CN-116008404-B - Heat supply pipeline damage identification and risk assessment method and system based on laser ultrasound

Abstract

The invention discloses a heat supply pipeline damage identification and risk assessment method based on laser ultrasound, which comprises the steps of establishing a heat supply pipeline digital twin model by adopting a mechanism modeling and data identification method, detecting a heat supply pipeline defect damage part by adopting laser ultrasound based on the heat supply pipeline digital twin model simulation to obtain heat supply pipeline damage simulation data, obtaining heat supply pipeline damage actual data by adding a laser ultrasound identification device in an actual scene of a heat supply pipeline to be identified, taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, denoising the samples and adopting a machine learning algorithm to establish a heat supply pipeline damage quantitative identification model, obtaining a heat supply pipeline damage predicted value based on the heat supply pipeline damage quantitative identification model, and carrying out heat supply pipeline damage risk assessment by adopting a fuzzy cloud theory method.

Inventors

• JIN HEFENG
• XIE JINFANG
• MU PEIHONG
• ZHAO QIONG

Assignees

• 浙江英集动力科技有限公司

Dates

Publication Date

20260508

Application Date

20221215

Claims (8)

1. 1. The heat supply pipeline damage identification and risk assessment method based on laser ultrasound is characterized by comprising the following steps of: Establishing a heating pipeline digital twin model by adopting a mechanism modeling and data identification method; detecting a defect damage part of the heat supply pipeline by adopting laser ultrasonic based on the heat supply pipeline digital twin model simulation to obtain heat supply pipeline damage simulation data; Obtaining actual data of heat supply pipeline damage by adding a laser ultrasonic recognition device in an actual scene of a heat supply pipeline to be recognized; taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, carrying out sample denoising and adopting a machine learning algorithm to establish a heat supply pipeline damage quantitative identification model, wherein the heat supply pipeline damage quantitative identification model comprises the following steps: taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, and performing sample data preprocessing including missing value processing, abnormal value processing and data standardization processing; The sample denoising is carried out on the preprocessed sample by adopting a method combining empirical mode decomposition and wavelet threshold value, namely, a high-frequency IMF component, a low-frequency IMF component and a residual signal are obtained by decomposing a noisy sample signal by adopting an empirical mode decomposition method, and the high-frequency IMF component is denoising processed by adopting a wavelet threshold value denoising method, and then the high-frequency IMF component, the low-frequency IMF component and the residual signal are reconstructed to obtain a denoised sample; extracting relevant characteristic quantity of a laser ultrasonic signal by a wavelet packet decomposition method from a denoised sample, excavating wavelet packet energy characteristics of the laser ultrasonic signal with strong correlation to the damage size of a heat supply pipeline, and constructing the wavelet packet energy characteristics, time domain characteristics and frequency domain characteristics of the laser ultrasonic signal into characteristic vectors; Inputting the feature vector into an ELM model optimized by a gray wolf optimization algorithm for learning and training, and then establishing a heat supply pipeline damage quantitative recognition model; Extracting relevant characteristic quantity of a laser ultrasonic signal by adopting a wavelet packet decomposition method, excavating wavelet packet energy characteristics of the laser ultrasonic signal with strong correlation to the damage size of a heat supply pipeline, and constructing the wavelet packet energy characteristics, time domain characteristics and frequency domain characteristics of the laser ultrasonic signal into characteristic vectors, wherein the method comprises the following steps: decomposing the laser ultrasonic signals through wavelet packets, generating wavelet packet coefficients on the corresponding frequency bands by the decomposed damage signals, and obtaining the energy distribution characteristics of the laser ultrasonic signals in the frequency bands by utilizing different energy distribution of the wavelet packet decomposed signals, namely the wavelet packet energy characteristics; The method comprises the steps of taking time differences between wave crests and wave troughs of reflected echoes of different-size lesions in a laser ultrasonic signal as a lesion time domain feature, carrying out frequency spectrum transformation on a laser ultrasonic transmission wave signal to obtain frequency spectrum energy of the different-size lesions as a lesion frequency domain feature; constructing the wavelet packet energy feature, the damage time domain feature and the damage frequency domain feature into feature vectors; The heat supply pipeline damage identification and risk assessment method further comprises the steps of obtaining a laser ultrasonic