A method for evaluating disturbance degree of ecological diving in coal mining

Abstract

The application discloses a coal mining ecological diving disturbance degree evaluation method, and relates to the technical field of coal mining ecological protection. The method comprises the following steps: collecting environmental data of a coal mining area through deployment of Internet of Things sensors; performing outlier rejection and missing value filling on the collected environmental data, and performing standardization processing on the environmental data; converting the standardized environmental data into time series data, and integrating the time series data into a feature data set; constructing a disturbance prediction model through a deep learning algorithm, inputting the feature data set into the disturbance prediction model, outputting a predicted disturbance risk index based on the input feature data set, classifying according to the predicted disturbance risk index, and dividing a risk interval; formulating corresponding measures according to the risk interval; and regularly feeding the latest real-time data to the disturbance prediction model, and updating and optimizing the disturbance prediction model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coal mining ecological protection, and particularly relates to a method for evaluating disturbance degree of ecological groundwater in coal mining. BACKGROUND

[0002] As one of main energy sources in China, coal has been concerned about ecological disturbance in the process of mining. With the development of technologies such as Internet of Things, big data and artificial intelligence, the evaluation method of disturbance degree of ecological groundwater in coal mining gradually develops from traditional manual observation and experience judgment to intelligent and data-driven direction. The wide application of Internet of Things sensors in coal mining area realizes real-time collection of key environmental data such as soil moisture content, groundwater depth and environmental humidity, which provides data basis for quantitative evaluation of disturbance degree of ecological groundwater. However, the existing evaluation method still has many deficiencies. First, the data preprocessing link often lacks effective processing of abnormal values and missing values, resulting in low data quality and affecting the accuracy of evaluation results. Second, the existing prediction model is mostly based on traditional statistical method, which is difficult to capture the nonlinear relationship between complex environmental factors, and the prediction accuracy is limited. In addition, the existing method lacks systematization and dynamics in risk classification and measure development, and cannot quickly respond and adjust according to real-time data changes. Therefore, it is urgent to develop a method that can comprehensively use Internet of Things technology and deep learning algorithm to realize accurate evaluation and dynamic management of disturbance degree of ecological groundwater in coal mining. SUMMARY

[0003] In view of the above existing problems, the present application is proposed.

[0004] Therefore, the present application provides a method for evaluating disturbance degree of ecological groundwater in coal mining, which solves the problem that the existing prediction model is mostly based on traditional statistical method, which is difficult to capture the nonlinear relationship between complex environmental factors, and the prediction accuracy is limited.

[0005] To solve the above technical problems, the present application provides the following technical solutions:

[0006] In the first aspect, the present application provides a method for evaluating disturbance degree of ecological groundwater in coal mining, which comprises,

[0007] collecting environmental data of coal mining area by deploying Internet of Things sensors;

[0008] removing abnormal values and filling missing values of collected environmental data, and standardizing the environmental data, converting the standardized environmental data into time series data, and integrating the time series data into feature data set;

[0009] A disturbance prediction model is constructed through a deep learning algorithm, a feature data set is input into the disturbance prediction model, a predicted disturbance risk index is output based on the input feature data set, classification is performed according to the predicted disturbance risk index, and a risk interval is divided;

[0010] According to the risk interval, corresponding measures are formulated;

[0011] Periodically feed the latest real-time data to the disturbance prediction model to update and optimize the disturbance prediction model.

[0012] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, the environmental data of the coal mining area is collected by deploying Internet of Things sensors, and the specific steps are as follows:

[0013] Sensors are arranged at the mining operation face, groundwater zone and climate monitoring point in the coal mining area;

[0014] The soil moisture content, groundwater level depth and environmental humidity of the coal mining area are collected by deploying Internet of Things sensors;

[0015] The Internet of Things sensors will periodically collect soil moisture content, groundwater level and environmental humidity, and send them to the central data platform through a wireless network;

[0016] The comprehensive disturbance index D of soil moisture content and groundwater level is calculated according to the collected soil moisture content and groundwater level.

