Coal mine equipment anomaly detection method based on multi-source data causal reconstruction

Abstract

The invention provides a coal mine equipment anomaly detection method based on multi-source data causal reconstruction, and the method comprises the steps: obtaining an initial time series data set of a measurement point of sample coal mine equipment, and obtaining a time series data set; performing causal discovery on the time sequence data set, and constructing a causal relationship graph of the time sequence data set; determining target time series data from the time series data set according to the causal relationship graph; and training a to-be-trained coal mine equipment anomaly detection model according to the target time sequence data to obtain a trained target coal mine equipment anomaly detection model, inputting the time sequence data of the to-be-detected coal mine equipment into the target coal mine equipment anomaly detection model to obtain an anomaly detection result, the causal relationship graph of the time sequence data set is constructed, the target time sequence data is determined according to the causal relationship graph, and anomaly detection is performed through the target coal mine equipment anomaly detection model, so that the accuracy and reliability of the coal mine equipment anomaly detection result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of deep learning, and particularly relates to a method for detecting abnormal conditions of coal mine equipment based on causal reconstruction of multi-source data. Background Art

[0002] In the field of coal production monitoring, there are significant pain points in the detection of abnormal conditions of coal mine equipment. Firstly, there is a problem of insufficient accuracy in predicting through single-source data, which is difficult to comprehensively reflect complex working conditions. Secondly, although multi-source data prediction can reflect abnormal working conditions through deep learning technology, its interpretability is poor, resulting in difficulty in effectively interpreting and trusting the detection results in practical applications. Summary of the Invention

[0003] This application aims to solve at least one of the technical problems in the related technologies to some extent.

[0004] According to the first aspect of this application, there is provided a method for detecting abnormal conditions of coal mine equipment based on causal reconstruction of multi-source data, including obtaining an initial time series data set of measuring points of a sample coal mine equipment, preprocessing the initial time series data set to obtain a time series data set; performing causal discovery on the time series data set to construct a causal relationship graph of the time series data set; determining target time series data from the time series data set according to the causal relationship graph; training a coal mine equipment abnormal condition detection model to be trained according to the target time series data to obtain a trained target coal mine equipment abnormal condition detection model; obtaining the time series data of the coal mine equipment to be detected, and inputting the time series data into the target coal mine equipment abnormal condition detection model to obtain the abnormal condition detection result of the coal mine equipment to be detected.

[0005] According to the second aspect of this application, there is provided a device for detecting abnormal conditions of coal mine equipment based on causal reconstruction of multi-source data, including: an acquisition module, configured to obtain an initial time series data set of measuring points of a sample coal mine equipment, and preprocess the initial time series data set to obtain a time series data set; a processing module, configured to perform causal discovery on the time series data set to construct a causal relationship graph of the time series data set; a determination module, configured to determine target time series data from the time series data set according to the causal relationship graph; a training module, configured to train a coal mine equipment abnormal condition detection model to be trained according to the target time series data to obtain a trained target coal mine equipment abnormal condition detection model; an abnormal condition detection module, configured to obtain the time series data of the coal mine equipment to be detected, and input the time series data into the target coal mine equipment abnormal condition detection model to obtain the abnormal condition detection result of the coal mine equipment to be detected.

[0006] To achieve the above objectives, a third aspect of the present application provides an electronic device, which is characterized by including: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction as described in the first aspect.

[0007] To achieve the above objectives, a fourth aspect of the present application provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction as described in the first aspect.

[0008] To achieve the above objectives, a fifth aspect of the present application provides a computer program product, including a computer program, which when executed by a processor, implements the method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction as described in the first aspect.

[0009] The technical solutions provided in the embodiments of the present application at least include the following beneficial effects:

[0010] The present application provides a method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction. According to the causal relationship diagram, the target time series data is determined, and based on the target time series data, the abnormal detection model of coal mine equipment is trained. Through the abnormal detection model of the target coal mine equipment, the abnormal detection of coal mine equipment is carried out. Through causal analysis, it can be ensured that the abnormal detection results of coal mine equipment are interpretable, and the accuracy and reliability of the abnormal detection of coal mine equipment are improved.

