A method for evaluating the water environment of a mining subsidence area

Abstract

The present invention discloses a water environment evaluation method for a mining subsidence area, belonging to the technical field of water resource evaluation in mining areas, including: collecting water quality parameters and movement parameters of the water environment in the mining subsidence area; performing data dimensionality reduction on the water quality parameters and movement parameters based on a stepwise regression model to obtain a water quality data set; constructing a long short-term memory network, initializing the parameters of the long short-term memory network through an ant colony algorithm, and then inputting a sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model; obtaining a weight factor of the water environment; and constructing a water environment evaluation system for the mining subsidence area based on the weight factor of the water environment. The present invention can improve the accuracy and prediction precision of the water environment in the mining subsidence area. By comprehensively applying LSTM and the ant colony algorithm, it can adapt to the water environments of different types of mining subsidence areas and has strong versatility.

Description

Technical Field

[0001] The present invention belongs to the field of water resource evaluation in mining areas, and particularly relates to a method for evaluating the water environment in a subsidence area of a mining area. Background Art

[0002] A subsidence area in a mining area refers to an area where the ground surface or underground undergoes subsidence or collapse due to underground resource mining activities. This phenomenon is usually related to the extraction of ore, minerals, or coal seams in mining. When underground cavities form, the supporting rock strata break or are squeezed, or surface collapse occurs due to mining, it may lead to subsidence in the mining area. Subsidence in mining areas is a common problem during the mining process, especially in coal mining or mineral mining, because the extraction of minerals can cause the formation of underground cavities, leading to surface collapse or subsidence. This may have a serious impact on the local environment, infrastructure, and community. For example, ground subsidence can cause damage to buildings, roads, and pipelines, and a decline in the groundwater level may affect water resource supply and the ecosystem.

[0003] Currently, subsidence in mining areas may cause the groundwater level to drop or rise, which may have a direct impact on the nearby groundwater system and related ecosystems. Changes in the water environment may have direct or indirect effects on the ecosystems in nearby wetlands, rivers, and other water bodies, including aquatic plants, fish, and other aquatic organisms. Therefore, there is an urgent need for a method for evaluating the water environment in a subsidence area of a mining area to provide a scientific means for on-site workers to monitor and conduct on-site evaluations. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for evaluating the water environment in a subsidence area of a mining area to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above purpose, the present invention provides a method for evaluating the water environment in a subsidence area of a mining area, including:

[0006] Collecting water quality parameters and movement parameters of the water environment in the subsidence area of the mining area;

[0007] Based on a stepwise regression model, performing data dimensionality reduction on the water quality parameters and movement parameters to obtain a water quality data set;

[0008] Constructing a long short-term memory network, initializing the parameters of the long short-term memory network through an ant colony algorithm, and then inputting a sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model;

[0009] Inputting the water quality data set into the water quality parameter weight calculation model for calculation to obtain a weight factor of the water environment;

[0010] Based on the weight factor of the water environment, constructing a water environment evaluation system for the subsidence area of the mining area.

[0011] Preferably, the process of collecting water quality parameters and motion parameters of the water environment in the subsidence area of the mining area includes:

[0012] Collect water samples from the subsidence area of the mining area, conduct chemical analysis on the water samples, and obtain the water quality parameters;

[0013] Based on motion sensors, conduct motion monitoring on the water environment in the subsidence area of the mining area, and obtain the motion parameters;

[0014] The water quality parameters include temperature, turbidity, conductivity, dissolved oxygen in water, ammonia nitrogen, nitrite, nitrate, total phosphorus, total nitrogen, and microorganism species;

[0015] The motion parameters include the velocity and direction of water flow, the groundwater level and the height of the water surface, the terrain of the subsidence area, and the average precipitation in different seasons.

[0016] Preferably, the process of reducing the dimensionality of the water quality parameters and motion parameters based on the stepwise regression model to obtain the water quality data set includes:

[0017] Conduct regression analysis on the water quality parameters and motion parameters to obtain a regression model affected by multiple parameters;

[0018] Based on the stepwise regression model, interpret and select the regression model affected by multiple parameters to obtain the water quality data set.

