Modular Design and Its Application to Overhead Traveling Crane Hong-Zhong Huang1* - Yanfeng Li1 - Zhonglai Wang1 - Li-Hua Xue2 - Yong-Fu Wang2 1 University of Electronic Science and Technology of China, School of Mechatronics Engineering, China 2 Dalian University of Technology, School of Mechanical Engineering, China In this paper, the method of modular design of a crane is presented. First, the main customer groups and their expectations of the new product are confirmed. Then, the main function of the new product is decomposed into smaller function modules. Based on these function modules, the relativity of function modules is analyzed and quantified. By the entropy method, the best scheme is selected through the evaluation of design, manufacture, assembly and maintenance, etc. Finally, a product platform is set up by analyzing the degree of influence of the function modules on customer needs. The storage format of a function module and the structure of the function module database are also investigated. The function of a model is set as the keyword in the function model database, and a module-selecting algorithm based on fuzzy pattern recognition is developed. The validity of the modular design method presented is verified by the design of an overhead traveling crane. © 2009 Journal of Mechanical Engineering. All rights reserved. Keywords: modular design, product planning, function decomposition, similarity theory, overhead traveling crane 0 INTRODUCTION A modern mechanical system often consists of power modules, transmission modules, execution modules, computer modules and control modules. As traditionally, products are designed and manufactured individually according to design requirements, the overall performance of the products may not be satisfactory. This leads to difficulties in maintenance, because function components are hard to change or upgrade. Modular design is a method proposed by Occident in the 1950's. The method mainly solves the conflicts between types, specifications, design cycles, and cost. In short, modular design is such a method that standardizes the connecting sizes of congeneric components that fulfill certain functions. These components can be interchanged easily, so it is convenient to replace and change the assemblies. The purpose of this method is to improve the design efficiency and to make the design experience transferred and inherited. The evident advantages of this method are not only the satisfied customer needs, but also the shortened design cycle and reduced development cost. The method therefore, ensures the enterprises to be competitive in the market. In 1900, a German company designed the so-called "ideal bookshelf" with modular principles. In this design, the bookshelf was decomposed into a base locator, shelf and upper plate. There are several shelves with different sizes. A customer could select different shelves to build up an "ideal bookshelf' according to his or her specific requirements. An "Ideal bookshelf' is one of the earliest products that were designed according to modular design. Afterwards, the idea and the theory of modular design was studied further and applied by scholars and engineers into every aspect of life. The theory of modular design has been evolving gradually. Around 1920, the modular principles began to be applied in the design of mechanical products in Europe, especially in Germany. In the 1950's, after a great deal of exploration, modular design was formally and systematically practiced as an advanced design method by Euro-American experts. With the improvements of the theory and techniques, modular design is generally used in the development of all kinds of products. Research on modular design is still progressing, and the theory of modular design continues to be enriched. For example, many researchers like Huang, Peter, Liang and Kusiak, are studying the object-oriented modular design [1] to [5]. Mok et al. [6] have also made much progress in the modularization of products. Combined with life cycle theory, research on the modular design of products has been conducted *Corr. Author's Address: School of Mechatronics Engineering, University of Electronic Science and Technology of China, Chengdu, Sichuan, 611731, China, hzhuang@uestc.edu.cn by Gu et al. [7]. From a different angle, Tong views modular design as a system problem. He believes that modular design is an integrated process where modular systems are set up and modules are combined to build up a product system through module decomposition and combination. He concludes that modularization is a superior form to standardization [8]. Many Chinese scholars have done considerable studies on modular design. For example, Xu et al. work on the modular design of numerically-controlled machine ([9] and [10]). Their contribution is the generalized modularization system and flexible modular system. Wang and Chen have developed the machine center using the method of modular design [11]. The theory of modular design has been used to develop the main drive system of the numerically-controlled machine center by Deng and Wang [12]. Wang et al. applies the method of modular design to the design of product families [13]. The focus of this work is to apply the method of modular design to crane design. The whole system and components of a crane are traditionally designed according to the main parameters and performance requirements. The use of some function components in span product series is seldom considered and therefore, the universal performance of components is not satisfied, since due to a lack of idea of standardization, experience is difficult to transfer and accumulate. Because the developing cycle can be shortened, the developing pace can be expediated, the cost can be lowered, new techniques can be employed, the manufacturing management is convenient and the modular design of a crane has been developed greatly. This paper is organized as follows. Section 1 illustrates the method of modular design. The product specification of the modular design is discussed in Section 2. The definition of customer groups and the planning of product families according to customer group requirements has been investigated. The construction and evaluation of a module system are discussed in Section 3. The establishment of module databases and the storage structure of modules are discussed in Section 4. In Section 5, with the example of the design of an overhead traveling crane, the validity and usability of the proposed method of modular design is verified. Section 6 concludes this paper with discussions of future work. 1 METHOD OF MODULAR DESIGN 1.1 Concepts of Modular Design 1.1.1 Relevant Concepts [14] (1) Product series: A group of products with the same functionality, similar structure and manufacturing process and different performance parameters. (2) Classification of product series: Product series can be classified into vertical, parallel and whole product series. (3) Vertical product series: A group of products with the same functionality, similar structure and manufacturing process, but different sizes and different major performances. (4) Parallel product series: A group of products with the same