基于VW-FDA的盾构机液压系统故障诊断方法研究

摘要

盾构机规模庞大,施工过程中检测变量众多,液压系统故障易发.为了实现在线监测,提高液压系统故障诊断的准确性,探究与盾构机适应的快速准确有效的故障识别方法,提出使用加权因子与传统费歇尔判别法相结合的加权费歇尔判别法(Variable-Weighted-Fisher Discriminant Analysis,VW-FDA).通过AMESim软件建立推进系统模型,仿真四类故障,运用仿真故障数据得到每一类故障的加权向量,建立VW-FDA模型.结果表明:对于盾构机液压推进系统,VW-FDA比传统FDA具有更强的故障诊断能力,能够将故障信息进行有效的分类,正确地做出故障诊断.%A shield machine has a large scale size and a complex structure,and numerous variables to be inspected,so faults of hydraulic system offen occur during an engineering construction process.In order to realize the online-monitoring of numerous variables and to improve the fault diagnosis accuracy of the hydraulic system,a fast,accurate,and effective fault diagnosis of the shield machine has a great significance.Combined the weighted factor method with the traditional Fisher discriminant method,a variable-weighted-Fisher discriminant analysis (VW-FDA) method is proposed for the fault diagnosis of the hydraulic system of shield machine.A model based on the hydraulic propulsion system is established by the software AMESim,and four types of faults of the propulsion system are simulated.With the simulation result data of fault,the weighted vector of each type of fault is obtained,and the VW-FDA model is built.The research result shows that,for the hydraulic propulsion system of shield machine,the VW-FDA has a stronger diagnosibility than the traditional FDA,and it can classify the fault information effectively and make faults diagnosis more correct.