Compressed air energy storage cavern surrounding rock deformation monitoring system

Abstract

The application provides a compressed air energy storage cavern surrounding rock deformation monitoring system, and relates to the technical field of cavern surrounding rock, comprising: a data acquisition module, which arranges concentric monitoring points in the radial direction of the pipeline at the compressed air inlet and outlet areas of the cavern, and obtains time series data of strain and displacement; a preliminary screening module calculates the displacement-strain change rate ratio, and screens out an abnormal point set A; a calculation module calculates the displacement rate and the cumulative displacement rate; a secondary screening module judges the overall or local deformation, and screens out significant abnormal points; a data prediction model construction module constructs a model based on LSTM, a simulation module performs prediction, and a weighted calculation and judgment module calculates the risk coefficient and judges the risk level. The application introduces strain parameters, quantifies the coupling relationship between the strain parameters and deformation, fits the operation condition of the energy storage cavern, accurately locates the local deformation area, and avoids the missed judgment of abnormal points caused by the monitoring blind area.

Description

Technical Field

[0001] The present invention relates to the technical field of cavern surrounding rock, and in particular to a compressed air energy storage cavern surrounding rock deformation monitoring system. Background Art

[0002] Compressed air energy storage vaults are core facilities for large-capacity physical energy storage. The stability of the surrounding rock is directly related to the safety and operational life of the energy storage system. During the inflation and deflation cycles, the surrounding rock is affected by air pressure loads, temperature field changes, and geological structures, which may cause crack expansion, plastic deformation, and even instability and failure. Therefore, real-time monitoring and risk assessment of surrounding rock deformation are key technical links to ensure vault safety.

[0003] In the prior art, there is a system for monitoring the deformation of surrounding rock in underground chambers, with publication number CN117056659A, which includes: an acquisition module, a transmission module, a processing module, and a monitoring module; the acquisition module is used to acquire the deformation parameters of the surrounding rock to be monitored; the transmission module is used to transmit the deformation parameters to the processing module; the processing module is used to construct a prediction model, and input the deformation parameters into the prediction model to predict the deformation trend and obtain the prediction results; the monitoring module is used to monitor the prediction results and perform remote early warning and control in real time. This method can accurately monitor the displacement, deformation, and stress of the surrounding rock in real time, helping engineers and mine operators understand the stability of the surrounding rock and take timely measures to prevent disasters and accidents. In addition, this system can also provide trend prediction and early warning functions for surrounding rock deformation through data analysis, model prediction, and other methods to ensure the safety and efficiency of underground projects and mines.

[0004] However, there are still the following deficiencies. As can be seen from the above statements, the existing technology only collects displacement, deformation and stress data, and does not involve parameters that are strongly related to strain and the operation of compressed air energy storage caverns; it only predicts the deformation trend by building a prediction model, and does not mention the hierarchical screening logic for abnormal points, making it difficult to distinguish between overall deformation and local deformation, which leads to misjudgment of abnormal points.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present invention is to provide a system for monitoring deformation of surrounding rock of a compressed air energy storage cavern to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A system for monitoring deformation of surrounding rock of a compressed air energy storage cavern, comprising:

[0009] The data acquisition module is used to arrange multiple concentric circles equidistantly along the radial direction of the compressed air inlet and outlet areas of the cavern with the center of the pipeline as the origin. Multiple monitoring points are set along the circumference of the same concentric circle. Before the pipeline is put into operation, the circumference of the concentric circle where each monitoring point is located is measured as the initial length of the pipeline. During the current monitoring period, the time series data of the strain and displacement at the monitoring point are obtained at equal time intervals. Based on the time series data, the strain and displacement at the monitoring point at the current time and the previous n times are obtained;

[0010] The preliminary screening module is used to calculate the displacement-strain change rate ratio of the same concentric circle based on the strain data and displacement data at the monitoring point at the current moment and the previous n moments, compare the change rate ratio with the preset change rate ratio threshold, and screen out monitoring points with abnormal change rate ratios to form an abnormal point set A. Each concentric circle corresponds to a set A;

[0011] The calculation module is used to calculate the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments based on the displacement data at the current moment and the previous n moments and the initial length of the pipeline;

[0012] The secondary screening module, for the same concentric circle, determines whether the surrounding rock of the cavern is deformed as a whole or locally through the concentration of abnormal points in set A and the consistency coefficient of displacement direction. For overall deformation, a strain-surrounding rock deformation model is constructed. The strain and displacement data measured at the previous n moments are used to train the strain-surrounding rock deformation model. The strain data at the current moment is input to obtain the displacement prediction value, which is compared with the safety threshold before performing subsequent operations. For local deformation, the Pearson correlation coefficient between the monitoring points on the same concentric circle is calculated and used to mark the local abnormal points to form set B. The intersection of sets A and B is taken to screen out significant abnormal points.

[0013] The data prediction model construction module is used to build a data prediction model based on LSTM, and train the data prediction model with the strain data, displacement data, displacement rate and cumulative displacement rate measured n moments before the significant abnormal point;

[0014] The simulation module is used to input the strain and displacement data of the significant abnormal point at the current moment into the trained data prediction model to predict the corresponding displacement rate and cumulative displacement rate;

[0015] The weighted calculation and judgment module is used to calculate the deformation risk coefficient of significant abnormal points based on the predicted displacement rate and cumulative displacement rate, and to judge the deformation risk level of the surrounding rock of the cavern based on the deformation risk coefficient.

[0016] Furthermore, with the center of the pipeline as the origin, multiple concentric circles are arranged equidistantly along the radial direction of the pipeline, and multiple monitoring points are set along the circumference of the same concentric circle. The specific method is as follows:

[0017] The surrounding rock strain shows a decreasing trend with increasing radial distance from the pipeline. Based on the scale of the cavern and the diffusion range of the surrounding rock strain, 3 to 5 layers of concentric circles are set. With the center of the pipeline as the origin, inner, middle and outer concentric circles are set in sequence along the radial direction of the pipeline, with a spacing of 1 to 3m.

