A multi-source mine water quality prediction method, device, equipment, medium and product

Abstract

The present application discloses a multi-source mine water quality prediction method, device, equipment, medium and product, relating to the technical field of coal mine water quality prediction. The method comprises obtaining a working face water source affected by mining fissures; collecting samples of the working face water source according to a set sample collection time interval; analyzing the collected water source samples using a Pearson algorithm to obtain a dataset of main characteristic indicators of mine water quality monitored at each source layer in the working face water source; based on the dataset of main characteristic indicators of mine water quality monitored at each source layer, using a trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to predict water quality for each source layer, thereby obtaining water quality prediction data for each source layer; the prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm. The method provided by the present application can improve the accuracy and reliability of water quality prediction.

Description

Technical Field

[0001] The present application relates to the technical field of coal mine water quality prediction, and in particular to a multi-source mine water quality prediction method, device, equipment, medium and product. Background Art

[0002] Mine water comes from diverse and complex sources, primarily including water accumulated in overlying goafs, water from overburden aquifers, and surface water. Water accumulated in goafs primarily originates from the fracturing of rock strata during coal mining, which in turn creates water channels between the goaf and the overburden aquifer. When groundwater seeps into the goaf, it causes water accumulation. This water, influenced by hydrochemical and microbiological factors, exhibits high heterogeneity and spatiotemporal complexity. It can also serve as a water source for influx into the working face after subsequent coal seam mining. Overburden aquifers are water-rich strata within the subsurface rock and unconsolidated formations, typically formed by the accumulation of water from groundwater layers. Their water quality and properties are influenced by multiple factors, including the mineral composition of the aquifer, hydrogeological conditions, and the surrounding environment. During mining operations at the working face, cracks and delaminations in the overburden create new water channels, allowing water from the overburden aquifer, affected by mining, and even surface water, to enter the goaf, becoming a significant component of mine water.

[0003] In recent years, researchers have attempted to use models and algorithms such as autoregressive averaging, support vector machines, and grey theory to predict water quality, achieving some success. However, research remains to be done on how to effectively address the chaotic nature of water quality data and conduct time-series predictions of water quality in mining areas, taking into account the source of mine water. Summary of the Invention

[0004] The purpose of this application is to provide a multi-source mine water quality prediction method, device, equipment, medium and product, which performs water quality prediction based on the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model, and can improve the accuracy and reliability of water quality prediction.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a multi-source mine water quality prediction method, the multi-source mine water quality prediction method comprising:

[0007] Obtain water sources for the working face affected by mining fissures; the water sources for the working face mainly originate from old goaf areas and several overburden aquifers.

[0008] Samples are collected from the working face water source according to the set sample collection time interval.

[0009] The collected water source samples were analyzed using the Pearson algorithm to obtain a data set of main characteristic indicators of mine water quality monitored in each source layer of the working face water source.

[0010] Based on the data set of main characteristic indicators of mine water quality monitored at each source layer, the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used to predict the water quality of each source layer to obtain the water quality prediction data of each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm.

[0011] In a second aspect, the present application provides a multi-source mine water quality prediction device, the multi-source mine water quality prediction device comprising:

[0012] The acquisition module is used to obtain the water source of the working face affected by mining fractures; the water source of the working face mainly comes from the old goaf area and several overburden aquifers.

[0013] The sample collection module is used to collect samples from the working face water source according to the set sample collection time interval.

[0014] The sample analysis module is used to analyze the collected water source samples using the Pearson algorithm to obtain a data set of main characteristic indicators of mine water quality monitored in each source layer of the working face water source.

[0015] The water quality prediction module is used to predict the water quality of each source layer based on the main characteristic indicator data set of mine water quality monitored at each source layer, using the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to obtain water quality prediction data for each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm.

[0016] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any one of the above-mentioned methods for predicting water quality of multi-source mine water.

[0017] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for predicting water quality of multi-source mine water.

[0018] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for predicting water quality of multi-source mine water.