performances or power, but different relationships between the main performances through only interchanging some executable components with the same kind of functions and different performance parameters. (5) Whole product series: A group of product series by adding or changing some function modules of the vertical product series. They therefore, consist of both vertical product series and parallel product series. (6) Product families: A product family consists of products that share a common product platform but have specific features and functionality to satisfy different sets of customers. From the above concepts, it is obvious that product series with the same overall functionality and different performances are formed through the combination of various function modules with different sizes. Corresponding to the above definitions, modular products can be classified into vertical product series, parallel product series and span product series. Modular design can be classified into modular design of parallel product series, vertical product series and span product series. One process of modular design is to abstract functions and construct modules according to the requirements. The other process is to select a group of modules rationally to constitute a product scheme. We define modular design as: Determine function modules of a product family through the requirement analysis and function decomposition. Construct the product platform of the product family through the relativity analysis of the function modules. Form the vertical, parallel, and span product series through adding auxiliary function modules and supporting function modules. 1.2 Classification of Function Modules Function modules help implement technical functions independently or are combined with other functions. Product modules are designed independently of their functions and are based only on production-based considerations. Function modules are classified as basic, auxiliary, adaptive, special, and nonmodules in [1] and [15]. Basic functions do not vary in principle and are fundamental to a product. A basic function can fulfill an overall function by itself or in combination with other functions. A basic module is a module implementing basic functions [15]. Auxiliary functions are complementary and task-specific sub-functions that need not appear in all overall function variants. An auxiliary module corresponds to auxiliary functions that are used in conjunction with the basic modules to create various products. Adaptive functions are necessary for the adaptation to other systems and to marginal conditions. An adaptive module is a module in which adaptive functions are implemented. Adaptive functions adapt a part or a system to other products or systems. Adaptive modules handle unpredictable constraints. Special function modules are specific complementary sub-function modules, which correspond to basic function modules and are required by the development of new products. Special function modules are only designed to perform special functions required by some products. Customer-specific functions not provided in the modular system will recur even in most careful development. A non-module implements customer-specific functions that do occur even in most careful design development. Non-modules have to be designed individually for specific tasks to satisfy customer needs. 1.3 Principles of Modular Design The modular design consists of product planning, general layout and function analysis, feasibility evaluation and the design of modular structure. The design of modular structure belongs to detailed design, which is beyond the scope of this paper. The product types will be specified at first in the design of product series or product families. The main contents of the specifications are the investigation and analysis of market requirements, the identification of product series, the analysis of parameter ranges and the determination of main parameters. The general layout of products is a general scheme of a group of products acquired through modular design, which includes main functions and structure features and further responds to the market requirements [16]. According to the coverage form and the extent of modular products in product series, the layout scheme is classified into the vertical, parallel, span, and whole series, etc. The product will be broken down into several modules at first when modular design is implemented. The design will continue with modules as basic units. The main action of the function analysis is to decompose the overall product function, to determine the correlation degree between function modules, and to reassemble function modules back in order to form the entire modular systems. In essence, the product can be classified in arbitrary levels to form modules. When determining the decomposition schemes of modules many factors such as design, assembly, maintenance and cost, etc. should be taken into account. Function analysis includes the following two aspects: the function decomposition and the clustering of function modules. Function decomposition specifies the overall functions of a product. The sub-functions that are required to perform the overall functions can be acquired. These sub-functions can be divided gradually until every sub-function has corresponding physical modules. The clustering of function modules determines the relativity between function modules and the form correlation matrices of function modules. Based on evaluation criteria, a group of schemes can be selected as the final concept. The design and manufacture complexity of modules increases with the decreased number of modules. But the assembly and maintenance complexity of products increases with the number of modules. Therefore, an optimal number of modules may exist. In modular design, every module partition scheme constitutes a system; and the information entropy about assembly, maintenance and cost exists in this system. Therefore, the algorithm entropy can be used to measure the complexity in [17] to [19]. An important feature of modular design is that function modules are convenient to interchange, which not only requires the connecting factors in connecting site to be consistent, but also requires that the performance of the whole machine does not decrease after the modules are changed. To meet these requirements, the key of modular design is to identify the relativity between modules, to construct the consistency of module interfaces and to build module structures by physical similarity principles [16]. 2 PRODUCT PLANNING Product planning includes the definition of product requirements, the development of product families and the construction of the modular system. In this section, the definition of product requirements and the development of product families are discussed. 2.1 Product Series Planning of Modular Design A product family is a collection of a series of products that share universal core components. They can meet multi-market and multi-customer needs through simple modifications or variations. The panning of product families includes the following hierarchies [20]: 1. Definition of product requirements: Acquiring function and parameter requirements through the analysis of customer needs and constructing the function structure of product. 