[0018] Set 4-8 monitoring points in a single concentric circle with an angular interval of 45°-90°;

[0019] Inner concentric circles: 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°, a total of 8 points. The inner concentric circles are the strain concentration areas;

[0020] Middle concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, forming a "cross orientation" grid. The middle concentric circles are the strain transition zone;

[0021] Outer concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, the outer concentric circles are the strain attenuation area.

[0022] Furthermore, based on the strain data and displacement data at the monitoring points at the current moment and the previous n moments, the displacement-strain rate of change ratio is calculated according to the following formula:

[0023] Combine the current moment and the previous n moments into a set , is the index of the current moment, is the index of the moment;

[0024] ;

[0025] in, For the The monitoring point is The displacement-strain rate of change ratio at each moment, For the The monitoring point is The displacement change at each moment, For the The monitoring point is The strain change at each moment, is the index of the monitoring point on the same concentric circle, , is the number of monitoring points on the same concentric circle;

[0026] The first The monitoring point is The displacement-strain rate of change ratio at each moment and the preset rate of change threshold For comparison, when When , the monitoring point is an abnormal point.

[0027] Furthermore, the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments are calculated according to the following formula:

[0028] ;

[0029] in, For monitoring point The displacement rate at a moment, For monitoring point The coordinates of a moment, For monitoring point The coordinates of a moment, is the adjacent time interval, is the cumulative displacement rate of the monitoring point, is the coordinate at the current moment, For the The coordinates of a moment, The coordinates of ), The monitoring points are The horizontal and vertical coordinates of the moment, is the initial length of the pipe.

[0030] Furthermore, the concentration of abnormal points in set A and the displacement direction consistency coefficient are used to determine whether the surrounding rock of the cavern is deformed as a whole or locally. The specific steps and the formulas are as follows:

[0031] The plane straight-line distance between any two outlier points in the same concentric circle is:

[0032] ;

[0033] in, In the same concentric circle The outlier and The plane distance of the outlier points, 、 Respectively The outlier and The horizontal and vertical coordinates of the abnormal points, 、 are the indices of outliers in the same concentric circle in set A. ;

[0034] ;

[0035] in, is the arithmetic mean of the plane distances between all the outliers in the same concentric circle. , is the number of outliers in set A;

[0036] ;

[0037] in, is the concentration coefficient, 、 are the maximum and minimum distances between all outliers respectively;

[0038] ;

[0039] in, For the The outlier and The displacement direction consistency coefficient of the abnormal point is 、 Respectively The abnormal point is 、 The displacement increment in the direction, 、 Respectively The abnormal point is 、 The displacement increment in the direction;

[0040] ;

[0041] in, is the average value of the displacement direction consistency coefficient;

[0042] when and , then the surrounding rock of the tunnel is deformed as a whole;

[0043] when or , then the surrounding rock of the tunnel is locally deformed;

[0044] in, is the preset concentration coefficient threshold, It is the preset threshold of the average value of the displacement direction consistency coefficient.

[0045] Furthermore, the Pearson correlation coefficient between outliers is calculated according to the following formula:

[0046] ;

[0047] in, In the same concentric circle Monitoring points and Pearson correlation coefficient of displacement data of each monitoring point;

[0048] Where, For the The monitoring point is The coordinates of a moment, The coordinates of ), Respectively The monitoring point is The horizontal and vertical coordinates of the moment, For the The monitoring point is The coordinates of a moment, For the The mean coordinate value of the monitoring points, For the The mean coordinate value of the monitoring points, is the index of the monitoring point on the same concentric circle, ;

[0049] Will and the preset Pearson correlation coefficient threshold When the Pearson correlation coefficients of a monitoring point and other monitoring points in the same concentric circle are all less than When , the monitoring point is determined to be a local abnormal point.

[0050] Furthermore, we take the intersection of sets A and B to filter out significant outliers. The specific steps are as follows:

[0051] 1) When there is one intersection between set A and set B, the point is identified as a significant outlier;

[0052] 2) When there are more than two intersections of set A and set B, extract the positions of all points in the intersection, including the vault, side walls, and bottom plate. Sidewall Prioritize the bottom plate and identify significant outliers;

[0053] If there are multiple points in the same position in the intersection, the significant outlier is determined based on the priority of maximum displacement rate > minimum Pearson correlation coefficient > maximum cumulative displacement rate > minimum concentration;

[0054] 3) When the intersection of set A and set B is an empty set, determine the significant outliers based on set A or set B separately. The specific steps are as follows:

[0055] Determine significant outliers based on set A:

[0056] Extract all outliers in set A and prioritize the vault > sidewall > floor as the most significant outliers.

[0057] If there are multiple outliers at the same location, the significant outlier is determined based on the priority of maximum displacement rate > maximum cumulative displacement rate > minimum concentration;

[0058] Determine significant outliers based on set B:

[0059] For all monitoring points in set B, calculate the average of the absolute values ​​of the Pearson correlation coefficients with other monitoring points on the same concentric circle, sort the average values ​​of the absolute values ​​of the Pearson correlation coefficients in ascending order, and select the top three candidate points;

[0060] If there are multiple arch vertices among the candidate points, the significant outlier points are determined based on the priority of the smallest average value of the absolute value of the Pearson correlation coefficient > the largest displacement rate > the largest cumulative displacement rate;

[0061] If there is no vault among the candidate points and the points include side walls and bottom plates, based on the priority of side walls > bottom plates, the point with the smallest average absolute value of the Pearson correlation coefficient in the same position is selected as the significant outlier.

