Study On The Method Of Manufacturability Evaluation Of Prismatic Parts For Virtual Manufacturing

A virtual manufacturing (VM) system is defined as a computer system that can generate the same information about manufacturing system's structure, states, and behaviors as those that can be observed in a real manufacturing system. Virtual manufacturing utilizes the computer emulation and virtual reality technology to simulate all the processes from design to manufacture. Manufacturability evaluation is to design the structure of product under the constraints of manufacturing resources, which can consider the factors concerned with product manufacturing in design stage. By evaluating the manufacturability of product effectively in design phase, the problems in design and manufacture of products can be found as soon as possible, the controllability and decision-making ability is increased. In this way, the manufacturing cost and the development period is reduced and the quality of product is improved. In recent years, the manufacturability has been one of the most interesting and important research fields in VM.Manufacturability evaluation is an important component of VM. This paper proposed the method of manufacturability evaluation for prismatic parts in VM system after analyzing the contents and the demands of information of manufacturability evaluation. The model of manufacturing resources and feature recognition are involved in the system of evaluation. The evaluation has two layers of meaning, one is to evaluate the manufacturability of products based on features to find out whether the products can be machined, and the other one is to assess the overall machining scheme synthetically to find the optimum machining scheme according to different demands. The study on manufacturability evaluation has important significance to shorten the development period and reduce manufacturing cost in theory and practical application.The feature recognition is the premise and basis of manufacturability evaluation based on features. To solve the problem of date exchanging and sharing between different solid modeling systems, the feature recognition based on STEP neutral file is studied. The manufacture feature recognition includes two parts: the extraction information of geometry and topology from the STEP neutral file and the feature recognition based on the attribute adjacent graph (AAG) by using the artificial neural network (ANN). STEP is an international standard designed to provide a complete, unambiguous, computer-related definition of physical and functional characteristics of a product throughout its lifecycle. Object oriented programming is used to build the mapping relationship of the data types between STEP and C++ and to design lexical analyzer to extract information from STEP file after analyzing the grammatical characteristics, geometry information and topological relation of STEP file, and the extracted information is converted into AAG. A series of heuristic rules is used to decompose the AAG into sub-graphs and the intersecting feature is decomposed into simple features. The sub-graph is represented as a set of vector which is composed of 1 and 0 by using the"problem"method. By summing up the characteristics of features, several sets of vector are used to represent the features of hole, blind hole, slot, blind slot, step, blind step respectively and these vectors are inputted into the three layers BP neural networks to train the network. Then the vectors according to different sub-graphs are inputted into the BP neural networks to recognize the features.The manufacturability of a product depends on the processing ability of specific manufacturing resource. In order to evaluate the manufacturability of part based on features, the model of manufacturing resources based on feature is studied. The information model of manufacturing resources is built by using the object-oriented method to meet the demands of information used in manufacturability evaluation and other activities in VM. The information model of manufacturing resources includes the basic information and the machining capability information of resource, and the association relation between resources (processing equipment, cutting tool, fixture, measuring tool and so on) and features is built. To better utilize the information of manufacturing resources and decrease the searching time of processing equipments in manufacturability evaluation, the hybrid algorithm of fuzzy c-means clustering algorithm (FCM) and genetic algorithm (GA) is implemented in this paper to group manufacturing resources based on manufacturing features. In order to make the partition more reasonable, the process capacity of component dimension and precision are also considered as the attributes of equipments. By this means, the optimum number of clusters and the optimum partition can be gotten at the same time. When evaluating, it only needs to search the corresponding group of equipments to find the processing equipment according to different features and the processing ability of equipment is evaluated according to the machining capability information. The corresponding cutting tools and fixtures can be found at the same time. By modeling the manufacturing resources, the efficiency of manufacturability evaluation is increased.To evaluate the manufacturability of part, the three-level manufacturability system is built. Manufacturability evaluation needs to solve two problems, one is whether the part can be processed, and the other is how to process the part. In the first level of evaluation, structural processing is evaluated. The purpose is to find whether the design of part has processing problem based on the rules which are summarized from the experiences in design and manufacturing and send the information back to designer. In the second level of evaluation, the machinability of features is evaluated. The method is to evaluate whether the design requirements of every feature can meet the constraints of manufacturing resources which includes processing equipments, cutting tools, fixtures, measuring tools and all the equipments which can be used to machine the product are recorded. Machining range and quality of equipments are the main contents of evaluation. The first and second level manufacturability evaluation and design of part are carried out alternatively. In this way, the problem in design can be found and solved as early as possible. The third level manufacturability evaluation is quantitative and general which mainly solves the problem how to manufacture, and the aim is to find the optimum machining scheme. In this paper, the genetic algorithm (GA) and the analytic hierarchy process (AHP) are used in the selection of machining scheme for the first time. Because a same manufactur feature could be processed by several different machining methods and machining equipments, one part maybe have a lot of different machining schemes. In conventional machining process, the selection of machining scheme is always based on knowledge and experiences that the effects of different factors are not considered fully and the selection is done only at the lowest cost. Because of the diversity of evaluation factors and the different demands of decision-maker, the machining scheme selection is a multi-objective decision-making problem and plays an important role in the manufacturability evaluation. In order to reflect the effects of different factors, AHP is introduced to assignment the weights to different factors. The multiple-objective optimization problem is to combine the individual objective functions into a single composite function and GA as a global search technology is used to optimize the machining scheme according to different demands of decision-maker.Finally, the main topics and conclusions of this thesis are summarized, and the future research direction has been proposed.