages. Finally, an experimental verification is presented. © 2017 CIRP.

Introduction Manufacturing companies of today operate on a highly competitive market that is characterized by increasingly diverging customer requirements. Throughout modern manufacturing history, a succession of manufacturing paradigms have sought to solve the different challenges to meet market demands. The challenge to provide customized products was initially met by the introduction of flexibility — later to be known as Flexible Manufacturing Systems (FMS) [1]. As a result of seeking flexibility, fixtures shifted from a dedicated to a modular nature where a fixture’s geometry was divided into simple sub-geometries that could be mechanically rebuilt to fit another workpiece and process [2]. However, the increased flexibility affected the performance of manufacturing systems in terms of cost and quality. In return, the operations to technologies utilized in manufacturing systems sought optimization of flexibility. This led to the birth of the concepts of agility and reconfigurability — later to be known as Agile Manufacturing Systems (AMS) and Reconfigurable Manufacturing Systems (RMS) respectively [3]. Consequently, fixtures were developed in a form where reconfigurability was met by a built-in flexibility by adjusting the internal parameters such as the length of an actuator in a kinematic structure [4]. Furthermore, the features such as adaptive (also known as active fixturing) were integrated to

* Corresponding author. E-mail address: [email protected] (I. Erdem).

increase the performance of flexible fixtures by means of sensorintegrated fixture elements [5]. Today, customers are even expecting more tailored products, resulting in a large variety of product variants that need to be developed and manufactured [6]. This requires a manufacturing system that can manage the product variety, which in return leads to a need for fixtures that are able to abide by the same principle to support responsiveness [7,8] to new customer requirements, while maintaining high quality across the product range. As a part of manufacturing system development, the fixtures need to be designed to meet performance requirements and envelop the product variety. This places a responsibility on fixture designers to meet criteria and variety challenges while addressing the responsivity demand. Stemming from the rapid development and deployment need, several systematic processes for fixture design have been developed. Initially, Trappey and Liu [9] described three major steps in fixture design as configuration, assembly and verification. Later, Rong and Bai [10] proposed a procedure in three steps: setup, fixture planning and detailed design. Bi and Zhang [11] also developed a design process where a fixture design was carried out in steps of problem description, analysis, synthesis and verification. Mervyn et al. [12] divided the fixture design process into conceptual and detailed design steps where their study offered a conceptual analysis based on cost and time. With the introduction of Computer-Aided Process Planning (CAPP), these fixture design processes were complemented with computerized automation — which was later coined as Computer-Aided Fixture Design (CAFD)

http://dx.doi.org/10.1016/j.cirpj.2017.06.004 1755-5817/© 2017 CIRP.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

2

I. Erdem et al. / NULL xxx (2017) xxx–xxx

Nomenclature

Wc Wa WT Vc Va VT Rcc Pa PT Ruc Pra PrT Cc Cf Ce Cs C wh TA CT Tc Ts Tt Dc Rec Rei Ret Mc Ns NT Coc

eo ei

wi

The weight efficiency metric of a reconfigurable fixture The achieved weight of a reconfigurable fixture The weight limit specified for a prospective fixture The physical volume efficiency metric of a reconfigurable fixture The achieved volume of a reconfigurable fixture The volume limit specified for a prospective fixture The reconfigurability efficiency metric of a reconfigurable fixture The number of satisfied products in a product family The number of products in the target product family The reusability efficiency metric of a reconfigurable fixture The number of satisfied processes The number of target processes The cost efficiency metric of a reconfigurable fixture The capital cost of a reconfigurable fixture The set-up cost of external equipment The software cost of a reconfigurable fixture The cost of software development per work-hour The allocated time for software development in hours The cost threshold specified for a reconfigurable fixture The time efficiency metric of a reconfigurable fixture during set-up The time spent for a set-up operation The time limit designated for a set-up operation The diagnosability metric of a reconfigurable fixture The reliability efficiency metric of a reconfigurable fixture The reliability of each standard component The reliability threshold for a system of standard components The modularity efficiency metric of a reconfigurable fixture The number of standard components in a reconfigurable fixture The total number of components in a reconfigurable fixture The convertibility metric of a reconfigurable fixture The overall efficiency of a reconfigurable fixture Single efficiency metric The individual weight designated for a single efficiency metric

[13]. Even though these fixture design processes utilize various analysis techniques to apply requirements on the fixture design and find the most feasible option, the majority of the research and application efforts are limited to the use of simple geometries in modular fixtures [14]. When the span of fixture design processes is extended to reconfigurable fixtures, the researchers in literature tend to follow a certain reasoning where the design procedure for reconfigurable fixtures is mainly conducted from mechanical, control and software perspectives. For example, Yu et al. [15] and Millar and Kihlman [16] demonstrated the development of kinematic structures to create a custom reconfigurable fixturing solution. Furthermore, Zhang et al. [17] and Li et al. [18] illustrated the control system formulation through a number of actuation and

process requirements such as adaptive control and smart assembly. Moreover, a formulation based on the control and process parameters can also be conducted to define the capabilities of fixture controller and software. For example, Erdem et al. [19] and Soetebier et al. [20] defined the software architecture in terms of function families, and demonstrated the integration with a graphical user interface. Although the fixture design solutions offered by literature show versatility, the aforementioned reasoning does not offer a complete and reliable insight into reconfigurable fixture design and its long-term financial impact — in particular, fixture designers as practitioners are left without concrete guidance in how to design reconfigurable fixtures with pertinent parameters in mind. Furthermore, the lack of formal procedure and computerized support in the development of reconfigurable fixtures creates ambiguity in design trade-offs [21–23]. Therefore, there is a need to develop an adapted design procedure that encapsulates the design complexity of reconfigurable fixtures and offers fixture designers a more comprehensive verification perspective. The objective of this paper is, therefore, to propose a comparative design and evaluation procedure for reconfigurable fixtures that unifies mechanical, control and software design perspectives. The focus of this procedure is limited to conceptual and detailed design/verification of kinematic units whereas preceding planning stages are assumed to be established. Finally, the structure of this paper is as follows: in Section “Research approach”, the research approach is presented. In Section “Theoretical framework”, a theoretical framework is given. Section “The proposed procedure” focuses on the synthesis and development of the proposed procedure whereas Section “Experimental study” offers the results of an empirical study to exemplify the use of the proposed procedure. Research approach In order to realize the design procedure, an inductive approach is formulated around two research questions. Firstly, this paper aims to determine which functions and constraints are pertinent in order to develop a reconfigurable fixture. The following research question is formulated: What parameters can be used as means of input to design and verification aspects of reconfigurable fixtures? Secondly, the utilization of these parameters in a systematic manner plays an important role in achieving this paper’s objective. Hence, the second question is formulated: How can these parameters be integrated and utilized systematically to design reconfigurable fixtures? The first question is answered with a review of the available literature on reconfigurable fixture design, where the expected outcome is to establish a set of theoretically motivated design parameters. To answer the second question, design and comparative verification procedure is synthesized in conjunction with the conversion of the design parameters into design functions and constraints. Finally, a lab-based experiment is conducted to exemplify and verify the proposed design procedure. Theoretical framework The underlying theory of the proposed fixture design procedure will be presented in this section. Initially, the established range of fixture design theory will be presented to identify design characteristics and fundamental fixturing parameters widely accepted in literature. Later, the second section will identify the evaluation of fixturing technologies from a manufacturing paradigm perspective.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx

Fundamental parameters of fixture design By definition, a fixture is a physical device that locates and secures a workpiece accurately and repeatedly against process forces. Within this broad definition, each stage of fixturing design requires an extensive amount of analyses. However, based on that definition, a certain set of expectations can be placed on the fixture body and locating unit. These will inevitably constitute stiffness expectations to withstand the internal and external forces, accuracy and repeatability demands so that workpiece can be correctly located. Furthermore, parameters such as weight and dimensions of a fixture are important from a manufacturing perspective so that parameters such as accessibility and ease of use can be ensured [23]. Finally, the capital and recurring costs constraint the fixture solution, and the evaluation phase is realized with respect to the return on investment [24]. Manufacturing systems and evaluation Historically, mass production environments and the use of batches forced manufacturing cells to handle variety. Hence, fixture designers were encouraged to optimize the fixture layout in order to minimize the capital cost of fixtures and time consumption for a set-up operation [25]. Through clustering fixtures with respect to occurrences of the features in a product family, the concept of Group Technology (GT) was developed [26–28]. The optimization process in GT was further elaborated in terms of setup time reduction with focus on reducing the idle time by the Single Minute Exchange of Dies (SMED) philosophy [29,30]. GT and SMED concepts focused mainly on cost and time perspectives where both concepts aimed to capture the direct and indirect impacts of different fixture solutions. With FMS, the performance parameters and overall design approach of fixtures shifted towards changeable manufacturing technologies. Thus, modular fixtures comprising reusable standardized blocks were developed to provide flexibility so that the changeability challenge could be met [31]. Hence, the flexibility and quality were incorporated to performance parameters along with cost and time [32–34]. After the application of FMS, the optimization of flexibility became an apparent driver as the performance of the technologies created for FMS did not yield expected results in terms of cost and time effectiveness [35]. Consequently, RMS and AMS were introduced [36,37]. In RMS, certain parameters and characteristics were expected from manufacturing technologies to cope with the

3

optimization challenge. These were mainly categorized as customized flexibility,convertibility, scalability, reusability, diagnosability and modularity [38–40] where AMS proposed similar characteristics on optimized flexibility from networking and business perspectives [41,42]. The evaluation of these characteristics for RMS and AMS was conducted based on parameters of flexibility, quality, cost and time [43] where the metrics of each parameter was in conjunction with what Performance Measurement Systems (PMS) offered [44–46]. As PMS covered a variety of metrics for each parameter, an adoption process of each metric with respect to an individual manufacturing system and technology was considered essential [47–49]. Fixture design, manufacturing systems and evaluation In order to establish a fixture design and evaluation from a manufacturing system perspective as customary in PMS, a redefinition process integrating the aforementioned characteristics and parameters is proposed by this study. Thus, we separate the parameters into four categories: cost, time, quality and flexibility. The cost represents the summation of capital and recurring costs whereas the time parameter represents the time consumption related to fixture’s set-up operation. The quality parameter aims to quantify the fixture robustness by analyzing the fixture capabilities throughout a manufacturing process. Finally, the flexibility parameter defines the fixture’s capability to adapt to different products and processes. Firstly, the adaptation of metrics for the flexibility focuses on reconfigurability, reusability and modularity. The reconfigurability metric is defined as the total degree-of-freedom and workspace required for a fixture to adapt to the relevant products in a product family. The reusability metric represents the capability of the fixture to be utilized for different processes. The modularity metric defines the capability of a fixture’s components to be reused for a different application. Secondly, the cost parameter is defined by the metrics as investment and setup costs. The investment cost represents the capital of the respective fixture in terms of hardware and software whereas set-up cost is the external hardware and software investment in order to implement a setup. Thirdly, the time parameter is represented by a set-up time metric defined as the time allowed for a set-up operation starting with process halt until restart. Finally, the fourth parameter – quality – is operationalized by diagnosability, reliability and convertibility metrics. The diagnosability metric is defined as the fixture’s capability to give feedback on the process; and the

Table 1 Manufacturing system level fixture design parameters. Parameters/metrics

Definition

Flexibility
Reconfigurability
Reusability
Modularity

Physical capability of a fixture to adapt to different products and processes
Capability to adapt products within a product family
Number of processes the fixture can be used
Capability of a fixture to be modularly rearranged for different applications

Cost
Investment Cost
Set-up cost

Summation of capital and recurring costs
Hardware and software procurement and development cost
External hardware and software investment for a set-up operation

Time
Set-up time

Timewise parametrization of fixture utilization during manufacturing
Time required to implement a set-up

Quality
Diagnosability
Reliability
Convertibility

Parametric quantification of process robustness
A flexible fixture’s capability to give/receive feedback on the process such as accuracy, workpiece deformation and process forces
Total reliability value of standard components within a fixture
Capability of the fixture to mount/remove or interact with external resources

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

4

I. Erdem et al. / NULL xxx (2017) xxx–xxx

reliability metric is the reliability value of the complete fixture. The convertibility metric represents the fixture’s capability to interact with resources such as metrology systems in a manufacturing process. In Table 1, the identified parameters and their metrics are summarized. The proposed procedure This section explains the proposed comparative procedure for design and verification of reconfigurable fixtures. The parameter conversion to design functions and constraints are described. Each design step and the corresponding solution groups are explained in detail. An important milestone is to recapture the parameters synthesized in the previous section in terms of meaningful functions and constraints the can be used in the conceptual and detailed design stages. The detailed procedure is illustrated in Fig. 1. It introduces four stages, two of which comprise several design steps. The order of the steps follows a logic chain of necessary design decisions.

