Search

CN-121502267-B - Intelligent diagnosis and trend prediction system for tearing evolution state of conveying belt

CN121502267BCN 121502267 BCN121502267 BCN 121502267BCN-121502267-B

Abstract

The invention relates to the technical field of intelligent monitoring and fault diagnosis of industrial conveying equipment, in particular to an intelligent diagnosis and trend prediction system for a tearing evolution state of a conveying belt, which comprises a data acquisition center, a physical prediction unit, a nerve observation unit, a fusion correction unit and a trend diagnosis unit, wherein the data acquisition center is used for obtaining real-time stress data, the physical prediction unit is used for carrying out dynamic evolution analysis on a system state at the last moment to obtain priori state estimation, the nerve observation unit is used for obtaining an observation state, the fusion correction unit is used for obtaining optimal posterior state estimation, the trend diagnosis unit is used for carrying out fracture risk assessment feedback analysis on crack parameters in the optimal posterior state estimation to obtain a safety margin coefficient, and carrying out discrimination processing on the safety margin coefficient to obtain a maintenance early warning signal, an emergency stop signal or a normal operation signal.

Inventors

  • FANG CHEN
  • MA XIAONA
  • CHEN ZIWEI

Assignees

  • 北京易玖智能科技有限公司

Dates

Publication Date
20260512
Application Date
20260114

Claims (4)

  1. 1. An intelligent diagnosis and trend prediction system for tearing evolution state of a conveying belt is characterized by comprising: the data acquisition center is used for calling basic operation data of the conveying belt and sending the basic operation data to the physical prediction unit for stress field calculation and analysis to obtain real-time stress data; The physical prediction unit is used for carrying out dynamic evolution analysis on the system state at the previous moment to obtain prior state estimation, wherein the prior state estimation comprises crack evolution parameters deduced based on a physical model; The nerve observation unit is used for carrying out nonlinear mapping analysis on the acquired belt surface image and acoustic emission characteristics to obtain an observation state, wherein the observation state comprises crack characterization parameters based on deep learning identification; The fusion correction unit is used for carrying out optimal estimation weighted analysis on the prior state estimation and the observed state to obtain an optimal posterior state estimation, wherein the optimal posterior state estimation is a system state obtained by carrying out closed-loop fusion on data and a physical model; the trend diagnosis unit is used for carrying out fracture risk assessment feedback analysis on the crack parameters in the optimal posterior state estimation to obtain a safety margin coefficient, and carrying out discrimination processing on the safety margin coefficient to obtain a maintenance early warning signal, an emergency stop signal or a normal operation signal; The dynamic evolution analysis process is as follows: Acquiring a system state at the last moment, wherein the system state comprises equivalent crack length, crack expansion rate and energy release rate; based on a preset sampling time interval, calculating a linear accumulated value of the equivalent crack length and the crack growth rate, and setting the linear accumulated value as the prior crack length; setting the crack growth rate as a priori crack growth rate; Acquiring real-time stress data and elastic modulus of the conveyor belt covering glue; Based on the real-time stress data, the priori crack length and the elastic modulus, carrying out fracture mechanical energy calculation to obtain the priori energy release rate; combining the prior crack length, the prior crack propagation rate and the prior energy release rate to generate prior state estimation; The optimal estimation weighted analysis process is as follows: Acquiring a priori error covariance matrix at the current moment and a current observation noise covariance matrix; calculating a Kalman gain based on the prediction error covariance matrix and the observed noise covariance matrix; calculating a difference value between the observed state and the prior state estimation, and setting the difference value as an innovation residual error; calculating the product of the Kalman gain and the innovation residual error to obtain a correction term; Adding the prior state estimation and the correction term to generate an optimal posterior state estimation; The system also comprises an energy mutation monitoring unit, and the analysis process of the energy mutation monitoring unit is as follows: extracting an observed energy release rate component in an observed state; extracting a priori energy release rate component in a priori state estimation; calculating the absolute value of the difference between the observed energy release rate component and the prior energy release rate component, and setting the absolute value as an energy mutation factor; comparing the energy mutation factor with a preset statistical threshold value for analysis; If the energy mutation factor is larger than a preset statistical threshold value, generating a tearing acceleration abnormal signal; if the energy mutation factor is smaller than or equal to a preset statistical threshold value, generating a steady operation signal; The fracture risk assessment feedback analysis process is as follows: extracting a corrected crack length in the optimal posterior state estimation; acquiring real-time stress data; calculating a real-time stress intensity factor based on the real-time stress data and the corrected crack length; Obtaining a fracture toughness threshold value of a material; calculating the ratio of the real-time stress intensity factor to the fracture toughness threshold value subtracted from the value 1 to obtain a calculation result; and setting the calculation result as a safety margin coefficient.
  2. 2. The intelligent diagnosis and trend prediction system for tearing evolution state of conveyor belt according to claim 1, wherein the stress field calculation and analysis process is as follows: collecting real-time longitudinal tension of a conveying belt and effective cross section area of the belt; Dividing the real-time longitudinal tension by the effective cross-sectional area of the belt to obtain real-time stress data; The real-time stress data characterizes the average tensile stress within the belt at the current time.
  3. 3. The intelligent diagnosis and trend prediction system for tear evolution state of conveyor belt according to claim 1, wherein the nonlinear mapping analysis process is as follows: acquiring a belt surface image matrix and an acoustic emission frequency domain feature vector acquired by a data acquisition center; Calling a deep neural network model, inputting the belt surface image matrix and the acoustic emission frequency domain feature vector into the deep neural network model for feature extraction and regression calculation to obtain an output vector; Extracting components corresponding to crack length, crack expansion rate and energy release rate from the output vector; The extracted components are combined to generate an observation state, which includes observation noise.
  4. 4. The intelligent diagnosis and trend prediction system for tearing evolution state of conveyor belt according to claim 3, wherein the discriminating process is as follows: comparing and analyzing the safety margin coefficient with a preset alarm threshold value and a preset safety threshold value; If the safety margin coefficient is smaller than or equal to a preset alarm threshold value, generating an emergency stop signal, wherein the emergency stop signal is used for triggering the braking operation of the conveying system; if the safety margin coefficient is larger than a preset alarm threshold value and smaller than the preset safety threshold value, generating a maintenance early warning signal; And if the safety margin coefficient is greater than or equal to a preset safety threshold value, generating a normal operation signal.

