Mine field geological disaster prediction method and system based on artificial intelligence

Abstract

The invention discloses a mine field geological disaster prediction method and system based on artificial intelligence. The method comprises the steps of data acquisition, data preprocessing, geological disaster classification model construction, disaster risk prediction model construction and mine field geological disaster prediction. The invention relates to the technical field of mine field geological data prediction, in particular to a mine field geological disaster prediction method and system based on artificial intelligence. A data preprocessing method of data cleaning, data coding and data normalization is adopted; a dynamic weight K-means clustering model is adopted to carry out preliminary classification, structured information is provided for subsequent geological time series data analysis, and the identification capability of the model is enhanced; according to the method for predicting the mine geological disasters by adopting the bidirectional long-short-term memory network model combined with bee colony optimization, potential modes in the mine geological disaster data can be fully mined, so that more accurate prediction is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of mine geological data prediction, and specifically refers to a mine geological disaster prediction method and system based on artificial intelligence. Background Technique

[0002] Mine geological disaster prediction refers to predicting possible geological disasters in a mining area by analyzing various factors such as the geological conditions of the mining area, historical disaster records, climate factors, and groundwater, combined with geological exploration data and modern information technology. It can effectively reduce the safety risks in mining production, reduce casualties and property losses, and ensure the safety of miners' lives.

[0003] However, traditional mine geological disaster prediction methods have the technical problems of directly processing all data and ignoring the structure of non-temporal data in geological data, resulting in difficulty in mining the relationship between geological temporal data and geological non-temporal data, and affecting the prediction accuracy; traditional mine geological disaster prediction methods have the technical problems that when optimizing model parameters, it is difficult to adapt to high-dimensional and complex geological data, and in a complex parameter space, it is impossible to fully explore the global optimal solution, resulting in insufficient model performance. Summary of the Invention

[0004] In view of the above situation, to overcome the defects of the prior art, the present invention provides a mine geological disaster prediction method and system based on artificial intelligence. Aiming at the technical problems of traditional mine geological disaster prediction methods that directly process all data and ignore the structure of non-temporal data in geological data, resulting in difficulty in mining the relationship between geological temporal data and geological non-temporal data, and affecting the prediction accuracy, this solution creatively uses a dynamic weight K-means clustering model for preliminary classification, which can effectively provide structured information for subsequent geological temporal data analysis, convert complex geological non-temporal data into simple category labels, and enhance the identification ability of the model; aiming at the technical problems of traditional mine geological disaster prediction methods that when optimizing model parameters, it is difficult to adapt to high-dimensional and complex geological data, and in a complex parameter space, it is impossible to fully explore the global optimal solution, resulting in insufficient model performance, this solution creatively uses a bidirectional long short-term memory network model combined with bee colony optimization, which can efficiently find the global optimal solution in a multi-dimensional parameter space and is not easily trapped in local optimal solutions, enabling the model to fully mine the potential patterns in mine geological disaster data, thereby achieving more accurate predictions.

[0005] The technical solution adopted by the present invention is as follows: The mine geological disaster prediction method based on artificial intelligence provided by the present invention includes the following steps:

[0006] Step S1: Data collection;

[0007] Step S2: Data preprocessing;

[0008] Step S3: Construction of geological disaster classification model;

[0009] Step S4: Construction of disaster risk prediction model;

[0010] Step S5: Prediction of geological disasters in the mine.

[0011] Further, in step S1, the data collection is used to collect the data required for predicting the occurrence risk of geological disasters in the mine. Specifically, through data collection, an original dataset for disaster prediction is obtained;

[0012] The original dataset for disaster prediction specifically includes a historical geological original dataset and a current geological original dataset. Both the historical geological original dataset and the current geological original dataset include geological time-series data and geological non-time-series data. The geological time-series data specifically includes mine meteorological data, dynamic data of mine soil humidity, variation data of mine groundwater level, disaster record data, slope displacement data, and geological activity record data. The geological non-time-series data specifically includes mine geographical spatial data and human activity data. The mine geographical spatial data specifically includes mine geographical coordinates, slope, aspect, terrain type, soil type, and stratigraphic characteristics. The human activity data specifically includes mining progress, mine depth, and disaster prevention project intensity.

[0013] Further, in step S2, the data preprocessing is used to preprocess the collected original data, specifically including the following steps:

[0014] Step S21: Data cleaning, which is used to clean the original data. Specifically, the missing values and duplicate values in the historical geological original dataset and the current geological original dataset are removed to obtain a historical preliminary dataset and a current preliminary dataset;

[0015] Step S22: Data encoding, which is used to encode the preliminary data. Specifically, the one-hot encoding method is used to encode the historical preliminary dataset and the current preliminary dataset to obtain a historical encoded dataset and a current encoded dataset;

[0016] Step S23: Data normalization, which is used to normalize the encoded data. Specifically, the min-max method is used to normalize the historical encoded dataset and the current encoded dataset to obtain a historical geological dataset and a current geological dataset;

[0017] Step S24: Perform preprocessing, specifically, through the data cleaning, the data encoding, and the data normalization, preprocess the historical geological original dataset and the current geological original dataset to obtain a historical geological dataset and a current geological dataset.

[0018] Further, in step S3, the construction of the geological disaster classification model is used to construct the model required for the preliminary classification of mine geological disasters. Specifically, a dynamic weight K-means clustering model is constructed, and the historical geological data set is preliminarily classified based on the types of mine geological disasters.

