Hydraulic support support efficiency evaluation and early warning method based on fuzzy evaluation-lstm

Abstract

The application discloses a kind of based on fuzzy evaluation-LSTM's hydraulic support supporting efficiency evaluation and early warning method, comprising: the time series data of hydraulic support is obtained by sensor;Evaluation index is filtered using Pearson correlation coefficient;The classification threshold of evaluation index is calculated;Hydraulic support supporting efficiency evaluation index system with target layer, criterion layer and index layer is constructed;Subjective weight is solved using interval analytic hierarchy process;Objective weight is solved using entropy weight method;By minimizing the deviation between subjective and objective weight, the comprehensive weight of Nash equilibrium is obtained using game theory;Fuzzy comprehensive evaluation model is constructed, membership function is dynamically adjusted according to classification threshold, and fuzzy comprehensive evaluation result is obtained according to maximum membership degree principle;LSTM model is used to predict the grade of hydraulic support supporting efficiency.The application can realize the real-time evaluation and intelligent early warning of hydraulic support supporting efficiency, and provide reference basis for the adjustment of hydraulic support supporting parameters in time.

Description

Technical Field

[0001] This invention relates to the field of intelligent early warning technology in coal mines, and in particular to a method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM. Background Technology

[0002] Hydraulic supports are in direct contact with the surrounding rock of the roadway, bearing the enormous mining pressure generated by rock movement. They are the core equipment in the coal mine roadway support system, and their support effect directly affects the stability of the surrounding rock, the safety of workers, and production efficiency. The cause of hydraulic support failure ultimately relates to insufficient support strength. Subsequent improvement measures aim to increase support strength, but in reality, many failed supports do not have low support strength. Rather, the failure stems from a somewhat one-sided analysis and evaluation of hydraulic support performance, focusing only on support strength. This leads to unnecessary and unlimited increases in hydraulic support strength, violating the design and selection principles of "scientific rationality and economic applicability" for hydraulic supports.

[0003] Traditional evaluation methods often focus on a single parameter of the hydraulic support (such as support strength) in isolation, neglecting the fact that the hydraulic support and the surrounding rock of the roadway are an interacting coupled system. This one-sided analysis leads to inaccurate judgments of support effectiveness, often simply attributing support failure to insufficient support strength, thus causing blind and excessive increases in support strength. This not only increases unnecessary costs but also violates the design principles of "scientific and reasonable, economical and applicable" hydraulic supports, failing to comprehensively reflect the true working state of the support in multiple dimensions such as strength, stiffness, and stability.

[0004] Furthermore, existing technologies lack in-depth analysis of the temporal correlation of multi-source monitoring data, making it difficult to achieve dynamic grading and trend prediction of support status. Traditional static models or simple threshold alarm methods have poor early warning timeliness and cannot provide forward-looking warnings in the early stages of stent performance deterioration. Therefore, it is difficult to provide timely and effective decision support for dynamic adjustment of support parameters and proactive avoidance of potential safety risks. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating and warning the support performance of hydraulic supports based on fuzzy evaluation-LSTM, which can realize real-time evaluation and intelligent warning of the support performance of hydraulic supports, and provide timely reference for adjusting the support parameters of hydraulic supports.

[0006] To achieve the above objectives, the present invention provides the following solution: A method for evaluating and providing early warning of hydraulic support performance based on fuzzy evaluation-LSTM includes the following steps: S1. Obtain time-series data of the hydraulic support through sensors, including the sensor acquisition date and corresponding support parameters; S2. Use Pearson correlation coefficient to perform correlation analysis on support parameters, and remove support parameters with correlation coefficients greater than the set threshold in order to determine the evaluation indicators. S3. Based on the coupling relationship between the support and the surrounding rock structure, calculate the classification thresholds of each evaluation index under the conditions of strength coupling, stiffness coupling and stability coupling respectively; S4. Construct a hydraulic support support performance evaluation index system with three levels: target layer, criterion layer, and index layer. The criterion layer includes strength coupling state, stiffness coupling state, and stability coupling state. S5. Calculate the subjective weights of each level of indicators in the hydraulic support support effectiveness evaluation index system using the interval analytic hierarchy process. S6. The objective weights of each indicator in the indicator layer of the hydraulic support support effectiveness evaluation index system are calculated using the entropy weight method. S7. Based on game theory, optimization is performed with the goal of minimizing the deviation between subjective weights and objective weights, and the comprehensive weights of the indicators under the Nash equilibrium are obtained by solving the problem. S8. Based on the hydraulic support support efficiency evaluation index system and the comprehensive weight of the index, a fuzzy comprehensive evaluation model is constructed. The membership function of each index is dynamically determined according to the classification threshold obtained in step S3. The real-time collected time series data is input into the fuzzy comprehensive evaluation model, and the support efficiency level at the current moment is obtained according to the principle of maximum membership. S9. Train an LSTM model based on the time series data with support effectiveness level labels output by S8, and then input the sensor data to be predicted into the trained LSTM model to predict the support effectiveness level of the hydraulic support at future moments.

