基于变分时频分解与随机森林的柴油机气门故障诊断方法

A diesel engine valve fault diagnosis method based on variational time-frequency decomposition and a random forest algorithm

柴油机气门间隙异常作为常见的柴油机故障,对其进行状态监测和诊断对柴油机的高效正常运行具有重要意义。针对柴油机气门间隙异常典型故障,提出一种基于变分时频分解与随机森林算法的气门故障诊断方法。首先,由于不同气门故障所引起的冲击信号存在冲击差异,利用变分时频分解方法可以自适应提取冲击信号中的单频率单冲击成分的特点,实现振动信号的分解并进行特征值提取;然后,利用随机森林算法对特征值进行训练;最后,在柴油机试验台上通过调整气门间隙开展模拟试验。实际试验结果表明,所提方法在气门间隙异常故障诊断中的整体准确率为96.42%,验证了所提方法的有效性。

Diesel engine valve clearance anomalies are commonly observed faults, and have a significant effect on the efficient and proper operation of diesel engines. In this work, we propose a valve fault diagnosis method based on variational time-frequency decomposition and a random forest algorithm, targeting a typical fault of abnormal valve clearance in diesel engines. Firstly, due to the differences in impact signals caused by different valve faults, the variational time-frequency decomposition method is used to adaptively extract the single-frequency single-impact components in the impact signal. This allows for the decomposition of the vibration signal and the extraction of characteristic values. The feature values are then trained using the random forest algorithm. Finally, simulation experiments are conducted on a diesel engine test bench by adjusting the valve clearance. The test results show that the overall accuracy of the proposed method in valve lash abnormality fault diagnosis is 96.42%, which verifies the effectiveness of the proposed method.