Fault diagnosis method for valve mechanism of diesel engine

Abstract

The invention discloses a fault diagnosis method for a valve mechanism of a diesel engine. The method comprises the following steps of: decomposing a vibration signal acquired from a cylinder cover of the diesel engine by using redundant second generation wavelet packet transform to extract a time domain statistical characteristic of the decomposed signal to constitute a fault characteristic vector; and constructing a multi-fault multi-class support vector machine model based on two classes of multi-class support vectors to realize classification identification of a fault state of the valve mechanism of the diesel engine. By using the method, the validity of extracted characteristic parameters can be improved, and the generalization capacity during fault classification can also be improved; and a practical new technology is provided or fault diagnosis of the valve mechanism of the diesel engine.

Description

technical field [0001] The invention relates to a power equipment fault diagnosis technology, in particular to a diesel engine valve mechanism fault diagnosis method. Background technique [0002] Diesel engine is a kind of common power equipment, which plays an important role in the national economy and military field. Once it breaks down, it will bring huge economic losses. Valve mechanism failure is the most common failure of diesel engine, which will reduce the output power of diesel engine, accelerate the wear of parts, and generate high noise at the same time. Therefore, effective fault diagnosis of diesel valve train has always been a research hotspot. [0003] The effective extraction of fault features is a key issue in the fault diagnosis of diesel engine valve trains. At present, the commonly used method is to use the first-generation wavelet or wavelet packet transform to decompose the vibration signal collected from the cylinder head of the diesel engine, and t...

AI Technical Summary

Problems solved by technology

However, the artificial neural network method is a learning method based on empirical risk minimization, which tends to fall into local minimum during the learning process, and over-learning phenomenon will occur. Therefore, when dealing with small-sample pattern recognition problems such as mechanical fault diagnosis The generalization ability is poor, which affects the accuracy of fault diagnosis

Images