[0007] Preferably, in S1, the sensors include a column pressure sensor, a top plate displacement monitor, and an inclination sensor; the support parameters include pressure-related parameters, displacement-related parameters, and inclination-related parameters. The pressure-related parameters include the support strength of the upper column, the support strength of the middle column, and the support strength of the lower column. The displacement-related parameters include the top plate displacement of the upper column, the top plate displacement of the middle column, and the top plate displacement of the lower column. The inclination-related parameters include the inclination angle of the top beam and the inclination angle of the base.

[0008] Preferably, in S2, Pearson correlation coefficient is used to perform correlation analysis on the support parameters, and support parameters with correlation coefficients greater than a set threshold are removed. Specifically, this includes: S2-1, Pearson correlation coefficient calculation formula:

[0009] in, Here, x1 and x2 are two distinct indicators, E(.) is the expected value, and D(.) is the variance. S2-2. A Pearson correlation coefficient greater than the set threshold indicates a correlation between evaluation indicators. Evaluation indicators with a Pearson correlation coefficient greater than the preset threshold are eliminated, and the final evaluation indicators of hydraulic support performance are determined for the subsequent construction of the hydraulic support performance evaluation indicator system.

[0010] Preferably, in S3, the classification thresholds for each evaluation index are calculated under the states of strength coupling, stiffness coupling, and stability coupling, specifically including: S3-1, Strength Coupling State: Calculation of the hydraulic support strength when the surrounding rock of the roadway is in the elastic limit state based on elasticity theory and the Mohr-Coulomb strength criterion.

[0011] Where P is the hydraulic support strength, γ is the rock mass unit weight, M is the tunnel depth, and C is the rock cohesion. The internal friction angle of the rock; S3-2, Stiffness Coupling State: Calculation of Top Plate Displacement Threshold under Support-Surrounding Rock Coupling State Based on Elastic Foundation Beam Theory: =

[0012] in, Let be the displacement of the top plate, k be the elastic coefficient of the support and the bottom plate, and L be half the length of the top plate; S3-3, Stability Coupling State: Calculation of Critical Tilting Angle of Hydraulic Support Based on Torque Limit Equilibrium Equation:

[0013] in, B represents the critical tilt angle of the support, and B represents the width of the support base. G is the pressure exerted by the top plate on the support, μ is the coefficient of friction, H is the height of the support, and h is the height of the support's center of gravity.

[0014] Preferably, in S5, the subjective weights of each level of indicators in the hydraulic support support effectiveness evaluation index system are calculated using the interval analytic hierarchy process, specifically including: S5-1. Based on the 1-9 proportional scale, the support effectiveness evaluation indicators of the same layer are compared pairwise to construct an interval decision matrix A composed of multiple interval numbers. The definition of A is:

[0015] in, and These are the definite matrices formed in A. Evaluation indicators obtained through the binary comparison method and The upper and lower limits of the intervals representing the relative importance of the indicators, where n is the number of evaluation indicators; S5-2: Calculate the correction coefficients of the interval decision matrix and perform a consistency test: ,

[0016] Where a and b are interval decision matrices respectively. and Correction factor; If a≤1 and b≥1, the interval decision matrix is ​​considered to have high consistency; otherwise, if a>1 or b<1, the interval decision matrix is ​​considered to have low consistency. In this case, the interval decision matrix A needs to be reconstructed until the obtained interval decision matrix has high consistency. S5-3. Calculate the subjective weights of each indicator:

[0017] in, To improve the subjective weights of each element in the analytic hierarchy process, and For matrix and The normalized eigenvectors corresponding to their respective largest eigenvalues.

