针对机械产品质量特性多样化和复杂性,关键质量特性识别存在主观性强、识别结果不准确等问题,提出了一种机械产品关键质量特性权重综合识别方法.通过建立机械产品质量特性层次模型,统计分析得到质量特性分层集合,采用可拓层次分析法计算主观权重,信息熵法计算客观权重,模糊Borda法计算综合主客观权重,并对结果绘制帕累托图,根据"20/80原则"识别出关键质量特性.对数控转台的齿轮转动元动作单元的实例分析,验证了该方法的有效性和合理性.

为描述数控机床运动构件的故障率随时间变化的情况,本文从元动作单元出发,建立了一个关于元动作单元的故障概率模型.首先,根据元动作单元故障发生的原因将故障分成两类,随机故障和老化故障.然后,根据这两种故障类型故障数据的不同特点,选用两个不同的概率分布函数分别进行描述,随机故障用泊松分布进行描述,老化故障用威布尔分布进行描述.接着,给出这两种故障概率模型中各参数的物理意义和估计方法.更进一步,工作负载和工作环境会分别对元动作单元的老化故障和随机故障的故障率造成影响.为比较这两者对故障率影响的大小,提出了工作负载参数R l和工作环境参数R e,并给出这两个参数的估计方法.最后,根据收集到的运动构件的故障数据作出频率直方图,同时,对故障数据进行参数估计得到概率密度函数,并将这两者画在同一幅图上,发现两者具有较好的拟合效果.表明提出的元动作单元故障概率模型适合于描述运动构件故障率随时间的变化,模型有效.

In order to improve the assembly process quality of mechanical products, a structuralized decomposition method of Function-Movement-Action (FMA) was used to obtain the meta-action (MA) of minimum assembly granularity of mechanical products. Secondly, taking the performance index of the moving layer as the object of assembly process quality analysis, the state space model of the hybrid type MA was established, and the relationship models between the quality of the serial, parallel and hybrid type assembly process and the internal influencing factors were given. Then, the Partial Least Squares Regression (PLSR) method was used to solve the problem of multicollinearity and small sample among the influencing factors of each MA, and the key risk control points in the assembly process were obtained by the importance index analysis of variable projection. Finally, the effectiveness of the proposed method is proved by taking a numerical control turntable as an example.