Mechanical design methods: using theory to guide practice

Thread Content

The rapid advancement of science and technology, along with increasing functional requirements for products, rising complexity, shorter lifespans, and faster rates of replacement. However, the design of products, especially the methods used for designing mechanical products, prove to be inadequate and fail to keep up with the demands of technological progress. At present, computer-aided design and drafting, design calculations, machining, and production planning for products have been the subject of extensive and in-depth research, with initial results already being seen; however, computer-aided design for the preliminary stages of product development is still far from sufficient to meet the needs of such design processes. To this end, based on the extensive review of relevant literature, the author summarized the methods employed by design scholars at home and abroad in scheme design, and discussed the organic relationships among these various methods as well as the development trends in the computer-based implementation of mechanical product scheme design. 　　Based on the main characteristics of the methods used by domestic and international design scholars in the conceptual design of mechanical products, modern design methods for such concepts can be summarized into the following four categories. 　　1. Systematic design method The main feature of the systematic design method is that it views design as a system composed of various design elements. Each design element is independent, yet there are organic connections between them, and a hierarchical structure exists. When all these design elements are combined, they enable the design system to accomplish the tasks required of it. 　　The systematic design concept was proposed in the 1970s by German scholars Professors Pahl and Beitz. Based on system theory, they developed a general model for design and advocated that design work should be organized in a structured manner. Based on this design philosophy, the German Engineers’ Association developed standard VDI 2221 “Methods for the development and design of technical systems and products”. 　　 The established design process model for mechanical products basically follows the design approach specified in the German standard VDI2221. In addition, many design scholars in our country draw on and incorporate the systematic design concepts of other developed countries when developing product designs. Among these, the most representative ones are: (1) Taking user needs as the foundation for conceptualizing product functions, designing structures and components, planning manufacturing processes, and controlling operations. Starting from the overall product development process, this approach uses quality function deployment methods to systematically transform user need information into technical objectives and operational control guidelines at various stages of product development. 　(2) Treat the product as a life system at the organismal level, and by utilizing life system theory, divide the product design process into the functional requirement level, the conceptual level for fulfilling those functional requirements, and the specific design level of the product. Meanwhile, life system icons were used to abstractly represent the functional requirements of the product, thereby forming the structural framework of the product’s functional system. 　(3) The application of systems science in mechanical design can be summarized into two basic issues: one is to treat the product to be designed as a system and determine its components (units) and their interrelationships in the most optimal way ； Second, treat the product design process as a system, and based on the design objectives, correctly and reasonably determine the tasks involved in various aspects of the design as well as the different design stages. 　　Due to the different perspectives each designer adopts when studying a problem and the different focuses they give to it, the specific research methods used in scheme design also vary. Below are some representative systematic design methods. 1.1 Design Element Method The “product solution” is described using five design elements (function, effect, effect carrier, shape element, and surface parameter); it is believed that once the values of these five design elements for a product are determined, all the characteristics and their values of that product are also determined. Design scholars in our country have also used similar methods to describe the principles behind products. 1.2 Graphical Modeling Method
The developed “Design Analysis and Guidance System” KALEIT uses well-structured graphics to describe the functional structure of products and related abstract information. It enables graphical modeling of system structures and functional relationships, as well as the interconnections between different functional layers. 　　The design is divided into two aspects: auxiliary methods and information exchange. By utilizing the Nijssen information analysis method, it is possible to employ graphic symbols, semantic model structures with rich content, mechanisms for describing integration conditions, ways to classify constraint types, and the ability to establish any combination of relationships. This enables the integration of design methods with information technology, allowing for the graphical modeling of information relationships between different abstraction levels during the design process. 　　Reference [11] uses a semantic design network as a design tool. In the Active Semantic Design Network ASK developed there, designs are described through a network composed of nodes and lines; nodes represent modular units such as design tasks, functions, components, or processing equipment, while lines are used to adjust and define the various semantic relationships between nodes. This approach enables the establishment of models for all activities and results throughout the design process, allowing the definition of early design requirements as well as the detailed description of each structure to be expressed through the definitions of these relationships, thus achieving a leap from abstraction to specificity in the computer-aided design process. 1.3 The “Conception”—“Design” method
