A risk assessment and interactive response method for hydraulic support in extremely thin coal seams

Abstract

The present invention discloses a risk assessment and interactive response method for an ultra-thin coal seam hydraulic support, including candidate decision behavior analysis, support movement trend prediction based on support support behavior, danger warning based on safety support and support decision assistance; the present invention adopts machine learning algorithms such as long short-term memory networks, which can predict the movement trend of the support based on real-time collected data, thereby achieving instant assessment of potential risks. The use of LSTM is particularly suitable for processing time series data, and can memorize long-term dependent information. By analyzing historical operation data and current risk conditions, using data mining techniques such as cluster analysis, a decision support model is constructed to provide operators with response suggestions based on risk ratings, and assist operators in making more accurate decisions.

Description

Technical Field

[0001] The present invention relates to a prediction method for an ultra-thin coal seam hydraulic support, in particular to a risk assessment and interactive response method for an ultra-thin coal seam hydraulic support, belonging to the technical field of intelligent coal mines. Background Art

[0002] In terms of the current status of the hazard prediction algorithm for hydraulic supports in extremely thin coal seams, the rapid development of technology in recent years has significantly improved the safety level of mines. In particular, the application of machine learning and deep learning algorithms, such as support vector machines, random forests, convolutional neural networks, and long short-term memory networks, can effectively learn from historical data and identify patterns and signals that may lead to failures and safety issues. In addition, the use of real-time monitoring technology and sensor networks, combined with the Internet of Things technology, has achieved real-time monitoring and data analysis of the status of hydraulic supports, thereby responding to potential risks immediately. The development of predictive maintenance models can not only predict potential failures, but also recommend maintenance plans and operational adjustments to avoid failures. Despite the challenges of data quality and complex environment response, through continuous technological innovation, the hazard prediction technology of hydraulic supports is moving towards higher accuracy and efficiency.

[0003] As for the current status of human-machine interaction in hydraulic supports, technological advances are also constantly driving the improvement of operational efficiency and safety. Modern hydraulic support systems are equipped with advanced graphical user interfaces, providing intuitive operation and real-time monitoring functions. The application of mobile devices and remote control technology increases the flexibility of operation and allows operators to remotely access the status information of hydraulic supports. The use of augmented reality and virtual reality technologies in training and operation guidance provides an immersive learning experience to help operators perform fault diagnosis and maintenance. In addition, the introduction of voice control and human-machine collaborative robots further reduces the workload of operators and improves the safety of operations. With the continuous integration of information technology and intelligent manufacturing technology, the human-machine interaction system of hydraulic supports is expected to become more intelligent and user-friendly, greatly improving the safety and efficiency of mine operations.

[0004] At present, in the aspect of hydraulic support danger prediction, a multi-hydraulic support load model training and multi-hydraulic support load prediction method disclosed in the publication number CN117575046A includes: obtaining column pressure data, top beam pitch angle data, top beam roll angle data, coal mining machine cutting position data, coal mining machine traction speed data and initial hydraulic support pressure regression prediction model; preprocessing the data to generate a data set; generating a training data set and a verification data set based on the data set; training the initial hydraulic support pressure regression prediction model based on the training data set and the verification data set to generate a target hydraulic support pressure regression prediction model for the target fully mechanized working face hydraulic support. By considering the top beam pitch angle data, top beam roll angle data, coal mining machine cutting position data and coal mining machine traction speed data to establish and train the target hydraulic support pressure regression prediction model, the support pressure prediction effect can be improved. At the same time, through the model prediction method, the prediction efficiency can be improved and the prediction cost can be reduced. Although machine learning and deep learning technologies have made significant progress, making risk prediction more accurate and timely, the main disadvantage of these advanced algorithms is the dependence on a large amount of high-quality data. Incomplete data collection or low-quality data will directly affect the prediction accuracy of the model. In addition, the complexity and "black box" nature of deep learning models make the model decision-making process difficult to understand, which may not be fully trusted in safety-critical mining environments. At the same time, although real-time monitoring and IoT technologies can provide a large amount of real-time data, they rely on stable network connections, which is particularly problematic in some areas with poor network environments. Network instability may cause delays in data transmission and processing, affecting the real-time nature of decision-making.

