Cost-effective concept development using functional modeling guidelines

ARTICLE IN PRESS

JID: RCM

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Contents lists available at ScienceDirect

Robotics and Computer–Integrated Manufacturing journal homepage: www.elsevier.com/locate/rcim

Full length Article

Cost-effective concept development using functional modeling guidelines Heungjae Cho, Jaeil Park∗ Industrial Engineering Department, Ajou University, Suwon City, Gyeonggi-do 443-749, Republic of Korea

a r t i c l e

i n f o

Keywords: Functional modeling Cost-effectiveness Design for assembly Concept development Functional modeling guidelines

a b s t r a c t Functional modeling provides a formal approach to early concept development by directly translating customer needs into desired product behavior. By first developing product functionality, the gap between customer needs and product form is lessened. 70–80% of the life-cycle costs of a product are determined by decisions made by designers during the early design stages. Design for assembly (DFA) is a process in which products are designed with ease of assembly in mind. If a product contains fewer parts it will take less time to assemble, thus reducing assembly costs. If functional models can be evaluated based on the three principles of DFA, cost-effective product concepts can be developed earlier in the design stage. Therefore, it is important to estimate the cost-effectiveness of functional models as early as possible. The main goal of this research is to employ the three principles of DFA to evaluate functional models and develop functional modeling guidelines for reducing and integrating functions. We introduce a case study that illustrates how the proposed method works and helps to reduce the number of functions in the assembly process. Conclusions are then discussed and future research is described. © 2018 Elsevier Ltd. All rights reserved.

1. Introduction The main goal in product design is to understand key customer requirements and use them to build a cost-effective product architecture in order to better serve current customers while simultaneously attracting new ones [1]. Functional modeling provides a formal approach to early concept development by directly translating customer needs into desired product behavior [2]. By first developing product functionality, the gap between customer needs and product form is lessened. Modeling the functions in existing designs is a key process in researching competing products, gathering ideas for new products, and learning about functional modeling. 70–80% of the life-cycle costs of a product are determined by the decisions made by designers during the early design stages. Therefore, it is important to evaluate design concepts in terms of “values” as early as possible [3]. Value engineering (VE) has provided a systematic method for improving the "value" of products by performing an examination of functional models [4]. VE describes a functional model by using a verb and noun, and classifying its functions as primary and secondary functions. It then uses intuitive logic to test the values of functions and graphically display them in diagram or model form. Due to the importance of cost-effective design concepts, during VE, a concept’s functional model should be evaluated in terms of cost until its target cost is met. Despite the importance of cost, developing a cost-effective functional model at this early stage is very difficult because the concept of a prod-

∗

uct is formed when the overall shape, major features, and materials are determined at a later stage. Design for assembly (DFA) is a process in which products are designed with ease of assembly in mind. If a product contains fewer parts it will take less time to assemble, thus reducing assembly costs [5]. Reduction of the number of parts in an assembly has the added benefit of reducing the total cost of parts required for assembly. This is typically where the major cost benefits of the application of DFA occur [6]. Although DFA assists in the design of a product and improves its assemblability, it does not have the ability to develop design concepts that may lead to cost reductions during functional modeling. In other words, the concept of a product is determined by functional models, including functions, functional relationships, and energy flows. However, this does not allow designers to review the cost-effectiveness of a functional model they are currently working on. Because functional modeling guidelines are not available at this stage, designers rely on their experience in their areas of engineering focus during the development of functional models. The main goal of this research is to employ the three principles of DFA to evaluate functional models and develop functional modeling guidelines for reducing and integrating functions. The scope of this article is limited to the functional modeling portion of conceptual mechanical design. We begin by reviewing the terminology and motivation for functional modeling in Section 2. Section 3 presents the development of functional modeling guidelines based on the three design principles from DFA. We then introduce a case study that illustrates how the pro-

Corresponding author. E-mail addresses: [email protected] (H. Cho), [email protected] (J. Park).

https://doi.org/10.1016/j.rcim.2018.01.007 Received 31 May 2017; Received in revised form 21 January 2018; Accepted 31 January 2018 Available online xxx 0736-5845/© 2018 Elsevier Ltd. All rights reserved.

Please cite this article as: H. Cho, J. Park, Robotics and Computer–Integrated Manufacturing (2018), https://doi.org/10.1016/j.rcim.2018.01.007

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

posed method works and helps to reduce the number of parts required for assembly in Section 4. We then provide conclusions and discuss future research.