[0018] Preferably, in S7, optimization is performed with the objective of minimizing the deviation between subjective and objective weights, and the comprehensive weight of the index under Nash equilibrium is obtained by solving the problem, specifically including: S7-1. Using game theory, minimize the range of subjective and objective weights, and establish an objective function for the coefficients of two linear combinations. , Optimize: i=1,2,…,m S7-2. According to the principle of differentiation, the condition for the first derivative of the optimization is:

[0019] S7-3. Normalize the obtained linear combination coefficients:

[0020] S7-4. Calculate the overall weight of the indicators: i=1,2,…,m in, This is the overall weight vector. With These are the normalized linear combination coefficients. The weight vector of the analytic hierarchy process (AHP). This is the weight vector for the entropy weight method.

[0021] Preferably, in step S8, the real-time collected time-series data is input into the fuzzy comprehensive evaluation model, and the support effectiveness level at the current moment is obtained according to the maximum membership principle, specifically including: S8-1. The set of comments used has 5 evaluation levels, and the rating matrix is ​​a 5-dimensional row vector G.

[0022] S8-2. Construct membership functions using a combination of triangles and semi-trapezoidal shapes, and dynamically adjust the position and shape of the membership functions for each indicator evaluation level based on the classification threshold obtained in step S3. S8-3. Substitute the time series data collected by the sensor into the corresponding membership function to obtain the corresponding fuzzy relation matrix; S8-4, Second-level fuzzy comprehensive evaluation expression:

[0023] in, This is a set of weights for three levels of indicators. This is the fuzzy judgment matrix under the i-th evaluation criterion, calculated using the membership function; First-level fuzzy comprehensive evaluation expression:

[0024] in, G for Q is the set of weights for secondary indicators. Ultimately, the decision is made based on the principle of maximum membership. G The fuzzy comprehensive evaluation results.

[0025] Preferably, in S9, the sensor data to be predicted is input into the trained LSTM model to predict the hydraulic support effectiveness level at future moments, specifically including: The time-series dataset with support effectiveness level labels obtained through fuzzy comprehensive evaluation is normalized, divided into training and test sets proportionally, and input into the LSTM model. The network parameters are set, and the LSTM model is trained using the backpropagation algorithm and parameter optimization algorithm. The LSTM model learns the long-term dependencies in the input sequence and outputs the predicted support effectiveness level, using the hydraulic support support effectiveness level as the output. Using the trained LSTM model, the data samples collected by the sensor are used as test samples to predict the support effectiveness level.

[0026] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements a method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM as described above.

[0027] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: This invention provides a method for evaluating and predicting the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM. By integrating column pressure sensors, roof displacement monitors, and tilt angle sensors, a full-parameter sensing network for the support status is constructed. This network allows for in-depth research into the interaction between the support and the surrounding rock structure. Furthermore, by combining fuzzy mathematical membership functions, a comprehensive evaluation model reflecting the support effectiveness of hydraulic supports is established, incorporating three core indicators: support strength, roof displacement, and support tilt angle. The LSTM model is used to analyze the coupling state of the support and surrounding rock structure in real time, predicting the support effectiveness level of the hydraulic supports. Compared to traditional methods, this technology significantly improves the comprehensiveness and timeliness of hydraulic support effectiveness evaluation, providing key technical support for intelligent mine construction. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart illustrating the hydraulic support performance evaluation and early warning method based on fuzzy evaluation-LSTM provided by the present invention. Figure 2 This is a schematic diagram of the hydraulic support performance evaluation index system of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1As shown, a method for evaluating and providing early warning of hydraulic support performance based on fuzzy evaluation-LSTM includes the following steps: S1. Obtain time-series data of the hydraulic support through sensors, including the sensor acquisition date and corresponding support parameters; S2. Use Pearson correlation coefficient to perform correlation analysis on support parameters, and remove support parameters with correlation coefficients greater than the set threshold in order to determine the evaluation indicators. S3. Based on the coupling relationship between the support and the surrounding rock structure, calculate the classification thresholds of each evaluation index under the conditions of strength coupling, stiffness coupling and stability coupling respectively; S4. Construct a hydraulic support support performance evaluation index system with three levels: target layer, criterion layer, and index layer. The criterion layer includes strength coupling state, stiffness coupling state, and stability coupling state. S5. Calculate the subjective weights of each level of indicators in the hydraulic support support effectiveness evaluation index system using the interval analytic hierarchy process. S6. The objective weights of each indicator in the indicator layer of the hydraulic support support effectiveness evaluation index system are calculated using the entropy weight method. S7. Based on game theory, optimization is performed with the goal of minimizing the deviation between subjective weights and objective weights, and the comprehensive weights of the indicators under the Nash equilibrium are obtained by solving the problem. S8. Based on the hydraulic support support efficiency evaluation index system and the comprehensive weight of the index, a fuzzy comprehensive evaluation model is constructed. The membership function of each index is dynamically determined according to the classification threshold obtained in step S3. The real-time collected time series data is input into the fuzzy comprehensive evaluation model, and the support efficiency level at the current moment is obtained according to the principle of maximum membership. S9. Train an LSTM model based on the time series data with support effectiveness level labels output by S8, and then input the sensor data to be predicted into the trained LSTM model to predict the support effectiveness level of the hydraulic support at future moments.