Thresholds for functions and constraints The initial step in conceptual design stage is to quantify each parameter defined in Section “Theoretical framework”, reflecting the impact that design choices have on the final efficiency of the reconfigurable fixture solution. Therefore, in this section we will convert these parameters into corresponding design functions and constraints by deriving their respective equations. Fixturing parameters The fundamental fixturing constraints stiffness, weight, dimensions, repeatability and accuracy are important aspects affecting the fitness of a fixturing solution to a process and manufacturing system. However, the proposed procedure uses only weight and dimension parameters for conceptual design mainly because stiffness, accuracy and repeatability requirements are directly correlated to process requirements. Therefore, these parameters are incorporated into reusability. As weight and dimensional requirements represent physical thresholds, they constrain the prospective fixture to comply with

Fig. 1. Proposed design procedure.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx

certain standards. Thus, the requirements set by the processes and manufacturing system can be utilized to quantify the performance of a prospective fixture where surpassing or reaching these constraints create a negative impact on the performance. As the weight is a single value metric, it can be easily quantified with respect to a threshold. However, the dimensions of a fixture are relatively more complicated as they might require individual metrics and coordinate transformations. Thus, a volume constraint can be utilized to simplify the design process. Subsequently, the relevant equations for weight and volume metrics can be defined as Wc ¼ 1 

Vc ¼ 1 

Wa WT

Va VT

process niche. However, the complete reusability value can only be achieved in the detailed design particularly after the stiffness analysis of the designed fixture. Hence, the reusability can be formally determined by Ruc ¼

Pra PrT

ð2Þ

where Ruc is the ratio of reusability, and PrT and Pra are the number of targeted and satisfied processes respectively. The reusability metric (Ruc) yields a value between 0 and 1 where an increase in the number of satisfied processes (with respect to the process range) contributes to the overall efficiency of a flexible fixture. Cost The cost efficiency of a flexible fixture can be based upon the investment savings made with respect to a threshold that can justify an investment from a manufacturing system perspective. This threshold can be identified with respect to the existing fixturing costs in the manufacturing system, or setting a relative target accordingly. Specifically, the cost of a reconfigurable fixture can be approached from two perspectives. First, the capital cost of a single reconfigurable fixture can be defined as the investment cost corresponding to the total capital required for the components spanning from hardware to software. Second, since a fixture can be reconfigured with external equipment (e.g. external automation tools and measurement systems), the capital for reconfiguration tools can be included in the set-up cost. Thus, the metric of the cost is Cc ¼ 1 

Reconfigurability Often, the products in a family may demand a large workspace for a fixture that may increase the capital cost of a reconfigurable fixture and require a trade-off. Therefore, an efficiency-based reconfigurability can be utilized as the ratio of the number of products within the feasible workspace of a reconfigurable fixture to the number of targeted products within a family. The equation can be defined as Pa PT

ð4Þ

ð1Þ

Wc, Wa, WT are the weight efficiency, realized weight and target weight respectively. Vc, Va, VT are volume efficiency, achieved volume and target volume of the fixturing solutions where the volume of a fixture is defined by the height-length-width of the reconfigurable fixture at home position. In Eqs. (1) and (2), for achieved values remaining below the targets, the efficiency metrics yield positive values, meaning that the prospective solution is contributing positively to the overall efficiency of the fixture. However, in a case where the achieved values surpass the targets, the efficiency metric becomes negative and reduces the fixture efficiency.

Rcc ¼

5

ð3Þ

where Rcc is the ratio of reconfigurability, PT and Pa are the total number of products in a family and the number of satisfied products within the feasible workspace of a reconfigurable fixture respectively. Eq. (3) proposes the reconfigurability efficiency to be proportional to the number of satisfied products in relation to a product family. Since the number of satisfied products cannot surpass the total number of products in a family, as motivated by customized flexibility in RMS, this metric always yields a positive result between 0 and 1. Finally, the efficiency value can be immediately calculated in the conceptual verification stage by conducting a workspace analysis. Reusability Similar to reconfigurability, we propose reusability as the capability of a fixture to be reutilized in various processes. Thus, setting a target process range and driving the design procedure to satisfy its requirements establishes the foundation for reusability in conjunction with the trade-off aspect of the reconfigurability metric. Therefore, a flexible fixture’s reusability efficiency can be identified by the ratio of the satisfied processes to the number of processes in the target range. Different processes might have different requirements on fixtures in terms of stiffness, repeatability and accuracy. The latter two properties can be immediately acquired from the outcome of the conceptual analysis stage and can be used to filter the target

Cf þ Ce þ Cs CT

ð5Þ

where Cc is the cost efficiency. Cf and Ce are the capital costs of the fixture and the set-up cost of external equipment respectively. Cs is the software cost defined by the activities in Section “Controller software design”, and CT is the total cost threshold dedicated to hardware investment. Eq. (5) proposes that the cost efficiency of a flexible fixture is negatively affected as the capital, set-up or software costs increase. When the threshold is surpassed, the cost efficiency yields a negative efficiency value — which further reduces the overall efficiency of the fixture. Time The activities to reconfigure a fixture are assessed with respect to the total time allowable for set-up operations. By considering the time spent as a loss of efficiency, the time metric can be formulated as Tc ¼ 1 

Ts Tt

ð6Þ

Tc is the time metric that is defined as the ratio of time saved to the total time threshold (Tt) where Ts is the achieved set-up time. The time for set-up is the sum of the time starting from machine stop, through reconfiguration and restart. Particularly for the time metric, the proposed procedure requires the designer to have knowledge of the time values for activities in cell preparation such as stopping and preparing the machine(s) for manufacturing after reconfiguration. For different processes that allow different set-up times, the minimum threshold can be utilized to define the total time threshold. Diagnosability As adding features and components to a fixture would affect the remaining metrics (e.g. cost), the impact of having value adding diagnosable capabilities can be quantified in in relatively simplified manner. Thus, we propose the diagnosability term to be

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

6

I. Erdem et al. / NULL xxx (2017) xxx–xxx

defined as a binary value that corresponds to the capability of fixture feedback on the workpiece. If the designed fixture has technical features that collect and analyze data, then the diagnosability (Dc) value is 1. If the feedback is irrelevant (or not possible), then the diagnosability value is 0. Reliability The reliability metric is here defined as the total reliability of a system of serially attached components where multiplication of the reliability of each component yields the system reliability [50]. Thus, the following formula can be applied to define the reliability metric: Y Rei i Rec ¼ ð7Þ Ret

i ¼ f1; . . . ; ng where n is the number of standard components and Rei is the reliability of each standard component in the system. Ret is the expected reliability from a manufacturing system-level perspective. In Eq. (7), the ratio of system reliability to the manufacturing system expectancy gives the reliability efficiency of the designed fixture. An efficiency value of 1 represents a sufficiently reliable fixture, while a value greater than 1 can enable the fixture designer to make a trade-off between the remaining metrics. Finally, this metric only includes the efficiency of standardized parts since custom components require data gathering to define reliability. Modularity Based on the reasoning that standardized components allow a fixture to adapt to different requirements through modularity, an increase in the number of standard components positively influences the efficiency of a flexible fixture. Thus, the modularity metric is proposed as the ratio of standard components to the total number of components in a fixture. Hence, modularity can be defined as Ns Mc ¼ NT