and components can be clearly defined. This is important information for the conceptual design process. These three methods provide a systematic approach for identifying the modules in a product from a functional model. Jiang and Xie [17] proposed a method to combine a functional representation tool known as System State Flow Diagram (a solution independent approach), a solution search tool referred to as Morphology Table, and a Design Structure Matrix (mainly a solution dependent tool). The proposed approach incorporates Multiple Domain Matrix (MDM) to integrate the knowledge of both solution independent and dependent analyses. Yu-liang Wei [18] proposed a product top-down design process model based on product functions and physical topologies. An integrated collaborative decision-making system based on task agents was developed. The system has functions of optimized decomposition of module tasks, reasoning, solving of parts tasks, design collaboration, and results release, which is verified by the transmission design case. Sa’ed and Kamrani [19] proposed a method for identifying components that can be developed in parallel. In the proposed method, requirements and product functional requirements are set up first. The product is then decomposed according to functional and physical characteristics. Next, a similarity index is introduced to measure the association between the base components. Finally, we use clustering technology to integrate base components into design modules based on the similarity indices. A design language, called a functional basis, was introduced in functional modelling, where product functionality is characterized in a verbobject (function-flow) format. The set of functions and flows is intended to comprehensively describe the functions of mechanical devices. Clear definitions are provided for each function and flow [20]. With such a basis, designers can communicate product functionality in a universal language [21]. The main value in using these methods is in how they include additional insights that inspire users and allow their minds to wander or make intuitive leaps. The ability to decompose a design task is fundamental to arriving at creative solutions [22]. Likewise, it is critical to represent abstract and incomplete information to make decisions early in the design process or product development. The development of a functional basis and a systematic approach to functional modeling offers the best means of providing additional information for developing product concepts. Boothroyd and Alting [23] analyzed existing designs of hundreds of products and suggested design improvements based on ease of manufacturing and assembly. Using the experience gained from these products, they then developed a very large set of guidelines on how to estimate the feasibility of a design (from a manufacturing point of view) and potential methods for improving the design. Since then, DFA has been widely used to evaluate ease of assembly. Sako and Murray [24] proposed a method for generating each sub-assembly or component that can be clearly linked to a particular sub-function while defining an assembly sequence for a product with a very high level of detail. The structure of a product can be defined by the assembly sequence for the types of connections between parts. Stone and McAdams developed a conceptual DFA by reducing the number of components without using a detailed model of the product during the conceptual design stage [25]. They highlighted the importance of determining the best assembly sequence and spatial module layout in order define appropriate product architecture. TRIZ provides the most structured and systematic approach, and is best used to augment the traditional design improvement methods as it is currently applied in DFA [26]. The practical method for providing engineering system ideality is trimming. In accordance with that methodology, trimming conditions are formulated for the engineering system components, meaning the functions of the components are transferred to other components and to super-system elements (if those elements have enough resources) [27]. As a result, a problem is formulated around the performance of super-system components and elements of removed component functionality. Typically, the components selected for trimming have a very low ratio of number and quality of functions

2. Background Developing product architectures is a key phase in the design and development processes. It encompasses the transformation of product functionality into alternative product layouts. Much of the existing research on improving products has focused on modular product architecture creation during the conceptual design stage. Pandremenos et al. [7] defined modular design as a technique capable of creating many different finished products from a limited set of modular components. Dalgleish et al. [8] compared integral architecture with three different types of modularity: “Modular-Slot”, which uses components that all have different interfaces; “Modular-Bus”, which has a central component linked to all other components; and “Modular-Section”, which is concerned with the interface between sections of a product. In the automotive industry, depending on the context, modularity has been defined from three perspectives [9,10]: Modularity In Design (MID), Modularity In Use (MIU), and Modularity In Production (MIP). MID is related to the modular architecture of products, MIU disassembles products from the viewpoint of customer requirements, and MIP aims to assemble modules as easily as possible. Shahzad and Hadj-Hamou [11] proposed a concept of sustainable mass customization to address the challenges where an economically infeasible product for a market segment is replaced by an alternative superior product variant nearly at the same cost of mass production. AlGeddawy and ElMaraghy [12] proposed automatically redesigning product variants using physical commonality, instead of evaluating alternate solutions provided by designers using commonality indices. The model innovatively balances two conflicting strategies: Design for Manufacturing and Assembly (DFMA), and products modularity. It hierarchically clusters the common components among product variants to define a core platform while combining as many of the common parts as possible into integral parts and modules using Cladistics. Tyagi et al. [13] has proposed a method to exploit lean thinking concepts in order to manage, improve and develop the product faster while improving or at least maintaining the level of performance and quality. Lean thinking concepts encompass a broad range of tools and methods intended to produce bottom line results. However, the value stream mapping (VSM) method is used to explore the wastes, inefficiencies, and non-value added steps in a single definable process out of the complete product development process (PDP). da Cunha and Dias [14] proposed a model for finding design data groups that represent each mechanical design phase, which will be called the phase’s design signature. In addition, current data should be an evolution of the geometric and non-geometric information of the previous design phase. In this paper, the purpose is to identify and model a set of design features that encapsulate the design data and the transformations that occurred during the mechanical design phases. Significant advancements have been made in integrating components and reducing part counts at the form-level of design. New concepts with a functional basis and time ordered function chains are used to formally derive functional models for products. Functional modeling, also known as functional decomposition, plays important role in breaking down the overall functionality of a product into smaller, more easily solvable sub-functions, and in developing functional models. With functional modeling, Stone et al. [15] proposed three heuristic modules for modulating functions by identifying sub-functions that can be grouped together as a module: 1) dominant modules, 2) branch modules, and 3) convert-transmit modules. These three heuristic modules are presented using formal functional decomposition and heuristic methods. This allows modular design to be performed earlier in the product development process. Yu et al. [16] proposed a functional decomposition method that defines the logical relationships between functions, functional levels, and functional hierarchies. The relationships between functional units 2

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Table 1 Three flow classes. Flow class

Flow basic

Material Signal Energy

Human, Gas, Liquid, Solid Status (Auditory, Olfactory, Tactile, Taste, Visual), Control Human, Acoustic, Biological, Chemical, Electrical, Hydraulic, Magnetic, Mechanical (Rotational, Translational, Vibrational) Table 2 Eight functions classes. Function class

