現代化生產對機械零組件的更換時機要求越來越高,工業生產中迴轉體零件是應用最廣泛的一種,而形狀誤差更是機械零組件及其互換性的重要指標,往往是產品質量的關鍵,它在評定機械零件產品質量中有重要作用,對形狀誤差測量的數據處理方法的選定,直接影響形狀誤差的計算精確度。因此研究出一種快速、準確評定形狀誤差方法具有重要的理論意義和經濟價值。本文研究了圓度誤差、正方體誤差、圓柱度誤差與圓錐度誤差並建構其四種形狀誤差數學模式。 本研究以提出之粒子群演算法(Particle Swarm Optimization,PSO)、免疫演算法(Immune Algorithm,IA)與基因演算法(Genetic Algorithm,GA)等三種人工智慧演算法對此形狀誤差問題進行求解。除此之外,為了確保求解品質,本研究採用統計檢定方法,以找出最佳參數來求解此形狀誤差問題。本研究分別對圓度誤差、正方體誤差、圓柱度誤差與圓錐度誤差等四種形狀誤差進行20個例題測試,數值結果顯示,免疫演算法的求解品質優於粒子群演算法與基因演算法。
Form error is important to to the quality of piece parts. The rotation parts are the most widely used components in industrial production, and the form error is an important indicator. As known, it is also the key to the product quality. There are various approaches to evaluate the form errors for objects. In this thesis, we apply artificial intelligence approaches to evaluate different form errors, including roundness error, cube error, cylindricity error and conicity error. In this thesis, we apply three heuristic algorithms for solving the form error problem, including particle swarm optimization, immune algorithm and genetic algorithm. In addition, in order to ensure that the solution quality, in this study, we use the statistical test method to compare the results of three heuristic algorithms. In this thesis, we experiment four types of test problems for roundness error, cube error, cylindricity error and conicity error. Numerical results show that the solutions by the immune algorithm are better than those by particle swarm optimization and genetic algorithm, respectively.