[0019] The construction of the geological disaster classification model specifically includes the following steps:

[0020] Step S31: Model initialization, specifically initializing the clustering centers and determining the iteration termination conditions;

[0021] The initialization of the clustering centers specifically means randomly selecting K sample data as the initial clustering centers. The sample data is specifically the non-temporal data part of the samples in the historical geological data set, and K represents the number of classification types;

[0022] The iteration termination conditions specifically include that the change amount of the clustering centers is less than the threshold and the maximum number of iterations is reached;

[0023] Step S32: Initial assignment of clusters, and the steps include:

[0024] Step S321: Calculate the distance from the sample data to the clustering centers, and the formula used is as follows:

[0025] ;

[0026] In the formula, represents the distance calculation function, represents the l-th sample data, represents the k-th clustering center;

[0027] Step S322: Assign data to clusters, and the formula used is as follows:

[0028] ;

[0029] In the formula, represents the cluster label of the l-th sample data, represents the value of k that makes the distance from the l-th sample data to the k-th clustering center the smallest;

[0030] Step S33: Subsequent assignment of clusters, and the steps include:

[0031] Step S331: Update the clustering centers, and the formula used is as follows:

[0032] ;

[0033] In the formula, represents the updated clustering center of the k-th cluster, denotes the set composed of all sample data in the k-th cluster, denotes the total number of sample data in the k-th cluster, denotes the -th sample data in the k-th cluster;

[0034] Step S332: Calculate the intra-cluster complexity, and the formula used is as follows:

[0035] ;

[0036] In the formula, denotes the complexity calculation function, denotes the mean calculation function, denotes the standard deviation calculation function;

[0037] Step S333: Calculate the intra-cluster complexity of the sample data, and the formula used is as follows:

[0038] ;

[0039] In the formula, sci denotes the intra-cluster complexity of the sample data, denotes the total number of sample data;

[0040] Step S334: Calculate the inter-cluster similarity, and the steps include:

[0041] Step S3341: Calculate the global and local cluster radii, and the formula used is as follows:

[0042] ;

[0043] In the formula, denotes the global cluster radius calculation function, denotes the local cluster radius calculation function;

[0044] Step S3342: Calculate the global and local inter-cluster distances, and the formula used is as follows:

[0045] ;

[0046] In the formula, denotes the global inter-cluster distance calculation function, denotes the local inter-cluster distance calculation function, denotes the -th set composed of all sample data in the denotes the -th set composed of all sample data in the denotes the -th -th sample data in the denotes the a clustering center;

[0047] Step S3343: Obtain the similarity between clusters, and the formula used is as follows:

[0048] ;

[0049] In the formula, represents the function for calculating the similarity between clusters, and e represents the base of the natural logarithm;

[0050] Step S335: Calculate the similarity between clusters of sample data, and the formula used is as follows:

[0051] ;

[0052] In the formula, ssi represents the similarity between clusters of sample data, represents the total number of sample data in the th cluster, represents the total number of sample data in the th cluster;

[0053] Step S336: Calculate the distance dynamic weight, which is used to participate in the calculation of the distance between the sample data and the updated clustering center. The formula for calculating the distance dynamic weight is as follows:

[0054] ;

[0055] In the formula, represents the distance dynamic weight, represents the adjustment weight, represents the normalized within-cluster complexity of the sample data, represents the normalized similarity between clusters of the sample data;

[0056] Step S337: Calculate the distance between the sample data and the updated clustering center and assign clusters. Specifically, assign the sample data to the updated clustering center with the closest distance. The formula for the distance between the sample data and the updated clustering center is as follows:

[0057] ;

[0058] In the formula, represents the distance calculation function with the participation of the distance dynamic weight;

[0059] Step S34: Iterative update, specifically, continuously iterate and update until the iteration termination condition is reached;

[0060] Step S35: Preliminary classification. Specifically, through the model initialization, the initial assignment clustering, the subsequent assignment clustering, and the iterative update, a dynamic weight K-means clustering model is constructed, and based on the types of mine geological disasters, the non-temporal data part of the samples in the historical geological dataset is preliminarily classified. The obtained labels are combined with the temporal data part of the samples in the historical geological dataset to obtain a risk prediction dataset.

[0061] Further, in step S4, the construction of the disaster risk prediction model is used to construct a model required for predicting the occurrence risk of mine geological disasters. Specifically, a bidirectional long short-term memory network model combined with bee colony optimization is constructed and used as the disaster risk prediction model.

[0062] The construction of the disaster risk prediction model specifically includes the following steps:

[0063] Step S41: Label and split the dataset. Specifically, the data in the risk prediction dataset is labeled as low risk, medium risk, and high risk and used as data labels. The labeled risk prediction dataset is split into a risk prediction training set and a risk prediction test set.

[0064] Step S42: Construct a bidirectional long short-term memory network model. The steps include:

[0065] Step S421: Design the activation function. The formula used is as follows:

[0066] ;

[0067] In the formula, represents the activation function, represents the period adjustment parameter, represents the phase adjustment parameter, represents the input nonlinear response adjustment parameter, represents the smoothing gain parameter, represents the linear gain parameter, and x represents the independent variable of the activation function.

[0068] Step S422: Construct the forward module. The formula used is as follows:

[0069] ;

[0070] In the formula, represents the output of the forward forgetting gate at time t, represents the weight of the forward forgetting gate, represents the forward hidden state at time t - 1, represents the model input at time t, represents the bias term of the forward forgetting gate, represents the output of the forward input gate at time t, Represents the positive input gate weight, Represents the positive input gate bias term, Represents the positive candidate cell state at time t, Represents the hyperbolic tangent function, Represents the weight used to calculate the positive candidate cell state, Represents the bias term used to calculate the positive candidate cell state, Represents the positive cell state at time t, Represents the positive cell state at time t-1, Represents the positive output gate output at time t, Represents the positive output gate weight, Represents the positive output gate bias term, Represents the positive hidden state at time t;

[0071] Step S423: Construct the reverse module, and the formula used is as follows:

[0072] ;

[0073] In the formula, Represents the reverse forget gate output at time t, Represents the reverse forget gate weight, Represents the reverse hidden state at time t-1, Represents the reverse forget gate bias term, Represents the reverse input gate output at time t, Represents the reverse input gate weight, Represents the reverse input gate bias term, Represents the reverse candidate cell state at time t, Represents the weight used to calculate the reverse candidate cell state, Represents the bias term used to calculate the reverse candidate cell state, Represents the reverse cell state at time t, Represents the reverse cell state at time t-1, Represents the reverse output gate output at time t, Represents the reverse output gate weight, Represents the reverse output gate bias term, Represents the reverse hidden state at time t;

[0074] Step S424: Construct the output module, and the formula used is as follows:

[0075] ;

[0076] In the formula, Represents the output of the model at time t, represents the softmax function, represents the model output weights, represents the model output bias term;

[0077] Step S425: Construct and train the model. Specifically, construct a bidirectional long short - term memory network model through the designed activation function, the constructed forward module, the constructed reverse module, and the constructed output module, train the model based on the risk prediction training set, and verify the model performance based on the risk prediction test set. The model loss function uses the cross - entropy loss function;