Function basic

Branch Channel Connect Control Magnitude Convert Provision Signal Support

Separate, Remove, Define, Distribute, Dissipate Import, Export, Transport, Transmit, Guide, Rotate, Allow DOF Couple, Mix Actuate, Regulate, Change, Form Convert Store, Supply Sense, Indicate, Measure Stop, Stabilize, Secure, Position

Functions and flows are combined in verb-object form to describe a sub-function. Subsequent functional models are expressed in the standard vocabulary of the functional basis using the definitions provided in Tables 1 and 2. An example sequential function chain model for a consumer power screwdriver is presented in Fig. 1. Note that the chosen system boundary treats the bit as an input flow.

versus negative factors. All of these various methodologies agree that early functional modeling is important. 3. Research method Design for assembly is a process in which products are designed with ease of assembly in mind. If a product contains fewer parts it will take less time to assemble, thus reducing assembly costs. DFA has three basic principles that evaluate the potential integration of two neighboring components. The first principle is to examine if two components differ with respect to their physical materials. The second is to examine if two components have mobility relative to each other. The third is to examine if the components must be separate to enable assembly [28]. If functional models can be evaluated using the three principles of DFA, they can provide better ideas for cost-effective products in the early design stage.

3.1.2. Functional modeling with FAST diagrams Typically, customers only need to identify input or output flows, ignoring flows internal to the product. The level of detail at which input/output flows are identified depends on the type of design. In redesign, flows are typically well defined and benefit from the use of precise descriptions. Redesign is well suited to the sequential function chain model. However, in a conceptual design problem, flows may be listed more generally (even as material, energy, or signal) and refined as the design concept develops. A Function Analysis System Technique (FAST) is a method used to analyze functions by identifying correlations between all functions using How-Why logic during the conceptual stage [31,32]. ‘Why’ is represented leftward, with ‘How’ being expressed rightward. It is not the end product or result, but rather a starting point that helps to find functionality that is missing at the conceptual design stage. Any function in the How or Why logic is a critical path function. If a function along the Why direction enters the basic function(s), then it is a major critical path; otherwise, it will terminate in an independent (supporting) function and be considered a minor critical path. Supporting functions are typically secondary. This paper uses both the FAST to construct a functional model and the sequential function chain model to cross check any missing functions. To begin the graphical representation, a FAST model is presented in Fig. 2, illustrating the sequential function chain model describing the sander griper.

3.1. Functional modeling 3.1.1. Sequential function chain model Functional modeling is a key step in the product design process. In order to clarify the functional definition of the components constituting the product, it is necessary to have a design language, called a functional basis, where product functions are characterized in a verb-object (function-flow) format. A verb-object (function-flow) format consists of three flow classes and eight function classes [15]. The set of functions and flows is intended to comprehensively describe the mechanical design space. The flow class is a change in material, energy, or signal with respect to time. Expressed as the object of a sub-function, a flow is the recipient of the function operation. The functional basis flows are provided in Table 1. A functional basis is a design language consisting of a set of functions and a set of flows that are used to form sub-functions. The function classes used in the functional basis are provided in Table 2. The first column lists the eight function classes. These classes are extended to include basic functions in the second column [29]. The first task in functional model derivation is to create a sequential function chain model, which is a graphical representation of product functionality with input/output flows. A functional model may require the addition of new sub-functions or their combinations, thus defining the interfaces of modules within the representation. If a flow is transformed into another type, then the process follows the operations of the transformed flow until it exits the product. The functional model is presented in chronological order.

3.1.3. Flow heuristic modules In the conceptual design phase, groups of sub-functions related by flows are observed to form subsystems or modules in the device. This observation leads to the formulation of two heuristics for identifying modules based on the two conditions that a flow may experience: 1) a flow may pass through a product unchanged, or 2) a flow may be converted into another type [14]. We classify the heuristic modules as non-conversion and conversion modules. They are displayed in the primary and secondary levels in the FAST diagram. The primary modules describe the characteristic or task, which, from the user’s point of view, is the primary reason for the existence of an item. This is the main purpose the product or process was designed to serve. Secondary 3

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 1. Sequential function chain model of a grip sander [30].

Fig. 2. The FAST model of a grip sander.

4

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 3. Flow heuristic modules in FAST.

modules are those designed inside functions are necessary to allow the primary modules to operate. Secondary models are any modules that directly contribute to accomplishing the goal of the primary module. These heuristic models are presented in Fig. 3, showing two conversion and six non-conversion modules and describing their paths.

Part 2 move relative to each other because mechanical energy (rotational energy) is involved. In the case of Fig. 5(B), part movement is not essential because mechanical conversions are not required. That is to say, electricity is converted into light through Parts 1 and 2. 3.2.2. Functional modeling guideline 1.2: Within a conversion module, try to remove intermediate functions if the starting and ending flow basics are the same According to guideline 1.1, the part moves relative to all other parts only if mechanical conversions are required. Then, whether mechanical conversions are essential needs to be examined based on their repetition, which causes complexity. The set of functions that a flow passes through, from entry of the flow to conversion of the flow, define a conversion module. Conversion functions accept a flow of material or energy and convert the flow into another form of material or energy. A conversion function appears as converting a starting flow A into an ending flow B. A conversion function exists in a chain and the chain presents an opportunity to form a module. A conversion module consists of a set of conversion functions. A conversion module is represented by a set of starting and ending flows with respect to overall flow. If the same starting and ending flow basic is created within a conversion module, the functions with the same flow basic can be represented by a subconversion module that coverts starting flow A into ending flow A. This means that the sub-conversion module can be reduced to a single function. That is to say, intermediate functions in the sub-module can be removed. Function modeling guideline 1.2 is as follows: within a conversion module, try to remove intermediate functions if the starting and ending flow basics are the same. Fig. 6(A) shows a Clipper 1 model. The conversion module has five conversion functions. Within the module, the functions of Parts 2, 3, and 4 are used sequentially to convert mechanical torque (flow A) into mechanical torque (flow A) and form a sub-conversion module. The starting and ending flows of the sub-conversion module are the same. Thus, it is possible to consider reducing Parts 2, 3, and 4 into a single part that converts mechanical torque into mechanical torque. Fig. 6(B) shows a Clipper 2 model containing a new wave-shaped part that directly trans-

