Shield pump fault mode identification method and system based on deep convolutional network

Abstract

The invention discloses a shield pump fault mode identification method and system based on a deep convolutional network, and the method comprises the steps: S1, sampling initial data during the operation of a shield pump, i.e., vibration sensor data of the upper and lower parts of the shield pump during the operation of the shield pump; S2, taking the initial data as an input parameter, and inputting the initial data into a shield pump deep learning model based on a one-dimensional deep convolutional network for model training; and S3, by adopting a trained model, carrying out shield pump fault mode identification on shield pump operation data acquired in real time, identifying 14 types of fault types and damage degree modes of the shield pump, and outputting the identification results fordisplay. In the constructed model, an inner product transformation step of solving features is converted into a one-dimensional convolution layer, and adaptive selection of a primary function is realized; and meanwhile, the feature screening process is realized by utilizing a complex K-MaxPooling layer. The shield pump fault mode identification method is high in shield pump fault mode recognitionprecision and efficiency.

Description

technical field [0001] The invention relates to the technical field of fault pattern recognition of shielded pumps, in particular to a method and system for recognizing fault patterns of shielded pumps based on a deep reconvolution network. Background technique [0002] With the continuous expansion of industrial scale, the number and severity of modern system equipment failures have risen to a certain extent, which puts forward urgent needs for real-time monitoring and intelligent fault diagnosis of mechanical equipment. Intelligent diagnosis of early faults is mainly through traditional feature extraction methods, such as Fourier transform, wavelet transform and other transform kernel functions Transform the original space of faults x to feature space Using the intelligent classification algorithm in the feature space to diagnose the faults of mechanical equipment, the formula is: [0003] Since the accuracy of the results of this method depends on the completeness...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is that in the prior art for shielded pump fault pattern recognition methods, the manual recognition method is highly subjective and has a large workload; the traditional deep learning shielded pump fault pattern recognition method based on parameter optimization has a large workload Or blind parameter adjustment leads to inaccurate identification of canned pump failure modes, low efficiency, and difficult work, etc.

The purpose of the present invention is to provide a shielded pump fault pattern recognition method and system based on a deep reconvolution network. The shielded pump fault pattern recognition method of the present invention adopts a fault pattern recognition method that does not need to extract features and has directional deep learning for parameter optimization. method to realize the intelligent and automatic monitoring of the failure mode of the canned pump and reduce the difficulty of model training, thereby solving the difficulty of establishing the feature space of the traditional machine learning-based fault diagnosis and the difficulty and blindness of the existing deep learning-based fault mode recognition and parameter adjustment. Canned pump failure mode identification is not accurate, the efficiency is not high, and the work is difficult, etc.

Images