洒布车液压泵典型故障检测的实验研究

摘要

In order to ensure the property and reliability of asphalt distributor,an experimental method about the change of squared prediction error based on Principal Component Analysis (PCA) of Q statistics,which is based on analysis of vibration signal of center spring fault of axial piston pump that is the core component of hydraulic driving system in half intelligent asphalt distributor,is presented to deal with fault detection in this paper.In this method,the feature frequency of fault from envelope spectrum diagram by filtering de-noising and envelope demodulation will be firstly analyzed,and the feature vector samples could be obtained based on wavelet packet energy method; after that,the fault detection could be processed by using normal work sample to establish principal component model; eventually,three different fault types of center spring could be obtained and studied through the experimental data to verify the proposed method.%为保证洒布车的洒布性能及工作的可靠性,以半智能型沥青洒布车液压驱动系统的核心元件——轴向柱塞泵为实验对象,对其中心弹簧失效故障的振动信号进行了分析,并提出了利用主元分析Q统计中平方预报误差的变化对该故障进行检测的实验方法.该方法首先通过滤波消噪和包络解调的信号处理方法,从包络谱图中分析出故障的特征频率,再通过小波包能量法进行特征提取,得到特征向量样本,然后利用泵正常工作样本建立的主元模型进行故障检测,最后根据中心弹簧失效三种不同失效程度的故障检测实验来验证该方法的有效性.