ð8Þ

where Ns and NT are the number of standard and total components respectively. As the modularity value cannot surpass the total number of components, Eq. (8) metric yields values between 0 and 1. A higher value of modularity symbolizes the efficiency in rapid rearrangement of a fixture for different applications from a mechanical perspective. Convertibility The interaction of a fixture with external equipment on the control and software level can be an enabling factor for automation. Since such a feature affects the design and component selection, the impact of having a convertible fixture as a unit in an assembly cell is proposed to be included in the final efficiency definition. Similar to diagnosability, the convertibility factor (Coc) is also a binary number where value 1 corresponds to having the interaction requirement fulfilled. Kinematic structure A kinematic structure can be categorized into one of three groups: serial, parallel and hybrid. Serial kinematic structures comprise of serially linked joints such as articulated kinematics machines. A parallel structure is the combination of same or different types of serially linked joints in a parallel configuration. A hybrid kinematics structure can be a combination of two or more

parallel and serial structures as a single unit e.g. two parallel [51], or one parallel and serial [52]. In this step, a designer selects a kinematic structure, and later creates a model based on the configuration of components, kinematics, and joint types with respect to the actuated joint. For example, a Stewart–Gough platform is a six-limb universalactuated prismatic-spherical (UPS) joint configuration [53]. This kinematic structure can be utilized in selection of components and workspace analysis. Actuation After the kinematic classification, the designer can select the actuation type. The reconfiguration may be conducted internally by means of motors or hydraulics, or externally by another automation resource such as an articulated robot reconfiguring a reconfigurable fixture [4]. Consequently, a certain group of components become feasible for the actuated joint in kinematic structure. Position-holding Position-holding is the reconfigurable fixture’s capability to maintain a certain position and orientation after reconfiguration. In this step, possible choices are divided into three groups. In the first group, a fixture’s locking capability is manually provided on the actuated joints by either integrated or external clamping tools such as a shaft collar. In the second category — extra-locking, position-holding can be provided by an integrated standard or custom component whereas the clamping or unclamping functionality is externally controlled. The last category – intra-locking – is defined as integrated and internally controlled position-holding equipment such as, motor brakes. The input of this step is independent of the previous steps and conducted in parallel to kinematic structure and actuation selection. Mechanical component selection The mechanical components are divided into two categories: standard (off-the-shelf) and custom components. By using the input from the previous three steps, the components in a kinematic structure can be classified as either of these categories. For example, a passive joint can be custom made for a certain application whereas an actuator can be purchased off-the-shelf. Control system component selection The control system of a reconfigurable fixture is divided into two categories. In the first category, a reconfigurable fixture has a dedicated controller to a single fixture. Hence, this type of controller is illustrated in the proposed procedure as independent. Depending on the type of actuation and earlier inputs, this type of controller includes motors, position-holding control elements such as pneumatic control units, drives, PC, programmable logic controllers (PLC) for safety and sensors. Second type of control system is a generic or a shared controller where such application is referred as dependent. This type of control system aims to capture the reconfiguration and positionholding by means of a controller that belongs to an external automation tool such as an articulated robot’s controller connected to the actuated joints of a reconfigurable fixture. In such a case, the reconfigurable fixture shares the controller with a non-fixturing equipment; and hence, it is subdivided in the procedure as shared controller. An example of such can be seen in Kihlman and Engstrom [54]. The other type of dependent controller is a stand-

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx

alone controller controlling multiple reconfigurable fixtures simultaneously or one at a time [19]. Consequently, the choices regarding the control system are made with respect to the output of the previous design steps. For example, if a designer prefers an internal actuation in actuation step the choice of controller is automatically limited to either stand-alone or independent controllers. On the other hand, if an external actuation is preferred the control system can only utilize a shared controller for both reconfiguration and positon holding. Controller software design The final step before the conceptual verification focuses on the choice of controller and software features. Depending on the input from earlier steps, a controller can be classified in terms of providing standard and custom functionalities. The robotics features such as jogging, working coordinate frame transformations and motion control for each motor can be standardized and rescaled for a reconfigurable fixture. The second type of standard features are process related capabilities such as communication, sensor integration, user interface and offline programming language. The custom functions of a controller are also divided into two categories. First, the robotics related group focuses on implementing kinematics, dynamics and workspace capabilities. Second, the active fixturing functionality of a controller focuses on the control of a workpiece in a specific process such as adaptive clamping scenarios [21] or force feedback assembly [55]. In the case of selecting an independent control system, a designer is expected to include all aspects of the software groups. On the other hand, in a scenario where motion is created externally and a shared controller is of preference, then the software creation part will be limited to off-line programming of that particular automation equipment. In all scenarios, the software capability will vary extensively with respect to earlier inputs. Thus, the impact of software functionality is quantified in terms of cost where time based costing derived from Kaplan and Anderson [56] is proposed. Thus the conceptual software cost (Cs) is C s ¼ C wh T A

ð9Þ

where Cwh and TA are cost per work-hour and allocated effort time for functionality development. The time-based costing of a software focuses on the standardization of software functionalities; and it assumes that the designer has already established knowledge regarding the software functionalities. In a case where this knowledge does not exist, other time allocation and accurate cost estimations can be used as widely reviewed in Ref. [57]. Conceptual analysis and verification The conceptual analysis and verification is comprised of two consecutive steps. The first step aims to calculate the parameters that are an integral part of the design functions presented in Section “Thresholds for functions and constraints”. Then, the Eqs. (1)–(8) are calculated, and through a weight distribution, a final efficiency value is realized. Furthermore, the fixturing parameters accuracy and repeatability are also introduced in this section to define reusability. Initially, a fixturing solution is analyzed based on weight and dimensional requirements so that total weight (Wa) and volume (Va) of the fixturing solution are computed. The next step is to determine the workspace boundaries so that the number of satisfied products in a product family within the feasible workspace (Pa) can be determined. The workspace computation can be conducted efficiently by several algorithms feasible for various types of kinematic structures [58]. Through estimations on