3.2. Function modeling guidelines for cost-effective concept development If a functional model can be evaluated with the three principles of DFA in the early design stage, the final product will be more reliable and cost-effective because there will be fewer connections [37]. The three principles of DFA are to examine: 1) part movement relative to all other parts that have been assembled, 2) parts of a different material or parts isolated from all other parts that have been assembled, and 3) parts that are separable to allow for disassembly for in-service adjustment or replacement. During the conceptual design phase, groups of sub-functions related by flows are observed to form conversion or nonconversion modules in the mechanical device. This observation leads to the formulation of three functional modeling guidelines for identifying cost-effective concepts based on the three DFA principles that a function-flow can experience, as shown in Fig. 4. The necessary starting point for a set of functional modeling guidelines is a well-defined functional model, such as the one derived in the previous section. 3.2.1. Functional modeling guideline 1.1: Within a conversion module, allow part movement if mechanical energy is involved The first principle of DFA starts with a question: “Does the part move relative to all other moving parts?” This leads to the following answer on a functional level. The part moves relative to all other parts only if mechanical conversions are required. In Table 1, only the energy flow class can create part movement. Among its flow basics, only mechanical energies, such as rotation, translation, or vibration are involved in physical movement. Part movement is essential when a conversion module is associated with mechanical energy. As shown in Fig. 5(A), Part 1 converts electricity to rotation and Part 2 transmits rotation. Part 1 and 5

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 4. Three functional modeling guidelines with three principles of DFA.

mits mechanical torque. It plays the same role as the three parts used in Clipper 1.

are not the same, however, the integration of parts depends on their assembly dependencies. That is to say, possible disassembly depends on the assembly dependencies of the part within the non-conversion module. If the part is independent in the assembly sequence, then reordering the part can allow other parts to be integrated. Case 2 in Table 3 illustrates this half of the guideline. Functional modeling guideline 3 is illustrated using the filter housing of an air conditioner in Fig. 8. There are five parts. The functions of Part ○ 1 and Part ○ 3 are “Secure solid” and “Stabilize Solid”, the functions of Part ○ 2 and Part ○ 5 are “Secure Solid “ and “Stabilize Solid”, and the function of Part ○ 4 is “Remove Solid”. Within the non-conversion module, four functions are chained to support the “Remove Solid” function. If the orders of the function chain and the assembly of parts are not the same, then the integration of parts depends on their assembly dependencies. Part ○ 4 is the case we will examine. Because assembling this part is dependent on the assembly sequence, reordering this part cannot be allowed. The part must be disassembled. However, Part ○ 1 and Part ○ 3 can be candidates for integration because the orders of their functions and the assembly of parts are same. The same applies to Part ○ 2 and Part ○. 5

3.2.3. Functional modeling guideline 2: Within non-conversion modules, use the part of the same material or do not isolate the part when the same function class is used under the same flow basic The second principle of DFA comes with a question: “Must the part absolutely be of a different material from the other parts?” This leads to the following answer on the functional level. Within non-conversion modules, use the part of the same material or do not isolate the part when the same function class is used under the same energy basic. The function of the part is determined by its flow basic (object) and function basic (verb). Whether the part must absolutely be of a different material from the other parts can be determined by the type of its flow basic and the function class of the function basic. If the function class of two parts is the same under the same flow basic and the function of the parts can be identical, then the parts can be candidates for integration due to the possibility of using the same material. As shown in Parts 1 and 2 in Fig. 7(A), the material of Part 1 can be the same as that of Part 2 because both their flow basic (solid) and function class (support) are the same. As a result, Part 3 is an integration of Parts 1 and 2. As shown in Fig. 7(B), the functions of Parts 4 and 5 are “stabilize solid” and “secure solid”, respectively. However, their function class (support) is the same as under the same energy basic (solid). Therefore, the integration of the parts can be considered.