[0019] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0020] This application provides a multi-source mine water quality prediction method, device, equipment, medium, and product. These methods obtain working face water sources affected by mining fissures, primarily from old goafs and multiple overburden aquifers. Samples are collected from these water sources according to a set sample collection time interval. The collected water source samples are analyzed using the Pearson algorithm to obtain a dataset of key characteristic indicators of mine water quality monitored at each source layer in the working face water source. Based on these datasets, water quality prediction is performed for each source layer using an ant colony optimization algorithm, convolutional neural network, and long-short-term memory prediction model, ultimately obtaining water quality prediction data for each source layer. This application utilizes an ant colony optimization algorithm, convolutional neural network, and long-short-term memory prediction model to predict water quality, effectively increasing computational speed and utilizing the variability and temporal nature of water quality data for prediction, thereby improving the accuracy of water quality data prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 This is an application environment diagram of a multi-source mine water quality prediction method in one embodiment of the present application.

[0023] Figure 2 A flow chart of a multi-source mine water quality prediction method provided in one embodiment of the present application.

[0024] Figure 3 A flowchart of multi-source mine water quality prediction based on ACO-CNN-LSTM is provided in one embodiment of the present application.

[0025] Figure 4 A water source determination diagram for a working face mining area provided in one embodiment of the present application.

[0026] Figure 5 This is a structural diagram of the ACO-CNN-LSTM model provided in one embodiment of the present application.

[0027] Figure 6 A schematic diagram of the functional modules of a multi-source mine water quality prediction device provided in one embodiment of the present application.

[0028] Figure 7 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0030] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0031] The water quality prediction method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the water source sample to be processed to the server 104. After the server 104 receives the water source sample to be processed, the server 104 obtains the working face water source affected by the mining fracture for the water source sample to be processed; the working face water source mainly comes from the old empty area and several overburden aquifers; according to the set sample collection time interval, the working face water source is sampled; the collected water source samples are analyzed by the Pearson algorithm to obtain the main characteristic index data set of the mine water quality monitored by each source layer in the working face water source; based on the main characteristic index data set of the mine water quality monitored by each source layer, the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used to predict the water quality of each source layer to obtain the water quality prediction data of each source layer. The server 104 can feed back the obtained water quality prediction data for each source layer to the terminal 102. In addition, in some embodiments, the water quality prediction method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly perform sample processing on the water source sample to be processed, or the server 104 can obtain the water source sample to be processed from the data storage system and perform sample processing on the water source sample to be processed.

[0032] Terminal 102 may include, but is not limited to, various desktop computers, laptops, smartphones, tablet computers, IoT devices, and portable wearable devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, and smart car devices. Portable wearable devices may include smart watches, smart bracelets, and head-mounted devices. Server 104 may be implemented as a standalone server or a server cluster consisting of multiple servers, or may be a cloud server.

[0033] In an exemplary embodiment, Figure 2 As shown, a multi-source mine water quality prediction method is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the process, including the following steps 201 to 204.

[0034] Step 201: Obtain the water source of the working face affected by mining fractures; the water source of the working face mainly comes from the old goaf area and several overburden aquifers.

[0035] Step 202: Collect samples from the working surface water source according to the set sample collection time interval.

[0036] Step 203: Use the Pearson algorithm to analyze the collected water source samples to obtain a data set of main characteristic indicators of mine water quality monitored in each source layer of the working face water source.

[0037] Step 204: Based on the data set of main characteristic indicators of mine water quality monitored at each source layer, the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used to predict the water quality of each source layer to obtain water quality prediction data for each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm.

[0038] In some embodiments, before obtaining water from the working face affected by mining fractures, Figure 3 Also shown are:

[0039] Collect the occurrence information of mine working faces, old goafs and overburden aquifers in the target area.

[0040] Based on the occurrence information, the mining damage theory is adopted to determine the information of the water-conducting fracture zone in the target area.

[0041] Based on the information of the water-conducting fracture zone, the overlap between the water-conducting fracture zone and the old goaf and aquifer is analyzed to determine the water source of the working face affected by the mining fracture.