[0017] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, the environmental data collected is subjected to outlier rejection and missing value filling, and the environmental data is subjected to standardization processing, and the standardized environmental data is converted into time series data, and the specific steps are as follows:

[0018] Abnormal data points deviating from the normal range in the environmental data are identified and rejected using an anomaly detection algorithm based on machine learning;

[0019] The missing values in the environmental data are supplemented by mean filling;

[0020] Z-score standardization is used to convert data of different sources and units into a unified standard;

[0021] Each sensor collects environmental data with a time stamp, and the environmental data is synchronized using the time stamp, and the environmental data collected by all sensors is aligned by time, and the collected environmental data is integrated into a two-dimensional matrix, where each row represents a time point and each column represents the environmental data of different sensors, obtaining time series data.

[0022] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, the time series data is integrated into a feature data set, and the specific steps are as follows:

[0023] The disturbance feature of the time series data is extracted, and the expression is as follows:

[0024]

[0025] Wherein, D' represents the disturbance feature value, W* represents the normalized soil moisture content, T* represents the normalized groundwater level depth, and H* represents the normalized environmental humidity.

[0026] Each D' represents a disturbance feature value calculated at a time point, and all disturbance features are collected to obtain a feature data set X.

[0027] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, the time series data is integrated into a feature data set, and the specific steps are as follows:

[0028] The long short-term memory network (LSTM) in the recurrent neural network (RNN) is selected as the basic architecture of the disturbance prediction model, and the adaptive dynamic weight distribution mechanism (ADWM) and attention mechanism (AM) are introduced to enhance the model's attention ability to key time points.

[0029] The adaptive dynamic weight distribution mechanism (ADWM) dynamically adjusts the weights of soil moisture content, groundwater level depth and environmental humidity in the prediction model according to the importance of time series features, and the weight updating formula is as follows:

[0030] w i (t)=w i (t-1)+η*(r i (t)-r i (t-1));

[0031] Wherein, w i (t) represents the weight of the i-th feature at time t, η represents the learning rate, r i (t) represents the correlation score of the i-th feature at time t, and i represents the index variable of the feature.

[0032] A plurality of LSTM layers are set, and each LSTM layer has a plurality of neurons.

[0033] An attention layer is added after the LSTM layer, so that the disturbance prediction model focuses on the most important time points for disturbance prediction.

[0034] A fully connected layer is added after the attention layer, which is used to integrate the features and output the prediction results.

[0035] The feature data set is input into the disturbance prediction model, and a predicted disturbance risk index is output based on the input feature data set, and the specific steps are as follows:

[0036] The feature data set is input into the disturbance prediction model to predict the disturbance risk index, and the expression is as follows:

[0037]

[0038] Wherein, t represents time, δ represents the period of time of collection, represents the predicted disturbance risk index at time t, W*(t) represents the soil moisture content at time t after standardization, T*(t) represents the groundwater level depth at time t after standardization, H*(t) represents the environmental humidity at time t after standardization, exp represents the exponential function, α(t) represents the attention weight at time t, and dt represents the time increment.

[0039] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, wherein: according to the predicted disturbance risk index, the risk interval is classified, and the specific steps are as follows:

[0040] Collect historical environmental data, set adjustment coefficients k1 and k2 according to historical environmental data;

[0041] Set risk threshold θ1 and θ2, and the calculation formula of the risk threshold is:

[0042]

[0043] Wherein, θ1 represents the threshold value of low risk and medium risk, θ2 represents the threshold value of medium risk and high risk, represents the mean value of the predicted disturbance risk index ; represents the standard deviation of the predicted disturbance risk index ;

[0044] A classification method based on probability distribution is adopted, and the interval is divided according to the predicted disturbance risk index;

[0045] When , it is represented as a low-risk interval;

[0046] When , it is represented as a medium-risk interval;

[0047] When , it is represented as a high-risk interval.

[0048] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, wherein: according to the risk interval, corresponding measures are formulated, and the specific steps are as follows:

[0049] When the low-risk interval, continue to maintain the current monitoring frequency, while performing routine environmental management and maintenance measures, the measures taken in the low-risk interval are evaluated, the expression is:

[0050]

[0051] Wherein, E1(t) represents the evaluation index of the low-risk interval, R(t) represents the real-time disturbance risk index;

[0052] When the medium-risk interval, increase the monitoring frequency, pay close attention to the changes of environmental data, take preventive measures, optimize the operation process, evaluate the measures taken in the medium-risk interval, the expression is:

[0053]

[0054] Wherein, E2(t) represents the evaluation index of the medium-risk interval;

[0055] When the high-risk interval, start the emergency response mechanism, immediately take repair measures, carry out soil remediation, water level recovery and vegetation reconstruction, if necessary, suspend the mining operation, carry out comprehensive rectification and repair, evaluate the measures taken in the medium-risk interval, the expression is:

[0056]

[0057] Wherein, E3(t) represents the evaluation index of the high-risk interval;

[0058] The treatment measures taken are continuously monitored.

