Mechanical fault monitoring and diagnosis system establishment method based on SDAE-RCmvMSE

Abstract

The invention discloses a mechanical fault monitoring and diagnosis system establishment method based on SDAE-RCmvMSE. The method comprises the following steps of: firstly, acquiring vibration signalsof equipment through n vibration sensors, training an SDAE model in a diagnosis model through acquired digital signals under different working conditions, and obtaining optimal parameters of the SDAEmodel; extracting an RCmvMSE value of the acquired digital signal to train an SVM classifier, and obtaining an optimal parameter of the SVM; and deploying the SDAE, the RCmvMSE and the SVM into an embedded industrial control all-in-one machine to complete fault diagnosis model deployment, and putting the fault diagnosis model into use on site. The diagnosis model established through the method ishigh in fault recognition accuracy and good in fault tolerance performance.

Description

technical field [0001] The invention belongs to the field of mechanical fault detection, and relates to a fault detection system based on deep learning, in particular to a method for establishing a mechanical fault monitoring and diagnosis system based on SDAE-RCmvMSE. Background technique [0002] Rotating machinery parts are the most widely used in mechanical equipment, and their failures are also relatively common. It is of great significance to monitor and accurately diagnose the faults of mechanical equipment to ensure the normal operation of the equipment. Due to the complex fault types of mechanical equipment in the actual environment, its vibration signals often have unsteady and nonlinear characteristics. It is difficult to accurately diagnose faults by traditional time-frequency analysis methods. [0003] Commonly used non-time-frequency analysis methods include fractal method, sample entropy, permutation entropy, multi-scale permutation entropy, compound multi-s...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problem of poor stability of mechanical equipment vibration signals and the difficulty of extracting signal features due to the large signal noise in the actual working environment, so as to more accurately extract the nonlinear characteristics of mechanical equipment in actual working conditions

Images