[0042] Specifically, relevant information on the occurrence of mine working faces, old goafs, and overburden aquifers is collected; based on the theory of coal mining subsidence, the spatial overlap of water-conducting fracture zones with old goafs and aquifers is analyzed to determine the source of mining water in the working face. This is done through the following methods:

[0043] 1) Collect the occurrence information of mine working face, old goaf and overburden aquifer, including: working face mining depth H, mining thickness M, overburden lithology f, overburden old goaf and aquifer (a1, a2, ... a i …、a n ) layer height (h1, h2, ...h i …、h n )wait.

[0044] 2) Calculating the height of the water-conducting fracture zone during coal seam mining is a prerequisite for determining the source of water in the working face. Based on the theory of coal mining damage, combined with information such as the lithology of the overlying strata on the working face and the thickness of the working face, the height of the water-conducting fracture zone is calculated. The calculation formula is as follows:

[0045] When the overburden lithology is hard (f>8),

[0046] When the overburden lithology is medium hard (3<f≤8),

[0047] When the overburden lithology is weak (f≤3),

[0048] Among them, H li is the maximum height of the water-conducting fracture zone, m; M is the mining thickness of the coal seam, m; f is the Proctor hardness coefficient of the rock formation.

[0049] 3) Analyze the overlap of water-conducting fracture zones, old goaf areas, overburden aquifers, etc., determine the range of overburden affected by mining fractures, and determine the source of mining water in the working face, such as Figure 4 shown.

[0050] It can be foreseen that the overlap of water-conducting fracture zones and old void areas, overburden aquifers, etc. includes: ① When the height of the water-conducting fracture zone is relatively small, the overburden fractures may only involve nearby old void areas and aquifers closer to the coal seams, and the leakage of these water bodies is the source of mine water; ② When the height of the water-conducting fracture zone is relatively large or the mining depth is relatively small, the height of the conduction may reach the surface, and there may even be leakage of surface water at this time. The source of mine water is composed of water from the old void area, water from the aquifer, and surface water.

[0051] Comparison of water-conducting fracture zone height H li Height h relative to old empty area, aquifer, surface, etc. i The size of the working face affected by the mining fissure is determined, that is, the old empty area a1, aquifer a2, ..., aquifer a n .

[0052] In some embodiments, when executing step 203, the specific steps may be as follows: using the Pearson algorithm to analyze the collected water source samples, and obtaining a data set of main characteristic indicators of the mine water quality monitored at each source layer in the working face water source.

[0053] By arranging monitoring and detection sensors, water quality index data of the working face mining and mine water source area over the past period of time can be obtained, which may include: total dissolved solids, ammonia nitrogen, heavy metals, sulfides, radioactive substances, pH value, total suspended solids, etc.

[0054] The sample collection interval is then determined. Since the actual monitoring frequency of the automatic water quality monitoring system is typically every four hours, during the data preprocessing phase, the water quality data is screened and adjusted to a uniform four-hour interval. Missing and outliers in the collected data are also processed. Outlier detection is performed on anomalous data, and outliers are treated as missing values. Missing data is processed using Lagrange interpolation.

[0055] The determination of characteristic indicators that have a greater impact on water quality prediction: when there are too many factors such as total dissolved solids, heavy metals, sulfides, chemical oxygen demand, radioactive substances, pH value, total suspended solids, etc., the accuracy of the prediction will be affected. Therefore, the Pearson correlation coefficient method is used to obtain the correlation between each factor and water quality prediction, thereby obtaining the Pearson correlation matrix. The details are as follows:

[0056]

[0057] The correlation between data is determined by the value of the Pearson correlation coefficient. The evaluation criteria are: 0.8-1.0 is a very strong correlation, 0.6-0.8 is a strong correlation, 0.4-0.6 is a moderate correlation, 0.2-0.4 is a weak correlation, and 0.0-0.2 is a very weak correlation or no correlation. The formula is as follows:

[0058]

[0059] Where |sim(X,Y)|≤1 is the correlation coefficient between variables X and Y, and X i and Y i The value of the i-th group of data, is the mean of the two factors, and m is the number of data sets. According to the correlation matrix and the size of the correlation coefficient, the features with higher correlation with water quality indicators are selected as the input of the model, and the normalized data set of the main water quality characteristic indicators Q=(Q1,Q2,...,Q n ). Among them, Q1, Q2, ..., Q n Indicators such as total dissolved solids, heavy metals, sulfides, chemical oxygen demand, radioactive substances, pH value, and total suspended solids can be taken.