[0078] Step S43: Hyperparameter optimization. Specifically, optimize the hyperparameters of the bidirectional long short - term memory network model based on the bee colony optimization algorithm. The model hyperparameters specifically include the parameters of the Af activation function, the number of long short - term memory layers, the number of long short - term memory units, the model learning rate, and the batch size;

[0079] The hyperparameter optimization steps include:

[0080] Step S431: Algorithm initialization. Specifically, initialize the search space and construct a bee unit set. The bee unit is used to represent the hyperparameter combination of the bidirectional long short - term memory network model. The fitness function of the bee unit is the loss function of the bidirectional long short - term memory network model;

[0081] Step S432: Global search. The formula used is as follows:

[0082] ;

[0083] In the formula, represents the velocity of the a - th bee unit at the (dt + 1)-th iteration, represents the inertia weight, represents the velocity of the a - th bee unit at the dt - th iteration, represents the weight used to control learning its own optimal solution, represents the own optimal solution of the a - th bee unit, represents the position of the a - th bee unit at the dt - th iteration, represents the weight used to control learning the global optimal solution, represents the global optimal solution at the dt - th iteration, represents the position of the a - th bee unit at the (dt + 1)-th iteration;

[0084] Step S433: Local search. The steps include:

[0085] Step S4331: Calculate the probability of selecting the neighboring solution. The formula used is as follows:

[0086] ;

[0087] In the formula, represents the probability that the a-th bee unit selects the b-th neighbor bee unit, represents the fitness calculation function, N represents the number of neighbor bee units, represents the position of the b-th neighbor bee unit at the (dt + 1)-th iteration;

[0088] Step S4332: Design the neighbor solution learning strategy. Specifically, the bee unit selects a neighbor bee unit based on the neighbor solution selection probability, and optimizes its own position based on the position of the selected neighbor bee unit. The formula used is as follows:

[0089] ;

[0090] In the formula, represents the position of the a-th bee unit at the (dt + 1)-th iteration in local search, de represents a decay factor that linearly decreases from 2 to 1, and rand represents a random number in the range [0, 1];

[0091] Step S4333: Design the neighbor solution comparison strategy. Specifically, after the bee unit optimizes its own position based on the position of the neighbor bee unit, it selects a neighbor bee unit again based on the neighbor solution selection probability, and compares the fitness with the selected neighbor bee unit to determine whether to update the position. The formula used is as follows:

[0092] ;

[0093] In the formula, represents the fitness comparison function, represents the final position of the a-th bee unit at the (dt + 1)-th iteration, represents the position of the c-th neighbor bee unit at the (dt + 1)-th iteration;

[0094] Step S434: Determine the global optimal solution. Specifically, continuously iterate and update until the iteration termination condition of the optimization algorithm is reached, and take the position of the bee unit with the lowest fitness as the global optimal solution. The global optimal solution is specifically the hyperparameter combination of the optimal bidirectional long short-term memory network model. The iteration termination condition of the optimization algorithm specifically includes that the fitness of the bee unit is less than the set threshold and the maximum number of iterations is reached;

[0095] Step S435: Optimize the model hyperparameters. Specifically, through the algorithm initialization, the global search, the local search, and the determination of the global optimal solution, perform hyperparameter optimization of the bidirectional long short-term memory network model to obtain a bidirectional long short-term memory network model combined with swarm optimization, and use it as a disaster risk prediction model.

[0096] Further, in step S5, the prediction of geological disasters in the mine field specifically involves using the current geological data set as the input of the dynamic weight K-means clustering model and the disaster risk prediction model, respectively obtaining reference data on the types of geological disasters in the mine field and reference data on the disaster risk levels, and comprehensively evaluating the risk of geological disasters occurring in the mine field based on the reference data on the types of geological disasters in the mine field and the reference data on the disaster risk levels.

[0097] The mine field geological disaster prediction system based on artificial intelligence provided by the present invention includes a data acquisition module, a data preprocessing module, a geological disaster classification model construction module, a disaster risk prediction model construction module, and a mine field geological disaster prediction module;

[0098] The data acquisition module is used for data acquisition. Through data acquisition, an original data set for disaster prediction is obtained, and the original data set for disaster prediction is sent to the data preprocessing module;

[0099] The data preprocessing module is used for data preprocessing. Through data preprocessing, a historical geological data set and a current geological data set are obtained, the historical geological data set is sent to the geological disaster classification model construction module, and the current geological data set is sent to the mine field geological disaster prediction module;

[0100] The geological disaster classification model construction module is used for the preliminary classification of geological disasters in the mine field. By constructing a dynamic weight K-means clustering model and performing preliminary classification, a dynamic weight K-means clustering model and a risk prediction data set are obtained, the dynamic weight K-means clustering model is sent to the mine field geological disaster prediction module, and the risk prediction data set is sent to the disaster risk prediction model construction module;

[0101] The disaster risk prediction model construction module is used for constructing a disaster risk prediction model. By constructing a bidirectional long short-term memory network model combined with bee colony optimization, a disaster risk prediction model is obtained, and the disaster risk prediction model is sent to the mine field geological disaster prediction module;

[0102] The mine field geological disaster prediction module is used for predicting geological disasters in the mine field. By using the dynamic weight K-means clustering model and the disaster risk prediction model to predict geological disasters in the mine field, reference data on the types of geological disasters in the mine field and reference data on the disaster risk levels are obtained.

[0103] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0104] (1) Aiming at the technical problem that traditional mine geological disaster prediction methods directly process all data, ignoring the structure of non-temporal data in geological data, making it difficult to mine the relationship between geological temporal data and geological non-temporal data and affecting the prediction accuracy, this solution creatively uses a dynamic weight K-means clustering model for preliminary classification, which can effectively provide structured information for subsequent geological temporal data analysis, convert complex geological non-temporal data into concise category labels, and enhance the identification ability of the model.

[0105] (2) Aiming at the technical problem that traditional mine geological disaster prediction methods are difficult to adapt to high-dimensional and complex geological data when optimizing model parameters, and cannot fully explore the global optimal solution in the complex parameter space, resulting in insufficient model performance, this solution creatively uses a bidirectional long short-term memory network model combined with swarm optimization, which can efficiently find the global optimal solution in the multi-dimensional parameter space and is not easily trapped in the local optimal solution, enabling the model to fully mine the potential patterns in mine geological disaster data and thus achieve more accurate prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 It is a schematic flow chart of the mine geological disaster prediction method based on artificial intelligence provided by the present invention;

[0107] Figure 2 It is a schematic module diagram of the mine geological disaster prediction system based on artificial intelligence provided by the present invention;

[0108] Figure 3 It is a schematic flow chart of constructing the geological disaster classification model in step S3;

[0109] Figure 4 It is a schematic flow chart of constructing the bidirectional long short-term memory network model in step S42;

[0110] Figure 5 It is a schematic flow chart of hyperparameter optimization in step S43.