pipeline damage visualization image through the laser ultrasonic identification device, denoising, enhancing, segmenting and extracting features of the image, and establishing a heat supply pipeline damage qualitative identification model after learning and training by adopting a machine learning algorithm; obtaining a heat supply pipe damage predicted value based on the heat supply pipe damage quantitative identification model, and performing heat supply pipe damage risk assessment by adopting a fuzzy cloud theory method, wherein the heat supply pipe damage risk assessment method comprises the following steps: Obtaining a heat supply pipeline damage predicted value based on the heat supply pipeline damage quantitative recognition model, combining the heat supply pipeline damage predicted value with a pre-calculated heat supply pipeline fatigue index, a pipeline geological index and a pipeline residual life index as a heat supply pipeline risk evaluation index, and drawing up a risk evaluation grading standard to construct a heat supply pipeline risk evaluation index system, wherein the heat supply pipeline risk evaluation index system comprises a 1-level health state, a 2-level better state, a 3-level general state, a 4-level worse state and a 5-level dangerous state; Subjective weight determination using fuzzy analytic hierarchy process Determining objective weight by using coefficient of variation method Then, the subjective weight and the objective weight of each evaluation index are synthesized by adopting a least square comprehensive weighting method to determine the combined weight of each evaluation index ; Based on cloud model theory, calculating cloud model parameters of each corresponding grade according to the grading standard of each evaluation index, and generating membership cloud of each evaluation index by using a forward cloud generator, wherein the cloud model parameters comprise expectations Entropy of And super entropy ; Substituting the measured data to be evaluated into an X-condition cloud generator, and solving to obtain membership degrees of each grade corresponding to each evaluation index Combining the combined weights Calculating to obtain a grade comprehensive evaluation membership vector Determining risk assessment grade S according to maximum membership principle, wherein the X-condition cloud generator is based on the fact that the quantitative value X is known, and the cloud digital characteristics are combined Calculating to obtain cloud drops ; ; Calculating fuzzy entropy E for comprehensively evaluating the risk grade of the heat supply pipeline according to a fuzzy entropy theory, analyzing the complexity of the comprehensive evaluation result of the grade, and taking the complexity as an auxiliary parameter for evaluating the grade to obtain a final risk evaluation result (S, E) of the heat supply pipeline, wherein the fuzzy entropy is used for evaluating the risk grade of the heat supply pipeline , ; As the total number of steps, The membership of the corresponding class i is assessed for the heat supply pipe risk, Is a normalized coefficient.
2. 2. The method for identifying damage and evaluating risk of heat supply pipeline according to claim 1, wherein the method for establishing a digital twin model of heat supply pipeline by adopting mechanism modeling and data identification comprises the following steps: constructing a physical entity model, a logical model and a simulation model of the heat supply pipeline, Establishing a controllable closed-loop logic model according to a logic mechanism relation of a physical entity of a heat supply pipeline, and mapping the physical model to the logic model; The simulation model is built based on the collected operation data, state data and physical attribute data of the heating pipeline, and parameters of the simulation model are optimized according to the error of the predicted value and the actual value output by the simulation model; Carrying out virtual-real fusion on the physical entity model, the logic model and the simulation model, and constructing a system-level digital twin model of the physical entity of the heat supply pipeline in a virtual space; And accessing the multi-working-condition real-time operation data of the heating pipeline into the system-level digital twin model, and carrying out self-adaptive identification correction on the simulation result of the system-level digital twin model by adopting a reverse identification method to obtain the identified and corrected heating pipeline digital twin model.
3. 3. The heat supply pipeline damage identification and risk assessment method according to claim 1 is characterized in that the laser ultrasonic identification device comprises a laser ultrasonic emission unit, a laser ultrasonic signal acquisition and receiving unit and a laser controller display unit, in the process of identifying damage to a heat supply pipeline to be detected, the laser ultrasonic emission unit and the laser ultrasonic signal acquisition and receiving unit are synchronously controlled through the laser controller display unit, triggering of a laser in the laser ultrasonic emission unit and light gathering and scanning of a laser source are controlled through an electric scanning small mirror, signals acquired through filtering, amplifying and A/D conversion in the laser ultrasonic signal acquisition and receiving unit are stored, and received ultrasonic signals are subjected to visual display of waveform data through inversion, the laser ultrasonic signal acquisition and receiving unit adopts an indirect contact type identification and detection method, laser ultrasonic identification and detection is achieved through contact of an ultrasonic probe on the surface of the heat supply pipeline to be detected, and signal filtering, amplifying and A/D conversion processing are conducted through a filter, an amplifier and an A/D converter.