[0005] In terms of human-machine interaction, a method for recommending a hydraulic support human-machine collaborative control strategy disclosed in announcement number CN116291659A includes: obtaining the current human-machine collaborative control strategy influencing factor set, the human-machine collaborative control strategy influencing factor set includes the working face geological environment state set, the hydraulic support system state set and the post work state set; inputting the human-machine collaborative control strategy influencing factor set into the pre-established hydraulic support human-machine collaborative modal decision AOG model; and recommending the hydraulic support human-machine collaborative control strategy according to the output result of the hydraulic support human-machine collaborative modal decision AOG model. The hydraulic support human-machine collaborative control strategy can be automatically recommended according to the current human-machine collaborative control strategy influencing factor set, which not only reduces the workload of the post workers, but also promotes the construction of intelligent coal mines. Modern graphical user interfaces (GUIs), mobile devices and remote control technologies improve the intuitiveness and flexibility of operations, but these technologies also have shortcomings. For example, complex GUIs may increase the cognitive burden of operators, especially for users who are not technically proficient. Mobile devices and remote control rely on the stability of the network, and any network delays or interruptions may affect the efficiency and safety of operations. In addition, although augmented reality (AR) and virtual reality (VR) technologies provide operators with an immersive interactive experience, their application is limited by high costs and user adaptation issues, such as eye fatigue or motion sickness that may result from long-term use.

[0006] Therefore, although the current hazard prediction algorithms and human-computer interaction technologies for hydraulic supports have made significant progress in improving the safety and efficiency of mine operations, there are still challenges in terms of data dependence, technical complexity, network dependence, and user adaptability. Solving these problems requires not only technological innovation, but also cost-effectiveness, user-friendliness, and environmental adaptability. Future research and development work should focus on improving the interpretability of the algorithm, enhancing the stability and reliability of the system, and improving the design of the user interface to make it more intuitive and easy to use, thereby providing stronger support for mining safety. Summary of the invention

[0007] The purpose of the present invention is to provide a risk assessment and interactive response method for an ultra-thin coal seam hydraulic support in order to solve at least one of the above technical problems.

[0008] The present invention achieves the above-mentioned purpose through the following technical solutions: a risk assessment and interactive response method for an ultra-thin coal seam hydraulic support, including an auxiliary system for realizing early warning information and performing decision-making auxiliary functions, the auxiliary system specifically includes:

[0009] The auxiliary system analyzes the dynamic mine environment and candidate decision behaviors based on the hydraulic support status information and external environment information;

[0010] The auxiliary system combines the hydraulic support status and environmental constraints to generate support safety limits for candidate behaviors;

[0011] The auxiliary system predicts the movement trend of the hydraulic support and infers the support intention; then, the auxiliary system determines whether the current hydraulic support state meets the warning condition according to the support corresponding to the operation intention expressed by the operator, and issues a warning signal if the warning condition is met;

[0012] The auxiliary system determines whether the hydraulic support can be made to re-enter the support safety limit corresponding to the support intention by adjusting the control command, and provides behavioral decision assistance on whether the hydraulic support can achieve safe support by pressurizing, unloading, or switching other support decision behaviors;

[0013] The risk assessment and interactive response method specifically includes the following steps:

[0014] Step 1: Analysis of candidate decision behaviors. The auxiliary system first collects and analyzes historical operation data, focusing on the response behavior of hydraulic supports under different working conditions in extremely thin coal seams, to quantify the degree of difficulty faced during the operation of hydraulic supports. The size of the collected values ​​reflects the complexity of the working environment and the challenge of the operation. Then, cluster analysis is used to identify the patterns of operation behaviors and classify them into high-risk and low-risk operations.