[0060] Normalization is performed on different indicators of the mine water quality data set: To reduce the impact of numerical differences between different detection items on the simulation weights and improve the accuracy and applicability of the neural network, the water quality data is normalized before entering the grid, and the range is controlled between 0 and 1. The normalization formula is as follows:

[0061]

[0062] Among them, x is the data to be normalized, x max and x min Represent the maximum and minimum values ​​in the data set respectively.

[0063] In some embodiments, when executing step 204, the specific steps may be as follows:

[0064] Step 301: Construct Figure 5 The ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model (ACO-CNN-LSTM model) shown in the figure is as follows:

[0065] The Ant Colony Optimization (ACO) algorithm is used to optimize the parameters of the CNN-LSTM prediction module to further reduce the error and achieve the best prediction effect of the model. The details are as follows:

[0066] 1) Initialize the neural network weights and thresholds.

[0067] 2) Initialize the pheromone concentration matrix T ij (t), heuristic function η ij , as well as the ant coordinates, the maximum number of iterations.

[0068] 3) Each ant selects the next feature indicator according to the probability of pheromone and heuristic function until all feature indicators are traversed.

[0069] 4) After the ants have completed all the characteristic indicators, the paths and pheromone concentration matrix T of the ants are updated. ij (t).

[0070] 5) Repeat steps 1) to 3) until all ants in the population have completed their journey or the number of iterations reaches a preset value. Otherwise, proceed to the next iteration.

[0071] 6) A positive feedback mechanism is formed based on the pheromone intensity to find the shortest path.

[0072] 7) The paths generated by the ant colony are used to optimize the neural network, and the optimal model parameters such as sliding window size, number of hidden layer nodes, number of network iterations, etc. are trained to achieve the prediction effect.

[0073] Among them, the probability formula for the ant to select the next feature indicator is:

[0074]

[0075] Pheromone update formula:

[0076] Among them, the pheromone factor α reflects the relative importance of the amount of pheromone accumulated on the path during the ant movement in guiding the ant search, and its value is between [1,4]. The heuristic function factor β reflects the relative importance of heuristic information in guiding the ant colony search, and the intensity of the role of priori and deterministic factors in the ant colony optimization process, and its value is between [0,5].

[0077] Where p is the pheromone volatilization rate, is the pheromone left by ant k on the path from i to j in iteration t, and λ is the number of ants.

[0078] Step 302: Based on the dataset of main characteristic indicators of mine water quality monitored at each source layer, the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used. The water quality data sequence of the old empty area a1, aquifer a2, and aquifer a3 monitored at the previous m times is input to predict the water quality indicators of each source layer at time m+t. The details are as follows:

[0079] 1) Taking time m as the current state, collect water quality indicators p of old empty area a1, aquifer a2, and aquifer a3 i ,q i ,r i s i The data sequences from the previous d moments to the current moment are expressed as:

[0080] Old empty area a1:p i (m)=[p i (m-d+1),...,p i (m-1),p i (m)].

[0081] Aquifer a2:q i (m)=[q i (m-d+1),...,q i (m-1),q i (m)].

[0082] Aquifer a3:r i (m)=[r i (m-d+1),...,r i (m-1),r i (m)].

[0083] Among them, d represents the size of the sliding window, i represents different water quality characteristic indicators, q iIt is the water quality index data set collected from aquifer a2, including q1, which can be the water quality data sequence of total dissolved solids; q2, which can be the water quality data sequence of heavy metals; and q3, which can be the water quality data sequence of sulfide. The output vector of the convolutional layer is The mathematical expression is:

[0084]

[0085] in, It is calculated based on the output vector of the aforementioned convolutional input layer. represents the offset of the j-th mapping feature, w is the kernel weight, z is the filtered index value, and σ is the ReLU activation function.