4. 4. The method for identifying and evaluating heat supply pipe damage according to claim 1, wherein the noise reduction of the high-frequency IMF component by using the wavelet threshold noise reduction method comprises: Selecting proper wavelet basis function to make wavelet transformation on high-frequency IMF component containing noise, making threshold value treatment on the wavelet coefficient after transformation, the coefficient smaller than threshold value is produced by noise signal, the coefficient larger than threshold value is produced by original signal, discarding the wavelet coefficient smaller than threshold value, retaining the coefficient larger than threshold value and making signal noise reduction treatment.
5. 5. The method for identifying and evaluating heat supply pipeline damage according to claim 1, wherein the step of inputting the feature vector into the ELM model optimized by the wolf optimization algorithm for learning and training, and then establishing a heat supply pipeline damage quantitative identification model comprises the following steps: Setting GWO initial parameters of the wolf optimization algorithm, including the number of wolves, the maximum iteration number and the value range of the optimization parameters, initializing 、 、 The position and objective function value of each grade of gray wolves; Randomly generating ELM model parameters, comparing the ELM model parameters with the target function values of the gray wolves of each level by taking the mean square error as the target function value, and replacing the target function with a larger target function value; Continuously updating 、 、 The position and objective function value of each grade of the wolves, the position of the optimal wolves is the optimal value of the input layer weight and the hidden layer threshold of the ELM model; And inputting the feature vector into the ELM model with optimized parameters for training, and then establishing a heat supply pipeline damage quantitative identification model.
6. 6. The method for identifying heat supply pipeline damage and evaluating risk according to claim 1, wherein the method for denoising the image is characterized by adopting a median denoising method based on wavelet threshold variation, and comprises the steps of performing median filtering treatment on an original image, performing wavelet layered transformation to obtain a plurality of subgraphs, setting a threshold according to a generated coefficient matrix, performing median filtering treatment on each subgraph again respectively, and performing image restoration through wavelet reconstruction operation according to a newly generated information matrix to obtain a denoised image; When the image is enhanced, a Retinex enhancement algorithm based on self-adaptive weights is adopted, and the method comprises the steps of calculating the same image by adopting a plurality of Retinex enhancement algorithms, carrying out weighted summation according to different weights according to calculation results to obtain an enhanced image, wherein the enhanced image is expressed as follows: , The output of the multi-channel i is the Retinex enhancement algorithm; pixel value for original image channel i; Is a gaussian surround function; to calculate the number of times of the surround function, i.e. the number of elements involved in the weighted summation; the weight of each element; When the image is segmented, an adopted image segmentation algorithm at least comprises image segmentation based on a threshold value, image segmentation based on edges, image segmentation based on areas, segmentation based on clusters and segmentation based on a neural network; When the image is subjected to feature extraction, a gray level co-occurrence matrix analysis is adopted to calculate a feature value of the gray level co-occurrence matrix to describe the image features, wherein the feature value at least comprises angular second moment, contrast, entropy, inverse moment, variance, mean value sum, variance sum, clustering shadow and significant clustering; After learning training by adopting a machine learning algorithm, establishing a qualitative recognition model of heat supply pipeline damage, comprising the following steps: penalty function factor C and kernel width of SFLA optimized SVM model by adopting mixed frog-leaping optimization algorithm : Initializing a frog population, dividing n frogs in the population into n model groups, calculating individual fitness values and sequencing, wherein frog i (i=1, 2,) corresponds to the i model group, and the (n+1) th frog enters the frog population of the 1 model group for grouping; Updating frog-leaping step length and position of the frog in the model factor group, which are expressed as: ; ; Is the distance moved by the frog; And Frog corresponding to the position with the optimal and the worst adaptability in the current module factor group; ; When the new solution is better, the worst individual is replaced, otherwise, the frog in the best position is used Instead of Continuously iterating to determine the position of the best frog, namely penalty function factor C and kernel width of the SVM model ; And inputting the extracted image features into the optimized SVM model for training, and establishing a qualitative recognition model of the heat supply pipeline damage.