[0011] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become easily understandable through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings are used to better understand the solution and do not constitute a limitation to the present application. Among them:

[0013] Figure 1 is a schematic flow chart of a method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction provided by an embodiment of the present application;

[0014] Figure 2 is a schematic flow chart of another method for abnormal detection of coal mine equipment based on multi-source data causal reconstruction provided by an embodiment of the present application;

[0015] Figure 3 is a schematic structural diagram of a device for abnormal detection of coal mine equipment based on multi-source data causal reconstruction provided by an embodiment of the present application;

[0016] Figure 4 A schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0017] The following describes exemplary embodiments of the present application with reference to the accompanying drawings. Various details of the embodiments of the present application are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, descriptions of well-known functions and structures are omitted below for clarity and conciseness.

[0018] The following uses embodiments to elaborate in detail on the coal mine equipment anomaly detection method based on multi-source data causal reconstruction of the present application.

[0019] Figure 1 A schematic flowchart of the coal mine equipment anomaly detection method based on multi-source data causal reconstruction provided by an embodiment of the present application.

[0020] As Figure 1 shown, the coal mine equipment anomaly detection method based on multi-source data causal reconstruction proposed in this embodiment specifically includes the following steps:

[0021] S101. Obtain the initial time series data set of the measurement points of the sample coal mine equipment, and preprocess the initial time series data set to obtain the time series data set.

[0022] In the embodiment of the present application, the time series operation data and time series working environment data of the sample coal mine equipment can be obtained, it is determined whether there is a collaborative device for the sample coal mine equipment, in response to the existence of a collaborative device for the sample coal mine equipment, the time series data of the collaborative device is obtained, and according to the time series operation data, time series working environment data and time series data of the collaborative device, the initial time series data set is obtained.

[0023] Among them, the time series operation data can be the measurement point data collected by the sample coal mine equipment itself, and the above data usually includes the real-time operation data recorded by various sensors.

[0024] For example, if the sample coal mine equipment is a shearer, the time series operation data of the shearer includes but is not limited to temperature time series data, pressure time series data, position time series data, voltage time series data, current time series data, and mechanical state time series data.

[0025] Among them, for the temperature time series data, the temperature data of each key component (such as the motor and hydraulic system) of the shearer can be monitored to prevent equipment damage caused by overheating.

[0026] Among them, for the position time series data, the position data of the shearer in the working face can be obtained in real time through a positioning system to ensure automated mining and equipment scheduling.

[0027] Among them, for the voltage time series data and current time series data, the current data and voltage data of the shearer motor can be monitored in real time to evaluate the electrical performance and prevent electrical faults.

[0028] Among them, for the mechanical state time series data, such as vibration time series data and noise time series data, to reflect the mechanical health status of the shearer.

[0029] It should be noted that the time series working environment data includes but is not limited to gas concentration data, dust level data, and geological condition data. Among them, the time series working environment data is crucial for ensuring the safety of the working face. By monitoring the gas concentration data and dust level data, it is beneficial to prevent explosions and occupational health problems. By monitoring the geological data, it helps to adjust the mining strategy to adapt to different coal seam conditions.

[0030] For example, if the sample coal mine equipment is a shearer, the collaborative equipment is a support. As an important support structure in coal mine mining, the spatial displacement data is crucial for ensuring operation safety. By analyzing the spatial displacement data of the support, the characteristics reflecting the displacement change can be constructed, providing an important basis for the safety monitoring and maintenance of the coal mine. Then, the spatial displacement data of the support can be used as the time series data of the collaborative equipment.

[0031] In the embodiment of the present application, after obtaining the time series operation data, time series working environment data, and time series data of the collaborative equipment, the time series operation data, time series working environment data, and time series data of the collaborative equipment can be summarized to obtain an initial time series data set.