7

accuracy and repeatability in the case of internal actuation [59], the achieved reusability (Pra) value can be determined by analyzing the process requirements in the target group. The capital cost analysis of the fixturing solution is conducted from two perspectives. The proposed solution assumes that the standard components already have an off-the-shelf price. However, for the custom parts the procedure estimates a cost by assuming that the knowledge of manufacturing prices of similar parts is available to the designer. On the other hand, there exist various cost estimation models based on feature and machining type analysis for a more accurate estimation [60]. Therefore, the parameters for internal (Ce) and external equipment (Cf) can be determined. Time analysis in the proposed procedure is calculated based on the output of actuation and position-holding steps. By determining the time span for machine stop-reconfigurationmachine start, set-up time (Ts) can be calculated. The binary terms diagnosability (Dc) and convertibility (Coc) are automatically inherited based on the design choices where the reliability (Rec) is computed based on Eq. (7). Subsequently, final efficiency of the fixturing solution is calculated by multiplying parameters with a set of weights (wi) determined by the needs of a manufacturing system through earlier studies, where the equation can be formulated as X10 e i wi i eo ¼ X ð10Þ wi

ei ¼ fW c ; V c ; Rcc ; Ruc ; Cc ; T c ; Dc ; Coc ; Rec ; Mc g i ¼ f1; 2; 3; . . . ; 10g where ei represents the metrics defined by Eqs. (1)–(8) and eo is the final efficiency of the reconfigurable fixture. Once the efficiency is computed, a comparison can be made with the expected efficiency and changes to the design can be implemented. Detailed design and analysis The third stage addresses the realization of conceptual design results. Similar to the conceptual design, the detailed design of mechanical and control is initially completed. Using the output of the detailed design, an analysis step is conducted where the reconfigurable fixture solution is analyzed with respect to kinematics, dynamics, singularity and rigidity. Finally, development of a controller software and integration with a graphical user interface are conducted. For mechanical design, designers may begin with the custom components (e.g. adapters between standard components and platforms). Later, the assembly design can be made. In parallel, the control system design can be conducted. In this step, the wiring is created with respect to specifications of the components selected in conceptual design, and the control unit assembly is realized. Once the parallel design steps are completed, relevant schematics and drawings are generated. In the second step, the detailed analysis is done with respect to a specific kinematic structure. With the mechanical assembly completed, the conceptual kinematics and workspace analyses are revised in case of a change in the conceptual dimensions. Later, the dynamical examination is done with respect to the transmitted forces and torques so that the corresponding results do not surpass the joint specifications. If the selected kinematic structure is part of the parallel kinematics family, the singularity analysis can be conducted either for the coordinates of the number of products achieved (Pra) at the conceptual level or the entire workspace.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

8

I. Erdem et al. / NULL xxx (2017) xxx–xxx

Finally, at specified points, rigidity analysis is conducted by using either stiffness [61] or finite element analysis methods [13]. In the software design step, the kinematic and dynamic analyses can be incorporated into the custom parts of the software. Furthermore, for active fixturing scenarios the specific development with respect to achieved products can be conducted. Then, the synthesis of customized and standardized parts of the controller software is finalized and the graphical user interface (GUI) is modified based on the process requirements. Final verification The final stage of the proposed procedure is based on the recalculation of the Eqs. (1)–(8) with respect to the output of the detailed analysis. Particularly, the achieved reusability (Ruc) term is revisited after the rigidity analysis is conducted. By redistributing the efficiency terms with respect to the weights in Eq. (10), the final efficiency term is recomputed. By making a comparison similar to the conceptual verification stage, certain minor changes in the selected components can be made to increase the efficiency. However, fundamental changes such as in kinematic structure or actuation type cannot be made, as these changes will have a major impact on detailed design and analysis levels. Experimental study In automotive industry, a major assembly operation is the spot welding process of Body-in-White (BiW) components. Based on an experimental setup on an industrial component data from an automotive manufacturer, the utilization of the procedure will be demonstrated with respect to the products in a family of the front side of BiW. The experimental workpiece is illustrated in Fig. 2 with exemplified locator and clamping points. For the experimental study, we assume that the thresholds of metrics are previously defined by studying the current manufacturing paradigm. For this workpiece, the products’ locator and clamping points vary within a work envelope of 20 mm in the XYZ-axes of a car coordinate frame — which corresponds to a product family of 10 products where the majority of products remain within 40 mm. Furthermore, four different spot welding processes for the product family have stiffness requirements of 0.2 mm deformation under 100 N at the locator points. The accuracy/repeatability expectations of all processes have a threshold of 0.1 mm. For all of the processes, the weight and dimensional restrictions for each reconfigurable fixture is 10 kg and 300 mm in a cubical form corresponding to 27 l. The cost threshold here is disclosed as nominal 100 USD ($) so that the fixture components can be relatively expressed. The time limitation for setup operations is maximally 60 s among the target processes. The reliability and modularity expectancy are 0.99 and 0.8 respectively. Table 2 summarizes the metrics in this experimental setup, along with their thresholds.

Fig. 2. Experimental workpiece with locating (denoted as L) and clamping (denoted as C) points.

Table 2 Efficiency metrics with corresponding symbols and thresholds. Parameter

Symbol

Threshold

Weight Volume Reconfigurability Reusability Cost Time Diagnosability Reliability Convertibility Modularity

Wa Vd Pt Prt Ct Tt Dc Ret Coc Mc

5 kg 27 l 10 4 100$ 60 s/locator N/A 0.99 N/A 0.8

In order to clarify the functionality of the procedure, we assume that the designer is interested in generating and comparing two types of kinematic solutions, both externally actuated by an articulated robot in the cell. In the first solution, a serial kinematics reconfigurable – namely Cartesian – is of interest. For the second solution, a parallel kinematics fixture – a Tsai manipulator [61] – will be the focus. In order to demonstrate the impact of different choices in the procedure, the position-holding functionality for the serial kinematics fixture is chosen as extra-locking whereas positon-holding for the parallel kinematics fixture will be manually provided. The standard mechanical component selection can be established with respect to the designer’s available resources. Specifically for this example, the common standardized components for both solutions are precision shafts, linear ball bearings and an endeffector for robot docking. The Cartesian solution, on the other hand, utilizes position-holding component that is pneumatically released. Uniquely for the parallel kinematics solution, standard shaft collars are used for position-holding. The custom components for both solutions are the adapters and platforms. For the control system, the Cartesian solution utilizes the controller on the articulated robot for position-holding control. Furthermore, the software creation is limited to the automatic offline programming generation for controlling reconfiguration with position-holding. The manipulator, however, is of a manual nature — which eliminates the need for a controller, yet programming of the robot for reconfiguration is still required. In the conceptual verification stage, the number of custom components are estimated by following reasoning: (I) For every couple of serially linked standard parts, there exists a custom adapter in between. (II) For every parallel assembly (meaning that the standard components are in a cluster), there are two adapters. For the Cartesian solution, each degree-of-freedom is assumed to have 2 precision shafts and 2 linear ball bearings in parallel assembly — which requires 4 custom components. Each pair of custom components forms a cluster for standard parts and connects to the next degree-of-freedom. Hence, the number of custom components is estimated 10 in addition to 14 standard components. For the Tsai manipulator, a previously established limb structure with 4 precision shafts forming a cluster is utilized. With linear bearings and universal joints included, the number of standard components for the Tsai manipulator is determined as 45 where the custom components are estimated as 7. Since there is previous knowledge about the span of prices, an average cost value for each standard component is utilized. The time values are expected to be available to the designer for different aspects of reconfiguration. In this example, machine stop and start times are assumed 10 s each. Furthermore, automated locking is averagely 0.5 s due to specifications in the standard