[0059] As a preferred scheme of the coal mining ecological groundwater disturbance degree evaluation method, wherein: the latest real-time data is fed back to the disturbance prediction model periodically, and the disturbance prediction model is updated and optimized, and the specific steps are:

[0060] The preprocessed real-time environmental data is input into the disturbance prediction model as a training sample, the model performance is continuously optimized, the adaptive learning rate of the disturbance prediction model is adjusted, and the expression is:

[0061]

[0062] Wherein, η(t) represents the learning rate at time t, ∫0 represents the initial learning rate, exp represents the exponential function, λ represents the learning rate adjustment coefficient, and R(t) represents the real-time disturbance risk index.

[0063] In a second aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program, when executed by the processor, implements any step of the coal mining ecological groundwater disturbance degree evaluation method according to the first aspect of the present application.

[0064] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the coal mining ecological groundwater disturbance degree evaluation method according to the first aspect of the present application.

[0065] The present application has the following beneficial effects: the present application collects data in real time through Internet of Things sensors, ensuring high quality and real-time of the data; the disturbance prediction model is constructed through deep learning algorithm, significantly improving the accuracy and reliability of disturbance prediction; corresponding measures are formulated through risk classification, realizing hierarchical management and improving the pertinence and effectiveness of the measures; the prediction efficiency and accuracy are maintained through continuous updating and optimization of the disturbance prediction model; overall, the present application forms a complete, efficient and intelligent evaluation system, significantly improving the scientificity and practicality of coal mining ecological protection, and providing a strong guarantee for ecological stability and sustainable development of the coal mining area. BRIEF DESCRIPTION OF DRAWINGS

[0066] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0067] Figure 1 The flowchart of the coal mining ecological groundwater disturbance degree evaluation method in embodiment 1.

[0068] Figure 2 The schematic diagram of risk interval division in embodiment 1. DETAILED DESCRIPTION

[0069] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification.

[0070] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0071] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, or characteristic under discussion. Thus, "at least one of A and B" means only A, only B, or both A and B. Further, "at least one of A or B" means only A, only B, or both A and B. It should be understood that any two of the above features, structures, or characteristics can be provided even though not all are provided in one particular embodiment.

[0072] Embodiment 1, Reference Figure 1 and Figure 2 The first embodiment of the present application provides a method for evaluating the disturbance degree of ecological groundwater in coal mining, comprising the following steps:

[0073] S1, collecting environmental data of the coal mining area by deploying Internet of Things sensors;

[0074] Sensors are arranged at the mining operation face, groundwater zone and climate monitoring point in the coal mining area;

[0075] S2, collecting soil moisture content, groundwater level depth and environmental humidity of the coal mining area by deploying Internet of Things sensors;

[0076] The Internet of Things sensors will periodically collect soil moisture content, groundwater level and environmental humidity, and send them to the central data platform through a wireless network;

[0077] S3, collecting soil moisture content by soil moisture sensors to monitor changes in water in the soil;

[0078] S4, monitoring changes in groundwater level in the coal mining area by groundwater level sensors;

[0079] S5, measuring changes in humidity in the air by temperature and humidity sensors;

[0080] S6, calculating the comprehensive disturbance index D of soil moisture content and groundwater level according to the collected soil moisture content and groundwater level, the expression is:

[0081]

[0082] Where D represents the comprehensive disturbance index of soil moisture content and groundwater level, W represents soil moisture content, T represents groundwater level depth, and H represents environmental humidity;

[0083] The value range of the comprehensive disturbance index D is [0, ∞), which represents the degree of groundwater disturbance. The greater the value of D, the greater the degree of disturbance of the region, the significant changes in soil humidity and groundwater level, and the unstable ecological environment;

[0084] The smaller the value of D, the smaller the disturbance, and the relatively stable ecological environment.