[0032] In the embodiment of the present application, after obtaining the initial time series data set, a preprocessing operation for the initial time series data set can be obtained. Among them, the preprocessing operation at least includes time alignment operation, outlier removal operation, and missing value filling operation. The preprocessing operation is performed on the initial time series data set to obtain a time series data set.

[0033] It should be noted that by obtaining the time series data set, the time series data sets together constitute a holographic portrait of the operating state of the sample coal mine equipment.

[0034] S102. Perform causal discovery on the time series data set to construct a causal relationship graph of the time series data set.

[0035] In an embodiment of the present disclosure, each type of time series data in the time series dataset can be used as a node in the initial undirected graph, and all nodes are connected to construct a fully connected initial undirected graph. An independence test is performed on each pair of nodes in the initial undirected graph. According to the independence test results, the initial undirected graph is updated to obtain an updated first undirected graph. A momentary conditional independence (MCI) test is performed on each pair of nodes in the first undirected graph to obtain the causal relationship and the corresponding causal direction between each pair of nodes. According to the causal relationship and the corresponding causal direction, the first undirected graph is updated to obtain a causal relationship graph.

[0036] S103. Determine target time series data from the time series dataset according to the causal relationship graph.

[0037] In an embodiment of the present application, the state of the sample coal mine equipment can be used as the analysis target. The target nodes having a causal relationship with the state of the sample coal mine equipment are determined from the causal relationship graph. The time series data corresponding to the target nodes is obtained from the time series dataset, and the time series data corresponding to the target nodes is used as the target time series data.

[0038] For example, if it can be seen from the causal relationship graph that there are arrows pointing to the state (S) of the shearer from vibration (V) and temperature (T), that is, vibration (V) and temperature (T) are the target nodes having a causal relationship with the state of the sample coal mine equipment, then the vibration time series data and the temperature time series data can be determined from the time series dataset as the target time series data.

[0039] It should be noted that an increase in temperature may cause the state of the shearer to deteriorate, an increase in vibration intensity may indicate damage to the shearer components, and a decrease in oil pressure may cause poor lubrication, thereby affecting the state of the shearer.

[0040] S104. Train the coal mine equipment anomaly detection model to be trained according to the target time series data to obtain a trained target coal mine equipment anomaly detection model.

[0041] In an embodiment of the present application, after obtaining the target time series data, the target time series data can be used as training data, the target time series data is input into the coal mine equipment anomaly detection model, the anomaly detection result of the sample coal mine equipment is obtained, the sample label of the sample coal mine equipment is obtained, and the coal mine equipment anomaly detection model to be trained is trained according to the sample label and the anomaly detection result, and cross-validation is performed to obtain a trained target coal mine equipment anomaly detection model.

[0042] In the embodiments of the present application, the loss function of the anomaly detection model of coal mine equipment can be determined according to the sample labels and anomaly detection results, and based on the loss function, the model parameters of the anomaly detection model of coal mine equipment are adjusted, and the adjusted anomaly detection model of coal mine equipment is continuously trained until the training is completed, so as to obtain the anomaly detection model of the target coal mine equipment.

[0043] Optionally, according to the loss function, the model parameters of the anomaly detection model of coal mine equipment are adjusted until the loss function meets the training end condition, and the anomaly detection model of coal mine equipment after the last adjustment of the model parameters is determined as the anomaly detection model of the target coal mine equipment that has completed training.

[0044] It should be noted that the setting of the training end condition in the present application is not limited and can be selected according to the actual situation.

[0045] Optionally, the training end condition can be set as the total loss function value being less than a preset loss threshold; optionally, the training end condition can be set as the number of adjustments of the parameters of the anomaly detection model of coal mine equipment reaching a preset number threshold.

[0046] S105. Obtain the time series data of the coal mine equipment to be detected, input the time series data into the target coal mine equipment anomaly detection model, and obtain the anomaly detection result of the coal mine equipment to be detected.