[0015] Step 2: Predict the movement trend of the stent based on the stent support behavior. The LSTM model is used to predict the movement trend of the stent. The training process uses back propagation and optimization algorithms to update the weights and bias terms of the model to reduce the value of the loss function. The trained model can predict the future stent position or state based on the input features. By analyzing the potential instability and abnormal movement in the prediction results, it can provide support for risk assessment.

[0016] Step 3: Based on the danger warning of safety support, a risk detection model based on support movement trend and decision-making behavior analysis is established in combination with the analysis of data and the trend prediction of support movement. The risk level of the impact of the operation on the safety support of the hydraulic support is evaluated through the risk detection model, and the warning mechanism is triggered when the risk condition is detected;

[0017] Step 4: Support decision-making assistance. After the risk level assessment in step 3, analyze historical data, identify the measures taken under similar risk levels and their success rates, and recommend a series of response measures based on the risk level of the year and the success rate of historical decisions. Sort the recommended measures according to the success rate in historical data, and finally recommend the best response measures after sorting to the operator.

[0018] As a further solution of the present invention: collecting and analyzing historical operation data including coal seam thickness, working difficulty index and pressure changes of hydraulic supports. The pressure changes of the hydraulic supports reflect the adaptability and stability of the hydraulic supports when facing different working conditions. The working difficulty coefficient is a comprehensive indicator, including but not limited to the geological conditions of the coal seam, the inclination angle of the coal seam, the depth of the mine, the environmental pressure around the support and the ventilation conditions of the mine.

[0019] As a further solution of the present invention: the work difficulty coefficient is determined by a combination of multiple influencing factors, assuming that the weight of each factor is ω G ,ω I ,ω D ,ω P ,ω V and ω S ; Its calculation can be in the form of weighted sum, and each influencing factor is assigned a corresponding weight according to its contribution to the work difficulty; then the calculation formula of the work difficulty coefficient can be expressed as:

[0020] WDC=ω G ×G+ω I ×I+ω D ×D+ω P ×P+ω V ×V+ω S ×S

[0021] Where: G ×G: weight and value of geological stability; ω I ×I: weight and value of coal seam inclination angle; ω D ×D: weight and value of mine depth; ω P ×P: weight and value of environmental pressure;

[0022] ω V ×V: weight and value of ventilation conditions; ω S ×S: weight and value of the complexity of the support system.

[0023] As a further solution of the present invention: the value of the work difficulty index reflects the complexity of the work environment and the challenge of the operation, which is specifically manifested as follows:

[0024] When the value is larger, it means that the working environment is more complex and challenging, including but not limited to unstable coal seam geological conditions, greater mine depth or greater inclination angle, all of which may increase the effort required for the hydraulic support to maintain stability, thereby increasing the difficulty of the work;

[0025] When the value is small, it means that the working environment is relatively simple and the challenges faced are relatively small; this is because the geological conditions of the coal seam are good, the mine is shallow, the inclination angle is small, or the environmental pressure around the support is low, making the operation and maintenance of the hydraulic support relatively easy.

[0026] As a further solution of the present invention: the cluster analysis algorithm can be expressed by the following formula:

[0027]

[0028] Where: J is the objective function, k is the number of clusters, C i is the point set in the i-th cluster, x is the point in C i Point, μ i It is C i The cluster center of ||x-μ i | 2 is the distance from point x to cluster center μ i The square of the Euclidean distance.

[0029] As a further solution of the present invention: the prediction of the movement trend of the bracket by the LSTM model specifically includes:

[0030] Let X be the input feature matrix, which contains the real-time monitored support position x t , pressure p t , speed v t , and environmental factors including the thickness d of the coal seam and the geological conditions g; Y is set as the target variable, i.e., the support position or movement state at a future time point;

[0031] X={x t ,p t ,v t ,d,g}

[0032] Y={x t+1}

[0033] The basic unit of the LSTM model is the LSTM cell, which contains three gates (input gate i t 、Forget gate t and output gate o t ) and a cell state C t ; The general form of the model can be expressed as:

[0034] Forget gate: f t =σ(W f ·[h t-1 ,x t ]+b f )

[0035] Input Gate:

[0036] Unit status update:

[0037] Output Gate:

[0038] Where: W and b are weights and biases respectively term, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, and h t is the output at time t;

[0039] Training an LSTM model typically involves minimizing the difference between the predicted output and the actual output. Common loss functions include mean squared error (MSE):

[0040]

[0041] Where: y t is the actual target value, is the model prediction value, and n is the number of samples.