[0085] S2, outlier rejection and missing value filling are performed on the collected environmental data, and the environmental data is standardized, the standardized environmental data is converted into time series data, and the time series data is integrated into a feature data set;

[0086] Anomaly detection algorithm based on machine learning is used to identify and eliminate abnormal data points in environmental data that deviate from the normal range;

[0087] Mean filling is used to supplement missing values in environmental data;

[0088] Z-score standardization is used to convert data from different sources and units into a unified standard, so that they have the same dimension;

[0089] Each sensor collects environmental data with a timestamp, and the timestamp is used to synchronize the environmental data, align the environmental data collected by all sensors by time, and integrate the collected environmental data into a two-dimensional matrix, where each row represents a time point and each column represents the environmental data of different sensors, obtaining time series data;

[0090] The disturbance feature of the time series data is extracted, and the expression is:

[0091]

[0092] Where D' represents the disturbance feature value, W* represents the normalized soil moisture content, T* represents the normalized groundwater level, and H* represents the normalized environmental humidity;

[0093] The value range of disturbance feature D' is [0, ∞), the larger the value of D', the stronger the disturbance, and the smaller the value of D', the smaller the disturbance;

[0094] Each D' represents a disturbance feature value calculated at a time point, and all disturbance features are collected to obtain a feature data set X, and the expression is:

[0095]

[0096] Where X represents the feature data set, which is a two-dimensional matrix, each row of which represents the feature data at a time point, W n represents the soil moisture content at the nth time point, T n represents the groundwater level at the nth time point, H n represents the environmental humidity at the nth time point, and D' n represents the disturbance feature value at the nth time point, and n represents the total number of time points in the feature data set.

[0097] S3. Build a disturbance prediction model using a deep learning algorithm, input the feature data set into the disturbance prediction model, output a predicted disturbance risk index based on the input feature data set, classify according to the predicted disturbance risk index, and divide the risk range;

[0098] We selected the Long Short-Term Memory (LSTM) network within the recurrent neural network (RNN) as the underlying architecture for the disturbance prediction model. We also introduced the adaptive dynamic weight allocation (ADWM) mechanism and the attention mechanism (AM) to enhance the model's focus on key time points. LSTM can effectively process and memorize long-term dependencies in time series data, making it suitable for prediction tasks.

[0099] The adaptive dynamic weight allocation mechanism (ADWM) dynamically adjusts the weights of soil moisture, groundwater depth, and ambient humidity in the disturbance prediction model based on the importance of time series features. By analyzing the trends and patterns of historical data, the mechanism identifies which disturbance features have a greater impact on phreatic disturbances within a specific time period and increases the weights of these disturbance features accordingly. This not only improves the accuracy of the prediction, but also enhances the adaptability of the disturbance prediction model to changes in different environmental conditions. The weight update formula is as follows:

[0100] w i (t) = w i (t-1)+η*(r i (t)-r i (t-1));

[0101] Among them, w i (t) represents the weight of the i-th feature at time t, η represents the learning rate, r i (t) represents the relevance score of the i-th feature at time t, and i represents the index variable of the feature;

[0102] The relevance score is specifically to use the random forest algorithm to directly output the importance score r of the perturbation feature i (t);

[0103] Set up multiple LSTM layers, each with several neurons, to capture long-term dependencies in time series data;

[0104] Add an attention layer after the LSTM layer to make the disturbance prediction model focus on the most important time points for disturbance prediction;

[0105] Add a fully connected layer after the attention layer to integrate features and output prediction results;

[0106] Input the feature data set into the disturbance prediction model, and output the predicted disturbance risk index based on the input feature data set. The specific steps are as follows:

[0107] The feature data set is input into the disturbance prediction model to predict the disturbance risk index, which is expressed as:

[0108]

[0109] Where t represents time, δ represents the period of acquisition time, represents the predicted disturbance risk index at time t, W*(t) represents the standardized soil moisture content at time t, T*(t) represents the standardized groundwater depth at time t, H*(t) represents the standardized ambient humidity at time t, exp represents the exponential function, α(t) represents the attention weight at time t, and dt represents the time increment;