[0033] Specifically, the hydraulic support performance evaluation and early warning method based on fuzzy evaluation-LSTM described in this embodiment includes the following steps: S1. Real-time data on hydraulic support force, top plate displacement, and hydraulic support tilt angle are obtained through sensors. Specifically, in S1, the sensors include a column pressure sensor, a top plate displacement monitor, and an inclination sensor; the hydraulic support force includes the upper column support strength, the middle column support strength, and the lower column support strength; the top plate displacement includes the upper column top plate displacement, the middle column top plate displacement, and the lower column top plate displacement; and the hydraulic support inclination data includes the top beam inclination angle and the base inclination angle.

[0034] S2. Pearson correlation coefficient analysis was performed on the support parameters, and support parameters with correlation coefficients greater than a set threshold were removed to determine the evaluation indicators; specifically including: S2-1, Pearson correlation coefficient calculation formula:

[0035] In the formula, Here, x1 and x2 are two distinct indicators, E is the expected value, and D is the variance. S2-2. A Pearson correlation coefficient greater than 0.8 indicates a high correlation between the evaluation indicators. Evaluation indicators with a Pearson correlation coefficient greater than 0.8 are removed to filter the evaluation indicators based on correlation analysis. The final evaluation indicators for the hydraulic support effectiveness are then determined and used to construct the subsequent hydraulic support effectiveness evaluation indicator system. S3. Based on the coupling relationship between the support and the surrounding rock structure, calculate the classification thresholds for each evaluation index under strength coupling, stiffness coupling, and stability coupling states, respectively; specifically including: S3-1. Solve for the optimal value of hydraulic support strength under the coupled strength state of support-surrounding rock structure;

[0036] In the formula, P is the hydraulic support strength, γ is the rock mass unit weight, M is the tunnel depth, and C is the rock cohesion. The internal friction angle of the rock; S3-2, Solve for the top plate displacement threshold under the stiffness coupling state of the support-surrounding rock structure; =

[0037] In the formula, γ is the unit weight of the rock mass, H is the tunnel burial depth, k is the elastic coefficient of the support and the floor, and L is half the length of the roof. S3-3, Solve for the hydraulic support instability threshold under the coupled stability state of the support-surrounding rock structure;

[0038] In the formula, B is the critical tilt angle of the support, and B is the width of the support base. G is the pressure exerted by the top plate on the support, μ is the coefficient of friction, H is the height of the support, and h is the height of the support's center of gravity. S4, such as Figure 2As shown, a three-level fuzzy factor comprehensive evaluation index system for hydraulic support performance is constructed. The first level is the target layer hydraulic support performance evaluation index system; the second level is the criterion layer, including strength coupling state, stiffness coupling state and stability coupling state; the third level is the decision layer, including upper column pressure, middle column pressure, lower column pressure, upper column top plate displacement, middle column top plate displacement, lower column top plate displacement, top beam inclination angle and base inclination angle. S5. Use the interval analytic hierarchy process (AHP) to solve for the subjective weights of each indicator in the hydraulic support performance evaluation index system, specifically including: S5-1. Based on the 1-9 proportional scaling method, see Table 1 (the range of interval values ​​and the definition of the importance scale for indicator comparison), compare the support effectiveness evaluation indicators at the same level pairwise to establish an interval decision matrix: Table 1 ; The interval decision matrix A is constructed as follows:

[0039] In the formula: and These are the definite matrices formed in A. Evaluation indicators obtained through the binary comparison method and The upper and lower limits of the intervals representing the relative importance of the indicators, where n is the number of evaluation indicators; ,

[0040] S5-2: Calculate the correction coefficients of the interval decision matrix and perform a consistency test: ,

[0041] In the formula: a and b are the interval decision matrices respectively. and Correction factor; If a≤1 and b≥1, then the decision matrix is ​​considered to have good consistency. Conversely, if a>1 or b<1 is satisfied, the consistency of the decision matrix is ​​considered poor. In this case, the interval decision matrix A needs to be reconstructed until satisfactory consistency is obtained. S5-3. Calculate the subjective weights of each indicator: ,

[0042] In the formula: -, + represents the normalized eigenvectors corresponding to the largest eigenvalues ​​of matrices A- and A+;

[0043] In the formula: The subjective weights of each element in the interval analytic hierarchy process; S6. Solve for the entropy weight method weights of the hydraulic support support effectiveness evaluation index, specifically including: S6-1. Construct a decision matrix using the original data of j indicators from i samples:

[0044] Standardize the decision matrix:

[0045] Calculate the information entropy for each indicator:

[0046] In the formula, The information entropy of the j-th column; The entropy weight method weight matrix is ​​obtained by solving:

[0047] S7. Using game theory, the comprehensive weights of the Nash equilibrium are obtained by minimizing the deviation between subjective and objective weights; specifically including: S7-1. Using game theory, minimize the range of subjective and objective weights, and establish an objective function for the coefficients of two linear combinations. , Optimize: i=1,2,…,m S7-2. According to the principle of differentiation, the condition for the first derivative of the optimization is: S7-3. Normalize the obtained linear combination coefficients:

[0048] S7-4. Calculate the overall weight: i=1,2,…,m In the formula: This is the overall weight vector. With For linear combination coefficients, The weight vector of the analytic hierarchy process (AHP). The weight vector is the entropy weight method weight vector; S8. Establish a fuzzy comprehensive evaluation model for the support effectiveness of hydraulic supports, specifically including: Step 8.1: The set of comments used includes five levels: excellent, good, average, fair, and poor; Step 8.2: Construct membership functions for each evaluation level using a combination of triangles and semi-trapezoidal methods, based on the optimal values ​​and thresholds obtained in steps 2, 3, and 4 above; Among them, the membership function for smaller-scale operations is:

[0049] Intermediate membership function:

[0050] Larger membership functions:

[0051] In the formula, The membership degree is represented by a, b, and c, which are the delimiting values. Step 8.3: Substitute the various index data collected by the sensor into the membership function to obtain the corresponding fuzzy relation matrix; Step 8.4: Second-level fuzzy comprehensive evaluation expression:

[0052] In the formula: B is the set of weights for the three-level indicators, and R is the fuzzy judgment matrix calculated by the membership function; First-level fuzzy comprehensive evaluation expression:

[0053] In the formula, Q is the set of weights for secondary indicators; T; Finally, the fuzzy comprehensive evaluation result of G is determined based on the principle of maximum membership. S9. On the other hand, this invention also introduces an LSTM model to predict the support effectiveness level of hydraulic supports: specifically including: S9-1. Normalize the dataset containing hazard level labels obtained through fuzzy comprehensive evaluation:

[0054] In the formula, For a certain original data x i After normalization, min(x) is the minimum value of all samples for this feature, and max(x) is the maximum value of all samples for this feature; S9-2. Divide the normalized dataset into training and test sets proportionally. S9-3. Use the preprocessed dataset as input data and input it into the LSTM model. Set the network parameters and train the LSTM model using historical data through backpropagation and parameter optimization algorithms. The LSTM model mainly consists of four parts: the forget gate (f), the input gate (i), the memory cell state (c), and the output gate (o), which can be expressed mathematically as follows: Forgotten Gate:

[0055] Input Gate:

[0056]

[0057] Memory cell state:

[0058] Output gate:

[0059]

[0060] In the formula: It is the output of the forget gate. Here, is the sigmoid activation function, W and b are the weight matrix and bias term, h and x are the input and output vectors of the LSTM neuron, respectively, and the subscripts t-1 and t represent different time steps. It is the output of the input gate. Let tanh be the current state of the memory cell, and let tanh be the activation function. It is an updated memory unit. It is the output of the output gate. The hidden state at the current time step; The LSTM model learns the long-term dependencies in the input sequence and outputs the predicted support effectiveness level using the hydraulic support support effectiveness level. The model is pre-trained and real-time data samples collected by sensors are input into the LSTM model to predict the support effectiveness level.