3.3. Validation of the functional modeling guidelines The necessary starting point is a well-defined functional model, such as the one defined in Fig. 3 in Section 3.1.3. Based on functional modeling guideline 1.1, part movement is allowed for conversion modules 1 and 2 because these modules are activated by mechanical energy. Based on functional modeling guideline 1.2, the conversion flows are essential because the starting and ending flow basics involved are not the same. Based on guideline 2, within non-conversion modules, one should use the part of the same material or avoid isolating the part when the same function class is used under the same energy basic. If the function class two of parts is the same under the same flow basic, then the function of the parts can be identical and the parts can be candidates for integration due to the possibility of using the same material. For nonconversion module 1 in Fig. 3 in Section 3.1.3, the material of Part 1 can be the same as that of Part 2 because both their flow basic (electricity) and function class (control magnitude) are the same. As a result, it is possible that Part 1 (controller terminal) and Part 2 (switch inner)

3.2.4. Functional modeling guideline 3: Within non-conversion modules, use a different part or allow possible disassembly if the orders of the function chain and assembly are not the same After identifying the material (guideline 2) of the part, its disassembly, including as adjustment or replacement, must be checked to determine if the part potentially requires disassembly from other parts. Within non-conversion modules, the part can be isolated from other parts even when guideline 2 is satisfied. In order to examine possible disassembly requirements, the orders of the function chain and the assembly of parts must be investigated. If the orders of the function chain and the assembly of parts are the same, then parts within non-conversion modules can be integrated. Case 1 in Table 3 illustrates this half of the guideline. If the orders of the function chain and the assembly of parts 6

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 5. Example of function modeling guideline 1 [33,34]. Table 3 Possible disassembly through functional modeling guideline 3.

Case

1

If the orders of the function chain and the assembly of parts are the same.

2

If the orders of the function chain and the assembly of parts are not the same.

Functional modeling guideline 3

7

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 6. Example of functional modeling guideline 1.2 [35,36].

can be integrated. For non-conversion modules 2, 3, and 4, the materials of the parts must be different due to having different flow basics and function classes. The results of functional modeling guideline 2 are presented in Table 5. As the final guideline, functional modeling guideline 3 must also be satisfied for possible part integration. Within non-conversion module 1, Part 1 is considered to be the same material as Part 2 based on guideline 2. If the orders of the function chain and the assembly of parts are the same, then the parts within non-conversion module 1 do not need to

be disassembled. Fig. 9 illustrates that the orders of the function chain and the assembly of parts are the same. As a result, parts within nonconversion module 1 can be integrated. After applying the functional modeling guidelines for cost-effective concept development, the total number of functions is reduced by 5% (20 ≥ 19). The total number of parts is reduced by 9.5% (21 ≥ 19). This tells us that design efficiency is improved because we have reduced the number of parts involved.

8

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Non-conversion module

Part 1

Part 2

Functional modeling guideline 2

Controller assembly 1

Flow basic

Solid

Solid

Function basic

Stabilize

Stabilize

G2: Same material due to same flow basic and function class.

Function class

Support

Support

Flow basic

Solid

Solid

Function basic

Stabilize

Secure

Function class

Support

Support

Controller assembly 2

G2: Same material due to same flow basic and function class.

Fig. 7. Example of functional modeling guideline 2. Table 4 Essential movement and conversion based on functional modeling guideline 1.

No. 1

2

Flow heuristic module Conversion module 2

Function chain

Function modeling guideline 1 G1.1: Movement essential due to mechanical energy involved. G1.2: Conversion essential due to different starting and ending flow basics (rotation, vibration). G1.1: Movement essential due to mechanical energy involved.

Conversion module 1

G1.2: Conversion essential due to different starting and ending flow basics (rotation, heat).

Fig. 8. Example of functional modeling guideline 3.

method can be used in any situation that can be described functionally. However, it is not a panacea; it is a tool that has limitations that must be understood if it is to be properly and effectively used. It is a system without dimensions – that is, it will display functions in a logical sequence, evaluate them according to functional modeling guidelines and test their reduction or combination, but will not tell how a function

The proposed method is a systematic analysis process primarily intended to reduce the assembly costs of a product by reviewing the product functions at the concept design stage. Therefore, it was developed with the assumption that the bulk of manufacturing costs are set in the design stage, before any manufacturing systems analysis and tooling development is undertaken. As an effective design tool, the proposed

9

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Table 5 Possible part integration based on functional modeling guideline 2. Flow heuristic module Non-conversion module 1 Non-conversion module 2 Non-conversion module 3 Non-conversion module 4

Flow basic Function basic Function class Flow basic Function basic Function class Flow basic Function basic Function class Flow class Function basic Function class

Part 1

Part 2

Part 3

Electricity Actuate Control magnitude Solid Secure Support Gas Guide Channel Solid Provision Provision

Electricity Regulate Control magnitude Solid Remove Branch Thermal energy Dissipate Branch Solid Support Support

Function modeling guideline 2 G2: Possible part integration (same material) due to same flow basic and function class. G2: Different material due to same flow basic and different function class. G2: Different material due to different flow basic and function class.

Solid Channel Channel

G2: Different material due to same flow basic and different function class.

Fig. 10. Height adjustment for an arm rest.

version flows are not essential because the starting and ending flow basic involved are the same. The energy basic of conversion module 2 is known as mechanical force. Table 7 depicts the function chain. It is possible to consider combining Part 1, Part 2, and Part 3 into a single part. The results of functional modeling guideline 1 are shown in Fig. 12. Based on guideline 2, within non-conversion modules, we should use parts of the same material or avoid isolating parts when the same function class is used under the same energy basic. If the function class of parts is the same under the same flow basic, the functions of the parts can be considered identical and the parts can be candidates for integration due to the possibility of using the same material. For nonconversion module 1, the material of the parts must be different due to them having different flow basics. The results of functional modeling guideline 2 are shown in Table 8. Functional modeling guideline 3 must also be satisfied for possible part integration. However, Part 1 is a different material from Part 2 according to guideline 2. Therefore, functional modeling guideline 3 is not investigated. Based on functional modeling guidelines 1, 2, and 3, the parts for the height adjustment of an armrest can be improved. After applying the functional modeling guidelines for cost-effective concept development, the total number of functions is reduced by 25% (8 ≥ 6). The number of parts is reduced by 28% (7 ≥ 5). This tells us that design efficiency has improved because fewer parts are involved.