[0111] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, but do not constitute a limitation to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0112] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0113] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc. indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0114] Example 1, referring to Figure 1 , the technical solution adopted by the present invention is as follows: The method for predicting geological disasters in a mine based on artificial intelligence provided by the present invention includes the following steps:

[0115] Step S1: Data collection;

[0116] Step S2: Data preprocessing;

[0117] Step S3: Construction of a geological disaster classification model;

[0118] Step S4: Construction of a disaster risk prediction model;

[0119] Step S5: Prediction of geological disasters in the mine.

[0120] Example 2, referring to Figure 1 and Figure 2 , in step S1, the data collection is used to collect the data required for predicting the risk of geological disasters in the mine. Specifically, through data collection, an original dataset for disaster prediction is obtained;

[0121] The original dataset for disaster prediction specifically includes an original historical geological dataset and an original current geological dataset. Both the original historical geological dataset and the original current geological dataset include geological time-series data and geological non-time-series data. The geological time-series data specifically includes mine meteorological data, dynamic data of mine soil humidity, data on changes in the mine groundwater level, disaster record data, slope displacement data, and geological activity record data. The geological non-time-series data specifically includes mine geographical spatial data and human activity data. The mine geographical spatial data specifically includes mine geographical coordinates, slope, aspect, terrain type, soil type, and stratigraphic characteristics. The human activity data specifically includes mining progress, mine depth, and disaster prevention project intensity.

[0122] Example 3, referring to Figure 1 and Figure 2 , based on the above example, in step S2, the data preprocessing is used to preprocess the collected original data, and specifically includes the following steps:

[0123] Step S21: data cleaning, for cleaning the original data, specifically removing missing values ​​and duplicate values ​​in the historical geological original data set and the current geological original data set to obtain a historical preliminary data set and a current preliminary data set;

[0124] Step S22: data encoding, which is used to encode the preliminary data, specifically, using a one-hot encoding method to encode the historical preliminary data set and the current preliminary data set to obtain a historical encoded data set and a current encoded data set;

[0125] Step S23: data normalization, which is used to normalize the coded data, specifically, using the minimum-maximum method to normalize the historical coded data set and the current coded data set to obtain a historical geological data set and a current geological data set;

[0126] Step S24: performing preprocessing, specifically, preprocessing the historical geological original data set and the current geological original data set through the data cleaning, the data encoding and the data normalization to obtain the historical geological data set and the current geological data set.

[0127] Example 4, see Figure 1 , Figure 2 and Figure 3 , this embodiment is based on the above embodiment. In step S3, the geological disaster classification model is constructed to construct a model required for preliminary classification of mine geological disasters, specifically to construct a dynamic weighted K-means clustering model, and to perform preliminary classification of the historical geological data set based on the types of mine geological disasters;

[0128] The construction of the geological disaster classification model specifically includes the following steps:

[0129] Step S31: model initialization, specifically initializing cluster centers and determining iteration termination conditions;

[0130] The initialization of cluster centers is specifically to randomly select K sample data as the initial cluster centers, wherein the sample data is specifically the geological non-time series data part of the sample in the historical geological data set, and K represents the number of classification types;

[0131] The iteration termination conditions specifically include that the change in cluster center is less than a threshold and the maximum number of iterations is reached;

[0132] Step S32: Initial cluster allocation, the steps include:

[0133] Step S321: Calculate the distance from the sample data to the cluster center. The formula used is as follows:

[0134] ;

[0135] In the formula, represents the distance calculation function, represents the l-th sample data, represents the k-th clustering center;

[0136] Step S322: Assign data to clusters, and the formula used is as follows:

[0137] ;

[0138] In the formula, represents the cluster label of the l-th sample data, represents the value of k that makes the distance from the l-th sample data to the k-th clustering center the smallest;

[0139] Step S33: Subsequent cluster assignment, and the steps include:

[0140] Step S331: Update the clustering center, and the formula used is as follows:

[0141] ;

[0142] In the formula, represents the updated clustering center of the k-th cluster, represents the set composed of all sample data in the k-th cluster, represents the total number of sample data in the k-th cluster, represents the -th sample data in the k-th cluster;

[0143] Step S332: Calculate the intra-cluster complexity, and the formula used is as follows:

[0144] ;

[0145] In the formula, represents the complexity calculation function, represents the mean calculation function, represents the standard deviation calculation function;

[0146] Step S333: Calculate the intra-cluster complexity of the sample data, and the formula used is as follows:

[0147] ;

[0148] In the formula, sci represents the intra-cluster complexity of the sample data, represents the total number of sample data;

[0149] Step S334: Calculate the inter-cluster similarity, and the steps include:

[0150] Step S3341: Calculate the global and local cluster radii, and the formula used is as follows:

[0151] ;

[0152] In the formula, represents the global clustering radius calculation function, represents the local clustering radius calculation function;

[0153] Step S3342: Calculate the distance between the global and local clusters. The formula used is as follows:

[0154] ;

[0155] In the formula, represents the global inter-cluster distance calculation function, represents the local inter-cluster distance calculation function, represents the th set composed of all sample data in the th cluster, represents the th set composed of all sample data in the th cluster, represents the th sample data in the th cluster;

[0156] Step S3343: Obtain the similarity between clusters. The formula used is as follows:

[0157] ;

[0158] In the formula, represents the inter-cluster similarity calculation function, and e represents the base of the natural logarithm;

[0159] Step S335: Calculate the inter-cluster similarity of the sample data. The formula used is as follows:

[0160] ;

[0161] In the formula, ssi represents the inter-cluster similarity of the sample data, represents the total number of sample data in the th cluster, represents the total number of sample data in the th cluster;

[0162] Step S336: Calculate the distance dynamic weight, which is used to participate in the calculation of the distance between the sample data and the updated cluster center. The formula for calculating the distance dynamic weight is as follows:

[0163] ;

