CN-121980952-A - Method and system for predicting residual life of equipment driven by dynamic working conditions

CN121980952ACN 121980952 ACN121980952 ACN 121980952ACN-121980952-A

Abstract

The invention discloses a method and a system for predicting the residual life of equipment driven by dynamic working conditions, which relate to the technical field of equipment operation and maintenance and life prediction, and are used for acquiring multi-source data and extracting features, quantifying coupling mechanisms of environment data for completing feature extraction, constructing a test scheme to fit a secondary response surface model, calculating coupling damage factors, constructing PINN networks containing mechanism constraints and coupling damage factors, establishing a trend evolution model by combining a neural network, acquiring a damage evolution trend, and calculating the residual life based on the damage evolution trend and a failure threshold. According to the invention, the working condition characteristics and the test design are directly bound, so that the coupling damage factor is calculated based on-site actual measurement data, and is cooperated with the multidimensional characteristics, the interpretability of the physical mechanism and the data driving precision are considered, and the accurate prediction of the service life of the equipment is realized.

Inventors

LI XIANRUI
LIU JIAN
LIU LEILEI
XU BIN
ZHAO HAOXU
Zhou Sanbo
LI BINGSHUAI
WANG ZHE
ZHANG LEI
LI ZHANGYUN

Assignees

交通运输部天津水运工程科学研究所

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (9)

1. The method for predicting the residual life of the equipment driven by the dynamic working condition is characterized by comprising the following steps of: s1, collecting environment data and working condition data of equipment, acquiring multi-source data, preprocessing the multi-source data, and extracting characteristics of the preprocessed multi-source data; S2, carrying out coupling mechanism quantification on the environmental data with the characteristics extracted, constructing a test scheme fitting secondary response surface model, and calculating a coupling damage factor; S3, constructing PINN networks containing mechanism constraints and the coupling damage factors, combining the neural networks to establish a trend evolution model, acquiring a damage evolution trend, and calculating the residual life based on the damage evolution trend and a failure threshold.
2. The method for predicting the residual life of equipment driven by dynamic working conditions according to claim 1, wherein the multi-source data comprises environment data and working condition data, the environment data is salt spray concentration data, temperature data and load data, and the working condition data is vibration acceleration data.
3. The method for predicting the residual life of equipment driven by a dynamic working condition according to claim 2 is characterized in that the feature extraction of the preprocessed multi-source data specifically comprises the steps of extracting time domain features and frequency domain features of vibration acceleration data and extracting salt spray concentration mean value, temperature change rate and load fluctuation coefficient.
4. The method for predicting the residual life of a device driven by a dynamic working condition according to claim 3, wherein the extracting the time domain feature and the frequency domain feature of the vibration acceleration data specifically comprises: Time domain characteristics of vibration signal peaks Kurtosis degree ; Frequency domain characteristics that the vibration signal is subjected to Fourier transform to obtain power spectral density The formula is: ; Wherein, the For vibration acceleration signals, t is a time variable, N is the number of signal sampling points, i is the serial number of the sampling points, For the sampling frequency of the vibration signal, As a function of the frequency variation, In imaginary units.
5. The method for predicting the residual life of equipment driven by dynamic working conditions according to claim 3, wherein the method for extracting the salt spray concentration mean value, the temperature change rate and the load fluctuation coefficient comprises the following steps of Rate of change of temperature Coefficient of load fluctuation , wherein, As the average value of the load, The original salt fog concentration measured value at the ith sampling point t, For the temperature measurement at the i-th sampling point t, Is the original measurement value of the load at the time of the ith sampling point t.
6. The method for predicting the residual life of equipment driven by a dynamic working condition according to claim 2, wherein the construction test scheme for fitting the secondary response surface model specifically comprises the steps of constructing a three-factor three-level test scheme by adopting a Box Behnke design, and quantifying a coupling relation by adopting a secondary response surface function, wherein the formula is as follows: ; Wherein, the Solving by least square method for response surface coefficient , The matrix is designed for the test to be performed, For the actual measurement of the vector of the damage, Is the average value of the salt fog concentration, In order to provide a rate of change of temperature, Is the load fluctuation coefficient; calculating coupling damage factor based on response surface model The formula is: ; Wherein, the For the reference damage amount under the no coupling effect, For characterizing the relative degree of damage under the coupling effect.
7. The method for predicting remaining life of a dynamic condition driven device according to claim 1, wherein said PINN network comprises extracting S1-extracted feature vectors , Coupling impairment factor for dominant frequency and S2 Splicing to form model input vector Obtained by normalization processing ; Based on Miner linear accumulation damage theory, the coupling damage factor is fused The formula is: ; Wherein, the Is a total injury including fatigue injury, corrosion injury, coupling injury; K, m is a material constant; In the event of a stress being applied to the substrate, Q is corrosion activation energy, R is gas constant, Is the average value of the salt fog concentration, Is the temperature; PINN network input layer is The hidden layer has 3 layers, 64 neurons in each layer, the activation function is ReLU, and the output layer is total injury The PINN loss function is a weighted sum of data loss and physical loss, where the physical loss includes a coupling impairment accumulation constraint.
8. The method for predicting the residual life of a dynamic condition driven device according to claim 1, wherein calculating the residual life based on the damage evolution trend and the failure threshold value specifically comprises outputting based on a trend evolution model Solving the residual life L by adopting a linear interpolation method, wherein the formula is as follows: ; Wherein, the For the current moment of time, Satisfy the following requirements , For the total damage failure threshold, the following interpolation is used: ; Wherein, the To meet the requirements of Is used for the time of maximum (a) and (b), For immediately subsequent moments.
9. A dynamic condition driven equipment remaining life prediction system, applying the dynamic condition driven equipment remaining life prediction method of any one of claims 1-8, comprising: the device comprises a characteristic extraction unit, a characteristic extraction unit and a control unit, wherein the characteristic extraction unit is used for acquiring environmental data and working condition data of the device, acquiring multi-source data, preprocessing the multi-source data and extracting characteristics of the preprocessed multi-source data; The damage calculation unit is used for quantifying a coupling mechanism on the environmental data with the characteristics extracted, constructing a test scheme fitting secondary response surface model and calculating a coupling damage factor; and the life calculation unit is used for constructing PINN networks containing mechanism constraints and the coupling damage factors, establishing a trend evolution model by combining a neural network to obtain a damage evolution trend, and calculating the residual life based on the damage evolution trend and a failure threshold.