[0110] Predicted disturbance risk index The range of is [0,∞), when The larger the value of is, the higher the predicted disturbance risk is, and the corresponding repair measures need to be taken. The smaller the value of , the lower the predicted disturbance risk and the relatively stable ecological environment;

[0111] Collect historical environmental data and set adjustment coefficients k1 and k2 based on the historical environmental data;

[0112] Set the risk thresholds θ1 and θ2, and determine the calculation formula for the risk threshold:

[0113]

[0114]

[0115] Among them, θ1 represents the threshold between low risk and medium risk, θ2 represents the threshold between medium risk and high risk, Represents the predicted disturbance risk index The mean of Represents the predicted disturbance risk index The standard deviation of

[0116] A classification method based on probability distribution is used to divide the intervals according to the predicted disturbance risk index;

[0117] when When , it indicates a low risk interval;

[0118] when When , it indicates a medium risk interval;

[0119] when When , it indicates a high risk interval;

[0120] This step realizes the accurate prediction of the ecological disturbance degree of coal mining by constructing a disturbance prediction model based on LSTM and attention mechanism. The multi-layer structure and neuron design of the disturbance prediction model enhance the feature extraction capability, and the attention mechanism improves the sensitivity to key time points, ensuring the accuracy and timeliness of the prediction. The classification method based on probability distribution scientifically divides the risk interval, facilitating the implementation of graded management. The real-time data feedback and model updating mechanism ensures the continuity and adaptability of the prediction, and the dynamically adjusted risk threshold enhances the flexibility and pertinence of risk control. Overall, the present application significantly improves the scientificity and practicality of the evaluation of the ecological disturbance degree of coal mining, providing strong support for ecological protection and sustainable development.

[0121] S4, form corresponding measures according to the risk interval;

[0122] When the low-risk interval, continue to maintain the current monitoring frequency, ensure the continuous collection of environmental data, and at the same time, perform routine environmental management and maintenance measures, and evaluate the measures taken in the low-risk interval, the expression is:

[0123]

[0124] Wherein, E1(t) represents the evaluation index of the low-risk interval, and R(t) represents the real-time disturbance risk index;

[0125] By maintaining the monitoring frequency and performing routine management, the continuity and integrity of the environmental data are ensured, providing a reliable data foundation for subsequent risk analysis, and the calculation formula of E1(t) is simple and clear, facilitating the rapid evaluation of the management effect of the low-risk interval;

[0126] In the low-risk interval, this step effectively avoids excessive intervention, saves resources, and at the same time ensures the stability of the environment;

[0127] When the medium-risk interval, increase the monitoring frequency, pay close attention to the changes in environmental data, take preventive measures, and optimize the operation process to reduce the potential disturbance impact, and evaluate the measures taken in the medium-risk interval, the expression is:

[0128]

[0129] Wherein, E2(t) represents the evaluation index of the medium-risk interval;

[0130] Increasing the monitoring frequency and taking preventive measures can timely discover and respond to potential risks, reduce the disturbance impact, and the formula of E2(t) considers the relative change of the risk index and the threshold, more accurately reflecting the management effect of the medium-risk interval;

[0131] In the medium risk interval, this step effectively prevents the escalation of risk, protects the ecological environment, and reduces potential economic losses;

[0132] When the high risk interval is triggered, the emergency response mechanism is activated, and immediate repair measures are taken to restore the soil, water level, and vegetation, to mitigate the impact of disturbance on the ecological environment. If necessary, mining operations are suspended for comprehensive rectification and repair. The measures taken in the medium risk interval are evaluated, and the expression is:

[0133]

[0134] where E3(t) represents the evaluation index of the high risk interval;

[0135] The emergency response mechanism and comprehensive rectification and repair measures can quickly mitigate the impact of disturbance and restore the ecological environment. The formula for E3(t) quantifies the response effect of the high risk interval by calculating the relative difference between the risk index and the highest threshold, providing a scientific basis for decision-making;

[0136] In the high risk interval, this step effectively avoids ecological disasters and ensures regional ecological safety and sustainable development;

[0137] Continuous monitoring of the measures taken.