Divide the conceptual design of a product into two phases: “conception” and “design”. “The task in the “conception” stage is to seek, select, and combine principles that meet the requirements of the design task. “The work in the “design” phase is to specifically implement the principles from the conceptualization phase. 　　The “concept” of the scheme is specifically described as: seeking principles that meet the requirements of the design task based on an appropriate functional structure. That is, the sub-functions within the functional structure are realized by “structural elements”, and the physical connections between these “structural elements” are defined as “function carriers”. The interactions between the “function carriers” and the “structural elements” in turn give rise to a functional schematic diagram (a simplified diagram of mechanical motion). The “design” of the scheme involves, based on the functional schematic diagram, first qualitatively describing all “functional carriers” and “structural elements”, and then quantitatively describing the shapes and positions of all “structural elements” and connectors (“functional carriers”) to obtain a structural schematic diagram. Roper, H., by utilizing graph theory and the concepts he defined such as \"General Design Element (GE)\\", \"Structural Element (KE)\\", \"Functional Structural Element (FKE)\\", \"Connecting Structural Element (VKE)\\", \"Structural Part (KT)\\", and \"Structural Element Part (KET)\\", as well as various diagrams that describe the relationships between the dimensions, positions of structural elements and their transmission parameters, formalized the methods used by design experts based on intuition. This approach enables the effective application of existing knowledge and can be utilized in the \"ideation\" and \"design\" phases. 　　From the perspective of design methodology, the design process after defining the design task is divided into three steps: 1) Obtaining functions and their structure (referred to as “functions” for short) ； 2) Search effect (abbreviated as “effect”) ； 3) Finding structures (abbreviated as “configuration rules”). And the workflow in the conceptualization stage of mechanical products is described using the following four strategies: Strategy 1: Consider “function,” “effect,” and “configuration rules” separately. Therefore, variant schemes can be created separately for each working step, thereby generating a wide range of principle spectra. Strategy 2: “Effects” are associated with “configuration rules” (including those created by the designer), while functionality is considered separately (usually related to the design task). At this point, distinguishing typical configuration rules and the effects to which they belong requires extensive experience, and the resulting set of solutions is far smaller than that of Strategy 1. Strategy 3: “Function,” “effect,” and “configuration rules” are closely related. It is suitable for fields where there is no choice among functional, performance, and design rules, and where special requirements apply, such as ultra-small machinery, very large machinery, high-value functional components, as well as parts with specific functional demands, and so on. Strategy 4: Conduct structured solutioning based on design requirements. This strategy starts with existing components, and by using different arrangements and connections between them, achieves the desired function. 1.4 Matrix Design Method: During the design process, a \"requirements—functions\" logic tree (\"AND/OR\" tree) is used to describe the relationships between requirements and functions, thereby obtaining a set of functional design solutions that meet the requirements and generating various design options. Then, a \"requirements–functions\" correlation matrix is established based on the \"requirements–functions\" logic tree, in order to describe the complex relationships among the functions required to meet the requirements, thereby showing the one-to-one correspondence between requirements and functions. 　　Kotaetal uses matrices as the basis for the conceptual design of mechanical systems, breaking down the design space of such systems into functional subspaces; each subspace represents a single module of the design. At the higher levels of the abstraction stage, each design module is represented by a motion transformation matrix and an operable constraint vector ； At the lower level of the abstract stage, each design module is represented by a parameter matrix and a motion equation. 1.5 Bond Graph Method The functions of the components constituting a system are classified into various types, such as energy generation, energy consumption, energy form conversion, and energy transmission. Bond graphs are utilized to represent these functional relationships among components. The aim is to integrate function-based models with bond graphs, thereby enabling the automatic generation of functional structures as well as automatic conversion between such structures and bond graphs. Additionally, efforts are made to find methods for generating multiple design solutions from bond graphs. 　　2. Modular design method for structures: From the perspective of product planning, it is proposed that when defining design tasks, a functional product structure should be used as a basis; existing product solutions (such as standard components) should be utilized to describe these tasks. In other words, when breaking down tasks, it is necessary to consider whether there are corresponding product solutions for each sub-task. This approach allows potential contradictions in the design tasks to be identified at the product planning stage, enabling early estimation of production capacity and costs, as well as assessing the flexibility of plans during the development process. As a result, design efficiency and reliability are improved, while the cost of new products is reduced. Feldmann divides the functional product structure of design tasks into four layers: (1) product → (2) functional components → (3) main functional components → (4) functional elements. It also employs an application-oriented structured feature catalog to provide more specific qualitative and quantitative descriptions of the functional components. At the same time, the tool software STRAT was developed for use in the early stages of product development and design. 　　