[0086] 2) After passing through the CNN input layer and multiple hidden layer outputs, the data enters the LSTM. LSTM is the subsequent layer of the CNN in the CNN-LSTM model, and its input is the output of the CNN layer. Assuming that the input gate of the LSTM unit at stage t is it, the forget gate is ft, the output gate is ot, and the hidden layer state is ht, then the relevant updates of the unit at stage t are:

[0087] Forget Gate: Choose to forget certain information

[0088] Input gate: memorize certain information

[0089] Output gate: output certain information

[0090] Among them, c t represents the unit state at stage t, σ represents the activation function, such as the tanh function, which is nonlinear and compresses the data between [-1,1]. w is the weight matrix of each unit, b is the corresponding offset vector, p t It represents the output of the key features of water quality prediction at time t after the CNN network passes through the pooling layer, and serves as the initial input of the LSTM.

[0091] 3) The formulas for LSTM unit state ct and hidden layer ht are:

[0092]

[0093] h t =o t *tanh(c t ).

[0094] 4) Let h l ={h1,h2,...,h l}, where l is the number of LSTM units. The output of LSTM is calculated as: Finally, the water quality prediction results of the old empty area a1, aquifer a2, and aquifer a3 are:

[0095] Old empty area a1:p i (m+t)=[p i (m+t-d+1),...,p i (m+t-1),p i (m+t)].

[0096] Aquifer a2:q i (m+t)=[q i (m+t-d+1),...,q i (m+t-1),q i (m+t)].

[0097] Aquifer a3:r i (m+t)=[r i (m+t-d+1),...,r i (m+t-1),r i (m+t)]. Where t is the time interval between the predicted value and the sampled value.

[0098] Among them, such as Figure 3 As shown, the training method of the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is:

[0099] Obtain a dataset of key characteristic indicators of sample mine water quality from each source layer. This dataset contains historical data collected at several time points.

[0100] The main characteristic indicator data of sample mine water quality at time t in the sample mine water quality main characteristic indicator dataset are input into the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to predict the water quality data of each source layer at time t+1.

[0101] The predicted water quality data of each source layer at time t+1 are compared with the main characteristic index data of sample mine water quality at time t+1 in the sample mine water quality main characteristic index dataset.

[0102] When the data fitting degree reaches the set threshold, the training of the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is completed.

[0103] If the data fitting degree does not reach the set threshold, the model parameters are adjusted and re-executed: the main characteristic indicator data of the sample mine water quality at time t is collected in the data set until the data fitting degree reaches the set threshold.

[0104] Specifically, the predicted data p of the ACO-CNN-LSTM output of the old empty area a1, aquifer a2, and aquifer a3 are i (m+t),q i (m+t),r i (m+t) and historical data collected from the mining face s i (m)=[s i (m-d+1),...,s i (m-1),s i (m)] as the initial input, and then input it into the ACO-CNN-LSTM prediction method to obtain the prediction result s of the mining face. i (m+t)=[s i (m+t-d+1),...,s i (m+t-1),s i (m+t)].

[0105] To judge whether the prediction model meets the accuracy requirements, the actual value and predicted value of the mining area water quality index are compared. The higher the degree of curve fitting between the two, the higher the accuracy.

[0106] The degree of fit is determined as follows:

[0107] Specifically, it includes: using the training set to train the model and outputting the water quality prediction results at the next moment.

[0108] Mean Square Error:

[0109] Root mean square error:

[0110] Mean absolute error:

[0111] Mean absolute percentage error:

[0112] Model fit:

[0113] Among them, Y i is the true value of each water quality index, is the predicted value of the model water quality index. i It is the deviation between the true value and the measured value. Different error judgment criteria are used for different data and problems. Generally, the smaller the MSE, Re, MAE, and MAPE values ​​are, the better the R2 The closer the value is to 1, the higher the prediction accuracy of the mine water quality model and the better the model fits the data. Generally, if the MSE is less than 1, the RMSE is less than 10% of the data range, the MAE is less than the standard deviation of the data, and the MAPE is less than 10%, the model is considered to have a good fit.