2. Definition of a product series: In terms of function and parameter requirements, specifying the proper product series parameters and function ranges, and identifying the product series spectrum. 3. Definition of a flexible module series: According to the product series spectrum, the planning of a flexible modular series can be performed. This mainly includes the partition of modules, the planning of module series, the construction of modules, the normalization and serialization of components, the planning and construction of modular structures, and similarly. Product requirements are the expression of customer needs. Product series and module series are the expression of design. The planning and the construction of the elements of product can be completed from defining product series to defining module series. The enlargement of product families is performed through the mapping from product series to module series. 2.2 Definition of Generalized Modular Products The definition of products generally represents normalized information of customer needs and all kinds of technical information needed in the design process. An exact definition of products is the foundation of enterprises to make strategies of product development, which is the basis and source of product design. For generalized modular design, the definition of products can be understood as follows. On one hand, it is the classification process of customer groups based on customer needs while on the other hand, the parameter specification ranges of products can be identified rationally in terms of customer needs and the structural design of products. Fig. 1 [20] describes the process of the definition of a generalized modular product. Fig. 1. Product definition of modular design 2.2.1 Acquisition and Analysis of Customer Needs Enterprises should gather raw data from customers through as many approaches and kinds of investigation as possible. The main ways to gather customer needs are as follows [22]. Acquiring the past and present customer needs by analyzing the order forms in the past two or three years. Acquiring the product value increment needs of customers by investigating and analyzing customer needs, and acquiring the potential customer needs indirectly. Acquiring the trend of future customer needs and other information by analyzing the effects of new techniques, new materials, the new manufacturing process, environment and zoology balance, etc. Acquiring the possible customer needs by analyzing economical and political situations. The main way of gathering customer needs is market investigation. The main content of market investigation is illustrated in [16] in detail. Moreover, the investigation and analysis of other factories that produce same or similar products should be conducted. 2.2.2 Data Analysis of Customer Needs After the raw data of customer needs are gathered, they need to be further generalized and interpreted on the premise of not distorting customer wills. Customer needs with the same or similar meanings will be incorporated. From the point of view of life cycle, customer needs consist of four elements: customer consultation, product purchase, product use and product abandonment and recovery. Fig. 2 depicts the classification tree of generalized customer needs based on life cycle [23]. 2.2.3 Mapping From Customer Needs to Function Modules After customer needs are gathered and customer groups are classified, customer needs can be ranked according to the importance of customer groups. After the investigation of customer needs and the classification of customer groups, the matrix of customer groups and customer needs can be identified as S = S11 S12 (1) where s^ is the requirement degree of the i customer group to the j1 customer need. Let the weights of customer groups be W= [wi, W2, ..., wm]. (2) By synthesizing customer needs, the weighted vector of customer group needs is U = W x S . (3) As shown in Fig. 3 [24], after the weighted vector of customer group needs is acquired, quality function deployment is used to describe the product groups-oriented relation between customer group needs and function modules. There are many approaches to realize the mapping between customer group needs and function modules. They are chosen according to different situations [25]. Generalized customer needs Consultant needs Similar products compare Similar service compare Expert consultant system I J= 8 ! ÄI8 of srt oe ep ra * ben e no dP £ g H* ^ s JZ in I Purchase needs . I X X =L Main function Physical dimension Working object Move- Auxiliary ment functions Work condition Color I -[ Shape Deliveiy j= Performance Function Adaptivity Appearance Economy Reliability Confidence level Validity Use needs I Maintenance Life span Once - life span Performance que support Multiple life span Abandonment needs -; i ;- Maintenance Service Upgrade Product replace Goods returned | Complaint e ce elu icrv vae se ylit aalv bia saf of e n n io re d ear ality d iba an ar n o pe u c ua e Re E Fig. 2. Tree of generalized customer needs S21 S22 s Engineering discription e a O Function modules Fig. 3. Mapping between customer group needs and function modules According to the structural features of products, customer group needs, function module needs of products and the range of parameter specification can be identified through qualitative analysis and quantitative comparison. 2.3 Definition of Modular Product Series A generalized modular product model is a parameterized model consisting of flexible modules from which a group of products with different parameters can be derived. Therefore, the emphasis of planning generalized product series is to analyze the main parameters that influence structural design. By analyzing the relation between functions, structures and parameters, a parameter that affects the whole design can mostly be selected as the basis to classify product series. And other main parameters are used to further perform the planning of module series. 2.3.1 Similarity Laws and Preferred Number Series Geometric similarity is widely accepted because of its simplicity. However, technical artifacts stepped up in geometric proportions are not satisfactory except in very rare cases. In modular design, similarity laws are usually the evaluation criteria. We talk about similarity if the relationship of at least one physical quantity in the basic and sequential designs is constant. It is possible to define basic similarities with the help of fundamental quantities, such as length, time, force, quantity of electricity, temperature and luminous intensity (Table 1) [15]. Thus, there is geometric similarity if the ratio of the length of any sequential design to all the lengths of the basic design is constant. Here, the non-dimensional parameter to be held constant is p * ? (7) pq ! V i=1 J { 1 p = ? where 3 is the regularization coefficient. To ensure rpqe {0, 1}, let ¡3 = 0.2, jm1 be the weight of customer group needs, ^2 be the weight of geometric correlation rules, ^3 be the weight of physical correlation rules, and Wi be the weight of ith customer group. The analytic hierarchy process (AHP) can be used to identify jm1, ^2 and ^3 ([19] and [30]). The identification of wt was discussed in Section 2. 