基于高斯混合模型的汽车车架纵梁故障监测方法研究

Study on fault monitoring method of vehicle frame longitudinal beam based on Gaussian hybrid model

针对现有汽车车架纵梁故障监测方法效果不佳的问题,提出基于高斯混合模型的汽车车架纵梁故障监测方法。首先利用传感器采集汽车车架纵梁数据,并对其进行预处理。然后基于主成分分析法完成数据降维,在T统计故障监测的基础上引入高斯混合模型设计了一种汽车车架纵梁故障监测方法,通过将T统计控制限、SPE统计量控制限与设置阈值比较,实现故障监测与识别。实验结果表明,该方法对汽车车架纵梁故障监测的准确率达到94%以上,并且不受月份条件的影响,优于对比方法,证明该方法实用性强、故障监测准确率高。

Aiming at the problem that the existing fault monitoring methods of automobile frame rail are not effective,a new method based on Gaussian mixture model for fault monitoring of automobile frame rail is proposed.It firstly uses the data collected by the sensor,and then completes the data dimensionality reduction based on the principal component analysis method.Based on the T statistical fault monitoring,it introduces the Gaussian mixture model to design a fault monitoring method for the car frame rail.By comparing the T statistical control limit and the SPE statistical control limit with the set threshold,it can realize the fault monitoring and identification.The final experimental results show that the accuracy of this method in monitoring the faults of car frame longitudinal beams reaches over 94%,and it is not affected by monthly conditions,which is superior to the comparison method.This proves that this method has strong practicality and high fault monitoring accuracy.