[0138] S5, regularly feed the latest real-time data to the disturbance prediction model for updating and optimization of the disturbance prediction model;

[0139] Through the deployment of Internet of Things sensors in the coal mining area, real-time data of soil moisture content, groundwater level depth, and environmental humidity are collected;

[0140] The collected real-time data are cleaned to remove outliers and missing values, and standardized to ensure data consistency and accuracy;

[0141] The preprocessed real-time environmental data are inputted into the disturbance prediction model as training samples to continuously optimize the model performance and adjust the adaptive learning rate of the disturbance prediction model, with the expression:

[0142]

[0143] where η(t) represents the learning rate at time t, ∫0 represents the initial learning rate, exp represents the exponential function, λ represents the learning rate adjustment coefficient, which controls the speed of learning rate reduction, and R(t) represents the real-time disturbance risk index;

[0144] By regularly feeding back real-time environmental data, the disturbance prediction model can continuously absorb the latest information, improving the accuracy and timeliness of the prediction. This step solves the problem that traditional models are difficult to adapt to environmental changes, and realizes the continuous optimization of the disturbance prediction model.

[0145] During coal mining, environmental conditions change constantly, and the real-time environmental data feedback mechanism ensures that the disturbance prediction model always matches the actual environment, providing a guarantee for accurate prediction.

[0146] The value range of the learning rate η(t) at time t is (0, η0). The smaller the value of η(t) is, the slower the learning speed of the disturbance prediction model at the current time point, and the more emphasis on the stability of the parameters. The larger the value of η(t) is, the faster the learning speed of the disturbance prediction model at the current time point, and the more emphasis on the adaptability of the parameters.

[0147] The embodiment also provides a computer device suitable for the coal mining ecological groundwater disturbance degree evaluation method, which comprises a memory and a processor.

[0148] The computer device can be a terminal, and the computer device comprises a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, operator network, NFC (near field communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device. In addition, the input device can be an external keyboard, touchpad or mouse, etc.

[0149] The embodiment also provides a storage medium on which a computer program is stored, the program being executed by a processor to implement the method for evaluating the disturbance degree of coal mining ecological groundwater proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk, or an optical disk.

[0150] To sum up, the present application realizes a complete, efficient and intelligent evaluation system by the following means: the Internet of Things sensors collect data in real time, ensuring the high quality and real-time nature of the data; the disturbance prediction model is constructed by a deep learning algorithm, significantly improving the accuracy and reliability of the disturbance prediction; the corresponding measures are formulated through risk classification, realizing hierarchical management and improving the pertinence and effectiveness of the measures; the prediction efficiency and accuracy are maintained through continuous updating and optimization of the disturbance prediction model. Overall, the present application significantly improves the scientificity and practicality of the ecological protection of coal mining and provides a strong guarantee for the ecological stability and sustainable development of the coal mining area.

[0151] Embodiment 2, referring to Table 1, is a second embodiment of the present application, which gives experimental simulation data of the method for evaluating the disturbance degree of coal mining ecological groundwater, for further verifying the technical solution of the present application.

[0152] In this embodiment, a certain coal mining area is selected as the test object, which has typical characteristics of ecological disturbance of coal mining. A total of 100 Internet of Things sensors are arranged at the mining operation face, groundwater zone and climate monitoring point, including soil moisture sensors, underground water level sensors and temperature and humidity sensors. The sensors periodically collect soil moisture content, underground water level depth and environmental humidity data and send them to the central data platform through a wireless network.

[0153] The data acquisition period is set to be once every 6 hours, and 1440 groups of data are obtained by continuously collecting for 30 days. The data are cleaned by using an abnormality detection algorithm based on machine learning, and abnormal values and missing values are removed. The missing values are supplemented by using a mean filling method. Subsequently, the data are converted into a unified standard by using a Z-score standardization method, so as to ensure the consistency and accuracy of the data.

[0154] According to the standardized data, the comprehensive disturbance index D and the disturbance characteristic value D' of the soil water content and the groundwater level are calculated, the characteristic data set X is obtained by collecting all the disturbance characteristic values, the LSTM network is selected as the basic framework of the disturbance prediction model, the attention mechanism is introduced to enhance the attention ability of the model to the key time points, the characteristic data set X is input into the disturbance prediction model, and the predicted disturbance risk index and the risk classification is carried out according to the predicted disturbance risk index.