3.2 Evaluation of Module Decomposition Scheme Each module decomposition scheme constitutes a system in which information entropy (also called system complexity) exists in design, manufacture, assembly, maintenance and cost. Based on this point of view, the realization of a modular system is determined by the complexity, viz. information entropy. The smaller the complexity, the more rational the module decomposition scheme [18] and [24]. 3.2.2 Assembly The smaller the granularity, and the greater the number of modules ka, the more assembly relations are identified, the longer the assembly time, and the assembly difficulty will increase, and vice versa. The definition of the assembly complexity is: Ha (a )=-±m4à ln niiH (9) n n where n(j) is the number of modules to compose the jth module. 3.2.3 Cost From the point of view of cost, during the process of function decomposition, the sub-function modules with high cost and additional value will be isolated to form different function modules. So during product maintenance, if one of the modules fails, only the failed one needs to be replaced; hence the maintenance cost can be reduced. As a whole, the average information entropy of cost should be decreased. The cost complexity is defined as 1 n Hc (a )=-1X (crl ln c„+(1 - c„)ln(1 - c„))(10) ka i=1 where cri is the relative cost of the ith sub-function and can be determined by AHP method. 3.2.4 Maintenance The vulnerable sub-function modules should be separated. Therefore, the products are prone to be subdivided under the consideration of maintenance. As a whole, the average information entropy in maintenance should be decreased. The definition of maintenance complexity is (a) = -1 £ (n ln n + (1 - n ) ln (1 - n )) ( 11) ka i=1 where n is the relative vulnerable probability (relative damage vulnerability) of the ith sub-function module and can be determined with AHP method. 3.2.5 Synthetic Evaluation Normalize the assembly, cost and maintenance complexity as follows [17] and [30]: Hp (ai ) Hp (a> ) = m SHp (ai) Ha (a, a \ i ! M H (a Hm (a, ) = £ H (a, ) i=1 Hc (a,) M Y.HC (a,) 1=1 Hm ( a, ) M S Hm (a,) (12) (13) (14) (15) where M is the number of module decomposition schemes. Ha(a,), Hc(a) and Hm(a) are the assembly, cost and maintenance complexity, respectively. They are decided by the granularity of module a, (i = 1, 2, ..., M). According to the complexities after the normalization, the model of unconstrained optimization is given by: min {H (a) = Hp (a) + Ha (a) + Hc (a) + Hm (a)} All the possible granularities of module at can be acquired from the definition of module granularity. If e is the number that makes H(ae) smallest, then the corresponding module decomposition scheme is optimal. 3.3 Construction of Product Family Platform For a product family, the product platform is a fixed and universal product core that can support the development and manufacture of product family by adding some function modules and corresponding constraint conditions. The construction of the product platform through the mathematical description is illustrated below [13]. Let K be the ultimate number of modules determined in modular system, m the customer group needs and W={wx, w2, ..., wm} the weight of customer group needs. The matrix F that describes the influence of the change of the requirements to function modules can be constructed as follows. r2 • • Rm 1 0 0 1 1 0 0 0 • 0 M. 1 (16) F= M2 1 1 " V 7 Mk where R (i = 1, 2, ..., m) is customer needs. Mj (j = 1, 2, ..., K) are modules. For example, if the change of R influences Mj, the number in crossing point of R\ and Mx is 1, which ensures the influence of Ri on Mj to be identified easily. Suppose that the influence matrix of function changes on modules is E, then E = W-F. (17) According to the functional principle diagram of product family, the number of connections between one module and others can be confirmed. Counting the connecting number ci of Module Mi and other modules to form Convex C = [cj. Let P = [ p> ] = E + C (18) where pt reflects the influence of Module Mt on the whole product family. The bigger pi is, the higher the influence of the function module on the product family and the more closely the function module combines with the product family. Associated with the importance of function modules, the modules can be rearranged. The first function modules after rearranging can be confirmed as the product platform. 4 DATA MANAGEMENT A module database is used to store module data, which are the working achievements in the process of module construction. The module i=1 1=1 database is also the basis of customization [31]. Therefore, the module database is the medium to relate the construction of function modules to redesign in modular design. Module management can be divided into design management and production management. The main purpose of design management is to utilize the current modules adequately to fulfil rational combinations, and make them satisfy the actual customers needs quickly and effectively. Moreover, if the current module combinations do not satisfy the customer needs, the determining indicating information can be presented to design departments through design management, such as which modules need to be designed and modified, etc. The main purposes of production management are the decomposition and summary of modules, the statistics of materials and the print of all kinds of summary sheets, and so on. In this section, the encoding and storage of modules, the structure of module database of design management are discussed and the module index mechanism based on fuzzy pattern recognition is set up. 4.1 Module Encoding Module encoding mainly reflects the function of modules and the features of correlative connections and interfaces. The rules of module encoding are uniqueness, extendibility, adaptability, implication, stability, identity, manipulability, brevity and format unification [32]. The encoding rules will affect designers to recognize the modules and to name the newly developed function modules. To adapt to the encoding rules of modules, an encoding rule of 9-bit codes is adopted in this section. 1 2 3 4 5 6 7 8 9 a b c d e Fig. 5. 9-bit codes In Fig. 5, a represents the primary recognition codes, which are the abbreviations of function module names and are used to identify the types of function modules. b contains the primary parameter codes, which are the main function parameters. c consists of the interface type codes. Different encoding of interface types corresponds to different interface types. d is the design edition codes, which are the design series of modules and indicate the number of module modifications. e is the additional codes, which are used to illustrate other important information of modules. In the process of module encoding, the 9 bit code must be complete. When the number in a bit is empty, the 0 will be used to fill-up. Additional codes can be added behind the encodement, which are some additional information of modules. The length of additional codes is confined within 7 bits. 4.2 Storage of Module Data The storage of module data means to store module data, construct the module database with certain structure and organization. The storage thus makes the management and utilization of modules more convenient. The key of accessing data with high efficiency is the encoding of modules, the organized form of module data and the structure of the module database. 