[0062] Furthermore, the deformation risk coefficient of the significant abnormal point is calculated based on the predicted displacement rate and cumulative displacement rate according to the following formula:

[0063] ;

[0064] in, is the deformation risk coefficient of the significant abnormal point;

[0065] Where, is the displacement rate of the predicted significant outlier point, is the cumulative displacement rate of the predicted significant outlier points;

[0066] Where, is the weight coefficient of the predicted displacement rate, is the weight coefficient of the predicted cumulative displacement rate, On the basis of .

[0067] Furthermore, the deformation risk level of the surrounding rock of the tunnel is determined based on the deformation risk coefficient. The specific process is as follows:

[0068] when , the deformation risk level of the surrounding rock of the cavern is low risk;

[0069] when , the deformation risk level of the surrounding rock of the cavern is medium risk;

[0070] when , the deformation risk level of the surrounding rock of the cavern is high risk;

[0071] in, To classify the critical value between low risk and medium risk, The critical value for dividing medium risk and high risk.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] The present invention arranges concentric circle monitoring points along the radial direction of the pipeline in the compressed air inlet and outlet areas through the data acquisition module to form a "radial-circumferential" structured point distribution mode, which can capture the deformation differences in different areas, accurately locate the local deformation area, and avoid missing abnormal points due to monitoring blind spots; through the preliminary screening module, the monitoring points with abnormal change rate ratios are screened out to form an abnormal point set A, eliminating misjudgment caused by data fluctuations; through the secondary screening module, for the same concentric circle, the concentration of abnormal points in set A and the displacement direction consistency coefficient are used to judge whether the surrounding rock of the cavern is deformed as a whole or locally, and the deformation of the surrounding rock is judged as follows: For overall deformation, a "strain-surrounding rock deformation model" is constructed to quantify the coupling relationship between strain and deformation, making up for the defect of existing technology that does not consider the influence of strain, making the monitoring results more in line with the actual operating conditions of the energy storage cavern, avoiding the limitations of the existing technology of "single model prediction", and greatly improving the accuracy of abnormal point identification. For local deformation, the Pearson correlation coefficient between monitoring points on the same concentric circle is calculated to form a set B, and significant abnormal points are screened out. Significant abnormal points correspond to areas with higher deformation risk coefficients, avoiding the problem of misjudgment of abnormal points by applying the prediction model of overall deformation to local abnormal scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0075] Figure 2 Schematic diagram of the fitting of displacement rate and deformation risk coefficient of the present invention;

[0076] Figure 3 Schematic diagram of the fitting of the cumulative displacement rate and the deformation risk coefficient of the present invention. DETAILED DESCRIPTION

[0077] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0078] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0079] Example 1:

[0080] See also Figure 1 , the present invention provides a technical solution:

[0081] A system for monitoring deformation of surrounding rock of a compressed air energy storage cavern, comprising:

[0082] The data acquisition module is used to arrange multiple concentric circles equidistantly along the radial direction of the compressed air inlet and outlet areas of the cavern with the center of the pipeline as the origin. Multiple monitoring points are set along the circumference of the same concentric circle. Before the pipeline is put into operation, the circumference of the concentric circle where each monitoring point is located is measured as the initial length of the pipeline. During the current monitoring period, the time series data of the strain and displacement at the monitoring point are obtained at equal time intervals. Based on the time series data, the strain and displacement at the monitoring point at the current time and the previous n times are obtained;

[0083] On the basis of the above embodiment, with the center of the pipeline as the origin, multiple concentric circles are arranged equidistantly along the radial direction of the pipeline, and multiple monitoring points are set along the circumference of the same concentric circle. The specific method is as follows:

[0084] The surrounding rock strain shows a decreasing trend with increasing radial distance from the pipeline. Based on the scale of the cavern and the diffusion range of the surrounding rock strain, 3 to 5 layers of concentric circles are set. With the center of the pipeline as the origin, inner, middle and outer concentric circles are set in sequence along the radial direction of the pipeline, with a spacing of 1 to 3m.

[0085] Set 4-8 monitoring points in a single concentric circle with an angular interval of 45°-90°;

[0086] Inner concentric circles: 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°, a total of 8 points. The inner concentric circles are the strain concentration areas;

[0087] Middle concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, forming a "cross orientation" grid. The middle concentric circles are the strain transition zone;

[0088] Outer concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, the outer concentric circles are the strain attenuation area.

[0089] On the basis of the above embodiment, within the current monitoring period, the time series data of strain and displacement at the monitoring point are obtained at equal time intervals. The specific method is as follows:

[0090] The monitoring point refers to the contact surface between the surrounding rock and the pipeline at each monitoring point or the surrounding rock area adjacent to the pipeline.

[0091] Strain refers to the relative deformation of the surrounding rock material at the monitoring point, that is, the ratio of the deformation to the initial length.

[0092] Displacement refers to the radial displacement of the monitoring point relative to its initial position, with the center of the pipeline as the origin. It reflects the deformation trend of the surrounding rock structure. A laser displacement meter measures radial displacement based on the principle of laser ranging. The transmitter is fixed to a stable reference point, and the receiver is installed on the surrounding rock surface corresponding to the monitoring point to obtain displacement data.

[0093] Based on the above embodiment, the collected strain and displacement data must be normalized. Through normalization, the data of different indicators are unified into the range of 0 to 1, eliminating the influence of dimension and value range, so that these data are comparable and consistent in subsequent analysis operations.