[0019] Preferably, the process of interpreting and selecting the regression model affected by multiple parameters based on the stepwise regression model to obtain the water quality data set includes:

[0020] Based on the AIC evaluation criterion and the forward stepwise regression method, add or delete features of the regression model affected by multiple parameters to obtain a set of processed models;

[0021] Conduct cross-validation on the set of processed models to obtain a validation result;

[0022] Based on the water quality parameters, set a threshold for the number of features, and based on the threshold for the number of features and the validation result, select the regression model affected by multiple parameters to obtain the best selection model;

[0023] Conduct data inversion on the best selection model to obtain the water quality data set.

[0024] Preferably, the process of constructing a long short-term memory network, initializing the parameters of the long short-term memory network through the ant colony algorithm, and then inputting the sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model includes:

[0025] Construct a long short-term memory network, optimize the parameters of the long short-term memory network based on the ant colony algorithm to obtain hyperparameters;

[0026] Import the hyperparameters into the long short-term memory network to obtain an optimized long short-term memory network, input the sample set of the water quality parameters into the optimized long short-term memory network for training to obtain a weight calculation model;

[0027] Verify the performance of the weight calculation model and output it to obtain the water quality parameter weight calculation model.

[0028] Preferably, the process of optimizing the parameters of the long short-term memory network based on the ant colony algorithm to obtain hyperparameters includes:

[0029] Set the number of ants, the number of iterations and the initial pheromone of the ant colony algorithm;

[0030] In each round, start all ants to obtain pheromone and heuristic information;

[0031] The ants in the ant colony construct a feasible solution for the optimization task of the long short-term memory network according to the heuristic information and the pheromone to obtain optimal parameters;

[0032] Construct an evaluation function, evaluate the optimal parameters based on the evaluation function to obtain the hyperparameters.

[0033] Preferably, the process of inputting the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factor of the water environment includes:

[0034] Input the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factors of each water quality parameter;

[0035] Map the weight factors of the water quality parameters to the importance range based on the hyperbolic tangent function mapping method to obtain standardized weight factors;

[0036] Interpret the standardized weight factors to obtain an interpretation result, and determine the weight factor of the water environment based on the interpretation result.

[0037] Preferably, the process of constructing a water environment evaluation system for the mining subsidence area based on the weight factor of the water environment includes:

[0038] Multiply the weight factor of the water environment by the values of its corresponding water quality parameters and motion parameters, and then add the results to obtain a weighted index value;

[0039] Construct a fuzzy evaluation model, input the weighted index value into the fuzzy evaluation model to obtain the water environment evaluation system of the mining subsidence area.

[0040] The technical effects of the present invention are as follows:

[0041] Since the long short-term memory network is more suitable for long sequence data of the water environment in the mining subsidence area, the present invention calculates the weights of water quality parameters through the long short-term memory network, and initializes the parameters of the long short-term memory network through the ant colony algorithm. Combining the long short-term memory network and the ant colony algorithm can more comprehensively consider the temporal and spatial information in the water environment data, which helps to reveal the complex dynamic change law of the water environment. At the same time, LSTM can capture temporal patterns, and the ant colony algorithm has global search ability. Combining the two can improve the accuracy and prediction accuracy of the water environment in the mining subsidence area. By comprehensively using LSTM and the ant colony algorithm, it can adapt to different types of water environments in the mining subsidence area and has strong versatility. Brief Description of the Drawings

[0042] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0043] Figure 1 It is a flowchart of the water environment evaluation method for the mining subsidence area in the embodiment of the present invention;

[0044] Figure 2 It is a schematic diagram of the long short-term memory network in the embodiment of the present invention. Detailed Embodiments

[0045] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0046] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0047] Embodiment 1

[0048] As Figure 1 shown, this embodiment provides a water environment evaluation method for a mining subsidence area, including:

[0049] Collect the water quality parameters and motion parameters of the water environment in the mining subsidence area;

[0050] Based on the stepwise regression model, perform data dimensionality reduction on the water quality parameters and motion parameters to obtain a water quality data set;

[0051] Construct a long short-term memory network. After initializing the parameters of the long short-term memory network through the ant colony algorithm, input the sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model; the ant colony algorithm can be used to optimize the hyperparameters of the LSTM, such as the learning rate, regularization parameter, etc. By searching in the hyperparameter space, the ant colony algorithm can help find a better hyperparameter configuration and improve the performance of the LSTM.

