基于故障关联度的数控系统可靠性建模

Reliability Modeling of CNC System Based on Failure Correlation Degree

为了研究高可靠长寿命高档数控系统可靠性,提高数控系统可靠性模型的拟合效果,提出一种基于故障关联度的数控系统可靠性建模方法。首先,基于关联规则,利用FP-Growth算法挖掘频繁项集,通过对故障数据的挖掘建立一个既满足最小支持度又满足最小置信度的关联规则;其次,提出基于故障关联度的数控系统建模方法并给出故障关联因子的计算公式,通过故障关联因子对Weibull分布模型进行优化,建立改进模型,并与Weibull分布模型进行了对比。结果表明,改进模型的均方误差远小于Weibull分布模型,且在超过400 h之后改进模型与故障数据的拟合效果优于Weibull分布模型,该模型完善了数控系统可靠性评估方法,提高了数控系统的故障预测能力。

In order to study the reliability of computer numerical control(CNC) system with high reliability and long life, and improve the fitting effect of CNC system reliability model, this study proposed a reliability modeling method of CNC system based on failure correlation degree.Firstly, based on association rules, this method uses FP-Growth algorithm to mine frequent itemsets, and an association rule satisfying both Min-support and Min-confidence is obtained by mining fault data.Then the modeling method of CNC system based on failure correlation degree is proposed and the calculation formula of failure correlation factor is given, the Weibull distribution model is optimized through failure correlation factor, and the improved model is established and compared with the Weibull distribution model.The results show that the mean-square error(MSE) of the improved model is much smaller than that of the Weibull distribution model, and the fitting effect between the improved model and the fault data is better than that of the Weibull distribution model after more than 400 h.The model improves the reliability evaluation method of CNC system and improves the fault prediction ability of the CNC system.