Design, Prediction And Control Of Product Assembly Quality And Its Application

Assembly is the key link which consumes much time and energy in the product manufacturing lifecycle, and it is also the final link to obtain the integral performance. Therefore, assembly is directly related to the quality of products. With the forming of global market, most parts can be obtained through outsourcing, so the enterprise production mode is changed. Assembly phase takes more and more proportion in the whole manufacturing process, and the influence of assembly quality on the quality of the products is also growing. Therefore, adjustment of assembly process according to the prediction results of assembly quality is highly significant to control the actual assembly quality and improve the product performance. Deep researches on Assembly precision establishment, assembly process analysis, assembly performance prediction and assembly precision control are carried out based on summary and analysis of the prediction and control technologies of assembly quality. According to the research results, software system of assembly quality prediction and control is developed, which is successful applied in the relevant scientific research projects.The main contents of this dissertation are as follows:Chapter1summerizes the development history and research status of the assembly process planning technology. The key technologies of assembly quality design, prediction and control are explored. The research significance of assembly quality design, prediction and control oriented to assembly process optimization is analyzed. Then the research contents and organization structure are introduced.Chapter2proposes the design method of assembly precision based on deviation propagation model of assembly sequence. According to the degree of dimensional variation, evaluation criteria of assembly precision are established. Deviation propagation model of assembly sequence is automatically created based on deviation cell and variable relation matrixes, while the model expression is decribed in the directed graph. Deviation propagation process is divided into propagation in the part and propagation among the parts. Accumulative deviation in the part is calculated based on analysis of deviation under tolerance restriction, and accumulative deviation among the parts is calculated based on analysis of assembly location. Prediction of assembly precision is realized by solving the degree of dimensional variation on basis of the deviation propagation process. The designer is able to obtain the assembly sequence with best assembly quality and improve the tolerance of part.Chapter3proposes collaborative concurrency control methed based on flexible conflict resolution of assembly behavior. According to the user’s intention, the concurrent assembly behavior is classified, and a flexible framework of concurrency control is established. For active concurrent assembly behavior, concurrency control method bassed on the degree of manipulation ramification is adopted oriented to deadlock joints, and concurrency control method bassed on adaptive motion of assembly unit is adopted oriented to free joints. For passive concurrent assembly behavior, negotiation optimization process of assembly program is completed using concurrency control based on dynamic allocation of operating authority and visual consultation. The proposed method dealing with concurrent acts effectively aids assembly quality prediction and control, which improves the efficiency of collaborative assembly and extends the application range of collaborative assembly.Chapter4proposes assembly performance prediction method based on multi-criteria modification of incomplete samples. Estimated value is solved by weighted calculation according to topological structure analysis of measurement data of chararteristic parameters, and confidence interval is determined by small sample information entropy calculation oriented to uncertainty of data distribution. Gross error is rejected according to the procedure metioned above, which improves the reliability of measurement data. The characteristic parameters are screened of by gray entropy correlation analysis, and the sacle of prediction model is reduced. The characteristic parameter which changes after assembly is modified by FEA method. Predicative value of assembly performance is composed of ideal value calculated by ideal model and performance error predicated by generalized regression neural network. The proposed method is suit for dealing with performance prediction of incomplete samples, which has better prediction effect.Chapter5proposes selective assembly method of oriented optimization based on assembly precision wave control. The grouping model of matching is pretreated according to sensitivity analysis of assembly parts on basis of improved signal noise ratio. The initial grouping scheme is generated using adaptive grouping based on central location and bi-directional division after transfinite analysis of manufacturing tolerance. The optimization grouping scheme is generated according to directed evolution based on grouping stability analysis. The proposed method ensures that the product has high fitting precision with wide manufacturing tolerance, which makes full use of the existing parts, and it is applicable to multi-part selective assembly.Chapter6develops the prototype system for assembly quality prediction and control oriented to precision mechanical and electrical products with research results. System architecture and function modules are described. The applications of servo valve and artificial-gem processing machine show the validity and feasibility of the new theory and method proposed in the dissertation.Chapter7summarizes the key research contents and achievements, and givens conclusions along with recommendations for future research.