[0155] According to the historical data and expert experience, the threshold value θ1 of low risk and medium risk is set to be 0.5, the threshold value θ2 of medium risk and high risk is set to be 0.8, the classification method based on probability distribution is adopted, the predicted disturbance risk index is divided into three intervals of low risk, medium risk and high risk, corresponding management measures are developed according to different risk intervals, the measures are evaluated, finally, the latest real-time data are fed back to the disturbance prediction model, and the model is updated and optimized.

[0156] Specifically, as shown in the following table 1:

[0157] Table 1 part of the experimental record table

[0158]

[0159] As can be seen from the experimental data table, under the condition that the soil water content, the groundwater level depth and the environmental humidity are the same, the comprehensive disturbance index D and the disturbance characteristic value D' calculated by the present application are lower than those of the prior art, which indicates that the evaluation of the disturbance degree of the present application is more accurate, and different levels of disturbance can be more effectively identified and distinguished.

[0160] Further analysis of the predicted disturbance risk index It can be seen that the prediction results of the present application in multiple test objects are lower than those of the prior art, which indicates that the disturbance prediction model constructed by the present application has higher accuracy and reliability, especially in the high risk interval, the prediction result of the present application is significantly lower than that of the prior art, which indicates that the present application can identify potential high risks earlier, and provides a scientific basis for timely taking repair measures.

[0161] By comparing the risk classification results, the present application realizes the down-regulation of the risk level in multiple test objects, that is, in the case of medium risk determined by the prior art, the present application determines low risk; in the case of high risk determined by the prior art, the present application determines medium risk, which shows that the measures taken by the present application are more effective and can better control the disturbance degree and protect the ecological environment.

[0162] In summary, through accurate data processing, scientific model construction and effective risk classification, the present application significantly improves the scientificity and practicability of the evaluation of the ecological disturbance degree of coal mining, and provides strong support for the ecological protection and sustainable development of the coal mining area.

[0163] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting, and although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A method for evaluating the disturbance degree of ecological diving in coal mining, characterized in that: The application relates to a coal mining area disturbance prediction method and device. environmental data of a coal mining area is collected through deployment of Internet of Things sensors; outlier values in the collected environmental data are removed, missing values are filled, and the environmental data is standardized, and the standardized environmental data is converted into time series data; the time series data is integrated into a feature data set, and the specific steps are as follows, a disturbance feature is extracted from the time series data, and the expression is as follows: wherein D' represents a disturbance feature value, W* represents standardized soil moisture, T* represents standardized groundwater depth, and H* represents standardized environmental humidity; each D ' represents a disturbance eigenvalue calculated at a time point, and all disturbance eigenvalues are collected to obtain a feature dataset X; a disturbance prediction model is constructed through a deep learning algorithm, the feature data set is input into the disturbance prediction model, and a predicted disturbance risk index is output based on the input feature data set, and the specific steps are as follows, a long short-term memory network (LSTM) in a recurrent neural network (RNN) is selected as the basic framework of the disturbance prediction model, and an adaptive dynamic weight distribution mechanism (ADWM) and an attention mechanism (AM) are introduced to enhance the attention ability of the model to key time points; the ADWM dynamically adjusts the weights of soil moisture, groundwater depth and environmental humidity in the prediction model according to the importance of the time series features, and the weight updating formula is as follows: w i (t) = w i (t - 1) + η * (r i (t) - r i (t - 1)); where w i (t) denotes the relevance score of the i-th feature at time t, i denotes the index variable of the feature; i (t) denotes the relevance score of the i-th feature at time t, i denotes the index variable of the feature; a plurality of LSTM layers are set, and each LSTM layer has a plurality of neurons; an attention layer is added after the LSTM layer, so that the disturbance prediction model focuses on the time points that are most important for disturbance prediction; a fully connected layer is added after the attention layer, which is used for integrating features and outputting prediction results; the feature data set is input into the disturbance prediction model, and a predicted disturbance risk index is output based on the input feature data set, and the specific steps are as follows: the feature data set is input into the disturbance prediction model to predict the disturbance risk index, and the expression is as follows: wherein t represents time, and δ represents a period of time of acquisition, represents the predicted disturbance risk index at time t, W*(t) represents the normalized soil moisture content at time t, T*(t) represents the normalized groundwater level depth at time t, H*(t) represents the normalized ambient humidity at time t, exp represents the exponential function, a(t) represents the attention weight at time t, and dt represents the time increment; classification is performed according to the predicted disturbance risk index, and a risk interval is divided; corresponding measures are formulated according to the risk interval; the latest real-time data is regularly fed back to the disturbance prediction model for updating and optimization of the disturbance prediction model.