Fig. 9. Part integration (possible disassembly not required) based on functional modeling guideline 3.

should be performed (specification), when (it is not time-oriented), by whom, or for how much. The targeted application area is the design phase in any industry where products are designed for assembly, but it has many application limitations. Since it is employed to assist up-front product conceptualization, it is often used for smaller and medium-sized products, or for the numerous sub-elements of larger systems such as manipulators, mobile robots, and machinery assemblies. 4. Case study 4.1. Redesign of height adjustment for an arm rest Two case studies are presented in order to illustrate how the functional modeling guidelines work and help to reduce the number of functions required in the assembly. The first pilot study presented is a redesign of the height adjustment mechanism for the arm rest shown in Fig 10. The seven parts used for the height adjustment are listed in Table 6, along with the functions of the parts that control the height of the arm rest. The first task of functional model derivation is to create a sequential function chain model, which is a graphical representation of product functions with input/output flows based on FAST (how-why logic). A non-conversion and a conversion module are displayed in the diagram, as shown in Fig. 11. Based on functional modeling guideline 1.1, part movement is allowed for conversion module 1 because the module is activated by mechanical energy. Based on functional modeling guideline 1.2, the con-

4.2. Redesign of built-in luminaire The second pilot study undertaken is the redesign of a built-in luminaire for installing a lamp inside the ceiling of a building, as shown in Fig. 13. The nine parts used in the product and their functions are listed in Table 9. The first task of functional model derivation is to create a sequential function chain model, which is a graphical representation of product functions with input/output flows based on FAST (How–Why logic). 10

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Table 6 Part and function list for height adjustment.

Part name

Arm post inner

Arm post outer

Locker

Axis

Button

spring

Cover

Number of Parts 7

Function Transmit human force

Change position

Convert Mechanical torque to Mechanical force

Convert Mechanical force to Mechanical torque

Convert Dissipate Human force to Mechanical force Transmit Mechanical force

Guide mechanical force

Number of Functions

8

Fig. 11. Flow heuristic modules of an armrest in FAST. Table 7 Movement and conversion based on functional modeling guideline 1.

No.

Flow heuristic module

1

Conversion module 1

Function chain

Functional modeling guideline

G1.1: Movement essential due to mechanical energy involved. G1.2: Conversion non-essential due to the same starting and ending flow basics (mechanical force, mechanical force).

11

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 12. Possible part integration based on functional modeling guideline 1.2.

12

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Table 8 Possible part integration based on functional modeling guideline 2. Flow heuristic module

Part 1

Part 2

Functional modeling guideline 2

Non-conversion module 1

Human force Transmit Channel

Position Change Channel

G2: Different materials due to different flow basics with the same function class.

Flow class Function basic Function class

Table 9 The parts and function list of the built-in luminaire.

Part name

Junction box

J/Box door 1

J/Box door 2

Frame

Housing

Dome

Sticker

Socket

Bulb

Number of Parts 9

Function

Support solid

Support solid

Support solid

Support solid

Support solid

Support solid

Signal status

Transmit electricity

Support Solid

Number of Functions

Convert elect. To Light energy

12

Convert human force to mechani cal force

Transmi t light

Table 10 Movement and conversion results based on functional modeling guideline 1.

No.

Flow heuristic module

1

Conversion module 1

Function chain

Function modeling guideline 1 G1.1: Movement not essential due to mechanical energy not being involved. G1.2: Conversion essential due to different flow basics (electricity, light energy).

2

Conversion module 2

G1.1: Movement essential due to mechanical energy being involved. G1.2: Conversion essential due to different flow basics (human force, mechanical force).

volved. Because the parts in the module can be integrated, they should be examined using functional modeling guideline 2. However, part movement is allowed for conversion module 2 because the module is activated by mechanical energy. Based on functional modeling guideline 1.2, both the conversion flows are essential because the starting and ending flow basics involved are not the same. Table 10 contains the function chain and the results of functional modeling guideline 1. Based on guideline 2, for non-conversion module 1, the material of Parts 1 (Frame), 2 (Junction Box), 3 (J/Box door 2), and 4 (J/Box door 1) can be the same because both their flow basics (solid) and function classes (support) are the same. As a result, it is possible to combine Parts 1, 2, 3 and Part 4 into a single part. Additionally, for non-conversion module 2, the material of Part 1 can be the same as that of Part 2 because both their flow basics (solid) and function classes (support) are the same. As a result, integrating Parts 1(Housing) and 2 (Dome) is possible. For non-conversion module 3 and conversion module 1, the material of the parts must be different due to them having different flow basics

Fig. 13. Built-in luminaire.

Three non-conversion and two conversion modules are displayed in the diagram, as shown in Fig. 14. Based on functional modeling guideline 1.1, part movement is not allowed for conversion module 1 because mechanical energy is not in13

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 14. Flow heuristic modules in FAST for built-in luminaire.

