Study On The Method Of Module Division For Complex Product Systems

The primary problem with modularity is module division.Traditional module division is mainly from the perspective of product function,and for the module division of complex product systems,due to the integrity of the complex product system itself,the hierarchy and complexity between modules,some modules usually have multiple attributes,such as technical characteristics,functional structure characteristics or itself is a complete small system,etc.Therefore,the traditional module division process and method does not apply.In addition,the current application of the module division algorithm,there are easy to fall into local search and other problems.In the evaluation of the classification results,there are also subjective giving of the weight of each evaluation index,which is difficult to guarantee the objectivity of the results.Based on the special project of major science and technology in Sichuan Province,"Modular Delivery of Aviation Large Parts"(project number 2018GZZZ0019),this thesis studies the method of module division for complex product systems.To this end,this thesis for complex product systems belongs in the complex network structure,with multi-level multi-attribute characteristics,first of all,based on the systemfunction-structure of the complex system layering model,based on customer demand,first establish the system layer,through the functional-behavior-structure(by functionbehavior-structure)The Function-Behavior-Structure,FBS mapping model straightens out the correlation between functional structures,reduces the complexity of division,guides the generation of module division principle based on functional structure,and uses the civil machine head electrical and electronic system to make an example of the construction of the hierarchical model.Secondly,on the basis of customer demand to expand the supplier demand and technical requirements,generate the functional structure-based module division principle,pre-division principle and life cycle-based division principle,so that the division principle is more perfect and reasonable.Under the guidance of the principle of division,the subrelated matrix is established,the weight is solved according to the hierarchical analysis method,and the comprehensive related matrix is established.The traditional clustering algorithm is improved,and the hybrid algorithm has a stronger ability to jump out of local search and converge by combining the improved firework algorithm(Firework Algorithm,FWA)with the fuzzy c-mean cluster algorithm(Fuzzy C-means,FCM).The effect of module division is improved,and the integrated related matrix method and IFWA-FCM algorithm are used as examples by the civil machine head electrical and electronic system.Third,for a variety of division schemes,the program set is evaluated by using the improved decision-making method of approaching the ideal solution.The evaluation semantics are converted by intuitive fuzziness,and the weight of the index attribute is determined by the main objective empowerment method to ensure the accuracy of the attribute weight.The evaluation method is applied by the electronic and electrical system of the civil machine head.Finally,the module division method proposed in this thesis is verified by taking the civil machine head flight control system as an example.Comparing the module scheme derived from this method with the actual scheme of the relevant enterprises,the validity of the method proposed in this thesis is proved.This study provides some basic research ideas for the module division method for complex product system,and provides method support for reducing the complexity of module division of complex product system,improving the performance of clustering algorithm and choosing the scheme more objectively.