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx

position-holding device. Manual locking is assumed 10 s for each limb. Thus, for a Tsai manipulator with three limbs the locking corresponds to 30 s. The cost estimation for controller software development is done with respect to the development time which is assumed to be completed within 50 work-hours with 0.2$ per work-hour. The reachability and dimensional estimations are done based on the kinematic calculations of each fixturing solution by using the standard component dimensions. For example, the standard shaft used in each limb is used in forward kinematics calculation, and the home position of the Tsai manipulator’s end-effector is calculated. Furthermore, the body of each limb is utilized to define

9

the travel length in order to conduct a workspace analysis. Similar to cost estimation, the weight and reliability values are computed based on the specifications of the standard components. The relevant parts of the design procedure are illustrated in Fig. 3. Furthermore, Table 3 shows the conceptual verification results for the Cartesian and Tsai manipulators. In order to evaluate the final fixturing efficiency, we assumed three theoretical manufacturing systems with different characteristics. To increase the visibility of different characteristics, the weight distribution in each system is assumed 200 in total. For the first production system, the emphasis is equally on the cost and time spent. Hence, the fixture designer receives the weights

Fig. 3. Design choices implemented in the procedure for (a) Cartesian (b) Tsai manipulator.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

10

I. Erdem et al. / NULL xxx (2017) xxx–xxx

Table 3 Efficiency metrics and their results for two fixturing solutions. Parameter

Cartesian solution Achieved

Eqs. (1)–(8)

Achieved

Eqs. (1)–(8)

Weight Volume Reconfigurability Reusability Cost Time Diagnosability Reliability Convertibility Modularity

8.457 kg 8l 8 4 55.45$ (Standard = 11.54$ + Custom = 34$ + Software = 10$) 20.5 s 0 0.7 0 Ns = 14 NT = 24

0.15 0.70 0.8 1 0.44 0.66 0 0.70 0 0.64

2.29 kg 26 l 10 4 42.2$ (Standard = 18.4$ + Custom = 29$ + Software = 10$) 50.5 s 0 0.96 0 Ns = 45 NT = 62

0.77 0.04 1 1 0.43 0.16 0 0.96 0 0.88

Parallel kinematics solution

Table 4 Final efficiency per manufacturing system. Parameter

Manufacturing system 1

Manufacturing system 2

Manufacturing system 3

Weight Volume Reconfigurability Reusability Cost Time Diagnosability Reliability Convertibility Modularity Cartesian solution Conceptual efficiency (eo) Parallel kinematics solution Conceptual efficiency (eo)

10 10 20 10 54 54 1 20 1 20 0.60

10 10 20 15 30 88 1 10 1 5 0.64

10 10 30 15 70 30 1 20 1 13 0.60

0.54

0.46

0.63

describing a highly automated manufacturing system with low variants in products and processes. Due to this lower variance, the flexibility metrics receive smaller weights whereas cost and time metrics dominate the distribution. In the second system, the emphasis shifts from cost to time. Having great similarities with the first system, the second manufacturing system aims to solve a critical process with a very limited flexibility requirement that often causes a delay. Thus, considering the time emphasis and very limited flexibility emphasis, the weights are primarily concentrated on the reconfiguration time. The final system emphasizes reconfigurability and cost. The main characteristics of this system represents relatively a low level of automation, which involves higher levels of manual work and product variants compared to other systems. Due to costs of dedicated fixtures caused by product variance, this manufacturing system aims to implement flexible fixtures where the cost metric is benchmarked to the existing fixtures. Consequently, by using the metrics, equation 10 is applied to compute conceptual efficiency (eo). The results of efficiency for respective fixture solutions along with the weight distributions for each manufacturing system are illustrated in Table 4. In a case where the manufacturing system characteristics coincide with system 2, a designer is assumed to choose higher efficiency solution without making changes in the conceptual design steps. Thus, the Cartesian solution is designed and analyzed in a detailed manner. In the first step of detailed design, the custom components are created and integrated with standard components to realize the assembly. Then, the detailed analysis step is conducted to verify the fitness of the solution within the workspace from kinematics, dynamics and rigidity perspectives. Finally, development of the software to control an articulated robot is realized. Each step and the developed Cartesian solutions are illustrated in Fig. 4 along with a laboratory prototype.

In the final verification stage, the efficiency metrics are recomputed to identify the deviation from conceptual verification. The outcome of the final design and the corresponding metrics are presented in Table 5. The deviation is observed in terms of weight, volume, cost and modularity values. The main reason behind the weight increase is the deviation in the weight of custom components. Furthermore, the increase in number of custom components further influences the overall efficiency in terms of cost and modularity. In the case of the Cartesian fixture, this impact is relatively small. However, in this example a rather rough estimation algorithm is utilized. Therefore, with established design knowledge the impact of detailed design can be easily minimized. Discussion The utilization of metrics to evaluate the efficiency of a reconfigurable fixture aims to capture the long-term impact by systematic introduction of the manufacturing parameters. Even if the design time might have an impact on the implementation of reconfigurable assembly fixtures, it is excluded from the procedure. The particular reason is that the design procedure cannot measure itself and such activities heavily rely on the designer’s capability and established knowledge. However, through established knowledge base and systematic procedure use the time to design can be reduced. Furthermore, the concept of lead-time based design where the designers are encouraged to select components from suppliers with smallest lead times is regarded as a part of so-called component library; hence, not included as part of efficiency. In the conceptual design stage, the mechanical design and controller software design are influenced by estimations that lead to deviations in the final efficiency of the fixturing solution. The

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx

11

Fig. 4. (a) Detailed design steps (b) final design.

main source of deviations in estimations result from custom and unique components where certain delimitations such as modularization and standardization can be utilized to minimize these deviations. Specifically, with the established knowledge and modularization approach the number of unique components can be reduced drastically. Secondly, the deviations in software cost are

also correlated to time estimations on adaptation of standard features and creation of unique functionalities. Even though the adaptation of these features can be standardized by parametrical approaches on kinematics and dynamics, the unique functions such as active fixturing scenarios remain to be limited to rough time estimations. Therefore, it becomes apparent that certain

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

12

I. Erdem et al. / NULL xxx (2017) xxx–xxx

Table 5 Efficiency results after detailed analysis. Cartesian solution Parameter

Achieved

Eqs. (1)–(8)