[0094] The preliminary screening module is used to calculate the displacement-strain change rate ratio of the same concentric circle based on the strain data and displacement data at the monitoring point at the current moment and the previous n moments, compare the change rate ratio with the preset change rate ratio threshold, and screen out monitoring points with abnormal change rate ratios to form an abnormal point set A. Each concentric circle corresponds to a set A;

[0095] On the basis of the above embodiment, the displacement-strain rate of change ratio is calculated based on the strain data and displacement data at the monitoring point at the current moment and the previous n moments, according to the following formula:

[0096] Combine the current moment and the previous n moments into a set , is the index of the current moment, is the index of the moment;

[0097] ;

[0098] ;

[0099] ;

[0100] in, For the The monitoring point is The displacement-strain rate of change ratio at each moment, For the The monitoring point is The displacement change at each moment, For the The monitoring point is The strain change at each moment, is the index of the monitoring point on the same concentric circle, , is the number of monitoring points on the same concentric circle;

[0101] Where, For the The monitoring point is The coordinates of a moment, For the The monitoring point is The coordinates of a moment, For the The monitoring point is The response of each moment, For the The monitoring point is The response of each moment;

[0102] The first The monitoring point is The displacement-strain rate of change ratio at each moment and the preset rate of change threshold For comparison, when When , the monitoring point is an abnormal point.

[0103] Among them, the change rate is greater than the threshold Determination method:

[0104] Collect the displacement-strain rate of change ratio data of the cavern during normal operation (no obvious deformation, stable air pressure), form a sample set, calculate the 95% or 99% quantile of the sample data, and use it as the threshold , ensuring that only 5% or 1% of normal data are judged as abnormal.

[0105] The calculation module is used to calculate the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments based on the displacement data at the current moment and the previous n moments and the initial length of the pipeline;

[0106] Based on the above embodiment, for pairs of monitoring points within the same cross section, the Pearson correlation coefficient of the displacement data is calculated for each pair. Based on the Pearson correlation coefficient, potential local abnormal points are marked to form a set B. The specific steps are as follows:

[0107] Calculate the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments according to the following formula:

[0108] ;

[0109] in, For monitoring point The displacement rate at a moment, For monitoring point The coordinates of a moment, For monitoring point The coordinates of a moment, is the adjacent time interval, is the cumulative displacement rate of the monitoring point, is the coordinate at the current moment, For the The coordinates of a moment, The coordinates of ), The monitoring points are The horizontal and vertical coordinates of the moment, is the initial length of the pipeline, is Based on the Obtained.

[0110] The secondary screening module, for the same concentric circle, determines whether the surrounding rock of the cavern is deformed as a whole or locally through the concentration of abnormal points in set A and the displacement direction consistency coefficient. For overall deformation, a strain-surrounding rock deformation model is constructed. The strain and displacement data measured at the previous n moments are used to train the strain-surrounding rock deformation model. The strain data at the current moment is input to obtain the displacement prediction value, which is compared with the safety threshold. When the displacement prediction value is greater than or equal to the safety threshold, it is determined to be a risky state and emergency monitoring is initiated; when the displacement prediction value is less than the safety threshold, it is determined to be a normal state and the regular monitoring frequency is maintained. For local deformation, the Pearson correlation coefficient between the monitoring points on the same concentric circle is calculated, and the local abnormal points are marked to form set B. The intersection of sets A and B is taken to screen out significant abnormal points.

[0111] On the basis of the above embodiment, the concentration of abnormal points in set A and the displacement direction consistency coefficient are used to determine whether the surrounding rock of the cavern is deformed as a whole or locally. The specific steps and the formula based on them are as follows:

[0112] The plane straight-line distance between any two outlier points in the same concentric circle is:

[0113] ;

[0114] in, In the same concentric circle The outlier and The plane distance of the outlier points, 、 Respectively The outlier and The horizontal and vertical coordinates of the abnormal points, , are the indices of outliers in the same concentric circle in set A. ;

[0115] ;

[0116] in, is the arithmetic mean of the plane distances between all the outliers in the same concentric circle. , is the number of outliers in set A;

[0117] ;

[0118] in, is the concentration coefficient, 、 are the maximum and minimum distances between all outliers respectively;

[0119] ;

[0120] in, For the The outlier and The displacement direction consistency coefficient of the abnormal point is 、 Respectively The abnormal point is 、 The displacement increment in the direction, 、 Respectively The abnormal point is 、 The displacement increment in the direction;

[0121] ;

[0122] in, is the average value of the displacement direction consistency coefficient;

[0123] when and , indicating that abnormal points are clustered, and the overall trend is consistent displacement. When the surrounding rock deforms as a whole, the displacement directions of the abnormal points are consistent, which conforms to the "overall deformation" feature. Therefore, the surrounding rock of the tunnel is deformed as a whole;

[0124] when or , indicating that the distribution of abnormal points is discrete and close, that is, the abnormal points are unevenly distributed on the concentric circles, suggesting that the deformation only occurs in the local area, rather than the overall synchronous displacement, and the displacement directions of the abnormal points are very different, that is, the deformation directions are inconsistent, which is a typical feature of local deformation. Therefore, the surrounding rock of the cavern is a local deformation;

[0125] in, is the preset concentration coefficient threshold, It is the preset threshold of the average value of the displacement direction consistency coefficient.

[0126] On the basis of the above embodiment, the strain-surrounding rock deformation model is constructed by a deep learning network based on a multilayer perceptron. The deep neural network of the multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer and an output layer. The first hidden layer, the second hidden layer and the third hidden layer each have at least two neurons and each use ReLU as an activation function.

[0127] The strain-displacement data from the first n measured moments are divided into a training set, a validation set, and a test set, with a typical ratio of 7:1:2. The training set is used to learn model parameters; the validation set is used to adjust hyperparameters during training to prevent overfitting; and the test set is used to evaluate the generalization ability of the model after training.