4.2.1 Data Structure of Single Module The storage structure of modules is different from ordinary components, which not only describes the information of function structure, performance and management of the module itself, but also expresses all the attributes and methods relevant to the module manipulation. The main information expressed by the storage structure of modules is as follows [33]: the management, structure, composition, interface and performance information of modules. All kinds of information should be stored in different data sub-modules according to types, which is beneficial in decreasing the recognition number in the index process of modules. 4.2.2 Construction of Module Database The purpose of constructing the module database is to organize and manage function modules better. Based on this point of view, the module database with the organization of function is adopted. For the function modules fulfilling the same function, they can be further managed systematically according to the basic realization principles of module function or they can be constructed directly and stored in the same function table. There can be different realization methods to the same function but the different realization methods must correspond to different structures and performances and should generate different function modules. In the early development stage of new product series, viz. conceptual design, all the constructed module databases are function structure models of product system [15]. 4.2.3 Additional Information Tables in Database The main parameter and interface of modules are designed according to the idea of serialization or product family in the module database. So the mapping method can be used to encode modules, which can simplify the encoding of modules. But after this mapping is adopted, the corresponding additional information tables in module database must be established. In terms of the process of constructing the module database system, the required additional information tables include the interface definition table, main parameter definition table, etc. If other information is needed in the construction of module database system, the corresponding additional tables must be established. 4.3 Index of Module Database In the computer-aided modular design of new products, the algorithm of module selection is very important. Much research on the algorithms has been conducted ([10] and [34]). An algorithm of module selection based on fuzzy pattern recognition is presented in this section. 4.3.1 Module Selection Policy In the development of new products, after the requirements of a function module is proposed, the designer will index the module database to find whether the existing function modules can satisfy the current design requirements. Fig. 7 is the detailed flow chart of module selection. 4.3.2 Module Index Based on Fuzzy Pattern Recognition Let R = {n, r2, r3, ..., rm, rm+x,..., rm+n} be a group of attributes of a design module, where ri (i = 1, 2, 3, m) are structure attributes and rm+j (j = 1, 2, 3, ..., n) are the function attributes. Suppose that D = {d1, d2, d3, ..., dm, dm+\, ..., dm+n} is the set of design requirement parameters and L = {A, h, h, ..., lm, L+\, ., L+n} is the set of module parameters in module database. Fig. 6. Storing framework of a function module N (A, i E W.(A (uJA Ms (u,[ 1 m+n 2 E (u,)+Ms « 2 i=1 (23) Offer main parameters Construct new modules r Add to module database After the closeness degrees of design modules and current modules in module database are acquired, the designer can select the module with the largest closeness degree or modify the existing module data or design requirements to meet design needs on the basis of the module with the largest closeness degree. 5 MODULAR DESIGN OF DOUBLE-GIRDER OVERHEAD TRAVELING CRANE Fig. 7. Flow chart of module selection A normal membership function is constructed as follows: f dd. Y (u ) = e ^ b J (20) where di and li are the module requirement parameter and the parameter of existing modules in module database, respectively. bi is constant corresponding to the parameter, which expresses the acceptance error of the parameter. Each recognition object (modules in module database) corresponding to a requirement module is a fuzzy set and can be expressed as B = ( (u1 ^ ^B iU2 ^ ' ' ', ^B iUm+n )) (2 11) where ¡u~~ (u.) denotes the membership degree of an attribute parameter to the requirement module attribute of the identification object. The correspondence between the module to be designed and itself is a fuzzy set. A = ( (U1 ^ Hi (u2 (Um+n )) . (22) It is obvious that the module to be designed is completely same as itself. So all the elements of A are 1. The closeness degree of two sets can be calculated by employing lattice closeness degree [16]. Because the importance of every module attribute and the corresponding weights are different, in this paper, the least weighted arithmetic mean method is used to calculate closeness degree. Let W = {w1, w2, w3, ..., wm+n} be the weight of module attributes. The closeness degree is given by An overhead traveling crane is an extensively used hoisting and transportation machine. It is a kind of bridge crane that moves in three dimensions. It has many advantages, such as large carrying capacity, high reliability, and simple manufacturing processes, etc. Based on the method of modular design discussed above, two aspects are addressed in this section: First, the product family planning and function decomposition of the overhead traveling crane is performed; second, the results are evaluated from assemblies, designs and manufacturing processes. 5.1 Identification of Crane The technical parameters determine the working ability of the crane and are the basis of the crane design [35]. The main technical parameters of the overhead traveling crane are lifting capacity Q, lifting height H, working speed of mechanism V, group classification code, etc. 5.1.1 Lifting Capacity The maximum lifting capacity of the crane is called the rated lifting capacity. The rated lifting capacity of overhead traveling crane is constant. Table 5 shows the series of maximum lifting capacity defined by the ISO 2374. These numbers are for all types of cranes. 5.1.2 Lifting Height The lifting height of the overhead traveling crane is the vertical distance from the ground to the maximum lifting position of load handling device. The lifting height is decided by the working requirements. Working speeds can be identified according to Table 8. 5.1.5 Group Classification The group classification of cranes is the basis of the steel structure and mechanism design. It provides a reference range to users and manufacturers. Table 9 is the group classification of overhead travelling cranes suggested by GB 3811-83. 