It is believed that most of the functions in specialized machinery can be addressed using existing products, with only a few specialized functions requiring novel solutions; therefore, adopting a functional product structure in the design of specialized machinery is highly beneficial for assessing the design and manufacturing risks associated with such machinery. 　　It is recommended to, based on the analysis of product functions, break down the product into one or several modular basic structures that possess certain functions, and then construct different products by selecting and combining these modular basic structures. These basic structures can be parts, components, or even a system. An ideal modular basic structure should have standardized interfaces (connection and mating parts); it should be serialized, universal, integrated, hierarchical, flexible, and cost-effective, while also possessing interchangeability, compatibility, and relevance. By integrating software component technology with CAD technology, our country combines deformation design with modular design. Based on the principle of hierarchical modularity, machining centers are divided, from largest to smallest, into product level, component level, sub-component level, and element level. Expert knowledge and CAD technology are then used to combine these elements into functional modules of various types and specifications, which in turn are used to form different overall designs for machining centers. 　　Using design catalogs as a tool for selecting variant mechanical structures, it is proposed to comprehensively and structurally organize the deconstructed elements of designs to form a solution-set design catalog. Additionally, additional information commenting on each solution is listed in the solution design catalog, which is very helpful for design engineers in selecting solution elements. 　　Based on the connection characteristics of mechanical components, they can be classified into four types: 1) Components with direct positioning between elements and self-adjusting capabilities ； 2) Assemblies with structural commonalities ； 3) Connection with a nested structure and nested components ； 4) A modular structure and the connection of modular components. Quasi-symbols are also used to represent typical components and the connection rules between them, thereby enabling the algorithmization of component connections and the visualization of concepts. 　　In the conceptual design of mechanical systems, the \"function establishment\" module is used to break down functions, with the optimal level of \"granularity\" for this decomposition being a one-to-one correspondence between functions and mechanism types. “The “Structure Establishment” module serves as the candidate for functional decomposition in order to implement the mapping algorithm. 　　3. Design method based on product feature knowledge The main feature of the design method based on product feature knowledge is that it uses a language understandable by computers to describe the features of products as well as the knowledge and experience of experts in the relevant design field. A corresponding knowledge base and inference engine are established, and computer-aided product design is then carried out by utilizing the stored domain knowledge and the established inference mechanisms. 　　The conceptual design of mechanical systems primarily involves making estimations and decisions based on the characteristics of the product, as well as the knowledge and experience of experts in the design field, thereby achieving a comprehensive determination of the type and quantity of mechanisms. To achieve computer-aided design at this stage, it is necessary to study the automatic acquisition, representation, integration, coordination, management, and utilization of knowledge. To this end, design scholars at home and abroad have conducted extensive research on the automated processing of knowledge related to the conceptual design of mechanical systems, and the methods employed can be summarized as follows. 3.1 Coding Method The mechanisms are classified based on the “motion transformation” function (referred to as a functional element), and codes are used to describe these functional elements and mechanism categories, thereby establishing a knowledge base for the “Expert System for Mechanism System Design”. On this basis, by combining binary logical reasoning with the principles of fuzzy comprehensive evaluation, the reasoning mechanism for this \"expert system\" was developed and applied in the design of specialized four-station machine tools. 　　By applying the theory of biological evolution, which relies on natural selection and sexual reproduction for the evolution of organisms, mechanism design utilizes graph theory to represent the structure of a mechanism as a topological graph. Encoding techniques are then used to convert the structure and properties of the mechanism into binary sequences representing individual chromosomes. Fitness values are determined based on design requirements, and the theory of biological evolution is applied to control the reproduction process. Through methods such as selection, crossing over, and mutation, individuals with low fitness are eliminated, thereby enabling the rapid evolution of individuals with the highest adaptability – that is, mechanism designs that best meet the design requirements. 3.2 Hybrid Representation of Knowledge For the conceptual design of complex mechanical systems, it is particularly appropriate to use a hybrid knowledge representation method to describe the various types of knowledge involved in such design, and this view has been shared by many design scholars in our country. 　　In the development of the intelligent decision support system DMDSS for complex product design, knowledge representation methods such as rules, frameworks, processes, and neural networks are integrated organically to accommodate the description of different types of knowledge in design. By combining various individual knowledge representation methods (rules, frameworks, and processes) in accordance with object-oriented programming principles—using framework slots to represent an object’s attributes, rules to represent its dynamic characteristics, and processes to represent the processing of knowledge—a hybrid knowledge representation format was developed. As a result, the \"Object-Oriented Intelligent System for the Design of CNC Gantry Milling Machine Gearboxes GBCDIS\" and the \"Expert System for Gearbox Structure Design GBSDES\" were successfully created. 