Diesel engine fault diagnosis method

Abstract

The invention provides a technical scheme of a diesel engine fault diagnosis method. The method comprises the steps of collecting a cylinder cover vibration signal when an engine operates; carrying out adaptive multi-scale morphological gradient filtering processing on the collected vibration signals, extracting fault information from the signals, carrying out PCA analysis on adaptive multi-scalemorphological gradient filtering processing results of each type of samples, and determining the rank r of a base vector with the requirement that the cumulative contribution rate of result feature values exceeds 99%; carrying out non-negative matrix factorization processing on a self-adaptive multi-scale morphological gradient filtering processing result, wherein the value taking method of the rank r of the non-negative matrix factorization base vector is to perform PCA (Principal Component Analysis) on the self-adaptive multi-scale morphological gradient filtering processing result of each type of samples, and determine the rank r of the base vector according to the requirement that the cumulative contribution rate of a result characteristic value exceeds 99%; dividing the non-negative matrix factorization processing result into a training set and a test set according to a K-fold cross validation method, training a deep belief network model by using the training set, inputting the test set into the trained deep belief network classification model, and then intelligently outputting a diagnosis result. By providing the self-adaptive multi-scale morphological gradient-NMF-DBN method, intelligent fault diagnosis is realized, and dependence on artificial experience is reduced.

Description

technical field [0001] The invention belongs to the technical field of mechanical engineering, and in particular relates to the research on the fault diagnosis method of mechanical equipment. Background technique [0002] Diesel engines have the characteristics of many and interrelated parts, complex motion process, harsh working environment and heavy load, so the possibility of failure is relatively high. In addition, due to these characteristics of the diesel engine, its vibration signal components are complex, the noise pollution is serious, and each component has no fault characteristic frequency, so the spectrum analysis is of little significance for fault diagnosis guidance. It is much more difficult to carry out fault diagnosis on it, and the general fault diagnosis method may not be very effective when applied to diesel engine. Therefore, it is necessary to study more suitable and effective fault diagnosis methods for the characteristics of diesel engines. [0003]...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to solve the technical problems in the diesel engine diagnosis method that rely heavily on expert experience, long calculation time, and low accuracy

Images