Table 11 Possible part integration based on functional modeling guideline 2. Flow heuristic module Non-conversion module 1

Non-conversion module 2

Non-conversion module 3

Conversion module 1

Part 1

Part 2

Part 3

Part 4

Function modeling guideline 2

Flow class

Solid

Solid

Solid

Solid

G2: Possible part integration (same material) due to same flow basics and function classes.

Function basic Function class Flow class

Stabilize Support solid

Stabilize Support solid

Secure Support

Secure Support

Function basic Function class Flow class

Stabilize Support Status

Stabilize Support solid

Function basic Function class Flow class

sense Signal electricity

stabilize Support Light

Function basic Function class

Convert convert

Transmit channel

G2: Possible part integration (same material) due to same flow basics and function classes.

G2: Different materials due to different flow basics and function classes.

G2: Different materials due to different flow basics and function classes.

and function classes. The results of functional modeling guideline 2 are shown in Table 11. Functional modeling guideline 3 must also be satisfied for possible part integration. Within non-conversion module 1, Parts 1, 2, 3, and 4 can be the same material based on guideline 2. If the orders of the function chain and the assembly of parts are the same, then the parts within non-conversion module 1 can be integrated. As shown in Fig. 15(A), the orders of the function chain and the assembly of Parts 4 and 5 are

different. Therefore, it is not possible to integrate Parts 1, 2, 3, and 4 within the non-conversion module. However, it is possible to integrate Parts 1, 2 and 3 because the orders of the function chain and the assembly of parts are the same. In the case of non-conversion module 2, Parts 6 and 7 can be the same material based on guideline 2. As shown in Fig. 15(B), the orders of the function chain and the assembly of parts are the same. As a result, the parts within non-conversion module 2 can be integrated.

14

ARTICLE IN PRESS

JID: RCM H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

Fig. 15. Possible part integration based on functional modeling guideline 3. Table 12 Possible part integration.

No. 1

Flow heuristic module Non-conversion module 1

2

Non-conversion module 2

Parts in the module

For non-conversion module 1, Parts 1(junction box), 2 (frame), and 3 (junction box door) can be integrated, and for non-conversion module 2, the Parts 1(housing) and 2(dome) can be combined into a single part, as shown in Table 12, respectively. Finally, after applying all the functional modeling guidelines for cost-effective concept development, the total number of functions is reduced by 33% (12 ≥ 8). The total number of parts is reduced by 33% (9 ≥ 6). This tells us that design efficiency has improved because fewer parts are involved.

Possible part integration

modeling provides a formal approach to early concept development by directly translating customer needs into desired products behavior. We proposed functional modeling guidelines by employing the three principles of DFA for reducing and integrating functions. Its procedures are as follows: 1) functions are defined with three flow classes and eight function classes. 2) The relationship between functions is graphically defined in terms of conversion module or non-conversion module. 3) The modules are reviewed in terms of the proposed design guidelines. The proposed method alone does not aid in assessing customer desires, setting product requirements, or technical functional specifications, or in conducting engineering functional analysis. However, it does assist up-front product conceptualization in a product development effort. It simply advises regarding the cost-effective function structure of a

5. Conclusion 70 to 80 percent of the life-cycle costs of a product are determined by the concepts designed during the early development stages. Functional 15

JID: RCM

ARTICLE IN PRESS

H. Cho, J. Park

[m5GeSdc;March 9, 2018;19:23]

Robotics and Computer–Integrated Manufacturing 000 (2018) 1–16

design, given that the design meets functional goals already. It can make radical changes in the product structure in terms of change of function. Further, it does assist design and manufacturing engineers in using technologies other than ones they are most comfortable with since it helps them visualize alternative approaches to realizing the product. The primary objective of traditional DFA is to minimize the number of parts that will transition to the production stage at a minimum cost while they are still in design. It specifically requires that design and manufacturing engineers work together. In the same way, the proposed method allows multi-disciplinary team members to contribute equally and communicate with one another while addressing function models cost-effectively without bias or preconceived conclusions. Some firms may choose to apply the analysis later in the design stage, while others may do so earlier when major design alternatives for the proposed new product are being evaluated. The proposed method specifically supports system-level applications. The case studies illustrate how the suggested method works and helps to reduce the number of functions required for assembly. Significant advancements were demonstrated for integrating components and reducing parts count at the form-level of design. In sum, it provides the cost effective concepts used to assist the design teams during the design stage itself and leads to the following primary expected results for the assembly: 1) reduced material cost, 2) reduced labor and/or automatic assembly cost, and 3) reduced assembly cycle times. Thus far, we have used functional modeling guidelines for developing cost-effective product concepts for existing products. This is because the proposed guidelines are a prototype for design strategy. In order to become a complete design methodology, more fine-tuned and succinct functional modeling guidelines will be required. An extension for this work is to identify additional functional modeling guidelines using a set of consumer products that have been developed sequentially. Our method also allows product architecture decisions to be made at a much earlier design stage (i.e., the functional modeling stage). Our future development plan is to apply functional modeling guidelines to new products development in order to aid development teams that are looking for new and better product concepts.