[0052] Input the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factor of the water environment;

[0053] Construct a water environment evaluation system for the mining subsidence area based on the weight factor of the water environment.

[0054] Preferably, the process of collecting the water quality parameters and motion parameters of the water environment in the mining subsidence area includes:

[0055] Collect water samples from the mining subsidence area and conduct chemical analysis on the water samples to obtain the water quality parameters;

[0056] Based on motion sensors, conduct motion monitoring on the water environment in the mining subsidence area to obtain the motion parameters;

[0057] The water quality parameters include temperature, turbidity, conductivity, dissolved oxygen in water, ammonia nitrogen, nitrite, nitrate, total phosphorus, total nitrogen, and microorganism species;

[0058] Specifically:

[0059] Physical property monitoring: including temperature, turbidity, conductivity, etc.

[0060] Chemical property monitoring: including parameters such as dissolved oxygen in water, ammonia nitrogen, nitrite, nitrate, total phosphorus, and total nitrogen.

[0061] Microorganism monitoring: Detect the presence of bacteria, algae, and other microorganisms in water.

[0062] Water quantity and water flow monitoring:

[0063] The motion parameters include the speed and direction of water flow, the groundwater level and the height of the water surface, the terrain of the subsidence area, and the average precipitation in different seasons.

[0064] Flow monitoring: Determine the speed and direction of water flow.

[0065] Water level monitoring: Determine the groundwater level and the height of the water surface.

[0066] Watershed analysis: Consider the terrain and precipitation of the watershed to evaluate runoff and groundwater recharge.

[0067] Preferably, the process of reducing the dimensionality of the water quality parameters and motion parameters based on the stepwise regression model to obtain a water quality data set includes:

[0068] Perform a regression analysis on the water quality parameters and motion parameters to obtain a regression model affected by multiple parameters;

[0069] Based on the stepwise regression model, interpret and select the regression model affected by the multiple parameters to obtain the water quality data set.

[0070] Preferably, the process of interpreting and selecting the regression model affected by the multiple parameters based on the stepwise regression model to obtain the water quality data set includes:

[0071] Based on the AIC evaluation criterion and the forward stepwise regression method, add or delete features of the regression model affected by the multiple parameters to obtain a set of processed models;

[0072] The core step of stepwise regression is to select or eliminate one feature in each step to maximize (or minimize) the performance of the model. There are usually two methods: forward stepwise regression and backward stepwise regression. In each step, the model decides whether to add or delete a feature according to the AIC evaluation criterion.

[0073] Perform cross-validation on the set of processed models to obtain a validation result;

[0074] Based on the water quality parameters, set a threshold for the number of features, and select the regression model affected by the multiple parameters based on the threshold of the number of features and the validation result to obtain an optimal selection model;

[0075] Perform data inversion on the optimal selection model to obtain the water quality data set.

[0076] Preferably, the process of constructing a long short-term memory network, initializing the parameters of the long short-term memory network through the ant colony algorithm, and then inputting a sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model includes:

[0077] Construct a long short-term memory network, optimize the parameters of the long short-term memory network based on the ant colony algorithm to obtain hyperparameters;

[0078] The long short-term memory network includes:

[0079] Cell State: LSTM introduces a cell state for transmitting and storing information along the sequence. The cell state can be regarded as the main carrier of the memory or information learned by the network and can be transmitted throughout the sequence.

[0080] Input Gate: The input gate determines which information should be updated into the cell state. It contains a Sigmoid activation function to control the degree of information update for each part.

[0081] Forget Gate: The forget gate determines how much previous information should be retained in the cell state. It uses a Sigmoid activation function to decide whether to discard or retain the information of each part.