[0128] In the strain-rock deformation model, the input features of the deep learning network of the multilayer perceptron include: strain data of the last moment among the previous n moments;

[0129] The structure of the deep learning network of multilayer perceptron is:

[0130] Input layer: used to receive strain data;

[0131] The first hidden layer has 128 neurons and uses ReLU as the activation function.

[0132] The second hidden layer has 64 neurons and also uses the ReLU activation function.

[0133] The third hidden layer has 32 neurons and uses the ReLU activation function.

[0134] Output layer: has 1 neuron, which is used to output the predicted value of displacement at the next moment.

[0135] The process of training the strain-rock deformation model is as follows:

[0136] The strain data of the last moment in the previous n moments is used as input, and the displacement of the next moment is used as the label output to train the strain-surrounding rock deformation model. The mean square error is used as the loss function. When the mean square error is When it is within the range, the training of the strain-rock deformation model is stopped.

[0137] Based on the above embodiment, the Pearson correlation coefficient between outliers is calculated according to the following formula:

[0138] ;

[0139] in, In the same concentric circle Monitoring points and Pearson correlation coefficient of displacement data of each monitoring point;

[0140] Where, For the The monitoring point is The coordinates of a moment, The coordinates of ), Respectively The monitoring point is The horizontal and vertical coordinates of the moment, For the The monitoring point is The coordinates of a moment, For the The mean coordinate value of the monitoring points, For the The mean coordinate value of the monitoring points, is the index of the monitoring point on the same concentric circle, ;

[0141] Will and the preset Pearson correlation coefficient threshold When the Pearson correlation coefficients of a monitoring point and other monitoring points in the same concentric circle are all less than When , the monitoring point is determined to be a local abnormal point.

[0142] Among them, the Pearson correlation coefficient threshold Determination method: Collect displacement data of similar monitoring points in different time periods under normal operation of the mine, calculate the Pearson correlation coefficient of all pairs of monitoring points to form a basic data set, perform statistical analysis on the data set, and determine the distribution range of the correlation coefficient under normal operating conditions;

[0143] Usually Set it as the upper limit of the normal distribution interval (such as mean + 1.5 times the standard deviation) to ensure that the correlation between monitoring point pairs will not be misjudged as "no significant correlation" under normal circumstances.

[0144] Based on the above embodiment, the intersection of sets A and B is taken to screen out significant outliers. The specific steps are as follows:

[0145] 1) When there is one intersection between set A and set B, the point is identified as a significant outlier;

[0146] 2) When there are more than two intersections of set A and set B, extract the positions of all points in the intersection, including the vault, side walls, and bottom plate. Sidewall Prioritize the bottom plate and identify significant outliers;

[0147] Because the vault bears the deadweight of the overlying rock and soil, it is the core structural component, and its stability directly impacts the overall safety of the vault. Deformation or collapse of the vault can easily trigger a chain reaction, leading to failure of the entire chamber structure. Side walls primarily bear lateral soil pressure, while the floor is subject to vertical loads and groundwater buoyancy, and their failure risk is generally lower than that of the vault. Therefore, prioritizing vault anomalies allows for rapid identification of high-risk areas. This is why these priorities are set.

[0148] If there are multiple points in the same position in the intersection, the significant outlier is determined based on the priority of maximum displacement rate > minimum Pearson correlation coefficient > maximum cumulative displacement rate > minimum concentration;

[0149] The displacement rate directly reflects the rate of deformation per unit time. A sudden increase in the rate indicates an immediate collapse risk and requires priority attention to buy time for emergency response. The Pearson correlation coefficient reflects the difference between the displacement trend and surrounding points, indicating local anomalies and hidden dangers. Local anomalies may develop into urgent dangers and are therefore prioritized for investigation. The cumulative displacement rate quantifies the cumulative effect of long-term deformation and reflects the degree of chronic structural damage. The risk develops relatively slowly and requires consideration of dynamic changes. The concentration rate measures the degree of clustering of anomalies, eliminates data interference and group misjudgment, and is used to accurately locate individual hidden dangers. Therefore, the above priorities are set.

[0150] 3) When the intersection of set A and set B is an empty set, determine the significant outliers based on set A or set B separately. The specific steps are as follows:

[0151] Determine significant outliers based on set A:

[0152] Extract all outliers in set A and prioritize the vault > sidewall > floor as the most significant outliers.

[0153] If there are multiple outliers at the same location, the significant outlier is determined based on the priority of maximum displacement rate > maximum cumulative displacement rate > minimum concentration;

[0154] Determine significant outliers based on set B:

[0155] For all monitoring points in set B, calculate the average of the absolute values ​​of the Pearson correlation coefficients with other monitoring points on the same concentric circle, sort the average values ​​of the absolute values ​​of the Pearson correlation coefficients in ascending order, and select the top three candidate points;

[0156] If there are multiple arch vertices among the candidate points, the significant outlier points are determined based on the priority of the smallest average value of the absolute value of the Pearson correlation coefficient > the largest displacement rate > the largest cumulative displacement rate;

[0157] The vault is a critical load-bearing component of a mine. Points with the lowest Pearson coefficient indicate that their displacement trends differ most from those of the surrounding area, making them most likely to be caused by localized geological defects (such as cracks) or structural damage (such as support failure). Prioritizing these points allows for precise identification of localized hazards independent of overall deformation, preventing the overall displacement trend from masking true risks. The displacement rate directly reflects the rate of deformation per unit time; a high rate indicates rapid deterioration at that point, posing a high risk of immediate collapse. After identifying localized anomalies, prioritizing the points with the highest displacement rate buys time for emergency response and prevents further damage. The cumulative displacement rate combines time and displacement to reflect the cumulative effects of long-term deformation. When the Pearson coefficient and displacement rate are indistinguishable, points with high Pearson coefficients indicate significant historical deformation, severe structural damage, and potential instability. Assigning these points as the third priority allows for a comprehensive risk assessment. Therefore, the above priorities are established.