[0047] Among them, the coal mine equipment to be detected can be any equipment that needs to be subjected to anomaly detection. For example, the coal mine equipment to be detected can be a shearer.

[0048] For example, in the process of anomaly detection of coal mine equipment based on causal reconstruction of multi-source data, if the target time series data includes the time series data corresponding to temperature and vibration, the temperature time series data and vibration time series data corresponding to the coal mine equipment to be detected can be used as the time series data of the coal mine equipment to be detected.

[0049] In the embodiments of the present application, the time series data can be input into the target coal mine equipment anomaly detection model, and the target coal mine equipment anomaly detection model performs anomaly detection on the time series data to obtain the anomaly detection result of the coal mine equipment to be detected.

[0050] The coal mine equipment anomaly detection method based on causal reconstruction of multi-source data provided by this application obtains the initial time series data set of the measurement points of the sample coal mine equipment, preprocesses the initial time series data set to obtain the time series data set, discovers the causality of the time series data set, constructs a causal relationship graph of the time series data set, determines the target time series data from the time series data set according to the causal relationship graph, trains the coal mine equipment anomaly detection model to be trained according to the target time series data to obtain the trained target coal mine equipment anomaly detection model, obtains the time series data of the coal mine equipment to be detected, inputs the time series data into the target coal mine equipment anomaly detection model, and obtains the anomaly detection result of the coal mine equipment to be detected. Thus, this application determines the target time series data according to the causal relationship graph, trains the coal mine equipment anomaly detection model according to the target time series data, and performs anomaly detection on the coal mine equipment through the anomaly detection model of the target coal mine equipment. Through causal analysis, it can ensure that the anomaly detection result of the coal mine equipment is interpretable, and improve the accuracy and reliability of the anomaly detection of the coal mine equipment.

[0051] Figure 2 It is a schematic flowchart of the coal mine equipment anomaly detection method based on causal reconstruction of multi-source data provided by an embodiment of this application.

[0052] As Figure 2 shown, the coal mine equipment anomaly detection method based on causal reconstruction of multi-source data proposed in this embodiment specifically includes the following steps:

[0053] S201. Obtain the initial time series data set of the measurement points of the sample coal mine equipment, and preprocess the initial time series data set to obtain the time series data set.

[0054] S202. Take each type of time series data in the time series data set as a node in the initial undirected graph, and connect all the nodes to construct a fully connected initial undirected graph.

[0055] S203. Perform an independence test on each pair of nodes in the initial undirected graph, and update the initial undirected graph according to the independence test result to obtain the updated first undirected graph.

[0056] Optionally, the independence between each pair of nodes in the graph can be tested through statistical methods such as the Pearson correlation coefficient or mutual information. According to the test result, if two nodes are considered independent, that is, there is no significant statistical correlation between the two nodes, then the edge between the two nodes is removed from the initial undirected graph, and only the edges between the node pairs with significant correlation are retained to obtain the updated first undirected graph.

[0057] S204. Perform a transient conditional independence test MCI on each pair of nodes in the first undirected graph to obtain the causal relationship and the corresponding causal direction between each pair of nodes.

[0058] S205. Update the first undirected graph according to the causal relationship and the corresponding causal direction to obtain a causal relationship graph.

[0059] Optionally, based on the first undirected graph, perform a transient conditional independence test. Among them, MCI is a more complex statistical test that considers the time order and conditional dependence of time series data and can reveal the causal relationship and its direction between variables. According to the test results of MCI, obtain the causal relationship and the corresponding causal direction between each pair of nodes. Update the first undirected graph according to the causal relationship and the corresponding causal direction to obtain a causal relationship graph.