5.2 Traditional Structure Decomposition of the Overhead Travelling Crane According to the traditional design, the overhead travelling crane consists of the following sections: electric control, trolley, crane travelling and bridge. Fig. 8 is a detailed decomposition of traditional overhead travelling crane. Table 5. Maximum lifting capacity series (t) 0.1 0.125 0.16 0.2 0.25 0.32 0.4 0.5 0.63 0.8 1 1.25 1.6 2 2.5 3.2 4 5 6.3 8 10 (11.2) 12.5 (14) 16 (18) 20 (22.5) 25 (28) 32 (36) (40) (45) 50 (56) 63 (71) 80 (90) 100 (112) 125 (140) 160 (180) 200 (225) 250 (280) 320 (360) 400 (450) 500 (560) 630 (710) 800 (900) 1000 Table 6. Lifting height series of overhead travelling crane (GB 791-65) Lifting capacity Q 3-50 80 100 125 160 200 250 Lifting height H Main hook 12 16 20 30 20 30 20 30 24 30 19 30 16 30 Auxiliary hook 14 18 22 32 22 32 22 32 26 32 21 32 18 32 Table 7. Crane span series of overhead travelling crane (GB 790-65) Workplace span 9 12 15 18 21 24 27 30 33 36 Crane span L Q = 3-50 t 7.5 10.5 13.5 16.5 19.5 22.5 25.5 28.5 31.5 - 7 10 13 16 19 22 25 28 31 - Q = 80-250 t - - - 16 19 22 25 28 31 34 working requirements. If there are no special requirements, the lifting height can be simply determined according to a design standard, e.g. the national standard of China GB 791-65 (Table 6). 5.1.3 Crane Span The horizontal distance between the crane rail centers is called crane span L. The crane span of overhead travelling crane should be less than the workplace span. Table 7 is the crane span series suggested by GB 790-65. 5.1.4 Working Speed Working speeds of a crane include main lifting speed, auxiliary lifting speed, trolley travelling speed and crane traversing speed. The actual working speeds are decided according to Table 8. Rated working speed of cranes Crane types Lifting speed (m/s) Crane travel speed (m/s) Main Auxiliary Trolley Crane General-purpose overhead travelling crane with hook A1, A2 0.016 - 0.05 0.133 - 0.166 0.166 - 0.332 005 - 0.667 A3, A4 0.033 - 0.2 0.133 - 0.332 0.332 - 0.667 0.667 - 1.5 A5, A6 0.133 - 0.332 0.3 - 0.332 0.667 - 0.833 1.167 - 2 Overhead travelling crane with magnet 0.3 - 0.332 0.332 - 0.416 0.667 - 0.833 1.667 - 2 Overhead travelling crane with grab 0.667 - 0.833 0.667 - 0.833 1.667 -2 Table 9. Group classification of overhead travelling cranes (GB3811-83) Application situation Group classification Hooks Instalment and maintenance of power A1-A3 Shop and warehouse A3-A5 Heavy duty, Shop and warehouse A6-A7 Brake — Hook block _| Pulley block Coupling Safety device Trolley Hoisting „ .. , Traveling mechanism Trolley frame mechanism Overhead traveling crane Crane traveling — Reducer Motor Reducer Brake Brake Buffer Buffer Wheel block Wheel block Bridge Main girder End carriage Conducive frame Fig. 8. Structure decomposition of traditional overhead travelling crane From Fig. 8, it is obvious that the traditional design process of the overhead travelling crane contains the idea of modular design. However, the idea of modular design does not carry out completely, and the following problems exist in traditional design: 1. Since the sizes of components are not in the same order of magnitudes, the design workload and complexities for components are different and some components are difficult to be integrated. 2. The connectivity among components is not high, and it is hard to achieve standardization and serialization. 3. The connections of components are complicated and are not beneficial to manufacturing and maintenance. 4. The exchangeability of components is weak. The traditional design method of cranes does not meet the market requirements. In this section, the modular design is employed to the design of overhead travelling crane. 5.3 Function Decomposition of Overhead Travelling cranes The main function of the overhead travelling crane is to safely and reliably move materials with a certain speed from one place to another. This process is performed under the action of control information within the specified field. In the three-dimension coordinates of Fig. 9, Axis X represents the length of the workplace, Axis Y represents the width of the workplace, Axis Z represents the lifting direction. According to the working cycle of the overhead travelling crane, the carrying function is performed by the following three basic actions: moving along the horizontal X and Y directions and moving along the vertical Z direction. The realization of the load handling function depends on certain tools. According to the working cycle diagram, the overall function of the overhead travelling crane can be divided into two sub-functions, viz. the movement along the horizontal axes X and Y and the movement along the vertical axis Z. Compared with the traditional structure diagram, the function carriers of these sub-functions can be found. The horizontal movement along X is realized by the crane travel module, the horizontal movement along Y is fulfilled by the trolley and the vertical movement along Z is accomplished by the lifting mechanism. In order to realize the horizontal movement, the realization form can be divided into pull-type and self-propelled type according to different types of power. The pull-type does not have a power source and the movement can be fulfilled by corresponding components. Self-propelled type completes the horizontal movement by combining the driving function, transmission function, control function and working function into a system. Among them, the driving function can be performed by a motor, hydraulic motor, etc. The transmission function is used to transfer torque and change speed to meet the requirements of working components. It can be further decomposed into the speed-reducing function, force-transfer function and coupling function i.e. a speed reducer, transmission shaft and coupling. The control function is used to control the beginning and the end of horizontal movements. The brake function can change the state of objects from moving to stationary. The travelling function can be fulfilled by wheels and other components. The vertical movement is lifting the loads up and down. The function of the crane is to transport heavy materials. In the process of horizontal movement, the vertical direction will undergo suspending heavy materials, which is an important function to be fulfilled by the crane. In the fulfilling process of this function, safety and reliability are very important. The decomposition of driving function, transmission function and control function is similar to the decomposition of horizontal movement function. The working function will be further decomposed into load handling, reducing force and lifting function. The brake function is used to suspend heavy objects. The protect function will be divided into limiting position and limiting force functions, which are used to limit the height and weight, respectively. Through function decomposition, we can find that the double-girder overhead travelling crane consists of five modules of Class II, sixteen modules of Class III and six modules of Class IV (Table 10). According to the realization principle of function modules, the modules of the overhead travelling crane can be classified into the following types: the supporting function module, driving function module, controlling function module, braking function module, working function module, reducing function module, coupling function module, transferring function module, lifting function module, reducing force function module, limiting position function module and limiting force function module. 