[0042] As a further solution of the present invention: the risk monitoring model specifically includes:

[0043] Assume that the movement trend of the hydraulic support can be expressed by the position change rate v t and the pressure change rate p t To describe; we can define the risk condition R t is a function based on the position change rate and pressure change rate exceeding the preset position change speed threshold v th and the pressure change rate p th ; Define a conditional function R t , used to evaluate whether there is a high-risk state; the output of this function depends on the current position change rate v t and the pressure change rate p t Whether it exceeds the preset threshold v th and p th ;

[0044] If v t >v th or p t >p th , that is, if the current speed exceeds the safety threshold of speed or the current pressure change rate exceeds the safety threshold of pressure, function R t The value of 1 indicates a high risk state;

[0045] If the above conditions are not met, that is, the speed and pressure change rates are within the safety threshold, function R t A value of 0 indicates a normal risk level;

[0046] Among them, Rt =1 indicates a high risk condition is detected, R t =0 indicates normal risk level.

[0047] As a further solution of the present invention: the position change speed threshold v th Set to 120% of the maximum velocity observed by the stent under normal operating conditions; the pressure change rate threshold p th Based on the typical pressure change of the bracket under different working conditions, a safety margin of 150% of the average pressure change rate is added; this means that when the system detects a high risk condition R t = 1, the early warning mechanism will be automatically triggered to alert the operator through sound, light signals and graphical user interface to indicate potential danger; if the risk level assessed by the system is normal, R t =0, no warning will be triggered.

[0048] The beneficial effects of the present invention are:

[0049] 1) The present invention uses machine learning algorithms such as long short-term memory networks (LSTM) to predict the movement trend of the stent based on real-time collected data, thereby achieving instant assessment of potential risks. The use of LSTM is particularly suitable for processing time series data and can memorize long-term dependent information. It is one of the advanced algorithms for solving such problems;

[0050] 2) It is a challenge to make effective decisions quickly in emergency situations. This invention analyzes historical operation data and current risk conditions, uses data mining techniques such as cluster analysis, builds a decision support model, and provides operators with response suggestions based on risk ratings. This method uses the knowledge contained in historical data to assist operators in making more accurate decisions through pattern recognition and predictive analysis;

[0051] 3) The present invention uses advanced data analysis methods to not only identify high-risk and low-risk operating modes, but also provide customized early warning and decision support based on this information. Cluster analysis, as an unsupervised learning method, can discover hidden patterns and structures in unlabeled data, providing an effective means for understanding operator behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the process flow of the present invention;

[0053] Figure 2 This is a block diagram of the human-machine collaborative decision-making algorithm of the present invention;

[0054] Figure 3 A schematic diagram of cluster analysis and classification of high-risk and low-risk operations according to the present invention;

[0055] Figure 4 This is a schematic diagram of the human-machine collaborative decision-making process of the present invention;

[0056] Figure 5 This is a block diagram of the algorithm for predicting the movement trend of the hydraulic support according to the present invention;

[0057] Figure 6 A decision-making assistance flow chart is provided to support the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] Embodiment 1, as Figures 1 to 6 As shown, a risk assessment and interactive response method for an ultra-thin coal seam hydraulic support includes an auxiliary system for realizing early warning information and performing decision-making auxiliary functions, and the auxiliary system specifically includes:

[0060] The auxiliary system analyzes the dynamic mine environment and candidate decision behaviors based on the hydraulic support status information and external environment information;

[0061] The auxiliary system combines the hydraulic support status and environmental constraints to generate support safety limits for candidate behaviors;