[0164] In the formula, represents the distance dynamic weight, represents the adjustment weight, represents the intra-cluster complexity of the normalized sample data, Represents the similarity between clusters of normalized sample data;

[0165] Step S337: Calculate the distance between the sample data and the updated cluster center and assign the cluster. Specifically, assign the sample data to the updated cluster center with the closest distance. The distance between the sample data and the updated cluster center is calculated as follows:

[0166] ;

[0167] In the formula, Indicates the distance calculation function with dynamic distance weight involved in the calculation;

[0168] Step S34: iterative updating, specifically, continuously iteratively updating until the iteration termination condition is reached;

[0169] Step S35: preliminary classification, specifically, constructing a dynamic weight K-means clustering model through the model initialization, the initial clustering allocation, the subsequent clustering allocation and the iterative update, and preliminarily classifying the geological non-time series data part of the samples in the historical geological data set based on the type of mining geological disasters, and combining the obtained labels with the geological time series data part of the samples in the historical geological data set to obtain a risk prediction data set.

[0170] By performing the above operations, the traditional mine geological disaster prediction method directly processes all data and ignores the structure of non-time series data in geological data, which makes it difficult to mine the relationship between geological time series data and geological non-time series data, affecting the prediction accuracy. This solution creatively uses the dynamic weight K-means clustering model for preliminary classification, which can effectively provide structured information for subsequent geological time series data analysis, convert complex geological non-time series data into concise category labels, and enhance the recognition ability of the model.

[0171] Example 5, see Figure 1 , Figure 2 , Figure 4 and Figure 5 , this embodiment is based on the above embodiment. In step S4, the disaster risk prediction model is constructed to construct a model required for predicting the risk of geological disasters in mining areas, specifically, a bidirectional long short-term memory network model combined with bee colony optimization is constructed as a disaster risk prediction model;

[0172] The disaster risk prediction model is constructed, specifically comprising the following steps:

[0173] Step S41: Label and split the dataset. Specifically, label the data in the risk prediction dataset as low risk, medium risk, and high risk, and use them as data labels. Then split the labeled risk prediction dataset into a risk prediction training set and a risk prediction test set;

[0174] Step S42: Construct a bidirectional long short-term memory network model. The steps include:

[0175] Step S421: Design the activation function. The formula used is as follows:

[0176] ;

[0177] In the formula, represents the activation function, represents the period adjustment parameter, represents the phase adjustment parameter, represents the input nonlinear response adjustment parameter, represents the smoothing gain parameter, represents the linear gain parameter, and x represents the independent variable of the activation function;

[0178] Step S422: Construct the forward module. The formula used is as follows:

[0179] ;

[0180] In the formula, represents the output of the forward forget gate at time t, represents the weight of the forward forget gate, represents the forward hidden state at time t-1, represents the model input at time t, represents the bias term of the forward forget gate, represents the output of the forward input gate at time t, represents the weight of the forward input gate, represents the bias term of the forward input gate, represents the forward candidate cell state at time t, represents the hyperbolic tangent function, represents the weight used to calculate the forward candidate cell state, represents the bias term used to calculate the forward candidate cell state, represents the forward cell state at time t, represents the forward cell state at time t-1, represents the output of the forward output gate at time t, represents the weight of the forward output gate, represents the bias term of the forward output gate, represents the forward hidden state at time t;

[0181] Step S423: Construct a reverse module, and the formula used is as follows:

[0182] ;

[0183] In the formula, represents the output of the forget gate at the reverse time step t, represents the weight of the reverse forget gate, represents the reverse hidden state at time step t - 1, represents the bias term of the reverse forget gate, represents the output of the input gate at the reverse time step t, represents the weight of the reverse input gate, represents the bias term of the reverse input gate, represents the candidate cell state at the reverse time step t, represents the weight used to calculate the reverse candidate cell state, represents the bias term used to calculate the reverse candidate cell state, represents the cell state at the reverse time step t, represents the cell state at the reverse time step t - 1, represents the output of the output gate at the reverse time step t, represents the weight of the reverse output gate, represents the bias term of the reverse output gate, represents the hidden state at the reverse time step t;

[0184] Step S424: Construct an output module, and the formula used is as follows:

[0185] ;

[0186] In the formula, represents the output of the model at time step t, represents the softmax function, represents the output weight of the model, represents the output bias term of the model;

[0187] Step S425: Construct and train the model. Specifically, construct a bidirectional long short-term memory network model by using the designed activation function, the constructed forward module, the constructed reverse module, and the constructed output module, train the model based on the risk prediction training set, verify the model performance based on the risk prediction test set, and use the cross-entropy loss function as the model loss function;

[0188] Step S43: Hyperparameter optimization. Specifically, optimize the hyperparameters of the bidirectional long short-term memory network model based on the bee colony optimization algorithm. The model hyperparameters specifically include the parameters of the Af activation function, the number of long short-term memory layers, the number of long short-term memory units, the model learning rate, and the batch size;

[0189] The hyperparameter optimization steps include:

[0190] Step S431: Algorithm initialization, specifically, initialize the search space and construct a bee unit set. The bee unit is used to represent the hyperparameter combination of the bidirectional long short-term memory network model, and the fitness function of the bee unit is the loss function of the bidirectional long short-term memory network model;

[0191] Step S432: Global search, and the formula used is as follows:

[0192] ;

[0193] In the formula, represents the velocity of the a-th bee unit at the (dt + 1)-th iteration, represents the inertia weight, represents the velocity of the a-th bee unit at the dt-th iteration, represents the weight for controlling learning its own optimal solution, represents the own optimal solution of the a-th bee unit, represents the position of the a-th bee unit at the dt-th iteration, represents the weight for controlling learning the global optimal solution, represents the global optimal solution at the dt-th iteration, represents the position of the a-th bee unit at the (dt + 1)-th iteration;

[0194] Step S433: Local search, and the steps include:

[0195] Step S4331: Calculate the probability of selecting a neighboring solution, and the formula used is as follows:

[0196] ;

[0197] In the formula, represents the probability that the a-th bee unit selects the b-th neighboring bee unit, represents the fitness calculation function, N represents the number of neighboring bee units, represents the position of the b-th neighboring bee unit at the (dt + 1)-th iteration;

[0198] Step S4332: Design a neighboring solution learning strategy, specifically, the bee unit selects a neighboring bee unit based on the probability of selecting a neighboring solution, and optimizes its own position based on the position of the selected neighboring bee unit. The formula used is as follows:

[0199] ;

[0200] In the formula, represents the position of the a-th bee unit at the dt+1-th iteration in the local search, de represents the decay factor whose size decreases linearly from 2 to 1, and rand represents a random number in the range [0,1];

[0201] Step S4333: Design a neighbor solution comparison strategy. Specifically, after the bee unit optimizes its own position based on the position of the neighbor bee unit, it selects a neighbor bee unit based on the neighbor solution selection probability again, and compares the fitness with the selected neighbor bee unit to determine whether to update the position. The formula used is as follows:

[0202] ;

[0203] In the formula, represents the fitness comparison function, represents the final position of the a-th bee unit at the dt+1-th iteration, represents the position of the cth neighbor bee unit at the dt+1th iteration;

[0204] Step S434: Determine the global optimal solution, specifically by continuously iterating and updating until the optimization algorithm iteration termination condition is reached, and take the position of the bee unit with the lowest fitness as the global optimal solution, the global optimal solution is specifically the optimal bidirectional long short-term memory network model hyperparameter combination, and the optimization algorithm iteration termination condition specifically includes that the fitness of the bee unit is less than the set threshold and the maximum number of iterations is reached;

[0205] Step S435: Optimizing model hyperparameters, specifically optimizing the hyperparameters of the bidirectional long short-term memory network model through the algorithm initialization, the global search, the local search and the determination of the global optimal solution, obtaining a bidirectional long short-term memory network model combined with swarm optimization, and using it as a disaster risk prediction model.

[0206] By performing the above operations, the traditional mine geological disaster prediction method has the technical problem of being difficult to adapt to high-dimensional and complex geological data when optimizing model parameters, and being unable to fully explore the global optimal solution in the complex parameter space, resulting in insufficient model performance. This solution creatively adopts a bidirectional long short-term memory network model combined with bee colony optimization, which can efficiently find the global optimal solution in the multi-dimensional parameter space and is not easy to fall into the local optimal solution, so that the model can fully explore the potential patterns in the mine geological disaster data, thereby achieving more accurate predictions.

[0207] Example 6, see Figure 1 and Figure 2, based on the above embodiment, in step S5, the prediction of mine geological disasters specifically involves using the current geological data set as the input of the dynamic weight K-means clustering model and the disaster risk prediction model to respectively obtain reference data on mine geological disaster types and reference data on disaster risk levels, and comprehensively evaluating the risk of mine geological disasters occurring based on the reference data on mine geological disaster types and the reference data on disaster risk levels.

[0208] Embodiment Seven, refer to Figure 1 and Figure 2 , based on the above embodiment, the mine geological disaster prediction system based on artificial intelligence provided by the present invention includes a data acquisition module, a data preprocessing module, a geological disaster classification model construction module, a disaster risk prediction model construction module, and a mine geological disaster prediction module;

[0209] The data acquisition module is used for data acquisition. Through data acquisition, an original data set for disaster prediction is obtained, and the original data set for disaster prediction is sent to the data preprocessing module;

[0210] The data preprocessing module is used for data preprocessing. Through data preprocessing, a historical geological data set and a current geological data set are obtained, and the historical geological data set is sent to the geological disaster classification model construction module, and the current geological data set is sent to the mine geological disaster prediction module;

[0211] The geological disaster classification model construction module is used for the preliminary classification of mine geological disasters. Through constructing a dynamic weight K-means clustering model and performing preliminary classification, a dynamic weight K-means clustering model and a risk prediction data set are obtained, and the dynamic weight K-means clustering model is sent to the mine geological disaster prediction module, and the risk prediction data set is sent to the disaster risk prediction model construction module;

[0212] The disaster risk prediction model construction module is used for constructing a disaster risk prediction model. Through constructing a bidirectional long short-term memory network model combined with bee colony optimization, a disaster risk prediction model is obtained, and the disaster risk prediction model is sent to the mine geological disaster prediction module;

[0213] The mine geological disaster prediction module is used for predicting mine geological disasters. Through using the dynamic weight K-means clustering model and the disaster risk prediction model to predict mine geological disasters, reference data on mine geological disaster types and reference data on disaster risk levels are obtained.

[0214] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0215] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention.

[0216] The above description of the present invention and its embodiments is not restrictive. What is shown in the drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. In general, if those of ordinary skill in the art are inspired by it and, without departing from the purpose of the present invention, design similar structural modes and embodiments to this technical solution without creative efforts, they shall fall within the protection scope of the present invention.

Claims

1. A mine geological disaster prediction method based on artificial intelligence, characterized by: The method comprises the following steps: Step S1: data collection, through which the original data set for disaster prediction is obtained, specifically including the original data set for historical geology and the original data set for current geology; Step S2: Data preprocessing, preprocessing the collected raw data to obtain historical geological data sets and current geological data sets; Step S3: Construction of a geological disaster classification model, which is used to construct a model required for preliminary classification of geological disasters in mines, specifically, to construct a dynamic weighted K-means clustering model and perform preliminary classification; Step S4: constructing a disaster risk prediction model, which is used to construct a model required for predicting the risk of geological disasters in mining sites, specifically constructing a bidirectional long short-term memory network model combined with bee colony optimization, and using it as a disaster risk prediction model; Step S5: predicting geological disasters in mines, specifically predicting geological disasters in mines through the dynamic weight K-means clustering model and the disaster risk prediction model, and obtaining reference data on the types of geological disasters in mines and reference data on the disaster risk levels.