[0082] Output Gate: The output gate determines which part of the cell state will be used as the output. Based on the current input and the previous hidden state, it uses a Sigmoid activation function and a Tanh activation function to generate the final output.

[0083] Hidden State: The hidden state is the main output of the LSTM network and is also one of the inputs for the next moment. It contains the information of the current moment, as well as partial information of the previous hidden state and the cell state.

[0084] Import the hyperparameters into the long short-term memory network to obtain an optimized long short-term memory network. Input the sample set of the water quality parameters into the optimized long short-term memory network for training to obtain a weight calculation model.

[0085] Verify the performance of the weight calculation model and then output it to obtain the water quality parameter weight calculation model.

[0086] Preferably, the process of optimizing the parameters of the long short-term memory network based on the ant colony algorithm to obtain hyperparameters includes:

[0087] Set the number of ants, the number of iterations, and the initial pheromone of the ant colony algorithm.

[0088] In each round, start all ants to obtain pheromone and heuristic information.

[0089] The specific process is as follows:

[0090] Initialize the ant colony parameters. Set the number of ants in the ant colony as m, the maximum number of iterations as Nmax, and the initial pheromone of all elements in the set as c. In each round, start all ants. For a certain ant k, at time t, when selecting an element from a certain set Ii, where 1 ≤ i ≤ 121, it will be selected according to the pheromone concentration of the elements in the set. Calculate the probability of the j-th element being selected and use the proportional selection method for probabilistic random selection.

[0091] The ants in the ant colony construct a feasible solution for the optimization task of the long short-term memory network according to the heuristic information and the pheromone to obtain the optimal parameters.

[0092] Construct an evaluation function, evaluate the preferred parameters based on the evaluation function, and obtain the hyperparameters.

[0093] Preferably, the process of inputting the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factor of the water environment includes:

[0094] Input the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factors of each water quality parameter;

[0095] Map the weight factors of the water quality parameters to the importance range based on the hyperbolic tangent function mapping method to obtain the standardized weight factors;

[0096] The hyperbolic tangent function is another commonly used mapping method that maps the input to the range (-1, 1):

[0097] [\text{Tanh}(x)=\frac{e^{2x}-1}{e^{2x}+1}];

[0098] where (x) is the original weight value.

[0099] Interpret the standardized weight factors to obtain the interpretation results, and determine the weight factors of the water environment based on the interpretation results.

[0100] Preferably, the process of constructing the water environment evaluation system for the mining subsidence area based on the weight factors of the water environment includes:

[0101] Multiply the weight factors of the water environment by the values of their corresponding water quality parameters and motion parameters, and then add the results to obtain the weighted index value;

[0102] Multiply the numerical value of each evaluation index by its corresponding weight factor, and then add the results to obtain the weighted index value. It is expressed by the following formula:

[0103] [\text{Weighted index value}=w_1\times x_1 + w_2\times x_2+\ldots+w_n\times x_n];

[0104] Construct a fuzzy evaluation model, input the weighted index value into the fuzzy evaluation model, and obtain the water environment evaluation system for the mining subsidence area.

[0105] Input the weighted index value into the evaluation model, and obtain the final water environment evaluation result according to the preset evaluation criteria. The evaluation model can be a simple weighted summation model or a more complex model, such as a fuzzy evaluation model or a neural network model. The specific choice depends on the complexity of the problem and the nature of the data. In this embodiment, a fuzzy evaluation model is adopted.