[0061] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements a method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM as described above.

[0062] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0063] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for evaluating and providing early warning of hydraulic support performance based on fuzzy evaluation-LSTM, characterized in that, Includes the following steps: S1. Obtain time-series data of the hydraulic support through sensors, including the sensor acquisition date and corresponding support parameters; S2. Use Pearson correlation coefficient to perform correlation analysis on support parameters, and remove support parameters with correlation coefficients greater than the set threshold in order to determine evaluation indicators. S3. Based on the coupling relationship between the support and the surrounding rock structure, calculate the classification thresholds of each evaluation index under the conditions of strength coupling, stiffness coupling and stability coupling respectively. S4. Construct a hydraulic support support performance evaluation index system with three levels: target layer, criterion layer, and index layer. The criterion layer includes strength coupling state, stiffness coupling state, and stability coupling state. S5. Calculate the subjective weights of each level of indicators in the hydraulic support support effectiveness evaluation index system using the interval analytic hierarchy process. S6. The objective weights of each indicator in the indicator layer of the hydraulic support support effectiveness evaluation index system are calculated using the entropy weight method. S7. Based on game theory, optimization is performed with the goal of minimizing the deviation between subjective weights and objective weights, and the comprehensive weights of the indicators under the Nash equilibrium are obtained. S8. Based on the hydraulic support support performance evaluation index system and the comprehensive weight of the index, a fuzzy comprehensive evaluation model is constructed. The membership function of each index is dynamically determined according to the classification threshold obtained in step S3. The real-time acquired time series data is input into the fuzzy comprehensive evaluation model, and the support performance level at the current moment is obtained according to the principle of maximum membership. S9. Train an LSTM model based on the time series data with support effectiveness level labels output by S8, and then input the time series data to be predicted into the trained LSTM model to predict the support effectiveness level of the hydraulic support at future moments.

2. The method for evaluating and early warning of hydraulic support performance based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In S1, the sensors include a column pressure sensor, a top plate displacement monitor, and an inclination sensor; the support parameters include pressure-related parameters, displacement-related parameters, and inclination-related parameters. The pressure-related parameters include the support strength of the upper column, the support strength of the middle column, and the support strength of the lower column. The displacement-related parameters include the top plate displacement of the upper column, the top plate displacement of the middle column, and the top plate displacement of the lower column. The inclination-related parameters include the inclination angle of the top beam and the inclination angle of the base.

3. The method for evaluating and early warning of hydraulic support performance based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S2, Pearson correlation coefficient is used to perform correlation analysis on the support parameters, and support parameters with correlation coefficients greater than a set threshold are removed. Specifically, this includes: S2-1, Pearson correlation coefficient calculation formula: in, Here, x1 and x2 are two distinct indicators, E(.) is the expected value, and D(.) is the variance. S2-2. A Pearson correlation coefficient greater than the set threshold indicates a correlation between evaluation indicators. Evaluation indicators with Pearson correlation coefficients greater than the preset threshold are removed, and the final evaluation indicators of hydraulic support performance are determined for the subsequent construction of the hydraulic support performance evaluation indicator system.

4. The method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S3, the classification thresholds for each evaluation index under the states of strength coupling, stiffness coupling, and stability coupling are calculated respectively, specifically including: S3-1, Strength Coupling State: Calculation of the hydraulic support strength when the surrounding rock of the roadway is in the elastic limit state based on elasticity theory and the Mohr-Coulomb strength criterion. Where P is the hydraulic support strength, γ is the rock mass unit weight, M is the tunnel depth, and C is the rock cohesion. The internal friction angle of the rock; S3-2, Stiffness Coupling State: Calculation of Top Plate Displacement Threshold under Support-Surrounding Rock Coupling State Based on Elastic Foundation Beam Theory: = in, Let be the displacement of the top plate, k be the elastic coefficient of the support and the bottom plate, and L be half the length of the top plate; S3-3, Stability Coupling State: Calculation of Critical Tilting Angle of Hydraulic Support Based on Torque Limit Equilibrium Equation: in, B represents the critical tilt angle of the support, and B represents the width of the support base. G is the pressure exerted by the top plate on the support, μ is the coefficient of friction, H is the height of the support, and h is the height of the support's center of gravity.

