基于正交最小二乘模型拟合的DPF故障诊断方法

根据柴油机颗粒过滤器(DPF)堵塞与泄漏故障型式开展了排气阻力试验和发动机台架试验,获取了排气流量、排气温度、排气压差等主要故障特征参数数据。通过与传统压降模型进行对比分析,将正交最小二乘方法应用于DPF的故障诊断,形成基于正交最小二乘模型拟合的DPF故障诊断方法,并用试验采集到的主要故障特征参数数据对模型拟合方法进行了可行性验证。通过定义一个错误系数向量,分析了广义故障空间中不同故障型式DPF所形成的故障区域,为DPF故障诊断技术的发展提供了理论依据。通过错误系数向量能够判断出故障程度及堵塞、泄漏等故障型式,基于该结果可以进行DPF故障诊断。

Research on DPF Fault Diagnosis Method Based on Orthogonal Least Squares Model Fitting

According to the diesel particulate filter(DPF) fault type of blockage and leakage, exhaust resistance test and engine bench test were carried out to get the major fault characteristic parameters including exhaust gas flow, exhaust temperature and exhaust pressure. By comparing with the traditional pressure drop model, an orthogonal least squares method was applied to the DPF fault diagnosis, which allowed to form a DPF fault diagnosis method based on the orthogonal least squares model fitting and to use the collected major fault characteristic parameters to verify the feasibility of the method. By defining an error coefficient vector, the different fault areas formed from different DPF fault types in the generalized failure space were analyzed to provide a theoretical basis for the development of DPF fault diagnosis technology. The degree and type of failure, such as blockage and leakage, can be judged by analyzing the error coefficient vector. On the basis of this feature, the DPF diagnostics strategy can be established.