[0158] If there is no vault among the candidate points and the points include side walls and bottom plates, based on the priority of side walls > bottom plates, the point with the smallest average absolute value of the Pearson correlation coefficient in the same position is selected as the significant outlier.

[0159] The data prediction model construction module is used to build a data prediction model based on LSTM, and train the data prediction model with the strain data, displacement data, displacement rate and cumulative displacement rate measured at the previous n moments;

[0160] Based on the above embodiment, the data prediction model is composed of a deep learning network based on a multilayer perceptron. The deep neural network of the multilayer perceptron includes an input layer, a first hidden layer, a second hidden layer, a third hidden layer, and an output layer. The first hidden layer, the second hidden layer, and the third hidden layer each have at least two neurons and each use ReLU (rectified linear unit) as an activation function.

[0161] The strain data, displacement data, displacement rate, and cumulative displacement rate measured at the first n moments are divided into training, validation, and test sets, with a typical ratio of 7:1:2. The training set is used to learn model parameters; the validation set is used to adjust hyperparameters during training to prevent overfitting; and the test set is used to evaluate the model's generalization ability after training.

[0162] In the data prediction model, the input features of the multilayer perceptron deep learning network include: strain data and displacement data of the last moment in the previous n moments;

[0163] The structure of the deep learning network of multilayer perceptron is:

[0164] Input layer: used to receive strain data and displacement data;

[0165] The first hidden layer has 128 neurons and uses ReLU as the activation function.

[0166] The second hidden layer has 64 neurons and also uses the ReLU activation function.

[0167] The third hidden layer has 32 neurons and uses the ReLU activation function.

[0168] Output layer: has 1 neuron, which is used to output the displacement rate and cumulative displacement rate of the next moment.

[0169] The process of training the data prediction model is as follows:

[0170] The strain data and displacement data of the last moment before the significant abnormal point are used as input, and the displacement rate and cumulative displacement rate of the next moment are used as labels to output the training data prediction model. The mean square error is used as the loss function. When the mean square error is When the data is within the range, stop training the data prediction model.

[0171] The simulation module is used to input the strain and displacement data of the significant abnormal point at the current moment into the trained data prediction model to predict the corresponding displacement rate and cumulative displacement rate;

[0172] The weighted calculation and judgment module is used to calculate the deformation risk coefficient of significant abnormal points based on the predicted displacement rate and cumulative displacement rate, and to judge the deformation risk level of the surrounding rock of the cavern based on the deformation risk coefficient.

[0173] Table 1. Changes in deformation risk coefficient with displacement rate and cumulative displacement rate

[0174]

[0175] Based on the above embodiment, according to the data in Table 1, the deformation risk coefficient is positively correlated with the displacement rate and the cumulative displacement rate, which conforms to the physical logic that "the greater the deformation degree, the higher the risk".

[0176] The contribution of the cumulative displacement rate to the deformation risk coefficient is more prominent (for example, in serial number 1-25, the cumulative displacement rate increases from 0.05 to 0.65, and the deformation risk coefficient increases from 0.1 to 1.3, which is faster than the displacement rate), suggesting that its weight coefficient ( ) is greater than the displacement rate weight ( ).

[0177] Depend on Figure 2-Figure 3 As shown in the figure, the deformation risk coefficient closely fits the fitting line, indicating that the two are strongly linearly positively correlated, that is, the greater the displacement rate, the higher the deformation risk coefficient; the deformation risk coefficient is also close to the fitting line, indicating that the cumulative displacement rate and the deformation risk coefficient are also strongly linearly positively correlated, and the greater the cumulative displacement rate, the higher the deformation risk coefficient.

[0178] In summary, the displacement rate and cumulative displacement rate are both significantly positively correlated with the deformation risk coefficient, and the correlation law can be intuitively reflected by the fitting line.

[0179] On the basis of the above embodiment, the deformation risk coefficient of the significant abnormal point is calculated according to the predicted displacement rate and cumulative displacement rate, according to the following formula:

[0180] ;

[0181] in, is the deformation risk coefficient of the significant abnormal point. The deformation risk coefficient is used to evaluate the deformation risk level of the surrounding rock of the tunnel by combining the two index parameters of displacement rate and cumulative displacement rate. The larger the deformation risk coefficient, the higher the deformation risk level of the surrounding rock of the tunnel.

[0182] On this basis, it should be noted that:

[0183] When the displacement rate When it increases, it indicates that the dynamic speed of the surrounding rock deformation is accelerating, and it may enter the stage of accelerated instability, which increases the deformation risk level of the mine surrounding rock, thereby increasing the deformation risk coefficient.

[0184] When the cumulative displacement rate When it increases, it indicates that the cumulative effect of historical damage to the surrounding rock deformation is aggravated. The closer it is to the ultimate bearing state of the material, the higher the deformation risk level of the cavern surrounding rock, which leads to an increase in the deformation risk coefficient.

[0185] In summary, the deformation risk factor and displacement rate and cumulative displacement rate Both are positively correlated.

[0186] In addition, the displacement rate reflects the "dynamic trend" and the cumulative displacement rate reflects the "historical damage". The two have orthogonal physical meanings and no significant causal overlap, which conforms to the independence assumption.

[0187] Furthermore, the effects of displacement rate and cumulative displacement rate on the deformation risk level of the surrounding rock of the mine are independent and additive. In other words, the impact of changes in each indicator parameter on the deformation risk level of the surrounding rock of the mine is independent of the other parameters, and their combined impact can be reflected by simple addition.

