Method and device for intelligent diagnosis of important service water pump

Abstract

The invention relates to an important service water pump intelligent diagnosis method and device, and the method comprises the steps: S1, collecting vibration signals of an important service water pump under different working conditions, and obtaining sample points of the vibration signals containing different fault states; s2, carrying out wavelet threshold denoising preprocessing on the sample points; s3, establishing a fault game model to provide an interactive environment of observation, action and award obtaining for a fault diagnosis agent; s4, dimensionality reduction and feature extraction are carried out on the multiple hidden layers through a stacked self-encoding neural network, initial parameters are optimized through a BP neural network, and a deep neural network model with feature extraction and mode recognition functions is obtained; and S5, outputting a diagnosis result. According to vibration signals in different states, the deep neural network model can be effectively established by designing a fault diagnosis game environment and combining reinforcement learning and deep learning, intelligent diagnosis is realized, and a good diagnosis effect is achieved.

Description

Technical field [0001] The present invention relates to mechanical fault diagnosis and computer artificial intelligence field, specifically to an important plant pump intelligent diagnosis method and apparatus. Background [0002] The Important Plant Water Pump is an important nuclear safety tertiary equipment for nuclear power plants, which is a device that serves the heat exchanger of the cooling water system of the nuclear power plant equipment. The seawater transported by the pumps of important plants transfers the heat from the heat exchanger to the natural world (the sea) to ensure the safe and reliable operation of various equipment in nuclear power plants. The failure of the pump will lead to a decrease in performance, which may cause huge economic losses or even catastrophic accidents. [0003] In the prior art, the commonly used rotating machinery fault diagnosis method preprocesses the data by acquiring signal data and performing empirical modal decomposition, local me...

AI Technical Summary

Problems solved by technology

However, there are still deficiencies in deep learning models: (1) It is impossible to establish a direct linear or nonlinear mapping between raw data and corresponding failure modes, and the performance of these fault diagnosis methods depends on the quality of building and training deep models

(2) The training mechanism of this method is mainly based on supervised learning or semi-supervised learning, which means that the diagnostic algorithm requires an expert system in order to specifically learn different failure modes

Images