加工中心早期故障的两重威布尔模型研究

Analysis of Machining Center by Models Involving two Weibull Distributions

为了对加工中心早期故障进行有效的预测和控制,采集了26台加工中心从试运行开始的102组故障数据,并以此为基础,列出并比较了3种较常用的两重威布尔分布函数模型,从中选取分段威布尔函数作为拟合模型。为提高分段威布尔模型特征参数估计精度,提出迭代寻址求最优解法,以拟合精度为目标优化量,有效避免了时间分界点选取错误,提高了拟合精度,计算得出早期故障期大约持续4个月,与厂家反馈一致,并对分析结果进行了拟合优度检验和Laplace检验,检验结果符合要求。

In order to decrease and predict the damage of early failures,the failure datas of 26 machining centers were collected. Three models involving two weibull distributions were studied. Depending on comparing the plots on Weibull plotting paper（ WPP）,the sectional model is selected as the best one to model the failure data. A newmethod of calculating parameters is introduced,which is useful to avoid the error of demarcation time selection and improve the accuracy of fitting. It can be calculated that earlier failure period is about 4 months,which is coincident with the consumer feedback. D test method and Laplace test method was used to examine the model,and the results validated the sectional model.