Manufacturing system 2

Efficiency

Weighta Volumea Reconfigurability Reusability Costa Time Diagnosability Reliability Convertibility Modularitya

10.541 kg 7.44 l 8 4 55.45$ (Standard = 11.54$ + Custom = 37.11$ + Software = 10.4$) 20.5 s 0 0.7 0 Ns = 14 NT = 26

0.05 0.72 0.8 1 0.41 0.66 0 0.70 0 0.54 Final verification (eo)

10 10 20 15 30 88 1 10 1 5

0.54 7.24 16.00 15.00 12.29 57.93 0.00 7.02 0.00 2.69 0.62

a

Deviated metrics.

activities on standardization of active fixturing scenarios would further enhance the efficiency of reconfigurable fixtures. Conclusion In this paper, we proposed a design procedure that introduces a parametric map of activities based on fixture design literature in order to enable efficiency-based design of reconfigurable fixtures. Through the utilization of manufacturing system requirements and characteristics, the fixture solution is first conceptually evaluated to estimate an efficiency where each design activity’s impact can be analyzed. Later, the detailed design and analysis steps are integrated to finalize the fixture design. In this way, the proposed procedure enables the industry to develop reconfigurable assembly fixtures based on unique characteristics of a manufacturing system. Furthermore, the procedure can also be utilized in computer-aided fixture design processes where optimization algorithms can be implemented to automate the design activities. Funding This work was supported by Sweden’s Innovation Agency VINNOVA [grant number 2016-01973]. The authors would like to thank Volvo Cars Corporation for their contribution in the experimental study. References [1] Gupta, Y.P., Goyal, S., 1989, Flexibility of Manufacturing Systems: Concepts and Measurements. European Journal of Operational Research, 43:119–135. [2] Grippo, P.M., Gandhi, M.V., Thompson, B.S., 1987, The Computer-aided Design of Modular Fixturing Systems. The International Journal of Advanced Manufacturing Technology, 2:75–88. [3] ElMaraghy, H., 2005, Flexible and Reconfigurable Manufacturing Systems Paradigms. International Journal of Flexible Manufacturing Systems, 17:261– 276. [4] Jonsson, M., Ossbahr, G., 2010, Aspects of Reconfigurable and Flexible Fixtures. Production Engineering, 4:333–339. [5] Shirinzadeh, B., 1995, Flexible and Automated Workholding Systems. Industrial Robot, 22:29–34. [6] Hu, S.J., 2013, Evolving Paradigms of Manufacturing: From Mass Production to Mass Customization and Personalization. Procedia CIRP, 7:3–8. [7] Bakker, O.J., Papastathis, T.N., Ratchev, S.M., Popov, A.A., 2013, Recent Research on Flexible Fixtures for Manufacturing Processes. Recent Patents on Mechanical Engineering, 6:107–121. [8] Michalos, G., Makris, S., Papakostas, N., Mourtzis, D., Chryssolouris, G., 2010, Automotive Assembly Technologies Review: Challenges and Outlook for a Flexible and Adaptive Approach. CIRP Journal of Manufacturing Science and Technology, 2:81–91. [9] Trappey, J.C., Liu, C.R., 1990, A Literature Survey of Fixture Design Automation. The International Journal of Advanced Manufacturing Technology, 5:240–255. [10] Rong, Y., Bai, Y., 1997, Automated Generation of Fixture Configuration Design. Journal of Manufacturing Science and Engineering-Transactions of The ASME, 119:208–219.

[11] Bi, Z.M., Zhang, W.J., 2001, Flexible Fixture Design and Automation: Review, Issues and Future Directions. International Journal of Production Research, 39:2867–2894. [12] Mervyn, F., Senthil Kumar, A., Nee, A.Y.C., 2006, Fixture Design Information Support for Integrated Design and Manufacturing. International Journal of Production Research, 44:2205–2219. [13] Hargrove, S.K., Kusiak, A., 1994, Computer-aided Fixture Design: A Review. International Journal of Production Research, 32:733–753. [14] Kang, X., Peng, Q., 2009, Recent Research on Computer-aided Fixture Planning. Recent Patents on Mechanical Engineering, 2:8–18. [15] Yu, H., Li, B., Wang, Y., Hu, Y., 2012, Conceptual Design and Workspace Analysis of Reconfigurable Fixturing Robots for Sheet Metal Assembly. Assembly Automation, 32:293–299. [16] Millar, A., Kihlman, H., 2009, Reconfigurable Flexible Tooling for Aerospace Wing Assembly. SAE Techni Paper. [17] Zhang, J., Yang, J., Li, B., 2009, Development of a Reconfigurable Welding Fixture System for Automotive Body. ASME/IFToMM International Conference on Reconfigurable Mechanisms and Robots (ReMAR, London), pp.736–742. [18] Li, X.N., Dang, X.G., Xie, B.Q., Hu, Y.L., 2015, Flexible Tooling Design Technology for Aircraft Fuselage Panel Component Pre-assembly. Assembly Automation, 35:166–171. [19] Erdem, I., Helgosson, P., Gomes, A., Engstrom, M., 2016, Automated Flexible Tooling for Wing Box Assembly: Hexapod Development Study. SAE Technical Paper 2016-01-2110. SAE International. [20] Soetebier, S., Müller, C., Mauser, N., Kock, S., Legeleux, F., 2008, Flexible Automation for Automotive Body Assembly. 4th IEEE Conference on Automation Science and Engineering CASE, 341–346. [21] Bakker, O.J., Papastathis, T.N., Popov, A.A., Ratchev, S.M., 2013, Active Fixturing: Literature Review and Future Research Directions. International Journal of Production Research, 51:3171–3190. [22] Boyle, I., Rong, Y., Brown, D.C., 2011, A Review and Analysis of Current Computer-aided Fixture Design Approaches. Robotics and Computer-Integrated Manufacturing, 27:1–12. [23] Pehlivan, S., Summers, J.D., 2008, A Review of Computer-aided Fixture Design with Respect to Information Support Requirements. International Journal of Production Research, 46:929–947. [24] Wang, H., Rong, Y., Li, H., Shaun, P., 2010, Computer Aided Fixture Design: Recent Research and Trends. CAD Computer Aided Design, 42:1085–1094. [25] Kusiak, A., Heragu, S.S., 1987, Group Technology. Computers in Industry, 9:83– 91. [26] Greene, T.J., Sadowski, R.P., 1984, A Review of Cellular Manufacturing Assumptions, Advantages and Design Techniques. Journal of Operations Management, 4:85–97. [27] Heragu, S.S., 1994, Group Technology and Cellular Manufacturing IEEE Transactions on Systems. Man and Cybernetics, 24:203–215. [28] Kusiak, A., 1987, The Generalized Group Technology Concept. International Journal of Production Research, 25:561–569. [29] Almomani, M.A., Aladeemy, M., Abdelhadi, A., Mumani, A., 2013, A Proposed Approach for Setup Time Reduction Through Integrating Conventional SMED Method with Multiple Criteria Decision-making Techniques. Computers & Industrial Engineering, 66:461–469. [30] McIntosh, R.I., Culley, S.J., Mileham, A.R., Owen, G.W., 2000, A Critical Evaluation of Shingo’s ‘SMED’ (Single Minute Exchange of Die) methodology. International Journal of Production Research, 38:2377–2395. [31] Peng, G., Wang, G., Liu, W., Yu, H., 2010, A Desktop Virtual Reality-based Interactive Modular Fixture Configuration Design System. CAD Computer Aided Design, 42:432–444. [32] De Toni, A., Tonchia, S., 1998, Manufacturing Flexibility: A Literature Review. International Journal of Production Research, 36:1587–1617. [33] Gerwin, D., 1993, Manufacturing Flexibility. A Strategic Perspective. Management Science, 39:395–410. [34] Sethi, A.K., Sethi, S.P., 1990, Flexibility in Manufacturing: A Survey. International Journal of Flexible Manufacturing Systems, 2:289–328.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004