5.4 Function Module Classification 5.4.1 Identification of Correlation Degree Between Functions 1. Function relativity determined by main parameters For the overhead travelling crane, the customized parameters include lifting capacity Q, lifting height, crane span, working speed and group classification. Different parameters required by different customers will lead to a change of parameters or performances in some or even all function modules. Therefore, in terms of the correlation between requirements and function modules, the correlation matrix between function modules under a certain requirement can be identified as the matrix in Fig. 10. 2. Function relativity The output of some functions can be the input to other modules. According to function relativity, some small modules can be classified into modules with strong function. 3. Geometric relativity Geometric correlation denotes the precision requirements of two modules in static position or relative motion locus, such as the size relation, verticality, parallelism and coaxiality of two modules, etc. 4. Physical relativity Physical correlation denotes the physical relations existing between two modules, such as the transmission of energy flow, information flow and material flow, etc. 5.4.2 Identification of Module Classification Schemes After identifying the correlation matrix between functions, different module classification schemes can be discovered. These schemes will be evaluated and the optimal scheme will be selected. Due to the limitation of data sources, only the assembly and design complexity evaluation will be performed in the design of the overhead traveling crane in this paper. Table 11 shows the results calculated with the corresponding formulas in Section 3. From the classification scheme table of function modules, it is obvious that when the threshold is 0.4950, the overall complexity has the least value of 0.1448. The total modules of overhead travelling crane consist of 11 modules. Support functions 1 and 2 are independent modules. Control functions 4, 7 and 11 are independent function modules. The limiting position function 20 is also an independent function module. The working function 8 of horizontal movement along the Axis X and the working " 1 0.468 0.134 0 0.269 0 0 0.134 0 0 0 0.134 0 0.269 0.269 0.269 0.269 0.269 0.269 0 0.269 0 0 0 0 0 0 0.468 1 0.134 0 0.269 0 0 0.134 0 0 0 0.134 0 0.269 0.269 0.269 0.324 0.311 0.269 0 0.269 0 0 0 0 0 0 0.134 0.134 1 0.371 0.722 0 0 0.134 0 0 0 0.134 0 0.393 0.435 0.393 0.357 0.357 0.357 0 0.134 0 0 0 0 0 0 0 0 0.371 1 0.371 0 0 0 0 0 0 0 0 0.223 0.223 0.223 0.223 0.223 0.223 0 0 0 0 0 0 0 0 0.269 0.269 0.722 0.371 1 0 0 0.134 0 0 0 0.134 0 0.527 0.527 0.527 0.495 0.495 0.495 0 0.269 0 0 0 0 0 0 0 0 0 0 0 1 0 0.371 0.579 0 0 0 0 0 0 0 0 0 0 0 0 0.249 0.249 0.505 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.134 0.134 0.134 0 0.134 0.371 0 1 0.371 0 0 0.134 0 0.134 0.134 0.134 0.134 0.134 0.134 0 0.134 0.223 0.223 0.404 0 0 0 0 0 0 0 0 0.579 0 0.371 1 0 0 0 0 0 0 0 0 0 0 0 0 0.249 0.249 0.249 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0.371 0.579 0 0 0 0 0 0 0 0 0 0 0 0.324 0.324 0.505 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.134 0.134 0.134 0 0.134 0 0 0.134 0 0.371 0 1 0.371 0.134 0.134 0.134 0.134 0.134 0.134 0 0.134 0 0 0 0.297 0.297 0.339 0 0 0 0 0 0 0 0 0 0.579 0 0.371 1 0 0 0 0 0 0 0 0 0 0 0 0.324 0.324 0.324 0.269 0.269 0.393 0.223 0.527 0 0 0.134 0 0 0 0.134 0 1 0.857 0.676 0.385 0.343 0.343 0 0.269 0 0 0 0 0 0 0.269 0.269 0.435 0.223 0.527 0 0 0.134 0 0 0 0.134 0 0.857 1 0.857 0.343 0.343 0.343 0 0.269 0 0 0 0 0 0 0.269 0.269 0.393 0.223 0.527 0 0 0.134 0 0 0 0.134 0 0.676 0.857 1 0.343 0.343 0.343 0 0.269 0 0 0 0 0 0 0.269 0.324 0.357 0.223 0.495 0 0 0.134 0 0 0 0.134 0 0.385 0.343 0.343 1 0.723 0.64 0 0.533 0 0 0 0 0 0 0.269 0.311 0.357 0.223 0.495 0 0 0.134 0 0 0 0.134 0 0.343 0.343 0.343 0.723 1 0.682 0 0.269 0 0 0 0 0 0 0.269 0.269 0.357 0.223 0.495 0 0 0.134 0 0 0 0.134 0 0.343 0.343 0.343 0.64 0.682 1 0 0.269 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0.269 0.269 0.134 0 0.269 0 0 0.134 0 0 0 0.134 0 0.269 0.269 0.269 0.533 0.269 0.269 0 1 0 0 0 0 0 0 0 0 0 0 0 0.249 0 0.223 0.249 0 0 0 0 0 0 0 0 0 0 0 0 1 0.398 0.579 0 0 0 0 0 0 0 0 0.249 0 0.223 0.249 0 0 0 0 0 0 0 0 0 0 0 0 0.398 1 0.579 0 0 0 0 0 0 0 0 0.505 0 0.404 0.249 0 0 0 0 0 0 0 0 0 0 0 0 0.579 0.579 1 0 0 0 0 0 0 0 0 0 0 0 0 0.324 0 0.297 0.324 0 0 0 0 0 0 0 0 0 0 0 1 0.398 0.579 0 0 0 0 0 0 0 0 0 0.324 0 0.297 0.324 0 0 0 0 0 0 0 0 0 0 0 0.398 1 0.579 0 0 0 0 0 0 0 0 0 0.505 0 0.339 0.324 0 0 0 0 0 0 0 0 0 0 0 0.579 0.579 1 Fig. 10. Correlation matrix Ai Huang, H.Z. - Li Y. - Wang, Z. - Xue, L.H. - Wang, Y.F. Table 11. All classification schemes of function modules Table 10. Function modules of overhead travelling crane Class II Support (I), Support (II), Brake (DRu), Control (CONu+), Brake (Bu) Class III Drive (DRx), Control (CONx), Working (Wx), Brake (Bx) Drive (DRy), Control (CONy), Working (Wy), Brake (By) Reduce speed (RBU), Coupling (COUU), Transfer force (TFU), Lifting (RIU), Reduce force (DBU), Load handling (SU), Limit position (LPU), Limit force (LFU) Class IV Reduce speed (RBX), Transfer force (TFX), Coupling (COUX) Reduce speed (RBy), Transfer force (tFy), Coupling (cOUy) Threshold Design complexity Regularization Assembly complexity Regularization Overall complexity 1.0000 1.0000 0.0409 3.2958 0.1042 0.1451 0.8566 1.1674 0.0478 3.1738 0.1003 0.1481 0.7232 1.3827 0.0566 3.1224 0.0987 0.1553 0.7222 1.3846 0.0566 3.0711 0.0971 0.1537 0.6818 1.4667 0.0600 3.0004 0.0949 0.1549 0.5789 1.7275 0.0707 2.6535 0.0839 0.1546 0.5334 1.8748 0.0767 2.5702 0.0813 0.1580 0.5272 1.8967 0.0776 2.4456 0.0773 0.1549 0.4950 2.0202 0.0826 1.9673 0.0622 0.1448 0.4684 2.1349 0.0873 2.0406 0.0645 0.1518 0.4036 2.4777 0.1014 1.7157 0.0542 0.1556 0.3710 2.6954 0.1103 1.3827 0.0437 0.1540 0.3108 3.2175 0.1316 1.1899 0.0376 0.1692 function 12 of the horizontal movement along Axis Y are independent function modules. Functions 14, 15, 16, 3, 5, 17, 18, 19, 21 are combined to form the lifting function module. Functions 6, 9, 22, 23, 24 are combined to form independent horizontal movement along the Axis X. Functions 10, 13, 25, 26, 27 are combined to form independent horizontal movement along the Axis Y. 6 CONCLUSIONS AND FUTURE WORK The method of modular design is mainly discussed in this paper. This method has been applied in the design of an overhead travelling crane. Promising results are obtained. Our main research work is as follows: 1. The design methodology is introduced into the modular design. 2. The acquisition of customer needs and the classification of customer groups in the process of modular design are studied. 3. The construction process, methods and the theoretical knowledge of modular system are investigated. 