7. 7. The method for identifying and evaluating heat supply pipeline damage according to claim 1, wherein subjective weight is determined by fuzzy analytic hierarchy process Comprising: constructing a fuzzy consistent matrix of heat supply pipeline risk assessment indexes: Constructing a fuzzy complementary judgment matrix C of a heat supply pipeline risk assessment index system, and summing the fuzzy complementary judgment matrix C according to rows to obtain For a pair of Processing to obtain The method has the requirements of fuzzy consistent matrix elements, and the calculation process of fuzzy consistent processing is as follows: ; the number of items for evaluating the index; According to Forming a fuzzy consistency judgment matrix of the heat supply pipeline risk assessment index, wherein the fuzzy consistency judgment matrix is expressed as: ; Calculating subjective weight of the risk assessment index of the ith heating pipeline, wherein the subjective weight is expressed as: ; summing the fuzzy consistency judgment matrix R according to rows; consistency test is carried out, wherein the calculation of the weight characteristic matrix is expressed as follows: ; And calculating compatibility indexes between the weight characteristic matrix and the set fuzzy consistency judgment matrix, wherein the compatibility indexes are expressed as follows: ; the closer to zero, the stronger the consistency of the judgment matrix; weight vector through consistency test Subjective weight of each risk assessment index; the objective weight is determined by using a coefficient of variation method Comprising: constructing a sample matrix by n items of evaluation index sample data of m groups of evaluation objects ; Calculating the mean value of each index for the measurement value of the ith evaluation object for the jth evaluation index Standard deviation of ; Calculating a weight value of the j-th heat supply pipeline risk assessment index, wherein the weight value is expressed as follows: ; obtaining weight vectors Objective weights for the risk assessment indicators; The combination weight of each evaluation index is determined by adopting the least square comprehensive weighting method to integrate the subjective weight and the objective weight of each evaluation index Comprising: setting a combination weight The satisfied optimization objective function is: ; ; solving an objective function by using a Lagrange method, and setting Lagrange multipliers And (3) carrying out equality constraint condition processing on the objective function, and constructing a Lagrangian function, wherein the Lagrangian function is expressed as: ; Establishing respective relations to Lagrangian functions And And let the partial derivative be zero to calculate the extremum of the objective function as: ; Solving the above to obtain the combination weight of each evaluation index 。
8. 8. Heating pipeline damage identification and risk assessment system based on laser ultrasound, characterized in that the heating pipeline damage identification and risk assessment system includes: the digital twin model building unit is used for building a digital twin model of the heating pipeline by adopting a mechanism modeling and data identification method; The damage simulation data acquisition unit is used for detecting a defect damage part of the heat supply pipeline based on the heat supply pipeline digital twin model simulation by adopting laser ultrasound to acquire heat supply pipeline damage simulation data; The damage actual data acquisition unit is used for acquiring the damage actual data of the heating pipeline by adding a laser ultrasonic recognition device in the actual scene of the heating pipeline to be recognized; The damage quantitative identification unit is used for taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, carrying out sample denoising and adopting a machine learning algorithm to establish a heat supply pipeline damage quantitative identification model, and comprises the following steps: taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, and performing sample data preprocessing including missing value processing, abnormal value processing and data standardization processing; The sample denoising is carried out on the preprocessed sample by adopting a method combining empirical mode decomposition and wavelet threshold value, namely, a high-frequency IMF component, a low-frequency IMF component and a residual signal are obtained by decomposing a noisy sample signal by adopting an empirical mode decomposition method, and the high-frequency IMF component is denoising processed by adopting a wavelet threshold value denoising method, and then the high-frequency IMF component, the low-frequency IMF component and the residual signal are reconstructed to obtain a denoised