3.3 Utilization of knowledge-based development tools In the CAD system for couplings, the knowledge-based development tool NEXPERT-OBJECT was used to create a database of object-oriented design methods, leveraging object-oriented approaches. This provides designers with a wide and reliable range of design methods for the conceptual and structural design of couplings. Thus, NEXPERT is utilized to describe the aspects of linear guide design that require knowledge-based approaches, thereby identifying knowledge-based solutions and developing an expert system for linear guide design. 3.4 Design Catalog Method A three-level hierarchical design catalog consisting of “functional modules,” “functional elements,” and “mechanical assemblies” was constructed, and this hierarchical design catalog was used as the knowledge base for the intelligent design system for mechanical transmission principle solutions, as well as an auxiliary tool for development and design. 3.5 Instance-based Methods In the knowledge base of design-oriented expert systems, basic predicates are used to describe design requirements, design conditions, and selected solutions; framework structures are employed to represent “engineering instances” and various “conceptual entities”. Instance-based reasoning techniques are then used to generate candidate solutions that meet the design requirements of the product. 　　4. Intelligent design methods The main characteristics of intelligent design methods are: based on the theories of design methodology, they utilize 3D graphics software, intelligent design software, and virtual reality technology, as well as multimedia and hypermedia tools, to develop and design products, express the concepts behind them, and describe their structure. 　　By utilizing mathematical system theory, along with systems engineering theory, product design techniques, and the system development methodology VDI2221, the multimedia development system software MUSE was developed for use in the early stages of product design. 　　In the design of ATM machines, the entire product development process is summarized into three stages: \"product planning,\" \"development,\" and \"production planning,\" making full use of advanced existing CAD technology—virtual reality technology. 1) Product planning—conceiving the product. Its task is to determine the external characteristics of the product, such as color, shape, surface quality, ergonomics, etc., and to represent the initial concepts through CAD 3D models. A simple model that reflects the overall appearance of the product is created; this model can be established in a virtual environment. With the help of data gloves and 3D mice, users can also participate to some extent in this environment and quickly generate different designs and colors. The 3D model serves as a basis for detecting the external shape effects, as well as a basis for design variables in geometric graphic display; it is also the foundation for various analyses during the development process. 2) Development—designing products. At this stage, the solution elements are arranged and integrated on the 3D model primarily based on the principle of \"system synthesis.\" These solution elements have different meanings depending on the design objectives: they can be basic elements such as bolts, shafts, or hub connections ； It can also be a composite system composed of mechanical, electrical, electronic components, control technologies, or software ； It can also be requirements, characteristics, shape, and so on. After assigning the key solution elements that enable the functions to the 3D model, it is possible to analyze the product configuration (the relationships between the solution elements in the design model). Product configuration analysis is an important method for integrating the results of \"product planning\" and \"development\". 3) Production planning—processing and assembling products. At this stage, the application of CAD technology in the assembly process is primarily discussed. It is proposed to use computer-generated images to display the assembly process of components at their respective positions; that is, by means of virtual assembly models, the relationship between design and assembly can be revealed, thereby identifying difficulties and problems and finding solutions to them. It is believed that integrating CAD technology across the three stages of product development enables continuous integration and analysis of the design process during \"product planning,\" \"development,\" and \"production planning.\" Therefore, problems at various stages can be identified early, allowing the product to be continuously refined and improved throughout the development process. 　　The use of virtual reality technology for design in our country is still in its infancy. By utilizing object-oriented techniques, an expert system for designing mechanism combinations synthesized in a sequential manner was studied in detail. With the help of OpenGL technology, which offers high-performance graphics and switching capabilities, it was possible to view the designs generated by the expert system from various angles in a 3D environment, to determine whether there were any conflicts in the way the mechanisms interacted during movement, among other things. 　　The development of standard modules, the overall design of the product, as well as the formulation of its manufacturing processes and usage instructions (see Figure 1) is referred to as rapid prototyping technology. It is recommended to integrate rapid prototyping technology, multimedia technology, as well as virtual representation and neural networks (used in stages where they are needed to solve problems) during the product development process. It is pointed out that as computer hardware and software continue to improve, multimedia graphics processing technologies should be utilized as much as possible in product development; for example, 3D graphics (three-dimensional models) can replace the need for imaginative mental representations of three-dimensional structures when assembling, disassembling, and designing connectors. 　　Using the intelligent CAD system SIGRAPH-DESIGN as a development platform, the product development process was divided into conceptual design, assembly design, and component design. Based on variable design techniques, a conceptual model of the cam linkage mechanism in offset printing machines was established. Based on the research work described in the literature, its conceptual model is established on the basis of defining the institutional and numerical aspects; by utilizing the variable design function provided by the software SIGRAPH-DESIGN, the schematic diagram changes as the structural parameters of the mechanism change, and the parameters of the conceptual model are passed on to the assembly model and part design at the next level. 　　5. Review of various design methods and their development trends