[0062] The auxiliary system predicts the movement trend of the hydraulic support and infers the support intention; then, the auxiliary system determines whether the current hydraulic support state meets the warning condition according to the support corresponding to the operation intention expressed by the operator, and issues a warning signal if the warning condition is met;

[0063] The auxiliary system determines whether the hydraulic support can be made to re-enter the support safety limit corresponding to the support intention by adjusting the control command, and provides behavioral decision assistance on whether the hydraulic support can achieve safe support by pressurizing, unloading, or switching other support decision behaviors;

[0064] The risk assessment and interactive response method specifically includes the following steps:

[0065] Step 1: Analysis of candidate decision behaviors. The auxiliary system first collects and analyzes historical operation data, focusing on the response behavior of hydraulic supports under different working conditions in extremely thin coal seams, to quantify the degree of difficulty faced during the operation of hydraulic supports. The size of the collected values ​​reflects the complexity of the working environment and the challenge of the operation. Then, cluster analysis is used to identify the patterns of operation behaviors and classify them into high-risk and low-risk operations.

[0066] Step 2: Predict the movement trend of the stent based on the stent support behavior. The LSTM model is used to predict the movement trend of the stent. The training process uses back propagation and optimization algorithms to update the weights and bias terms of the model to reduce the value of the loss function. The trained model can predict the future stent position or state based on the input features. By analyzing the potential instability and abnormal movement in the prediction results, it can provide support for risk assessment.

[0067] Step 3: Based on the danger warning of safety support, a risk detection model based on support movement trend and decision-making behavior analysis is established in combination with the analysis of data and the trend prediction of support movement. The risk level of the impact of the operation on the safety support of the hydraulic support is evaluated through the risk detection model, and the warning mechanism is triggered when the risk condition is detected;

[0068] Step 4: Support decision-making assistance. After the risk level assessment in step 3, analyze historical data, identify the measures taken under similar risk levels and their success rates, and recommend a series of response measures based on the risk level of the year and the success rate of historical decisions. Sort the recommended measures according to the success rate in historical data, and finally recommend the best response measures after sorting to the operator.

[0069] The auxiliary decision-making algorithms needed are shown in Table 1:

[0070] Table 1 shows some source code and results of the decision-making support algorithm

[0071]

[0072] Based on the simulated data, when faced with high risk, the system recommends a range of possible response measures based on the current risk level and the success rate of response measures in historical data. These recommendations are sorted by the success rate of historical data to help operators make the best decision.

[0073] Embodiment 2: In addition to all the technical features of Embodiment 1, this embodiment also includes:

[0074] The historical operation data collected and analyzed include coal seam thickness, working difficulty index and pressure changes of hydraulic supports. The pressure changes of hydraulic supports reflect the adaptability and stability of the hydraulic supports when facing different working conditions. The working difficulty coefficient is a comprehensive indicator, including but not limited to the geological conditions of the coal seam, the inclination angle of the coal seam, the depth of the mine, the environmental pressure around the support and the ventilation conditions of the mine.

[0075] The work difficulty coefficient is determined by multiple influencing factors. Assume that the weight of each factor is ω G ,ω I ,ω D ,ωP ,ω V and ω S ; Its calculation can be in the form of weighted sum, and each influencing factor is assigned a corresponding weight according to its contribution to the work difficulty; then the calculation formula of the work difficulty coefficient can be expressed as:

[0076] WDC=ω G ×G+ω I ×I+ω D ×D+ω P ×P+ω V ×V+ω S ×S

[0077] Where: G ×G: weight and value of geological condition stability, reflecting the impact of geological conditions on operation difficulty;

[0078] ω I ×I: weight and value of the coal seam inclination angle. The larger the inclination angle, the higher the difficulty of operation.

[0079] ω D ×D: weight and value of the mine depth. The greater the depth, the more complex the working conditions.

[0080] ω P ×P: weight and value of environmental pressure, reflecting the influence of surrounding rock pressure on support stability;

[0081] ω V ×V: weight and value of ventilation conditions, which directly affect the safety and comfort of the working surface;

[0082] ω S ×S: The weight and value of the bracket system complexity, which involves the difficulty of bracket design, installation and maintenance.