sample; extracting relevant characteristic quantity of a laser ultrasonic signal by a wavelet packet decomposition method from a denoised sample, excavating wavelet packet energy characteristics of the laser ultrasonic signal with strong correlation to the damage size of a heat supply pipeline, and constructing the wavelet packet energy characteristics, time domain characteristics and frequency domain characteristics of the laser ultrasonic signal into characteristic vectors; Inputting the feature vector into an ELM model optimized by a gray wolf optimization algorithm for learning and training, and then establishing a heat supply pipeline damage quantitative recognition model; Extracting relevant characteristic quantity of a laser ultrasonic signal by adopting a wavelet packet decomposition method, excavating wavelet packet energy characteristics of the laser ultrasonic signal with strong correlation to the damage size of a heat supply pipeline, and constructing the wavelet packet energy characteristics, time domain characteristics and frequency domain characteristics of the laser ultrasonic signal into characteristic vectors, wherein the method comprises the following steps: decomposing the laser ultrasonic signals through wavelet packets, generating wavelet packet coefficients on the corresponding frequency bands by the decomposed damage signals, and obtaining the energy distribution characteristics of the laser ultrasonic signals in the frequency bands by utilizing different energy distribution of the wavelet packet decomposed signals, namely the wavelet packet energy characteristics; The method comprises the steps of taking time differences between wave crests and wave troughs of reflected echoes of different-size lesions in a laser ultrasonic signal as a lesion time domain feature, carrying out frequency spectrum transformation on a laser ultrasonic transmission wave signal to obtain frequency spectrum energy of the different-size lesions as a lesion frequency domain feature; constructing the wavelet packet energy feature, the damage time domain feature and the damage frequency domain feature into feature vectors; The heat supply pipeline damage identification and risk assessment system further comprises the steps of obtaining a laser ultrasonic pipeline damage visualization image through the laser ultrasonic identification device, denoising, enhancing, segmenting and extracting features of the image, and establishing a heat supply pipeline damage qualitative identification model after learning and training by adopting a machine learning algorithm; the damage risk assessment unit is used for obtaining a heat supply pipe damage predicted value based on the heat supply pipe damage quantitative identification model, and performing heat supply pipe damage risk assessment by adopting a fuzzy cloud theory method, and comprises the following steps: Obtaining a heat supply pipeline damage predicted value based on the heat supply pipeline damage quantitative recognition model, combining the heat supply pipeline damage predicted value with a pre-calculated heat supply pipeline fatigue index, a pipeline geological index and a pipeline residual life index as a heat supply pipeline risk evaluation index, and drawing up a risk evaluation grading standard to construct a heat supply pipeline risk evaluation index system, wherein the heat supply pipeline risk evaluation index system comprises a 1-level health state, a 2-level better state, a 3-level general state, a 4-level worse state and a 5-level dangerous state; Subjective weight determination using fuzzy analytic hierarchy process Determining objective weight by using coefficient of variation method Then, the subjective weight and the objective weight of each evaluation index are synthesized by adopting a least square comprehensive weighting method to determine the combined weight of each evaluation index ; Based on cloud model theory, calculating cloud model parameters of each corresponding grade according to the grading standard of each evaluation index, and generating membership cloud of each evaluation index by using a forward cloud generator, wherein the cloud model parameters comprise expectations Entropy of And super entropy ; Substituting the measured data to be evaluated into an X-condition cloud generator, and solving to obtain membership degrees of each grade corresponding to each evaluation index Combining the combined weights Calculating to obtain a grade comprehensive evaluation membership vector Determining risk assessment grade S according to maximum membership principle, wherein the X-condition cloud generator is based on the fact that the quantitative value X is known, and the cloud digital characteristics are combined Calculating to obtain cloud drops ; ; Calculating fuzzy entropy E for comprehensively evaluating the risk grade of the heat supply pipeline according to a fuzzy entropy theory, analyzing the complexity of the comprehensive evaluation result of the grade, and taking the complexity as an auxiliary parameter for evaluating the grade to obtain a final risk evaluation result (S, E) of the heat supply pipeline, wherein the fuzzy entropy is used for evaluating the risk grade of the heat supply pipeline , ; As the total number of steps, The membership of the corresponding class i is assessed for the heat supply pipe risk, Is a normalized coefficient.