2. The method for predicting geological disasters in mines based on artificial intelligence according to claim 1, characterized in that: In step S3, the geological disaster classification model is constructed to construct a model required for preliminary classification of mine geological disasters, specifically, a dynamic weighted K-means clustering model is constructed, and the historical geological data set is preliminarily classified based on the types of mine geological disasters; The construction of the geological disaster classification model specifically includes the following steps: Step S31: model initialization, specifically initializing cluster centers and determining iteration termination conditions; The initialization of cluster centers is specifically to randomly select K sample data as the initial cluster centers, wherein the sample data is specifically the geological non-time series data part of the sample in the historical geological data set, and K represents the number of classification types; The iteration termination conditions specifically include that the change in cluster center is less than a threshold and the maximum number of iterations is reached; Step S32: Initial cluster allocation, the steps include: Step S321: Calculate the distance from the sample data to the cluster center. The formula used is as follows: ； In the formula, represents the distance calculation function, represents the lth sample data, represents the kth cluster center; Step S322: assign data to clusters using the following formula: ； In the formula, represents the cluster label of the lth sample data, Indicates the value of k that minimizes the distance from the lth sample data to the kth cluster center; Step S33: Subsequent allocation of clusters, the steps include: Step S331: Update the cluster center. The formula used is as follows: ； In the formula, represents the updated cluster center of the kth cluster, represents the set of all sample data in the kth cluster, represents the total number of sample data in the kth cluster, represents the kth cluster in the kth cluster Sample data; Step S332: Calculate the intra-cluster complexity using the following formula: ； In the formula, represents the complexity calculation function, represents the mean calculation function, Represents the standard deviation calculation function; Step S333: Calculate the intra-cluster complexity of the sample data, using the following formula: ； In the formula, sci represents the intra-cluster complexity of sample data, Indicates the total number of sample data; Step S334: Calculate the similarity between clusters, the steps include: Step S3341: Calculate the global and local clustering radii using the following formula: ； In the formula, represents the global clustering radius calculation function, Represents the local cluster radius calculation function; Step S3342: Calculate the global and local inter-cluster distances using the following formula: ； In the formula, represents the global inter-cluster distance calculation function, represents the distance calculation function between local clusters, Indicates The set of all sample data in a cluster is Indicates The set of all sample data in a cluster is Indicates The first Sample data, Indicates Cluster centers; Step S3343: Obtain the similarity between clusters. The formula used is as follows: ； In the formula, represents the similarity calculation function between clusters, and e represents the base of the natural logarithm; Step S335: Calculate the similarity between clusters of sample data. The formula used is as follows: ； In the formula, ssi represents the similarity between clusters of sample data. Indicates The total number of sample data in the clusters, Indicates The total number of sample data in the clusters; Step S336: Calculate the distance dynamic weight, which is used to participate in the calculation of the distance between the sample data and the updated cluster center. The formula used for calculating the distance dynamic weight is as follows: ； In the formula, represents the distance dynamic weight, represents the adjustment weight, represents the intra-cluster complexity of the normalized sample data, Represents the similarity between clusters of normalized sample data; Step S337: Calculate the distance between the sample data and the updated cluster center and assign the cluster. Specifically, assign the sample data to the updated cluster center with the closest distance. The distance between the sample data and the updated cluster center is calculated as follows: ； In the formula, Indicates the distance calculation function with dynamic distance weight involved in the calculation; Step S34: iterative updating, specifically, continuously iteratively updating until the iteration termination condition is reached; Step S35: preliminary classification, specifically, constructing a dynamic weight K-means clustering model through the model initialization, the initial clustering allocation, the subsequent clustering allocation and the iterative update, and preliminarily classifying the geological non-time series data part of the samples in the historical geological data set based on the type of mining geological disasters, and combining the obtained labels with the geological time series data part of the samples in the historical geological data set to obtain a risk prediction data set.

3. The method for predicting geological disasters in mines based on artificial intelligence according to claim 1, characterized in that: In step S4, the disaster risk prediction model is constructed to construct a model required for predicting the risk of geological disasters in mining areas, specifically, a bidirectional long short-term memory network model combined with bee colony optimization is constructed as a disaster risk prediction model; The disaster risk prediction model is constructed, specifically comprising the following steps: Step S41: labeling and segmenting the data set, specifically labeling the data in the risk prediction data set as low risk, medium risk and high risk, and using the labeled data labels as data labels, and segmenting the labeled risk prediction data set into a risk prediction training set and a risk prediction test set; Step S42: constructing a bidirectional long short-term memory network model, the steps include: Step S421: Design an activation function, the formula used is as follows: ； In the formula, represents the activation function, represents the period adjustment parameter, represents the phase adjustment parameter, represents the input nonlinear response adjustment parameter, represents the smoothing gain parameter, represents the linear gain parameter, and x represents the independent variable of the activation function; Step S422: construct a forward module, the formula used is as follows: ； In the formula, represents the output of the forget gate at time t, represents the weight of the forward forget gate, represents the forward hidden state at time t-1, represents the model input at time t, represents the forward forget gate bias term, represents the output of the positive input gate at time t, represents the weight of the forward input gate, represents the forward input gate bias term, represents the candidate cell state at time t, represents the hyperbolic tangent function, represents the weight used to calculate the positive candidate cell state, represents the bias term used to calculate the positive candidate cell state, represents the cell state at the positive time t, represents the cell state at time t-1, It indicates the output of the positive output gate at time t. represents the forward output gate weight, represents the forward output gate bias term, represents the hidden state at the positive time t; Step S423: Construct a reverse module, the formula used is as follows: ； In the formula, Represents the output of the forget gate at the reverse time t, represents the weight of the reverse forget gate, represents the reverse hidden state at time t-1, represents the reverse forget gate bias term, Represents the output of the input gate at the reverse time t, represents the reverse input gate weight, represents the reverse input gate bias term, represents the candidate cell state at the reverse time t, represents the weight used to calculate the reverse candidate cell state, represents the bias term used to calculate the reverse candidate cell state, represents the cell state at the reverse time t, represents the cell state at the reverse time t-1, Represents the output of the output gate at time t in the reverse direction, represents the reverse output gate weight, represents the reverse output gate bias term, Represents the hidden state at the reverse time t; Step S424: construct an output module, the formula used is as follows: ； In the formula, represents the output of the model at time t, represents the softmax function, represents the model output weight, Represents the model output bias term; Step S425: constructing a model and training it, specifically, constructing a bidirectional long short-term memory network model by designing an activation function, constructing a forward module, constructing a reverse module, and constructing an output module, and training the model based on the risk prediction training set, and verifying the model performance based on the risk prediction test set, and the model loss function adopts a cross entropy loss function; Step S43: Hyperparameter optimization, specifically optimizing the hyperparameters of the bidirectional long short-term memory network model based on the bee colony optimization algorithm, the model hyperparameters specifically including the parameters of the Af activation function, the number of long short-term memory layers, the number of long short-term memory units, the model learning rate and the batch size; The hyperparameter optimization steps include: Step S431: initializing the algorithm, specifically initializing the search space and constructing a bee unit set, wherein the bee unit is used to represent a hyperparameter combination of a bidirectional long short-term memory network model, and the fitness function of the bee unit is a loss function of the bidirectional long short-term memory network model; Step S432: global search, the formula used is as follows: ； In the formula, represents the speed of the ath bee unit at the dt+1th iteration, represents the inertia weight, represents the speed of the a-th bee unit at the dt-th iteration, represents the weight used to control the learning of its own optimal solution, represents the optimal solution of the a-th bee unit. represents the position of the a-th bee unit at the dt-th iteration, represents the weight used to control the learning of the global optimal solution, represents the global optimal solution at the dtth iteration, represents the position of the ath bee unit at the dt+1th iteration; Step S433: local search, the steps include: Step S4331: Calculate the neighbor solution selection probability, the formula used is as follows: ； In the formula, represents the probability that the a-th bee unit selects the b-th neighbor bee unit, represents the fitness calculation function, N represents the number of neighbor bee units, represents the position of the bth neighbor bee unit at the dt+1th iteration; Step S4332: Design a neighbor solution learning strategy, specifically, the bee unit selects a neighbor bee unit based on the neighbor solution selection probability, and optimizes its own position based on the position of the selected neighbor bee unit. The formula used is as follows: ； In the formula, represents the position of the a-th bee unit at the dt+1-th iteration in the local search, de represents the decay factor whose size decreases linearly from 2 to 1, and rand represents a random number in the range [0,1]; Step S4333: Design a neighbor solution comparison strategy. Specifically, after the bee unit optimizes its own position based on the position of the neighbor bee unit, it selects a neighbor bee unit based on the neighbor solution selection probability again, and compares the fitness with the selected neighbor bee unit to determine whether to update the position. The formula used is as follows: ； In the formula, represents the fitness comparison function, represents the final position of the a-th bee unit at the dt+1-th iteration, represents the position of the cth neighbor bee unit at the dt+1th iteration; Step S434: Determine the global optimal solution, specifically by continuously iterating and updating until the optimization algorithm iteration termination condition is reached, and take the position of the bee unit with the lowest fitness as the global optimal solution, the global optimal solution is specifically the optimal bidirectional long short-term memory network model hyperparameter combination, and the optimization algorithm iteration termination condition specifically includes that the fitness of the bee unit is less than the set threshold and the maximum number of iterations is reached; Step S435: Optimizing model hyperparameters, specifically optimizing the hyperparameters of the bidirectional long short-term memory network model through the algorithm initialization, the global search, the local search and the determination of the global optimal solution, obtaining a bidirectional long short-term memory network model combined with swarm optimization, and using it as a disaster risk prediction model.