[0083] The value of the work difficulty index reflects the complexity of the work environment and the challenge of the operation, which is specifically manifested as follows:

[0084] When the value is larger, it means that the working environment is more complex and challenging, including but not limited to unstable coal seam geological conditions, greater mine depth or greater inclination angle, all of which may increase the effort required for the hydraulic support to maintain stability, thereby increasing the difficulty of the work;

[0085] When the value is small, it means that the working environment is relatively simple and the challenges faced are relatively small; this is because the geological conditions of the coal seam are good, the mine is shallow, the inclination angle is small, or the environmental pressure around the support is low, making the operation and maintenance of the hydraulic support relatively easy.

[0086] Among them, the basic data set for candidate decision behavior analysis is shown in Table 2.

[0087] Table 2 is the basic data set for candidate decision behavior analysis

[0088]

[0089] The clustering analysis algorithm can be expressed by the following formula:

[0090]

[0091] Where J is the objective function, k is the number of clusters, and C i is the point set in the i-th cluster, x is the point in C i Point, μ i It is C i The cluster center of ∥x-μ i | 2 is the distance from point x to cluster center μ i The square of the Euclidean distance.

[0092] Embodiment 3, in addition to all the technical features of Embodiment 1, this embodiment also includes: In step 2, the prediction of the movement trend of the bracket by the LSTM model specifically includes:

[0093] Let X be the input feature matrix, which contains the real-time monitored support position x t , pressure p t , speed v t , and environmental factors including the thickness d of the coal seam and the geological conditions g; Y is set as the target variable, i.e., the support position or movement state at a future time point;

[0094] X={x t ,p t ,v t ,d,g}

[0095] Y={x t+1}

[0096] The basic unit of the LSTM model is the LSTM cell, which contains three gates (input gate i t 、Forget gate t and output gate o t ) and a cell state C t ; The general form of the model can be expressed as:

[0097] Forget Gate:

[0098] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0099] Input Gate:

[0100] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0101]

[0102] Unit status update:

[0103]

[0104] Output Gate:

[0105] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0106] h t =o t *tanh(C t )

[0107] Where: W and b are weights and biases respectively term, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, and h t is the output at time t;

[0108] Training an LSTM model typically involves minimizing the difference between the predicted output and the actual output. Common loss functions include mean squared error (MSE):

[0109]

[0110] Among them, y t is the actual target value, is the model prediction value, and n is the number of samples.

[0111] Embodiment 4, in addition to all the technical features of Embodiment 1, this embodiment further includes: In step 3, the risk monitoring model specifically includes:

[0112] Assume that the movement trend of the hydraulic support can be expressed by the position change rate v t and the pressure change rate p t To describe; we can define the risk condition R t is a function based on the position change rate and pressure change rate exceeding the preset position change speed threshold v th and the pressure change rate p th ; Define a conditional function R t, used to evaluate whether there is a high-risk state; the output of this function depends on the current position change rate v t and the pressure change rate p t Whether it exceeds the preset threshold v th and p th ;

[0113] If v t >v th or p t >p th , that is, if the current speed exceeds the safety threshold of speed or the current pressure change rate exceeds the safety threshold of pressure, function R t The value of 1 indicates a high risk state;

[0114] If the above conditions are not met, that is, the speed and pressure change rates are within the safety threshold, function R t A value of 0 indicates a normal risk level;

[0115] Among them, R t =1 indicates a high risk condition is detected, R t =0 indicates normal risk level.

[0116] Position change speed threshold v th Set to 120% of the maximum velocity observed by the stent under normal operating conditions; the pressure change rate threshold p th Based on the typical pressure change of the bracket under different working conditions, a safety margin of 150% of the average pressure change rate is added; this means that when the system detects a high risk condition R t = 1, the early warning mechanism will be automatically triggered to alert the operator through sound, light signals and graphical user interface to indicate potential danger; if the risk level assessed by the system is normal, R t =0, no warning will be triggered.