5. The method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S5, the subjective weights of each level of indicators in the hydraulic support support effectiveness evaluation index system are calculated using the interval analytic hierarchy process, specifically including: S5-1. Based on the 1-9 proportional scale, the support effectiveness evaluation indicators of the same layer are compared pairwise to construct an interval decision matrix A composed of multiple interval numbers. The definition of A is: in, and These are the definite matrices formed in A. Evaluation indicators obtained through the binary comparison method and The upper and lower limits of the intervals representing the relative importance of the indicators, where n is the number of evaluation indicators; S5-2: Calculate the correction coefficients of the interval decision matrix and perform a consistency test: , Where a and b are interval decision matrices respectively. and Correction factor; If a≤1 and b≥1, the interval decision matrix is ​​considered to have high consistency; otherwise, if a>1 or b<1, the interval decision matrix is ​​considered to have low consistency. In this case, the interval decision matrix A needs to be reconstructed until the obtained interval decision matrix has high consistency. S5-3. Calculate the subjective weights of each indicator: in, To improve the analytic hierarchy process, the subjective weights of each element are... and For matrix and The normalized eigenvectors corresponding to their respective largest eigenvalues.

6. The method for evaluating and early warning of hydraulic support performance based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S7, optimization is performed with the objective of minimizing the deviation between subjective and objective weights, and the comprehensive weight of the indicators under Nash equilibrium is obtained by solving for the following: S7-1. Using game theory, minimize the range of subjective and objective weights, and establish an objective function for the coefficients of two linear combinations. , Optimize: ,i=1,2,…,m S7-2. According to the principle of differentiation, the condition for the first derivative of optimization is: S7-3. Normalize the obtained linear combination coefficients: S7-4. Calculate the overall weight of the indicators: i=1,2,…,m in, This is the comprehensive weight vector. and These are the normalized linear combination coefficients. The weight vector of the analytic hierarchy process (AHP). This is the weight vector for the entropy weight method.

7. The method for evaluating and early warning of hydraulic support performance based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S8, the real-time collected time-series data is input into the fuzzy comprehensive evaluation model, and the support effectiveness level at the current moment is obtained according to the maximum membership principle, specifically including: S8-1. The set of comments used has 5 evaluation levels, and the rating matrix is ​​a 5-dimensional row vector G. S8-2. Construct membership functions using a combination of triangles and semi-trapezoidal shapes, and dynamically adjust the position and shape of the membership functions for each indicator's evaluation level based on the classification threshold obtained in step S3. S8-3. Substitute the time series data collected by the sensor into the corresponding membership function to obtain the corresponding fuzzy relation matrix; S8-4, Second-level fuzzy comprehensive evaluation expression: in, This is a set of weights for three levels of indicators. This is the fuzzy judgment matrix under the i-th evaluation criterion, calculated using the membership function; First-level fuzzy comprehensive evaluation expression: in, G for Q is the set of weights for secondary indicators. T Ultimately, the decision is made based on the principle of maximum membership. G The fuzzy comprehensive evaluation results.

8. The method for evaluating and early warning of hydraulic support performance based on fuzzy evaluation-LSTM according to claim 1, characterized in that, In step S9, the sensor data to be predicted is input into the trained LSTM model to predict the hydraulic support efficiency level at future moments, specifically including: The time series dataset with support effectiveness level labels obtained through fuzzy comprehensive evaluation is normalized, divided into training and test sets according to a set ratio, and input into the LSTM model. The network parameters are set, and the LSTM model is trained through backpropagation and parameter optimization algorithms. The LSTM model learns the long-term dependencies in the input sequence and outputs the predicted support effectiveness level with the hydraulic support support effectiveness level as the output. Using the trained LSTM model, the data samples collected by the sensor are used as test samples to predict the support effectiveness level.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a method for evaluating and warning the support effectiveness of hydraulic supports based on fuzzy evaluation-LSTM as described in any one of claims 1-8.

Images