[0114] Based on the same inventive concept, the present application also provides a prediction device for implementing the aforementioned water quality prediction method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more water quality prediction device embodiments provided below can be found in the above-mentioned limitations of the water quality prediction method and will not be repeated here.

[0115] In an exemplary embodiment, Figure 6 As shown, a multi-source mine water quality prediction device is provided, and the multi-source mine water quality prediction device includes:

[0116] The acquisition module 601 is used to obtain the water source of the working face affected by the mining fractures; the water source of the working face mainly comes from the old goaf area and several overburden aquifers.

[0117] The sample collection module 602 is used to collect samples from the working surface water source according to the set sample collection time interval.

[0118] The sample analysis module 603 is used to analyze the collected water source samples using the Pearson algorithm to obtain a data set of main characteristic indicators of mine water quality monitored in each source layer of the working face water source.

[0119] The water quality prediction module 604 is used to predict the water quality of each source layer based on the main characteristic indicator data set of mine water quality monitored at each source layer, using the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to obtain water quality prediction data for each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm.

[0120] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 7As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store water quality prediction data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a water quality prediction method is implemented.

[0121] Those skilled in the art will understand that Figure 7 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application and does not constitute a limitation on the computer device to which the solution of the present application is applied. A specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the above-mentioned method embodiments when executing the computer program.

[0122] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0123] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0124] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0125] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0126] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0127] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A multi-source mine water quality prediction method, characterized in that: The multi-source mine water quality prediction method comprises: Obtaining water sources at the working face affected by mining fractures; the water sources at the working face mainly come from old goaf areas and several overburden aquifers; Collect samples from the working face water source according to the set sample collection time interval; The collected water source samples were analyzed using the Pearson algorithm to obtain a dataset of key characteristic indicators of mine water quality monitored at each source layer of the working face water source. The dataset consists of total dissolved solids, heavy metals, sulfide, chemical oxygen demand, radioactive substances, pH value, and total suspended solids. Based on the main characteristic indicator data set of mine water quality monitored at each source layer, a trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used to predict the water quality of each source layer to obtain water quality prediction data for each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm; Before obtaining water sources from the working face affected by mining fissures, it also includes: Collect the occurrence information of mine working faces, old goaf and overburden aquifers in the target area; Based on the occurrence information, the mining damage theory is used to determine the water-conducting fracture zone information in the target area; Based on the information of the water-conducting fracture zone, the overlap between the water-conducting fracture zone and the old goaf and aquifer is analyzed to determine the water source of the working face affected by the mining fracture; The reservoir information is composed of the mining depth and thickness of the working face, the lithology of the overburden, the layer height of the old goaf and the layer height of the overburden aquifer; According to the occurrence information, the mining damage theory is used to determine the water-conducting fracture zone information in the target area, specifically including: When the hardness of the overburden lithology is greater than the first set threshold, according to the formula Determine the height information of the water-conducting fracture zone in the target area; When the hardness of the overburden lithology is less than or equal to the first set threshold and greater than the second set threshold, according to the formula Determine the height information of the water-conducting fracture zone in the target area; When the hardness of the overburden lithology is less than or equal to the second set threshold, according to the formula Determine the height information of the water-conducting fracture zone in the target area; Based on the dataset of main characteristic indicators of mine water quality monitored at each source layer, the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is used to input the water quality data sequence of the old empty area a1, aquifer a2, and aquifer a3 monitored at the previous m moments to predict the water quality indicators of each source layer at time m+t. Specifically, the following are the results: 1) Taking time m as the current state, collect water quality indicators p of old empty area a1, aquifer a2, and aquifer a3 i ,q i ,r i s i The data sequences from the previous d moments to the current moment are expressed as: Old empty area a1:p i (m)=[p i (m-d+1),...,p i (m-1),p i (m)]; Aquifer a2:q i (m)=[q i (m-d+1),...,q i (m-1),q i (m)]; Aquifer a3:r i (m)=[r i (m-d+1),...,r i (m-1),r i (m)]; Among them, d represents the size of the sliding window, i represents different water quality characteristic indicators, and q i represents the water quality index data set collected from aquifer a2; q2 represents the water quality data sequence of heavy metals; q3 represents the water quality data sequence of sulfides; the output vector of the convolutional layer The mathematical expression is: in, represents the offset of the j-th mapping feature, w represents the kernel weight, z represents the filtered index value, and σ represents the ReLU activation function; 2) After passing through the CNN input layer and multiple hidden layer outputs, the data will enter the LSTM. LSTM is the subsequent layer of CNN in the CNN-LSTM model, and its input is the output of the CNN layer. The input gate of the LSTM unit at stage t is i t , the forget gate is f t , the output gate is o t , the hidden layer state is h t , then the relevant update of the unit in stage t is: Forget Gate: Choose to forget certain information Input gate: memorize certain information Output gate: output certain information Among them, c t represents the unit state at stage t, σ represents the activation function; w is the weight matrix of each unit, b is the corresponding offset vector, p t Represents the output of the key features of water quality prediction at time t after the CNN network passes through the pooling layer; 3) LSTM cell state c t and hidden layer h t The formula is: h t =o t *tanh(c t ); 4) Let h l ={h1,h2,...,h l }, where l is the number of LSTM units; the output of LSTM is calculated as follows: The water quality prediction results of the old empty area a1, aquifer a2, and aquifer a3 are: Old empty area a1:p i (m+t)=[p i (m+t-d+1),...,p i (m+t-1),p i (m+t)]; Aquifer a2:q i (m+t)=[q i (m+t-d+1),...,q i (m+t-1),q i (m+t)]; Aquifer a3:r i (m+t)=[r i (m+t-d+1),...,r i (m+t-1),r i (m+t)]; where t is the time interval between the predicted value and the sampled value.