[0060] For example, take temperature (T), vibration (V), oil pressure (O), current (I), and the state of the shearer (S) as nodes in the initial undirected graph. At the initial stage, assume that there are potential associations between all nodes and connect all nodes to construct a fully connected initial undirected graph. If the independence test results show that there is a significant correlation between temperature (T) and vibration (V), oil pressure (O), but no significant correlation with current (I), then the edge between temperature (T) and current (I) will be removed. If through the transient conditional independence test MCI, it is obtained that vibration (V) and temperature (T) are the causes of the state of the shearer (S), and there are arrows from vibration (V) and temperature (T) pointing to the state of the shearer (S), then the first undirected graph can be updated according to the causal relationship and the corresponding causal direction to obtain a causal relationship graph. The causal relationship graph clearly shows how temperature (T), vibration (V), oil pressure (O), and current (I) affect the state of the shearer, providing strong support for subsequent fault detection.

[0061] S206. Determine the target time series data from the time series dataset according to the causal relationship graph.

[0062] S207. Train the coal mine equipment anomaly detection model to be trained according to the target time series data to obtain the trained target coal mine equipment anomaly detection model.

[0063] S208. Obtain the time series data of the coal mine equipment to be detected, input the time series data into the target coal mine equipment anomaly detection model, and obtain the anomaly detection result of the coal mine equipment to be detected.

[0064] S209. In response to the anomaly detection result of the coal mine equipment to be detected being abnormal, generate an anomaly warning prompt message for the equipment to be detected.

[0065] It should be noted that the present application does not limit the specific form of the abnormal warning prompt information. For example, the abnormal warning prompt information can be "acoustic and optical warning prompt information", "voice warning prompt information", "text warning prompt information", etc.

[0066] In summary, the coal mine equipment abnormal detection method based on multi-source data causal reconstruction provided by the present application discovers causality for a multi-source time series data set, constructs a causal relationship graph, determines target time series data according to the causal relationship graph, trains a coal mine equipment abnormal detection model through the target time series data, and performs abnormal detection of coal mine equipment based on the trained abnormal detection model of the target coal mine equipment, ensuring that the abnormal detection result of the coal mine equipment is interpretable, improving the accuracy and reliability of the abnormal detection of the coal mine equipment, being able to timely discover potential problems of the coal mine equipment to be detected according to the abnormal detection result of the coal mine equipment to be detected, and being able to timely remind relevant personnel through the abnormal warning prompt information, avoiding the occurrence of faults and production accidents of the coal mine equipment to be detected, and improving the efficiency and safety of coal mine production.

[0067] To implement the above embodiments, the present embodiment provides a coal mine equipment abnormal detection device based on multi-source data causal reconstruction. Figure 3 It is a structural schematic diagram of a coal mine equipment abnormal detection device based on multi-source data causal reconstruction provided by an embodiment of the present application.

[0068] As Figure 3 shown, the coal mine equipment abnormal detection device 1000 based on multi-source data causal reconstruction includes: an acquisition module 110, a processing module 120, a determination module 130, a training module 140, and an abnormal detection module 150.

[0069] The acquisition module 110 is configured to acquire an initial time series data set of the measurement points of the sample coal mine equipment, and preprocess the initial time series data set to obtain a time series data set.

[0070] The processing module 120 is configured to perform causal discovery on the time series data set and construct a causal relationship graph of the time series data set.

[0071] The determination module 130 is configured to determine target time series data from the time series data set according to the causal relationship graph.

[0072] The training module 140 is configured to train a coal mine equipment abnormal detection model to be trained according to the target time series data to obtain a trained target coal mine equipment abnormal detection model.

[0073] Anomaly detection module 150 is configured to obtain time series data of a coal mine equipment to be detected, input the time series data into a target coal mine equipment anomaly detection model, and obtain an anomaly detection result of the coal mine equipment to be detected.

[0074] According to an embodiment of the present application, the processing module 120 is further configured to: use each type of time series data in the time series data set as a node in an initial undirected graph, connect all the nodes, and construct a fully connected initial undirected graph; perform an independence test on each pair of nodes in the initial undirected graph, and update the initial undirected graph according to the independence test result to obtain an updated first undirected graph; perform a transient conditional independence test MCI on each pair of nodes in the first undirected graph to obtain a causal relationship and a corresponding causal direction between each pair of nodes; and update the first undirected graph according to the causal relationship and the corresponding causal direction to obtain the causal relationship graph.