[0106] As described above, the above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for evaluating the water environment in a mining subsidence area, characterized in that: The following steps are involved: Collect water quality parameters and movement parameters of the water environment in the mining subsidence area; Performing data dimension reduction on the water quality parameters and motion parameters based on a stepwise regression model to obtain a water quality data set; Constructing a long short-term memory network, initializing the parameters of the long short-term memory network by using an ant colony algorithm, and then inputting a sample set of water quality parameters into the long short-term memory network for training to obtain a water quality parameter weight calculation model; Inputting the water quality data set into the water quality parameter weight calculation model for calculation to obtain a weight factor of the water environment; Constructing a water environment evaluation system for mining subsidence areas based on the weight factors of the water environment; The process of constructing a long short-term memory network, initializing the parameters of the long short-term memory network by using an ant colony algorithm, inputting a sample set of water quality parameters into the long short-term memory network for training, and obtaining a water quality parameter weight calculation model includes: Constructing a long short-term memory network, optimizing the parameters of the long short-term memory network based on an ant colony algorithm to obtain hyperparameters; Importing the hyperparameters into the long short-term memory network to obtain an optimized long short-term memory network, inputting the sample set of water quality parameters into the optimized long short-term memory network for training, and obtaining a weight calculation model; The performance of the weight calculation model is verified and then outputted to obtain the water quality parameter weight calculation model; The process of optimizing the parameters of the long short-term memory network based on the ant colony algorithm to obtain the hyperparameters includes: Set the number of ants, number of iterations and initial pheromone of the ant colony algorithm; In each round, all ants are activated to obtain pheromone and heuristic information; Ants in the ant colony construct feasible solutions for the optimization task of the long short-term memory network according to the heuristic information and the pheromone to obtain optimal parameters; Constructing an evaluation function, and evaluating the preferred parameters based on the evaluation function to obtain the hyperparameters; The water quality data set is input into the water quality parameter weight calculation model for calculation, and the process of obtaining the weight factor of the water environment includes: Inputting the water quality data set into the water quality parameter weight calculation model for calculation to obtain the weight factor of each water quality parameter; Based on the hyperbolic tangent function mapping method, the weight factor of the water quality parameter is mapped to the importance range to obtain a standardized weight factor; Interpreting the standardized weight factor to obtain an interpretation result, and determining the weight factor of the water environment based on the interpretation result; The process of constructing a water environment evaluation system for mining subsidence areas based on the weight factors of the water environment includes: Multiplying the weight factor of the water environment by the values ​​of the corresponding water quality parameter and motion parameter, and then adding the results to obtain a weighted index value; A fuzzy evaluation model is constructed, and the weighted index value is input into the fuzzy evaluation model to obtain a water environment evaluation system for the mining subsidence area.

2. The method for evaluating the water environment of a mining subsidence area according to claim 1, characterized in that: The process of collecting water quality parameters and movement parameters of the water environment in the mining area subsidence area includes: Collecting water samples from the subsidence area of ​​the mining area, performing chemical analysis on the water samples, and obtaining the water quality parameters; Performing motion monitoring on the water environment of the mining area subsidence area based on a motion sensor to obtain the motion parameters; The water quality parameters include temperature, turbidity, conductivity, dissolved oxygen, ammonia nitrogen, nitrite, nitrate, total phosphorus, total nitrogen and microbial species; The movement parameters include the speed and direction of water flow, the height of the groundwater level and the surface of the water body, the topography of the subsidence area and the average precipitation in different seasons.

3. The water environment assessment method for a mining subsidence area according to claim 1, characterized in that: The process of performing data dimension reduction on the water quality parameters and motion parameters based on the stepwise regression model to obtain a water quality data set includes: Performing regression analysis on the water quality parameters and movement parameters to obtain a regression model of the influence of multiple parameters; The water quality data set is obtained by interpreting and selecting a regression model for the influence of the multiple parameters based on a stepwise regression model.

4. The method for evaluating the water environment of a mining subsidence area according to claim 3, characterized in that: The process of interpreting and selecting the regression model for the influence of the multiple parameters based on the stepwise regression model to obtain the water quality data set includes: Adding or deleting features to the regression model affected by the multiple parameters based on the AIC evaluation standard and the forward stepwise regression method to obtain a set of processing models; Cross-validating the set of processing models to obtain validation results; Setting a threshold value of the number of features based on the water quality parameters, selecting a regression model for the influence of the multiple parameters based on the threshold value of the number of features and the verification results, and obtaining the best selected model; Performing data inversion on the best selection model to obtain the water quality data set.

Images