[0117] By adopting machine learning algorithms such as long short-term memory networks (LSTM), it is possible to predict the movement trend of the bracket based on real-time collected data, thereby achieving instant assessment of potential risks. The use of LSTM is particularly suitable for processing time series data and can remember long-term dependent information. By analyzing historical operation data and current risk conditions, using data mining techniques such as cluster analysis, a decision support model is constructed to provide operators with response recommendations based on risk ratings.

[0118] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0119] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A risk assessment and interactive response method for an ultra-thin coal seam hydraulic support, including an auxiliary system for realizing early warning information and performing decision-making assistance functions, characterized in that: The auxiliary system specifically includes: The auxiliary system analyzes the dynamic mine environment and candidate decision behaviors based on the hydraulic support status information and external environment information; The auxiliary system combines the hydraulic support status and environmental constraints to generate support safety limits for candidate behaviors; The auxiliary system predicts the movement trend of the hydraulic support and infers the support intention. Then, the auxiliary system determines whether the current state of the hydraulic support meets the warning condition according to the support corresponding to the operation intention expressed by the operator. If the warning condition is met, a warning signal is issued. The auxiliary system determines whether to adjust the control command to make the hydraulic support re-enter the support safety limit corresponding to the support intention, and provides behavioral decision assistance on whether the hydraulic support can achieve safe support by pressurizing, unloading, or switching other support decision behaviors; The risk assessment and interactive response method specifically includes the following steps: Step 1: Candidate decision-making behavior analysis The auxiliary system collects and analyzes historical operation data, focusing on the response behavior of hydraulic supports under different working conditions in extremely thin coal seams, in order to quantify the degree of difficulty faced during the operation of hydraulic supports. The size of the collected values ​​reflects the complexity of the working environment and the challenge of the operation. Cluster analysis is then used to identify the patterns of operation behavior and classify them into high-risk and low-risk operations. Step 2: Prediction of stent movement trend based on stent support behavior The LSTM model is used to predict the trend of the stent movement. The training process uses back propagation and optimization algorithms to update the weights and bias items of the LSTM model to reduce the value of the loss function. The trained LSTM model can predict the future stent position or state based on the input features, and provide support for risk assessment by analyzing the potential instability and abnormal movement in the prediction results. Step 3: Danger warning based on safety support Combined with the analysis of data and the trend prediction of support movement, a risk detection model based on support movement trend and decision-making behavior analysis is established. The risk level of the impact of the operation on the safe support of the hydraulic support is evaluated through the risk detection model, and an early warning mechanism is triggered when risk conditions are detected; Step 4: Support decision-making assistance After the risk level assessment in step three, analyze historical data, identify the measures taken under similar risk levels and their success rates, and recommend a series of response measures based on the risk level of the year and the success rate of historical decisions. Rank the recommended measures based on the success rate in historical data, and finally recommend the best response measures after ranking to the operator.

2. The risk assessment and interactive response method according to claim 1, characterized in that: In the step 1, historical operation data including coal seam thickness, working difficulty index and pressure change of hydraulic support are collected and analyzed; The pressure change of the hydraulic support reflects the adaptability and stability of the hydraulic support when facing different working conditions; The work difficulty index is a comprehensive indicator, including but not limited to the geological conditions of the coal seam, the inclination angle of the coal seam, the depth of the mine, the environmental pressure around the support and the ventilation conditions of the mine.

3. The risk assessment and interactive response method according to claim 2, characterized in that: In the step 1, the work difficulty coefficient is determined by a combination of multiple influencing factors; Assume that the weight of each influencing factor is ω G ,ω I ,ω D ,ω P ,ω V and ω S ; The calculation adopts the form of weighted sum, and each influencing factor is assigned a corresponding weight according to its contribution to the difficulty of the work; then the calculation formula of the work difficulty coefficient is expressed as: WDC=ω G ×G+ω I ×I+ω D ×D+ω P ×P+ω V ×V+ω S ×S Where: G ×G: weight and value of geological stability; ω I ×I: weight and value of coal seam inclination angle; ω D ×D: weight and value of mine depth; ω P ×P: weight and value of environmental pressure; ω V ×V: weight and value of ventilation conditions; ω S ×S: weight and value of the complexity of the support system.