2. A multi-source mine water quality prediction method according to claim 1, characterized in that: The training method of the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is: Obtaining a dataset of main characteristic indicators of water quality of sample mine water monitored at each source layer; the dataset of main characteristic indicators of water quality of sample mine water includes historical data collected at several time points; The main characteristic index data of sample mine water quality at time t in the sample mine water quality main characteristic index data set are input into the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to predict the water quality prediction data of each source layer at time t+1; Compare the predicted water quality data of each source layer at time t+1 with the main characteristic index data of sample mine water quality at time t+1 in the main characteristic index data set of sample mine water quality; When the data fitting degree reaches a set threshold, the training of the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is completed; If the data fitting degree does not reach the set threshold, the model parameters are adjusted and re-executed: the main characteristic indicator data of the sample mine water quality at time t is collected in the data set until the data fitting degree reaches the set threshold.

3. A multi-source mine water quality prediction device, used to implement a multi-source mine water quality prediction method according to any one of claims 1-2, characterized in that: The multi-source mine water quality prediction device comprises: An acquisition module is used to obtain water sources at the working face affected by mining fractures; the water sources at the working face mainly come from old goaf areas and several overburden aquifers; The sample collection module is used to collect samples from the water source of the working face according to the set sample collection time interval; The sample analysis module is used to analyze the collected water source samples using the Pearson algorithm to obtain a data set of main characteristic indicators of mine water quality monitored at each source layer in the working face water source; The water quality prediction module is used to predict the water quality of each source layer based on the main characteristic indicator data set of mine water quality monitored at each source layer, using the trained ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model to obtain water quality prediction data for each source layer; the ant colony optimization algorithm-convolutional neural network-long short-term memory prediction model is a model obtained by optimizing the parameters of the convolutional neural network-long short-term memory prediction model using the ant colony optimization algorithm.

4. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement a multi-source mine water quality prediction method according to any one of claims 1-2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a multi-source mine water quality prediction method according to any one of claims 1 to 2 is implemented.

6. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, a multi-source mine water quality prediction method according to any one of claims 1 to 2 is implemented.

Images