Description

Method and system for predicting residual life of equipment driven by dynamic working conditions Technical Field The invention relates to the technical field of equipment operation and maintenance and life prediction, in particular to a method and a system for predicting the residual life of equipment driven by dynamic working conditions. Background The equipment residual life prediction is a core support technology of an industrial intelligent operation and maintenance system, is directly related to operation safety of key equipment, operation and maintenance cost control and production efficiency improvement, and has an irreplaceable function in the complex working condition service fields of port machinery, engineering machinery, energy equipment and the like. The existing prediction method focuses on single working condition or single damage factor, ignores the coupling effect of salt spray corrosion, day and night temperature change, dynamic load and other factors in actual service, and causes poor adaptability of a prediction model and the actual working condition. Therefore, how to realize accurate life prediction of a device is a problem to be solved by those skilled in the art. Disclosure of Invention In view of the above, the present invention provides a method and a system for predicting the residual life of a device driven by dynamic working conditions, which are used for solving the problems in the background art. In order to achieve the above purpose, the present invention adopts the following technical scheme: a method for predicting the residual life of equipment driven by dynamic working conditions comprises the following steps: s1, collecting environment data and working condition data of equipment, acquiring multi-source data, preprocessing the multi-source data, and extracting characteristics of the preprocessed multi-source data; S2, carrying out coupling mechanism quantification on the environmental data with the characteristics extracted, constructing a test scheme fitting secondary response surface model, and calculating a coupling damage factor; S3, constructing PINN networks containing mechanism constraints and the coupling damage factors, combining the neural networks to establish a trend evolution model, acquiring a damage evolution trend, and calculating the residual life based on the damage evolution trend and a failure threshold. Preferably, the multi-source data comprises environment data and working condition data, wherein the environment data is salt fog concentration data, temperature data and load data, and the working condition data is vibration acceleration data. Preferably, the feature extraction of the preprocessed multi-source data specifically comprises the steps of extracting time domain features and frequency domain features of vibration acceleration data, and extracting salt spray concentration mean value, temperature change rate and load fluctuation coefficient. Preferably, the extracting the time domain feature and the frequency domain feature of the vibration acceleration data specifically includes: Time domain characteristics of vibration signal peaks Kurtosis degree; Frequency domain characteristics that the vibration signal is subjected to Fourier transform to obtain power spectral densityThe formula is: ; Wherein, the For vibration acceleration signals, t is a time variable, N is the number of signal sampling points, i is the serial number of the sampling points,For the sampling frequency of the vibration signal,As a function of the frequency variation,In imaginary units. Preferably, the extracting salt spray concentration mean value, the temperature change rate and the load fluctuation coefficient specifically comprise the salt spray concentration mean valueRate of change of temperatureCoefficient of load fluctuation, wherein,As the average value of the load,The original salt fog concentration measured value at the ith sampling point t,For the temperature measurement at the i-th sampling point t,Is the original measurement value of the load at the time of the ith sampling point t. Preferably, the construction of the test scheme fitting secondary response surface model specifically comprises the steps of constructing a three-factor three-level test scheme by adopting a Box Behnke design, and quantifying a coupling relation by adopting a secondary response surface function, wherein the formula is as follows: ; Wherein, the Solving by least square method for response surface coefficient,The matrix is designed for the test to be performed,For the actual measurement of the vector of the damage,Is the average value of the salt fog concentration,In order to provide a rate of change of temperature,Is the load fluctuation coefficient; calculating coupling damage factor based on response surface model The formula is: ; Wherein, the For the reference damage amount under the no coupling effect,For characterizing the relative degree of damage under the coupling eff