4. The method for predicting geological disasters in mines based on artificial intelligence according to claim 1, characterized in that: In step S1, the data collection is used to collect data required for predicting the risk of geological disasters in mining areas, specifically, obtaining an original data set for disaster prediction through data collection; The disaster prediction original data set specifically includes a historical geological original data set and a current geological original data set. Both the historical geological original data set and the current geological original data set include geological time series data and geological non-time series data. The geological time series data specifically includes mine meteorological data, mine soil moisture dynamic data, mine underground water level change data, disaster record data, slope displacement data and geological activity record data. The geological non-time series data specifically includes mine geographic space data and human activity data. The mine geographic space data specifically includes mine geographic coordinates, slope, slope direction, terrain type, soil type and stratigraphic characteristics. The human activity data specifically includes mining progress, mine depth and disaster prevention project intensity.

5. The method for predicting geological disasters in mines based on artificial intelligence according to claim 1, characterized in that: In step S2, the data preprocessing is used to preprocess the collected raw data, and specifically includes the following steps: Step S21: data cleaning, for cleaning the original data, specifically removing missing values ​​and duplicate values ​​in the historical geological original data set and the current geological original data set to obtain a historical preliminary data set and a current preliminary data set; Step S22: data encoding, which is used to encode the preliminary data, specifically, using a one-hot encoding method to encode the historical preliminary data set and the current preliminary data set to obtain a historical encoded data set and a current encoded data set; Step S23: data normalization, which is used to normalize the coded data, specifically, using the minimum-maximum method to normalize the historical coded data set and the current coded data set to obtain a historical geological data set and a current geological data set; Step S24: performing preprocessing, specifically, preprocessing the historical geological original data set and the current geological original data set through the data cleaning, the data encoding and the data normalization to obtain the historical geological data set and the current geological data set.

6. The method for predicting geological disasters in mines based on artificial intelligence according to claim 1, characterized in that: In step S5, the mine geological disaster prediction is specifically to use the current geological data set as the input of the dynamic weight K-means clustering model and the disaster risk prediction model to obtain mine geological disaster type reference data and disaster risk level reference data respectively, and comprehensively evaluate the risk of mine geological disasters based on the mine geological disaster type reference data and the disaster risk level reference data.

7. A mine geological disaster prediction system based on artificial intelligence, used to implement the mine geological disaster prediction method based on artificial intelligence as described in any one of claims 1 to 6, characterized in that: It includes data acquisition module, data preprocessing module, geological disaster classification model building module, disaster risk prediction model building module and mine geological disaster prediction module.

8. The mine geological disaster prediction system based on artificial intelligence according to claim 7 is characterized by: The data acquisition module is used for data acquisition, and obtains the original disaster prediction data set through data acquisition, and sends the original disaster prediction data set to the data preprocessing module; The data preprocessing module is used for data preprocessing, and obtains a historical geological data set and a current geological data set through data preprocessing, and sends the historical geological data set to the geological disaster classification model construction module, and sends the current geological data set to the mine geological disaster prediction module; The geological disaster classification model construction module is used for preliminary classification of geological disasters in mines. By constructing a dynamic weight K-means clustering model and performing preliminary classification, a dynamic weight K-means clustering model and a risk prediction data set are obtained, and the dynamic weight K-means clustering model is sent to the mine geological disaster prediction module, and the risk prediction data set is sent to the disaster risk prediction model construction module; The disaster risk prediction model construction module is used to construct a disaster risk prediction model, obtain the disaster risk prediction model by constructing a bidirectional long short-term memory network model combined with bee colony optimization, and send the disaster risk prediction model to the mine geological disaster prediction module; The mine geological disaster prediction module is used for mine geological disaster prediction. It predicts mine geological disasters by adopting the dynamic weight K-means clustering model and the disaster risk prediction model to obtain mine geological disaster type reference data and disaster risk level reference data.

Images