Description

Intelligent diagnosis and trend prediction system for tearing evolution state of conveying belt Technical Field The invention relates to the technical field of intelligent monitoring and fault diagnosis of industrial conveying equipment, in particular to an intelligent diagnosis and trend prediction system for a tearing evolution state of a conveying belt. Background In the field of heavy-duty industrial conveying, a conveying belt is used as key bearing equipment, the structural integrity of the conveying belt directly determines the continuity and safety of a production line, and in the long-term high-load operation process, the belt is influenced by alternating stress and material impact, and is extremely easy to crack and evolve into a longitudinal tearing fault; In the prior art, although detection schemes based on physical models or machine vision exist, the methods have limitations generally, a pure physical model is difficult to accurately adapt to complex and changeable working condition parameters, and a single data driving model is insufficient in robustness when facing nonlinear noise interference, meanwhile, the prior art is difficult to realize deep fusion of physical priori knowledge and multisource observation data, so that the estimation precision of key evolution parameters such as crack expansion rate, energy release rate and the like is not high, and a quantitative evaluation and grading early warning mechanism of a system fracture risk trend is lacked, therefore, how to provide a system capable of fusing physical dynamics evolution and nerve observation characteristics and realizing accurate estimation and trend quantitative diagnosis of a tearing evolution state of a conveying belt is a problem which needs to be solved by a person in the art. Disclosure of Invention In order to solve the technical problems, the invention provides an intelligent diagnosis and trend prediction system for a tearing evolution state of a conveying belt, and specifically, the technical scheme of the invention comprises the following steps: the data acquisition center is used for calling basic operation data of the conveying belt and sending the basic operation data to the physical prediction unit for stress field calculation and analysis to obtain real-time stress data; The physical prediction unit is used for carrying out dynamic evolution analysis on the system state at the previous moment to obtain prior state estimation, wherein the prior state estimation comprises crack evolution parameters deduced based on a physical model; The nerve observation unit is used for carrying out nonlinear mapping analysis on the acquired belt surface image and acoustic emission characteristics to obtain an observation state, wherein the observation state comprises crack characterization parameters based on deep learning identification; The fusion correction unit is used for carrying out optimal estimation weighted analysis on the prior state estimation and the observed state to obtain an optimal posterior state estimation, wherein the optimal posterior state estimation is a system state obtained by carrying out closed-loop fusion on data and a physical model; and the trend diagnosis unit is used for carrying out fracture risk assessment feedback analysis on the crack parameters in the optimal posterior state estimation to obtain a safety margin coefficient, and carrying out discrimination processing on the safety margin coefficient to obtain a maintenance early warning signal, an emergency stop signal or a normal operation signal. Preferably, the stress field calculation and analysis process is as follows: collecting real-time longitudinal tension of a conveying belt and effective cross section area of the belt; Dividing the real-time longitudinal tension by the effective cross-sectional area of the belt to obtain real-time stress data; The real-time stress data characterizes the average tensile stress within the belt at the current time. Preferably, the kinetic evolution analysis process is as follows: Acquiring a system state at the last moment, wherein the system state comprises equivalent crack length, crack expansion rate and energy release rate; based on a preset sampling time interval, calculating a linear accumulated value of the equivalent crack length and the crack growth rate, and setting the linear accumulated value as the prior crack length; setting the crack growth rate as a priori crack growth rate; Acquiring real-time stress data and elastic modulus of the conveyor belt covering glue; Based on the real-time stress data, the priori crack length and the elastic modulus, carrying out fracture mechanical energy calculation to obtain the priori energy release rate; And combining the prior crack length, the prior crack propagation rate and the prior energy release rate to generate prior state estimation. Preferably, the nonlinear mapping analysis process is as follows: acquiring a belt surface image matrix and an