Font Size: a A A

Fuzzy-Set-Based Manufacturing Cell Formation And Its Application In Transformer Company

Posted on:2003-02-15Degree:DoctorType:Dissertation
Country:ChinaCandidate:J LiFull Text:PDF
GTID:1116360065460097Subject:Motor and electrical appliances
Abstract/Summary:PDF Full Text Request
Group Technology has been recognized as the prerequisite and integrative philosophy of modern manufacturing systems. As the core of Group Technology, grouping method is the research focus of integrated manufacturing in the international academic world.Based on the review of Group Technology, this dissertation researches fuzzy set based manufacturing cell formation approach systematically. The major research works are listed as follows:1. The fuzzy set based group technology approaches are analyzed, and the major problems of existing approaches are pointed out.2. After giving the fuzzy model and Fuzzy C-Means algorithm for manufacturing cell formation, this dissertation researches the influence of algorithm parameters on the clustering performance, put forwards the parameter values. The large-scale data tests testify the effect of the given parameter values.3. The performance of manufacturing cell formation with FCM is researched in depth, the major problems of FCM for manufacturing cell formation are put forward, and the reasons for the problems are analyzed, which is the bases for the problem solving.4. Based on the major problems of FCM, the modification methods are put forward, and the large-scale data experiments test and analyze the improvement effects of different modification methods. The major improvements include:Firstly, a modified Subtractive Clustering algorithm is put forward to generate a good initial clustering center matrix, which can improve the clustering performance of FCM.Secondly, the effect of distance function for clustering performance is researched, the disadvantages of Euclidean distance for part family formation are pointed out, a more suitable part-machine processing feature related distance function is put forward, with which the clustering errors for parts and machines are basically eliminated. ?Thirdly, the influence of cluster center function for clustering performance is researched, two different center functions are compared, the applicability of different center functions for different problems is put forward, on which the suitable center function can be selected for specific clustering problems.Fourthly, because of the fluctuation of grouping efficacy during the FCM iteration process, the iteration solution selecting procedure is put forward, with which the infeasible solution of last iteration will not be selected.5. Based on the systematic research of FCM and manufacturing cell formation, the fuzzy set based manufacturing cell formation algorithm SDFCM is put forward. In the new algorithm, the applications of improved Subtractive Clustering for initialization, part feature basedindistance function and the best solution selection procedure eliminate the clustering errors of FCM, and the clustering efficacy is improved dramatically. In addition, numerical example is given for algorithm illustration. 20 literature data sets and 90 random generated data sets are used to test the performance of SDFCM. The statistical analysis of the experiment proves that the SDFCM is better than FCM significantly, and the solutions of SDFCM are consistent with the best control solutions statistically.6. Based on the Tianwei Baobian Company's application example, the usage of membership degree information given by SDFCM is researched; the method of recognizing bottleneck machines with membership matrix is put forward. The application result shows that SDFCM can provide the device duplication solutions according to the machine workloads, which can help to find and add the bottleneck machines to increase the productivity of production system.7. The clustering performance of SDFCM is researched; the advantages of SDFCM compared with traditional methods, FCM and other artificial intelligent methods are pointed out.To sum up, the main creative results include:1. Put forward a part processing based distance function to eliminate the clustering errors caused by the application of Euclidean distance function.2. Put forw...
Keywords/Search Tags:group technology, fuzzy c-means, SDFCM, part family formation, manufacturing cell formation, subtractive clustering algorithm
PDF Full Text Request
Related items