作　　者： ; ;

机构地区： 华南理工大学机械与汽车工程学院

出　　处： 《华南理工大学学报（自然科学版）》 2012年第12期30-34,40,共6页

摘　　要： 形状误差的智能评定结果稳定性较差,掌握智能评定结果的概率分布特性及拟合方法对进一步提高该方法的可靠性有重要意义.文中以平面度误差粒子群算法评定为例,基于智能评定结果的概率分布特性分析,提出β分布统示法来对评定结果进行拟合,并利用K-S法检验拟合效果.选取平板实测三坐标数据,并利用粒子群算法对其进行100次智能评定,采用β分布统示法对评定结果样本进行拟合,结果表明:提取参数取值在[20%,70%]时,β分布统示法拟合效果较好,拟合形状参数均大于1,且拟合概率分布右偏,符合智能评定结果的概率分布特性. As the intelligent evaluation results of form error are of poor stability, it is necessary to investigate the probability distribution characteristics of the evaluation results and the corresponding fitting method to improve the reliability of the intelligent evaluation method. In this paper, by taking the flatness error evaluation based on the particle swarm optimization （PSO） algorithm as an example, the probability distribution characteristics of the inte- lligent evaluation results was analyzed, and a β-distribution uniform expression was proposed to fit the intelligent evaluation results, followed by a fitting performance test via the K-S method. Then, the three-coordinate data of a plate were measured, the plate flatness was evaluated for 100 times by using the PSO algorithm, and the intelligent evaluation results were fitted by employing the β-distribution uniform expression method. It is demonstrated that, when the intercept percentile Qp is in the interval of [ 20%, 70% ], the β-distribution uniform expression method is of high fitting performance for the intelligent evaluation results, the fitted shape parameters are both more than 1, and the fitted probability distribution deviates to the right, which is consistent with the probability distribution characteristics of the intelligent evaluation results.

关 键 词： 平面度误差 粒子群优化 概率分布拟合 分布

领　　域： [自动化与计算机技术] [自动化与计算机技术]