[0075] According to an embodiment of the present application, the determination module 130 is further configured to: use the state of the sample coal mine equipment as an analysis target, determine a target node in the causal relationship graph that has a causal relationship with the state of the sample coal mine equipment; obtain the time series data corresponding to the target node from the time series data set, and use the time series data corresponding to the target node as the target time series data.

[0076] According to an embodiment of the present application, the training module 140 is further configured to: input the target time series data into the coal mine equipment anomaly detection model to obtain an anomaly detection result of the sample coal mine equipment; obtain a sample label of the sample coal mine equipment; and train the coal mine equipment anomaly detection model to be trained according to the sample label and the anomaly detection result to obtain a trained target coal mine equipment anomaly detection model.

[0077] According to an embodiment of the present application, the acquisition module 110 is further configured to: obtain time series operation data and time series working environment data of the sample coal mine equipment; determine whether there is a collaborative equipment for the sample coal mine equipment; in response to the sample coal mine equipment having a collaborative equipment, obtain the time series data of the collaborative equipment; and obtain the initial time series data set according to the time series operation data, the time series working environment data, and the time series data of the collaborative equipment.

[0078] According to an embodiment of the present application, the acquisition module 110 is further configured to: obtain a preprocessing operation for the initial time series data set, where the preprocessing operation at least includes a time alignment operation, an outlier removal operation, and a missing value filling operation; and perform the preprocessing operation on the initial time series data set to obtain the time series data set.

[0079] According to an embodiment of the present application, the device is further configured to: in response to the abnormal detection result of the coal mine equipment to be detected being abnormal, generate an abnormal warning prompt message for the equipment to be detected.

[0080] The coal mine equipment abnormal detection device based on multi-source data causal reconstruction provided by the present application obtains an initial time series data set of the measurement points of the sample coal mine equipment, preprocesses the initial time series data set to obtain a time series data set, performs causal discovery on the time series data set, constructs a causal relationship graph of the time series data set, determines target time series data from the time series data set according to the causal relationship graph, trains a coal mine equipment abnormal detection model to be trained according to the target time series data to obtain a trained target coal mine equipment abnormal detection model, obtains the time series data of the coal mine equipment to be detected, inputs the time series data into the target coal mine equipment abnormal detection model to obtain the abnormal detection result of the coal mine equipment to be detected. Thus, the present application determines the target time series data according to the causal relationship graph, trains the coal mine equipment abnormal detection model according to the target time series data, and performs abnormal detection of the coal mine equipment through the abnormal detection model of the target coal mine equipment. Through causal analysis, it can ensure that the abnormal detection result of the coal mine equipment is interpretable, and improve the accuracy and reliability of the abnormal detection of the coal mine equipment.

[0081] To implement the above embodiment, the present application also proposes an electronic device 2000, as Figure 4 shown, including: a memory 210, a processor 220, and a computer program stored on the memory 210 and executable on the processor 220. When the processor executes the program, it implements the coal mine equipment abnormal detection method based on multi-source data causal reconstruction as described in the first aspect.

[0082] To implement the above embodiment, the present application proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the coal mine equipment abnormal detection method based on multi-source data causal reconstruction as described in the first aspect.

[0083] To implement the above embodiment, the present application also proposes a computer program product, including a computer program, where the computer program implements the coal mine equipment abnormal detection method based on multi-source data causal reconstruction as described in the first aspect when executed by a processor.

[0084] It should be understood that the various forms of processes shown above can be used, with steps reordered, added or deleted. For example, the steps described in the present application can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in the present application can be achieved, and no limitations are imposed herein.

[0085] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles disclosed in the present application shall be included within the protection scope of the present application.

