A method for diagnosing faults of an engine of equipment based on RFE-ICCO-BP

Abstract

The application discloses an equipment engine fault diagnosis method based on RFE-ICCO-BP, and belongs to the technical field of engine fault diagnosis. The method comprises the following steps: acquiring original data of equipment engine oil monitoring, pre-processing the original data, screening features of the pre-processed data by using a recursive feature elimination method, extracting key sensitive features, optimizing a BP neural network based on an improved cuckoo catfish optimization algorithm, constructing an equipment engine fault diagnosis model, training the equipment engine fault diagnosis model based on a preset data set, and diagnosing the engine state based on the trained equipment engine fault diagnosis model. The application reduces data redundancy through RFE feature screening, optimizes BP network parameters in combination with a multi-strategy improved ICCO algorithm, significantly improves the precision and stability of fault diagnosis, is higher in accuracy, faster in convergence, stronger in generalization, and meets the engineering diagnosis demand under complex working conditions.

Images