In summary, systematic design methods divide the design tasks into hierarchical levels, ranging from the abstract to the concrete (from the requirements of the design task to the plans or structures for fulfilling it). For each level, the desired objectives and corresponding methods are determined. These levels are then organically linked together, progressing from the simple to the complex and from the abstract to the concrete. This makes the entire design process systematic, providing clear rules and methods to follow; it also facilitates the computer-aided implementation of the design process. 　　The structural modular design method regards the implementation of a certain function as a structural module, and through the combination of such structural modules, the conceptual design of the product is achieved. For certain types of mechanical products, since the functions of their components are well-defined and relatively stable, it is easier to divide them into structural modules; therefore, using a structural modularization approach for design is more appropriate. Since there is no one-to-one relationship between entities and functions, a single entity can typically perform multiple functions, and a single function can often be carried out by multiple entities. Therefore, if the structural modular design method is applied to product design in a general sense, it becomes quite difficult to divide and select structural modules, and it requires designers to have extensive design experience as well as broad knowledge across multiple disciplinary fields. 　　The conceptual design of mechanical products generally cannot be carried out using pure mathematical calculations, nor can it be fully described by mathematical models; instead, a formal description based on the product’s characteristics is required, with reasoning and decision-making relying on the knowledge and experience of design experts. Therefore, to achieve computer-aided product design, it is necessary to address issues related to computer storage and the use of product design knowledge as well as expert design decisions, thereby developing a design method based on product feature knowledge. 　　Currently, intelligent design methods primarily rely on 3D graphics software and virtual reality technology for design; they offer good visual clarity, and users can participate directly in the design process to a certain extent during the initial stages of development. However, these methods lack systematicity, and determining the structure, shape, size, and position of components accurately requires software with a high level of intelligence or the involvement of designers with extensive experience. 　　It is worth noting that the various methods mentioned above are not entirely isolated; there are certain connections between them. For example, in the structural modular design method, the division of structural modules involves systematic thinking. When establishing knowledge bases for product features and design methods as well as inference engines, systematic and structural modular approaches are also typically utilized. Moreover, design based on product feature knowledge is one of the foundations for intelligent design of solutions. In the conceptual design of mechanical products, generic parts, components, or common mechanisms that can fulfill specific functions are regarded as structural modules, which are then applied in the detailed design at various levels of systematic design. By integrating this modular structural approach into systematic design methods, it is possible to ensure standardization in design, simplify the design process, improve design efficiency and quality, and reduce design costs. The rapid development of network technology has made remote collaborative design and manufacturing possible, as well as the implementation of parallel engineering from users’ functional requirements for products to design, processing, assembly, and the final product. However, one of the key prerequisites for achieving these goals is the three-dimensional visualization of the product design. To this end, not only are 3D graphics software and intelligent design software being used more and more in the conceptual design of products, but virtual reality technology as well as multimedia and hypermedia tools are also beginning to play a role in such design processes. Currently, developed countries such as Germany are focusing on researching the application of hypermedia technology, the product data exchange standard STEP, and the standard virtual reality modeling language VRML (a standard exchange format for virtual environments on the Internet) in product design. 　　The conceptual design of mechanical products is moving towards computer-aided implementation, intelligent design, and meeting the requirements of collaborative design and manufacturing across different locations. However, since research on computer-based methods for product conceptual design started relatively late, there are currently no mature design tool software programs capable of achieving these objectives. The author believes that a comprehensive application of the four design methods discussed in the text is an effective way to achieve this goal. Although the integrated application of these methods involves a wide range of fields – not only knowledge related to mechanical design but also theories of systems engineering, artificial intelligence, computer hardware and software engineering, network technology, and more – it remains a direction that must be pursued in product design. Research in this area abroad has already begun to yield results, and design scholars in our country have also realized the importance of CAD technology and international exchanges and cooperation, as well as the measures that should be taken.