G Model CIRPJ 427 No. of Pages 13

I. Erdem et al. / NULL xxx (2017) xxx–xxx [35] Elkins, D.A., Huang, N., Alden, J.M., 2004, Agile Manufacturing Systems in the Automotive Industry. International Journal of Production Economics, 91:201– 214. [36] Gunasekaran, A., Yusuf, Y.Y., 2002, Agile Manufacturing: A Taxonomy of Strategic and Technological Imperatives. International Journal of Production Research, 40:1357–1385. [37] Koren, Y., Heisel, U., Jovane, F., Moriwaki, T., Pritschow, G., Ulsoy, G., et al, 1999, Reconfigurable Manufacturing Systems. CIRP Annals Manufacturing Technology, 48:527–540. [38] Bi, Z.M., Lang, S.Y.T., Verner, M., Orban, P., 2008, Development of Reconfigurable Machines. The International Journal of Advanced Manufacturing Technology, 39:1227–1251. [39] Koren, Y., 2013, The Rapid Responsiveness of RMS. International Journal of Production Research, 51:6817–6827. [40] Mehrabi, M.G., Ulsoy, A.G., Koren, Y., 2000, Reconfigurable Manufacturing Systems: Key to Future Manufacturing. Journal of Intelligent Manufacturing, 11:403–419. [41] Gunasekaran, A., 1998, Agile Manufacturing: Enablers and an Implementation Framework. International Journal of Production Research, 36:1223–1247. [42] Gunasekaran, A., 1999, Agile Manufacturing: A Framework for Research and Development. International Journal of Production Economics, 62:87–105. [43] Wiendahl, H.P., ElMaraghy, H.A., Nyhuis, P., Zäh, M.F., Wiendahl, H.H., Duffie, N., et al, 2007, Changeable Manufacturing – Classification, Design and Operation. CIRP Annals Manufacturing Technology, 56:783–809. [44] De Toni, A., Tonchia, S., 2001, Performance Measurement Systems Models, Characteristics and Measures. International Journal of Operations and Production Management, 21:46–70. [45] Neely, A., Gregory, M., Platts, K., 1995, Performance Measurement System Design: A Literature Review and Research Agenda. International Journal of Operations and Production Management, 15:80–116. [46] White, G.P., 1996, A Survey and Taxonomy of Strategy-related Performance Measures for Manufacturing. International Journal of Operations and Production Management, 16:42–61. [47] Folan, P., Browne, J., 2005, A Review of Performance Measurement: Towards Performance Management. Computers in Industry, 56:663–680. [48] Hon, K.K.B., 2005, Performance and Evaluation of Manufacturing Systems. CIRP Annals Manufacturing Technology, 54:675–690.

13

[49] Nudurupati, S.S., Bititci, U.S., Kumar, V., Chan, F.T.S., 2011, State of the Art Literature Review on Performance Measurement. Computers & Industrial Engineering, 60:279–290. [50] Kusiak, A., Lee, G.H., 1997, Design of Parts and Manufacturing Systems for Reliability and Maintainability. The International Journal of Advanced Manufacturing Technology, 13:67–76. [51] Abdullah, C.Z., Saadat, M., Rakhodaei, H., 2014, A Methodology for Workspace Identification of Parallel Robots Using Parametric Sweep Search Method. ASME 2014 12th Biennial Conference on Engineering Systems Design and Analysis ESDA, . [52] Shang, M., Butterfield, J., Armstrong, C., Gibson, R., 2011, A Flexible Fixture for Aircraft Wing Assembly Based on a Parallel Kinematic Machine (Exechon). SAE International Journal of Aerospace, 4:839–849. [53] Cirillo, P., Maria, G.D., Natale, C., 2015, Customization of Low-cost Hexapod Robots Based on Optimal Design Through Inverse Dynamics Computation. 2015 20th International Conference on Methods and Models in Automation and Robotics (MMAR), 1110–1115. [54] Kihlman, H., Engstrom, M., 2010, Flexapods – Flexible Tooling at SAAB for Building the NEURON Aircraft. SAE International. [55] Jonsson, M., Stolt, A., Robertsson, A., Murray, T., Nilsson, K., 2011, Force Controlled Assembly of a Compliant Rib. SAE Techni Paper. [56] Kaplan, R.S., Anderson, S.R., 2004, Time-driven Activity-based Costing. Harvard Business Review, 82. 131-8+50. [57] Jorgensen, M., Shepperd, M., 2007, A Systematic Review of Software Development Cost Estimation Studies. IEEE Transactions on Software Engineering, 33:33–53. [58] Merlet, J.P., Gosselin, C.M., Mouly, N., 1998, Workspaces of Planar Parallel Manipulators. Mechanism and Machine Theory, 33:7–20. [59] Shiakolas, P.S., Conrad, K.L., Yih, T.C., 2002, On the Accuracy, Repeatability, and Degree of Influence of Kinematics Parameters for Industrial Robots. International Journal of Modelling and Simulation, 22:245–254. [60] Jung, J.Y., 2002, Manufacturing Cost Estimation for Machined Parts Based on Manufacturing Features. Journal of Intelligent Manufacturing, 13:227–238. [61] Tsai, L.-W., Joshi, S., 1999, Kinematics and Optimization of a Spatial 3-UPU Parallel Manipulator. Journal of Mechanical Design, 122:439–446.

Please cite this article in press as: I. Erdem, et al., A novel comparative design procedure for reconfigurable assembly fixtures, NULL (2017), http://dx.doi.org/10.1016/j.cirpj.2017.06.004