2. The method for evaluating the degree of disturbance of ecological diving in coal mining according to claim 1, characterized in that: The environmental data of the coal mining area is collected through deployment of Internet of Things sensors, and the specific steps are as follows: sensors are arranged on a mining operation surface, a phreatic zone and a climate monitoring point in the coal mining area; soil moisture, groundwater depth and environmental humidity in the coal mining area are collected through deployment of Internet of Things sensors; the Internet of Things sensors periodically collect soil moisture, groundwater depth and environmental humidity, and send the data to a central data platform through a wireless network; a comprehensive disturbance index D of soil moisture and groundwater depth is calculated according to the collected soil moisture and groundwater depth.

3. The method according to claim 2, wherein the method is characterized by: The collected environmental data is subjected to outlier value removal and missing value filling, and the environmental data is standardized, and the standardized environmental data is converted into time series data, and the specific steps are as follows: an abnormal data point deviating from a normal range in the environmental data is identified and removed through an abnormality detection algorithm based on machine learning; missing values in the environmental data are supplemented through mean filling; data of different sources and units are converted into a unified standard through Z-score standardization. The environmental data collected by each sensor is time-stamped, and the environmental data is synchronized using the time stamp, aligning the environmental data collected by all sensors by time, integrating the collected environmental data into a two-dimensional matrix, where each row represents a time point and each column represents the environmental data of different sensors, obtaining time series data.

4. The method according to claim 1, wherein the method is characterized by: The step of classifying according to the predicted disturbance risk index and dividing the risk interval is specifically as follows: Collect historical environmental data, and set adjustment coefficients k1 and k2 according to the historical environmental data; Set risk threshold values θ1 and θ2, and the calculation formula of the risk threshold values is: wherein θ1 represents the threshold value between low risk and medium risk, and θ2 represents the threshold value between medium risk and high risk, represents the mean of the predicted perturbation risk index , represents the standard deviation of the predicted perturbation risk index . A classification method based on probability distribution is adopted to divide the interval according to the predicted disturbance risk index; When low risk interval; When moderate risk interval; When a high risk interval is indicated.

5. The method for evaluating the degree of disturbance of coal mining ecological diving according to claim 1, characterized in that: The step of formulating corresponding measures according to the risk interval is specifically as follows: When the low-risk interval, continue to maintain the current monitoring frequency, and perform routine environmental management and maintenance measures, and evaluate the measures taken in the low-risk interval, and the expression is: Wherein, E1(t) represents the evaluation index of the low-risk interval, and R(t) represents the real-time disturbance risk index; When the medium-risk interval, increase the monitoring frequency, pay close attention to the change of the environmental data, take preventive measures, and optimize the operation process, evaluate the measures taken in the medium-risk interval, and the expression is: Wherein, E2(t) represents the evaluation index of the medium-risk interval; When the high-risk interval, start the emergency response mechanism, immediately take repair measures, perform soil remediation, water level recovery and vegetation reconstruction, and if necessary, suspend the mining operation, perform comprehensive rectification and repair, and evaluate the measures taken in the medium-risk interval, and the expression is: Wherein, E3(t) represents the evaluation index of the high-risk interval; The taken treatment measures are continuously monitored.

6. The method for evaluating the degree of disturbance of coal mining ecological diving according to claim 1, characterized in that: The step of periodically feeding the latest real-time data to the disturbance prediction model for updating and optimizing the disturbance prediction model is specifically as follows: The preprocessed real-time environmental data is input into the disturbance prediction model as a training sample to continuously optimize the model performance, and the adaptive learning rate of the disturbance prediction model is adjusted, and the expression is: Wherein, η(t) represents the learning rate at time t, η0 represents the initial learning rate, exp represents the exponential function, λ represents the learning rate adjustment coefficient, and R(t) represents the real-time disturbance risk index. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to realize the steps of the coal mining ecological groundwater disturbance degree evaluation method according to any one of claims 1-6.

8. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the coal mining ecological groundwater disturbance degree evaluation method according to any one of claims 1-6.

Images