4. The proposed method has been successfully applied to the design of the overhead travelling crane by accounting for the design characteristics of the overhead travelling crane. Methods of modular design will continue to advance with technical innovations, and there are many challenging research issues to be solved [37]. Examples of the future research work include: 1. Improve the theory of generalized modular design [38]. 2. Explore the construction method of flexible platforms of product families. 3. Investigate virtual modular design. 4. Integrate collaborative design and concurrent design with modular design. 5. Apply modular design widely to practical engineering design. 7 ACKNOWLEDGEMENT This research is partially supported by the National Natural Science Foundation of China under the contract number 50775026, and the Specialized Research Fund for the Doctoral Program of Higher Education of China under the contract number 20060614016. Constructive suggestions and comments from the referees and editors are very much appreciated. The authors would also like to acknowledge constructive revision suggestions and help by Dr. Christopher Hoyle, Department of Mechanical Engineering, Northwestern University, USA. 8 REFERENCES [1] Huang, C.C. (2000) Overview of modular product development. Proceedings of the National Science Council, ROC (A), vol. 24, no. 3, p. 149-165. [2] Peter, O.G., Liang, W.Y. (1998) An object oriented approach to design with modules. Computer Integrated Systems, vol. 11, no. 4, p. 267-283. [3] Peter, O.G., Liang, W.Y. (2001) OODS: An object-oriented design system - an example of electronics assembles. Internet Lab Technical Report TR, 02. [4] Kusiak, A., Huang, C.C. (1996) Development of modular products. IEEE Transactions on Components, Packaging and Manufacturing Technology-Part A, vol. 19, no. 4, p. 523-537. [5] Huang, C.C., Kusiak, A.(1998) Modularity in design of products and systems. IEEE Transactions on Systems, Man and Cybernetics Part A: System and Humans, vol. 28, no. 1, p. 66-77. [6] Mok, H.S., Yang, T.I., Moon, K.S., etc. (2000) Modularization of a product. Proceedings of the 5th Annual International Conference on Industrial Engineering-Theory, Applications and Practice. Hsinchu, Taiwan, p. 1-5. [7] Gu, P., Sosale, S.,(1999) Product modularization for life cycle engineering. Robotics and Computer Integrated Manufacturing, 15, p. 387-401. [8] Tong, S.Z.(1997) The implementary ways of modular design. Mechanical Manufacturing & Automation, (1), p. 14-17. [9] Hou, L., et al. (2001) Template module based modularizing design of frame structure of hydraulic press. Metalforming Machinery, (4), p. 30-33. [10] Gao, G.D., Xu, Y.S. (2002) Algorithm of module selection in modularized design of products. Journal of Mechanical Design, (8), p. 5-7. [11] Wang, X.J., Chen, Z.C. (2001) Modularization design of XH754 Series machining center. Machine Tool & Hydraulics, (3), 46. [12] Deng, Y.Y., Wang, H.P.(2001) Main Transmission system CAMD in numerical control center based on the design theory of modularization. Journal of Chongqing Institute of Technology, vol. 15, no. 2, p. 8589. [13] Wang, A.M., Meng, M.C., Huang, J.Y. (2003) Research on modular product family design methodology under multipurpose constraint environment. Mechanical Science and Technology, vol. 22, no. 4, p. 521-524. [14] Hu, W.G. (1994)Research and practice of the modular design of machine tools and the intelligent support system [Doctoral dissertation]. Huazhong University of Technology. [15] Pahl, G., Beitz, W. (1984) Engineering Design. Spring-Verlag, The Design Council, London. [16] Shi, J.F., You, L.H., Liang, X.C.(1997) Mechanical modularity. Chongqing Press, Chongqing. [17] Pan, S.X., Gao, F., Feng, P.E.(2003) Module partition method research under the circumstances about mass customization. Chinese Journal of Mechanical Engineering, vol. 39, no. 7, p. 1-6. [18] Gao, G.D. (2001) Research on virtual modular design and its application in numerical control machine [Doctoral dissertation]. Tianjin University. [19] Wang, G.Y., Wang, Y.Z. (2003) Application of information entropy fuzzy matter-element model based on AHP for projects assessment of foundation treatment. Journal of Shandong University, vol. 33, no. 4, p. 457460. [20] Hou, L., Xu, Y.S. (2003) Research on product family modeling for generalized modular design. China Mechanical Engineering, vol. 14, no. 7, p. 596-599. [21] Stone, R.B. (1997) Towards a theory of modular design [Doctoral dissertation]. The University of Texas at Austin. [22] Sun, W., Liu, X.B., Han, Y.S., etc. (2001) Research on the methods of product design and product planning for mass customization. Mechanical Science and Technology, vol. 20, no. 3, p. 315-317. [23] Wang, J.J., et. al. (2002) Research of the system of customer requirement classification. Journal of Dalian University, vol. 23, no. 12, p. 48-54. [24] Wang, A.M., Meng, M.C., Huang, J.Y.(2002) Mathematical description and determining method of product platform for developing product families. Journal of Tsinghua University, vol. 42, no. 8, p. 10491052. [25] Gu, Y.K., Huang, H.Z. (2004) Fuzzy mapping between physical domain and function domain in design process. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 12(1). [26] Su, Q., Wang, K.L., Sun, J.C. (2001) Modularization of product in agile manufacturing. Modular Machine Tool & Automatic Manufacturing Technique, (4), p. 1-4. [27] Du, T.J., Huang, H. (2003) Research on the class and level of module decomposition in modular design. Development & Innovation of Machinery & Electrical Products, 2, p. 5053. [28] Salhieh, S.M., Kamram, A.K. (1999) Macro level product development using design for modularity. Robotics and Computer Integrated Manufacturing, (15), p. 319-329. [29] Jiang, H., Xu, Y.S., Xie, Y. (1999) Research on modular dividing method of machinery products. Manufacturing Technology & Machine Tool, (3), p. 7-9. [30] Huang, H.Z. (1999) Fuzzy design. China Machine Press, Beijing. [31] Ren, S.L. (2002) Research on the modular design of artillery and the database [Master's dissertation]. Nanjing University of Science and Technology. [32] Guo, Y. (2000) Research on mass customization and key supporting techniques [Master's dissertation]. Dalian University of Technology. [33] Chen, Y.L., Xu, Y.S., Jiang, H., etc. (2000) Computer aided coding tools for modules of machining centers based on MDT. Mechanical Science and Technology, vol. 19, no. 3, p. 491-493. [34] Ying, J., Jiang, L., Pan, S.X. (1999) A solving model of cased based reasoning design problem. China Mechanical Engineering, vol. 10, no. 5, p. 555-558. [35] Zhang, Z.W., et al. (1998) Design manual of crane. China Railway Publishing House. Beijing. [36] Han, L. (2003) A new design system of overhead traveling crane based on AutoCAD [Master's dissertation]. Dalian University of Technology. [37] Kušar, J., Duhovnik, J., Tomaževič, R., Starbek, M. (2007) Finding and evaluating customers' needs in the product-development proces. Strojniški vestnik - Journal of Mechanical Engineering, vol. 53, no. 2, p. 78-104. [38] Xu, Y.S., Hou, L., Zhang, L.H., etc. (2001) Theory of the generalized modular design of hydraulic press and its application. Journal of Machine Design, (7), p. 1-3.