[8] G.F. Dalgleish, G.E.M. Jared, K.G. Swift, Design for assembly: influencing the design process, J. Eng. Des. 11 (1) (2000) 17–29. [9] Lucas Engineering Systems Ltd., University Of Hull, Design for Assembly/Manufacture Analysis Practitioners Manual, 10.5 edition, CSC Manufacturing, Solihull, 1995. [10] K. Ulrich, The role of product architecture in the manufacturing firm, Res. Policy 24 (1995) 419–440. [11] K.M. Shahzad, K. Hadj-Hamou, Integrated supply chain and product family architecture under highly customized demand, J. Intell. Manuf. (2013) 1–14. [12] T. AlGeddawy, H. ElMaraghy, Reactive design methodology for product family platforms, modularity and parts integration, CIRP J. Manuf. Sci. Technol. 6 (1) (2013) 34–43. [13] S. Tyagi, A. Choudhary, X. Cai, K. Yang, Value stream mapping to reduce the lead– time of a product development process, Int. J. Prod. Econ. 160 (2015) 202–212. [14] R.R.M. da Cunha, A. Dias, A feature-based database evolution approach in the design process, Robo. Comput. Integr. Manuf. 18 (3) (2002) 275–281. [15] R.B. Stone, K.L. Wood, R.H. Crawford, A heuristic method for identifying modules for product architectures, Des. Stud. 21 (1) (2000) 5–31. [16] F. YU, et al., An improved functional decomposition method based on FAST and the method of removal and operation, in: System Science and Engineering (ICSSE), 2012 International Conference on, IEEE, 2012, pp. 487–492. [17] W.W. Jiang, Z.B. Xie, Research of Design Method of Product Function Analysis and Combination, in: Applied Mechanics and Materials, 201, Trans Tech Publications, 2012, pp. 886–889. [18] L. Yu-liang, Z. Wei, Development of an integrated-collaborative decision making framework for product top-down design process., Rob. Comput. Integr. Manuf. 25 (3) (2009) 497–512. [19] M.S. Sa’ed, A.K. Kamrani, Macro level product development using design for modularity., Rob. Comput. Integr. Manuf. 15 (4) (1999) 319–329. [20] K.L. Wood, Development of a functional basis for design, J. Mech. Des. 122 (2000) 359–370. [21] R.B. Stone, D.A. McAdams, A product architecture-based conceptual DFA technique, Des. Stud. 25 (2004) 301–325. [22] D.G. Ullma, Mechanical Design Process, 2nd edition, McGraw-Hill, New York, 1997. [23] G. Boothroyd, L. Alting, Design for assembly and disassembly, CIRP Ann. Manuf. Technol. 41 (2) (1992) 625–636. [24] M. Sako, F. Murray, Modules in design, production and use: implications for the global automotive industry, MIT IMVP Annual Sponsors Meeting, Cambridge, Massachusetts, 1999. [25] B. Agard, M. Tollenaere, Design of product families: methodology and application, in: International Conference on Engineering Design, ICED’03, Stockholm, 2003, pp. 209–210. [26] D.W. Clarke, Integrating TRIZ with value engineering: discovering alternatives to traditional brainstorming and the selection and use of ideas, in: SAVE International Proceedings, 1999, pp. 42–51. [27] D.D. Sheu, C.T. Hou, TRIZ-based systematic device trimming: theory and application, Proc. Eng. 131 (2015) 237–258. [28] A.H. Redford, J. Chal, Design for Assembly: Principles and Practice, McGraw-Hill, London, 1994. [29] R.B. Stone, K.L. Wood, Development of a functional basis for design, J. Mech. Des. 122 (4) (2000) 359–370. [30] R.B. Stone, Towards a Theory of Modular Design, Doctoral Dissertation, University of Texas at Austin, 1998. [31] R.B. Stewart, Fundamentals of Value Methodology, Xlibris Corporation, 2005. [32] U. Ibusuki, P.C. Kaminski, Product development process with focus on value engineering and target-costing: a case study in an automotive company, Int J Prod. Econ. 105 (2) (2007) 459–474. [33] M.H. Kwon, Portable electric fan, Korea Patent No. 2003524710000, 2004. [34] J.J. Corp., An incandescent lamp for traffic signal lamp, Korea Patent No. 2020000021897, 2000. [35] T.J. Oh, Hair clipper, Korea Patent No. 1005761680000, 2004. [36] T.J. Oh, Hair clipper, Korea Patent No. 2003816670000, 2004. [37] G. Boothroyd, P. Dewhurst, W.A. Knight, Product Design For Manufacture and Assembly, CRC press, New York, 2010.

References [1] K.T. Ulrich, S.D. Eppinger, Product Design and Manufacturing, 2000. [2] W.Y. Zhang, S.Y. Tor, G.A. Britton, Managing modularity in product family design with functional modeling, Int. J. Adv. Manuf. Technol. 30 (7) (2006) 579–588. [3] J. Ebrahimi, E. Babaei, G.B. Gharehpetian, A new multilevel converter topology with reduced number of power electronic components, IEEE Trans. Ind. Electron. 59 (2) (2012) 655–667 2012. [4] R. Cooper, R. Slagmulder, Target Costing and Value Engineering, Productivity Press, Portland, OR, 1997. [5] G. Boothroyd, Design for Manufacture and Assembly: The Boothroyd–Dewhurst Experience, in: Design For X, Springer, Netherlands, 1996, pp. 19–40. [6] Design Profit, Inc, Design Profit Training Manual, Design Profit, Inc., Williamston, MI, 2010. [7] J. Pandremenos, J. Paralikas, K. Salonitis, G. Chryssolouris, Modularity concepts for the automotive industry: a critical review, CIRP J. Manuf. Sci. Technol. 1 (3) (2009) 148–152.

16