Geometric Error Evaluation And Assembly Precision Prediction For Complex Products

With the development of miniaturization、intellectualization and lightweight in aero and spaceflight, assembly precision protection has become one of the most difficult problems in the assembly of complex products. The geometric error is one of the key issues affecting the quality of product. To overcome these problems, we have investigated the reliability of the methods of geometric error evaluation and the algorithm of the assembly precision prediction in this thesis. The main contents of the thesis are concluded below:(1) The related conception and classification of the geometric error evaluation, assembly prediction and tolerance optimization are discussed, the existing problems and shortcomings are pointed out, the purpose and originates are given as well.We have reviewed some recent studies in terms of geometric error evaluation, assembly prediction and tolerance optimization to show the shortcomings and problems in the existing research method. Through this, we talk about the meaning and purpose of our work.(2) In order to achieve the data fitting of the measured clouded data, a method using the convex hull in computation geometry to construct the datum is proposed. Here, according to the definition of tolerance datum, we first get the constrained freedom from the priority and the geometric feature of the datum. Then the convex hull and the constrained freedom are incorporated to compute the location and direction of the datum.(3) To improve the accuracy of the geometric error evaluation, a method based on the minimum zone is proposed. The minimum zone is constructed using the definition of geometric tolerance and the measured clouded data. After taking the constrained freedom between the datum geometry and the function geometry into fully consideration, the geometric error is determined by comparing the minimum zone with the relevant requirement.(4) In order to evaluate the composite position error, a method using overlapping inclusive theory is proposed. First, the real region of the composite position error is computed by using the measured cloud data. Then, through rotating and translating this region, we build the mathematical model of pattern-locating error and feature-relating error with the help of the definition of composite position tolerance. At last, the composite error is obtained using the minimum circle and the bisection method in the model.(5) In order to get the propagation paths of tolerance and assembly constraint in the assembly, a method is proposed based on the constrained freedom between the geometric features. First, a unified model is obtained by combining the tolerance and assembly constraint. Second, we compute the actual freedom between the geometric features using their relative position in space. Third, we draw the diagram of the geometric features using the depth first search. Last but not least, we use the hierarchical searching method to get the path between the datum feature and the functional feature.(6)To predict and optimize the assembly precision,a method based on the small displacement vectors and the niche PSO algorithm is proposed. The mathematical model of vectors is constructed by the real region of geometric error, while the Monte-carol statistical method is used to simulate the vectors of error. The precision of assembly is deduced from the Jacobian Matrix and the vectors. For a multi-objective model of tolerance optimization, we take the repair cost and quality loss cost as design objectives to improve the traditional particle swarm optimization based on the idea of niche and Pareto. The Pareto front of uniform distribution is obtained using the niche PSO algorithm to achieve the multi-objective model of tolerance optimization.(7) The Geometric error evaluation and Precision Prediction system is developed based on the research result of this paper and engineering needs. The related technologies and algorithms are verified with concrete examples.Finally, the summing-up of the whole dissertation is made as well as the predictions of the future research and the directions of relevant techniques.