[0188] In summary, the above function form is used to express the functional relationship between the deformation risk coefficient and the displacement rate and cumulative displacement rate.

[0189] Where, is the weight coefficient of the predicted displacement rate, is the weight coefficient of the predicted cumulative displacement rate;

[0190] The cumulative displacement rate reflects the irreversible damage accumulation of the surrounding rock due to long-term deformation, and directly determines the degree to which the material approaches the damage threshold. In engineering practice, historical damage cannot be eliminated through short-term regulation, and the cumulative effect may cause chain damage (such as stress concentration → local instability → overall collapse). Therefore, it contributes most directly to the risk level and has the most far-reaching impact. Therefore, setting maximum.

[0191] The displacement rate characterizes the speed of deformation and reflects whether the surrounding rock has entered the stage of accelerated instability. Although the dynamic trend can be monitored and warned in real time, its changing speed may directly determine the urgency of risk evolution, so its weight is higher than the current state abnormality. .

[0192] Furthermore, the weighting principle is: risk assessment should prioritize irreversible historical damage, followed by accelerating damage processes. This approach not only conforms to the physical laws of surrounding rock deformation (damage accumulation → dynamic instability), but also aligns with the priority of long-term monitoring over trend warning in engineering safety management, thereby more scientifically quantifying the deformation risk level.

[0193] Therefore, in On the basis of .

[0194] As an implementation method, The value range is 0-0.5, The value range is 0.5-1. The specific value is set by technical personnel according to actual conditions and is not limited here.

[0195] On the basis of the above embodiment, the deformation risk level of the surrounding rock of the cavern is determined according to the deformation risk coefficient. The specific process is as follows:

[0196] when , the deformation risk level of the surrounding rock of the cavern is low risk;

[0197] when , the deformation risk level of the surrounding rock of the cavern is medium risk;

[0198] when , the risk level of deformation of the surrounding rock of the cavern is high risk.

[0199] in, To classify the critical value between low risk and medium risk, The critical value for dividing medium risk and high risk.

[0200] By statistically analyzing the deformation risk coefficients of similar caverns during normal operation, the mean plus the standard deviation is taken as , collect the deformation risk coefficient data before the accident in the historical accident cases of similar mines (surrounding rock collapse, severe deformation leading to shutdown), extract the critical value of risk out of control, as .

[0201] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0202] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by computer software, electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0203] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0204] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A system for monitoring deformation of surrounding rock of a compressed air energy storage chamber, characterized by: include: The data acquisition module is used to arrange multiple concentric circles equidistantly along the radial direction of the compressed air inlet and outlet areas of the cavern with the center of the pipeline as the origin. Multiple monitoring points are set along the circumference of the same concentric circle. Before the pipeline is put into operation, the circumference of the concentric circle where each monitoring point is located is measured as the initial length of the pipeline. During the current monitoring period, the time series data of the strain and displacement at the monitoring point are obtained at equal time intervals. Based on the time series data, the strain and displacement at the monitoring point at the current time and the previous n times are obtained; The preliminary screening module is used to calculate the displacement-strain change rate ratio of the same concentric circle based on the strain data and displacement data at the monitoring point at the current moment and the previous n moments, compare the change rate ratio with the preset change rate ratio threshold, and screen out monitoring points with abnormal change rate ratios to form an abnormal point set A. Each concentric circle corresponds to a set A; The calculation module is used to calculate the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments based on the displacement data at the current moment and the previous n moments and the initial length of the pipeline; The secondary screening module, for the same concentric circle, determines whether the surrounding rock of the cavern is deformed as a whole or locally through the concentration of abnormal points in set A and the consistency coefficient of displacement direction. For overall deformation, a strain-surrounding rock deformation model is constructed. The strain and displacement data measured at the previous n moments are used to train the strain-surrounding rock deformation model. The strain data at the current moment is input to obtain the displacement prediction value, which is compared with the safety threshold before performing subsequent operations. For local deformation, the Pearson correlation coefficient between the monitoring points on the same concentric circle is calculated and used to mark the local abnormal points to form set B. The intersection of sets A and B is taken to screen out significant abnormal points. The data prediction model construction module is used to build a data prediction model based on LSTM, and train the data prediction model with the strain data, displacement data, displacement rate and cumulative displacement rate measured n moments before the significant abnormal point; The simulation module is used to input the strain and displacement data of the significant abnormal point at the current moment into the trained data prediction model to predict the corresponding displacement rate and cumulative displacement rate; The weighted calculation and judgment module is used to calculate the deformation risk coefficient of significant abnormal points based on the predicted displacement rate and cumulative displacement rate, and to judge the deformation risk level of the surrounding rock of the cavern based on the deformation risk coefficient.

2. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 1 is characterized in that: With the center of the pipeline as the origin, multiple concentric circles are arranged equidistantly along the radial direction of the pipeline, and multiple monitoring points are set along the circumference of the same concentric circle. The specific method is as follows: The surrounding rock strain shows a decreasing trend with increasing radial distance from the pipeline. Based on the scale of the cavern and the diffusion range of the surrounding rock strain, 3 to 5 layers of concentric circles are set. With the center of the pipeline as the origin, inner, middle and outer concentric circles are set in sequence along the radial direction of the pipeline, with a spacing of 1 to 3m. Set 4-8 monitoring points in a single concentric circle with an angular interval of 45°-90°; Inner concentric circles: 0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°, a total of 8 points. The inner concentric circles are the strain concentration areas; Middle concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, forming a "cross orientation" grid. The middle concentric circles are the strain transition zone; Outer concentric circles: 0°, 90°, 180°, 270°, a total of 4 points, the outer concentric circles are the strain attenuation area.

3. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 1 is characterized in that: Based on the strain data and displacement data at the monitoring point at the current moment and the previous n moments, the displacement-strain rate of change ratio is calculated according to the following formula: Combine the current moment and the previous n moments into a set ,in, is the index of the current moment, is the index of the moment; ； in, For the The monitoring point is The displacement-strain rate of change ratio at each moment, For the The monitoring point is The displacement change at each moment, For the The monitoring point is The strain change at each moment, is the index of the monitoring point on the same concentric circle, , is the number of monitoring points on the same concentric circle; The first The monitoring point is The displacement-strain rate of change ratio at each moment and the preset rate of change threshold For comparison, when When , the monitoring point is an abnormal point.

4. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 3 is characterized in that: Calculate the displacement rate and cumulative displacement rate of the monitoring point at the current moment and the previous n moments according to the following formula: ； in, For monitoring point The displacement rate at a moment, For monitoring point The coordinates of a moment, For monitoring point The coordinates of a moment, is the adjacent time interval, is the cumulative displacement rate of the monitoring point, is the coordinate at the current moment, For the The coordinates of a moment, The coordinates of ), The monitoring points are The horizontal and vertical coordinates of the moment, is the initial length of the pipe.

5. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 4 is characterized in that: Through the concentration of abnormal points in set A and the consistency coefficient of displacement direction, it is judged whether the surrounding rock of the cavern is deformed as a whole or locally. The specific steps and the formula are as follows: The plane straight-line distance between any two outlier points in the same concentric circle is: ； in, In the same concentric circle The outlier and The plane distance of the outlier points, 、 Respectively The outlier and The horizontal and vertical coordinates of the abnormal points, , are the indices of outliers in the same concentric circle in set A. ; ； in, is the arithmetic mean of the plane distances between all the outliers in the same concentric circle. , is the number of outliers in set A; ； in, is the concentration coefficient, 、 are the maximum and minimum distances between all outliers respectively; ； in, For the The outlier and The displacement direction consistency coefficient of the abnormal point is 、 Respectively Anomalies in 、 The displacement increment in the direction, 、 Respectively Anomalies in 、 The displacement increment in the direction; ； in, is the average value of the displacement direction consistency coefficient; when and , then the surrounding rock of the tunnel is deformed as a whole; when or , then the surrounding rock of the tunnel is locally deformed; in, is the preset concentration coefficient threshold, It is the preset threshold of the average value of the displacement direction consistency coefficient.

6. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 5 is characterized in that: The Pearson correlation coefficient between monitoring points on the same concentric circle is calculated based on the following formula: ； in, In the same concentric circle Monitoring points and Pearson correlation coefficient of displacement data of each monitoring point; Where, For the The monitoring point is The coordinates of a moment, The coordinates of ), Respectively The monitoring point is The horizontal and vertical coordinates of the moment, For the The monitoring point is The coordinates of a moment, For the The mean coordinate value of the monitoring points, For the The mean coordinate value of the monitoring points, is the index of the monitoring point on the same concentric circle, ; Will and the preset Pearson correlation coefficient threshold When the Pearson correlation coefficients of a monitoring point and other monitoring points in the same concentric circle are all less than When , the monitoring point is determined to be a local abnormal point.

7. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 6 is characterized in that: Take the intersection of sets A and B and filter out significant outliers. The specific steps are as follows: 1) When there is one intersection between set A and set B, the point is identified as a significant outlier; 2) When there are more than two intersections of set A and set B, extract the positions of all points in the intersection, including the vault, side walls, and bottom plate. Sidewall Prioritize the bottom plate and identify significant outliers; If there are multiple points in the same position in the intersection, the significant outlier is determined based on the priority of maximum displacement rate > minimum Pearson correlation coefficient > maximum cumulative displacement rate > minimum concentration; 3) When the intersection of set A and set B is an empty set, determine the significant outliers based on set A or set B separately. The specific steps are as follows: Determine significant outliers based on set A: Extract all outliers in set A and prioritize the vault > sidewall > floor as the most significant outliers. If there are multiple outliers at the same location, the significant outlier is determined based on the priority of maximum displacement rate > maximum cumulative displacement rate > minimum concentration; Determine significant outliers based on set B: For all monitoring points in set B, calculate the average of the absolute values ​​of the Pearson correlation coefficients with other monitoring points on the same concentric circle, sort the average values ​​of the absolute values ​​of the Pearson correlation coefficients in ascending order, and select the top three candidate points; If there are multiple arch vertices among the candidate points, the significant outlier points are determined based on the priority of the smallest average value of the absolute value of the Pearson correlation coefficient > the largest displacement rate > the largest cumulative displacement rate; If there is no vault among the candidate points and the points include side walls and bottom plates, based on the priority of side walls > bottom plates, the point with the smallest average absolute value of the Pearson correlation coefficient in the same position is selected as the significant outlier.

8. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 1 is characterized in that: According to the predicted displacement rate and cumulative displacement rate, the deformation risk coefficient of the significant abnormal point is calculated according to the following formula: ； in, is the deformation risk coefficient of the significant abnormal point; Where, is the displacement rate of the predicted significant outlier point, is the cumulative displacement rate of the predicted significant outlier points; Where, is the weight coefficient of the predicted displacement rate, is the weight coefficient of the predicted cumulative displacement rate, On the basis of .

9. The system for monitoring deformation of surrounding rock of compressed air energy storage chamber according to claim 8, characterized in that: According to the deformation risk coefficient, the deformation risk level of the surrounding rock of the cavern is judged. The specific process is as follows: when , the deformation risk level of the surrounding rock of the cavern is low risk; when , the deformation risk level of the surrounding rock of the cavern is medium risk; when , the deformation risk level of the surrounding rock of the cavern is high risk; in, To classify the critical value between low risk and medium risk, The critical value for dividing medium risk and high risk.

Images