Claims

1. A method for abnormal detection of coal mine equipment based on causal reconstruction of multi-source data, characterized in that The method includes: Obtaining an initial time series data set of measurement points of a sample coal mine device, and preprocessing the initial time series data set to obtain a time series data set; Performing causal discovery on the time series data set to construct a causal relationship graph of the time series data set; Determining target time series data from the time series data set according to the causal relationship graph; Training a coal mine device anomaly detection model to be trained according to the target time series data to obtain a trained target coal mine device anomaly detection model; Obtaining the time series data of the coal mine device to be detected, inputting the time series data into the target coal mine device anomaly detection model, and obtaining the anomaly detection result of the coal mine device to be detected.

2. The method according to claim 1, wherein The performing causal discovery on the time series data set to construct a causal relationship graph of the time series data set includes: Taking each type of time series data in the time series data set as a node in an initial undirected graph, and connecting all the nodes to construct a fully connected initial undirected graph; Performing an independence test on each pair of nodes in the initial undirected graph, and updating the initial undirected graph according to the independence test result to obtain an updated first undirected graph; Performing a transient conditional independence test MCI on each pair of nodes in the first undirected graph to obtain the causal relationship and the corresponding causal direction between each pair of nodes; Updating the first undirected graph according to the causal relationship and the corresponding causal direction to obtain the causal relationship graph.

3. The method according to claim 2, characterized in that The determining target time series data from the time series data set according to the causal relationship graph includes: Taking the state of the sample coal mine device as an analysis target, and determining target nodes in the causal relationship graph that have a causal relationship with the state of the sample coal mine device; Obtaining the time series data corresponding to the target nodes from the time series data set, and taking the time series data corresponding to the target nodes as the target time series data.

4. The method according to claim 1, characterized in that, The training a coal mine device anomaly detection model to be trained according to the target time series data to obtain a trained target coal mine device anomaly detection model includes: Inputting the target time series data into the coal mine device anomaly detection model to obtain the anomaly detection result of the sample coal mine device; Obtaining the sample label of the sample coal mine device; Training the coal mine device anomaly detection model to be trained according to the sample label and the anomaly detection result to obtain a trained target coal mine device anomaly detection model.

5. The method according to claim 1, wherein The obtaining an initial time series data set of measurement points of a sample coal mine device includes: Obtaining the time series operation data and time series working environment data of the sample coal mine device; Judging whether there is a collaborative device for the sample coal mine device; In response to the existence of a collaborative device for the sample coal mine device, obtaining the time series data of the collaborative device; Obtaining the initial time series data set according to the time series operation data, the time series working environment data, and the time series data of the collaborative device.

6. The method according to claim 5, wherein Preprocessing the initial time series dataset to obtain a time series dataset, including: Obtaining preprocessing operations for the initial time series dataset, where the preprocessing operations at least include time alignment operations, outlier removal operations, and missing value filling operations; Performing preprocessing operations on the initial time series dataset to obtain the time series dataset.

7. The method according to claim 1, wherein The method further includes: In response to the abnormal detection result of the coal mine equipment to be detected being abnormal, generating an abnormal warning prompt message for the equipment to be detected.

8. An abnormal detection device for coal mine equipment based on causal reconstruction of multi-source data, characterized in that The device includes: An acquisition module, configured to acquire an initial time series dataset of measurement points of a sample coal mine equipment, and preprocess the initial time series dataset to obtain a time series dataset; A processing module, configured to perform causal discovery on the time series dataset and construct a causal relationship graph of the time series dataset; A determination module, configured to determine target time series data from the time series dataset according to the causal relationship graph; A training module, configured to train a coal mine equipment abnormal detection model to be trained according to the target time series data to obtain a trained target coal mine equipment abnormal detection model; An abnormal detection module, configured to acquire time series data of a coal mine equipment to be detected, input the time series data into the target coal mine equipment abnormal detection model, and obtain an abnormal detection result of the coal mine equipment to be detected.

9. An electronic device, comprising: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1-7.

Images