Pump station unit fault diagnosis method based on noise signal A-weighted analysis

Abstract

The invention discloses a pump station unit fault diagnosis method based on noise signal A-weighted analysis, and relates to the technical field of pump station unit fault diagnosis. The method comprises the following steps: monitoring a noise sound pressure signal of a pump station unit, performing spectral analysis on the sound pressure signal, performing A-weighted network correction on a sound pressure level of each frequency component, performing energy superposition on the sound pressure levels to obtain an A-weighted noise sound pressure level, and when the sound pressure level of the A-weighted noise exceeds an alarm value, performing 1 / 3 octave spectrum analysis on the sound pressure signal to extract energy characteristics of each octave frequency band. A deep extreme learning machine is utilized to quickly and effectively perform characteristic learning and extract implicit fault information of each characteristic, so that intelligent diagnosis of faults of the water pump unit is realized.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis of pumping station units. By monitoring the noise sound pressure signal of the pumping station unit, the frequency spectrum of the sound pressure signal is analyzed, and the sound pressure level of each frequency component is corrected by an A-weighting network. The energy superposition of the A-weighted noise sound pressure level is obtained. When the A-weighted noise sound pressure level exceeds the alarm value, a 1 / 3 octave spectrum analysis is performed on the sound pressure signal to extract the energy characteristics of each octave band, and deep limit learning is used. machine, quickly and effectively perform feature learning, and extract the hidden fault information of each feature, so as to realize the intelligent diagnosis of pump unit faults. Background technique [0002] The pumping station unit is the main component of the pumping station. It is widely used in urban drainage, ...

AI Technical Summary

Problems solved by technology

Although there have been a lot of theoretical research in this area, due to the complexity of the pumping station unit itself and the mutual interference between different sound sources, there are still many problems to be solved in specific applications. Among them, the effective processing of the collected noise signals and feature extraction is the key

[0004] Due to the influence of hydraulic, mechanical, electromagnetic and other factors on the pumping station unit, a single fault symptom cannot accurately represent the fault characteristics. In fact, the diagnosis result often depends on the quality of the extracted features. If the extracted features are not good, the whole It is difficult for the model to achieve a good classification effect

Secondly, traditional fault diagnosis often uses shallow classification methods such as support vector machines and decision trees, which are brought into the classification algorithm of machine learning. This shallow learning method often cannot automatically identify the weight of each feature, resulting in poor diagnostic accuracy. high

Images