Description

Heat supply pipeline damage identification and risk assessment method and system based on laser ultrasound Technical Field The invention belongs to the technical field of heat supply pipeline health management, and particularly relates to a heat supply pipeline damage identification and risk assessment method based on laser ultrasound. Background With the continuous development of social economy, central heating, which is one of important infrastructures in China, is also vigorously developed. In order to realize clean heat supply, the waste heat of the thermal power plant at the suburb outside the city is used for heat supply, so that the long-distance heat supply pipeline is increasingly applied. Therefore, the safe, efficient and energy-saving operation of the heating pipeline is an important problem, and the damage condition of the heating pipeline seriously affects the civil guarantee of the heating system. When the heat supply pipeline is used, the heat supply pipeline is damaged by factors such as external uneven stress, internal fluid corrosion and the like, cracks, pits and the like are generated in the heat supply pipeline, and if the heat supply pipeline is not overhauled in time for a long time, the damage state is expanded, so that the production and life safety is threatened. In addition, in order to more intuitively see the damage risk state of the heating pipeline during use, risk assessment of the heating pipeline is required. The traditional damage identification technology has lower dimensionality, can only realize the identification of the damage on the surface of the pipeline, can not identify the depth level of the damage, has the advantages of real-time and dynamic detection of the damage of the heat supply pipeline, high identification sensitivity, higher identification degree of tiny damage and the like compared with the traditional ultrasonic identification detection mode, so that the application of the laser ultrasonic technology to the field of the damage identification of the heat supply pipeline has great research significance, and the rapid and accurate quantitative analysis of the damage condition of the heat supply pipeline by using the laser ultrasonic technology and the reliability risk assessment of the damage condition of the heat supply pipeline by combining an assessment model according to the damage condition of the heat supply pipeline are the problems to be solved urgently at present. Based on the technical problems, a heating pipeline damage identification and risk assessment method based on laser ultrasound needs to be designed. Disclosure of Invention The invention aims to solve the technical problems of overcoming the defects of the prior art, and provides a heat supply pipeline damage identification and risk assessment method based on laser ultrasound, which can be used for quantitatively analyzing the damage condition of a heat supply pipeline through laser ultrasound signal characteristics and adopting a machine learning algorithm, so that the damage identification efficiency of the heat supply pipeline is effectively improved, and in addition, the risk assessment of the heat supply pipeline is carried out through adopting a fuzzy cloud theory method, the damage quantitative analysis condition of the heat supply pipeline is fully utilized, other assessment indexes are combined, subjective factors and objective factors of all indexes in an assessment system are fused, so that all assessment index weights are more in line with engineering practice, the ambiguity and randomness existing in the assessment process are considered, and the reliability of the risk assessment of the heat supply pipeline is ensured. In order to solve the technical problems, the technical scheme of the invention is as follows: the invention provides a heat supply pipeline damage identification and risk assessment method based on laser ultrasound, which comprises the following steps: Establishing a heating pipeline digital twin model by adopting a mechanism modeling and data identification method; detecting a defect damage part of the heat supply pipeline by adopting laser ultrasonic based on the heat supply pipeline digital twin model simulation to obtain heat supply pipeline damage simulation data; Obtaining actual data of heat supply pipeline damage by adding a laser ultrasonic recognition device in an actual scene of a heat supply pipeline to be recognized; Taking the heat supply pipeline damage simulation data and the heat supply pipeline damage actual data as pipeline damage identification samples, denoising the samples, and establishing a heat supply pipeline damage quantitative identification model by adopting a machine learning algorithm; and obtaining a heat supply pipe damage predicted value based on the heat supply pipe damage quantitative identification model, and performing heat supply pipe damage risk assessment by adopting a fuzzy cloud theory method. Further, the method