4. The risk assessment and interactive response method according to claim 3, characterized in that: In step 1, the value of the work difficulty index reflects the complexity of the work environment and the challenge of the operation, which is specifically manifested as follows: When the value of the work difficulty index is larger, it means that the working environment is more complex and challenging, including but not limited to unstable coal seam geological conditions, greater mine depth or greater inclination angle, all of which may increase the effort required for the hydraulic support to maintain stability, thereby increasing the difficulty of the work; When the work difficulty index value is smaller, it means that the working environment is relatively simple and the challenges faced are smaller; this is because the geological conditions of the coal seam are good, the mine is shallow, the inclination angle is small, or the environmental pressure around the support is low, making the operation and maintenance of the hydraulic support relatively easy.

5. The risk assessment and interactive response method according to claim 1, characterized in that: In step 1, the cluster analysis algorithm is expressed by the formula: Where: J is the objective function, k is the number of clusters, C i is the point set in the i-th cluster, x is the point in C i Point, μ i It is C i The cluster center of ||x-μ i || 2 is the distance from point x to cluster center μ i The square of the Euclidean distance.

6. The risk assessment and interactive response method according to claim 1, characterized in that: In step 2, the prediction of the movement trend of the bracket by the LSTM model specifically includes: Let X be the input feature matrix, which contains the real-time monitored support position x t , pressure p t , speed v t , and environmental factors including the thickness d of the coal seam and the geological conditions g; Y is set as the target variable, i.e., the support position or movement state at a future time point; X={x t ,p t ,v t ,d,g} Y={x t+1 } The basic unit of the LSTM model is the LSTM cell, which contains an input gate i t 、Forget gate t , output gate o t and a cell state C t ; The form of the LSTM model is expressed as: Forget gate: f t =σ(W f ·[h t-1 , x t ]+b f ) Input Gate: Unit status update: Output Gate: Where: W and b are weights and bias terms, σ is the sigmoid activation function, tanh is the hyperbolic tangent activation function, h t is the output at time t; Training an LSTM model typically involves minimizing the difference between the predicted output and the actual output. Common loss functions include mean squared error: Where: yt is the actual target value, is the model prediction value, and n is the number of samples.

7. The risk assessment and interactive response method according to claim 1, characterized in that: In step 3, the risk monitoring model specifically includes: Assume that the movement trend of the hydraulic support is determined by the position change rate v t and the pressure change rate p t to describe; Define the risk condition R t is a function based on the position change rate and pressure change rate exceeding the preset position change speed threshold v th and the pressure change rate p th ; Define a conditional function R t , used to evaluate whether there is a high-risk state. The output of this function depends on the current position change rate v t and the pressure change rate p t Whether it exceeds the preset threshold v th and p th ; If v t >v th or p t >p th , that is, if the current speed exceeds the safety threshold of speed or the current pressure change rate exceeds the safety threshold of pressure, function R t The value of 1 indicates a high-risk state; If the above conditions are not met, that is, the speed and pressure change rates are within the safety threshold, function R t A value of 0 indicates a normal risk level; Among them, R t =1 indicates a high risk condition is detected, R t =0 indicates normal risk level.

8. The risk assessment and interactive response method according to claim 7, characterized in that: The position change speed threshold v th Set to 120% of the maximum speed observed by the stand under normal operating conditions; Pressure change rate threshold p th Based on the typical pressure changes of the bracket under different working conditions, plus a safety margin of 150% of the average pressure change rate, it means that when the system detects a high risk condition R t =1, the early warning mechanism will be automatically triggered, alerting the operator to potential dangers through sound, light signals and graphical user interface; If the risk level assessed by the system is normal R t =0, no warning will be triggered.

Images