Remote intelligent operation and maintenance system and method for coal mine equipment

Abstract

The invention discloses a remote intelligent operation and maintenance system and method for coal mine equipment, and relates to the technical field of fault diagnosis, and the system comprises a local client which is used for obtaining fault sample data of the coal mine equipment and training a local model; the cloud server is used for aggregating the parameters of the local model to obtain a diagnosis model; and the controller is used for acquiring the real-time state data and the real-time operation data in the operation process of the coal mine equipment, and obtaining an early warning result of the fault of the coal mine equipment based on the diagnosis model. According to the method, the diagnosis model is constructed in a federated learning mode, and the diagnosis model is written into the controller of the coal mine equipment, so that fault early warning can be carried out in the operation process of the coal mine equipment, the downtime of the equipment is greatly shortened, and the benefits of coal mine production are improved.

Description

Technical Field

[0001] The present application relates to the technical field of fault diagnosis, and particularly relates to a remote intelligent operation and maintenance system and method for coal mine equipment. Background Art

[0002] In coal mine production, many mechanical equipment are used. These equipment usually work continuously for a long time. Therefore, it is very important to monitor the state of coal mine equipment in real time and accurately. By monitoring data such as vibration, oil temperature, oil pressure, and current of coal mine equipment, early warnings can be given for possible faults, or rapid and accurate diagnosis can be carried out after a fault occurs, so as to reduce the downtime of the equipment and improve the efficiency of coal mine production.

[0003] For underground coal mine mining, due to the influence of underground rock strata, water, etc., wireless communication technology cannot be used to monitor coal mine equipment in real time. And because there are many underground roadways with large lengths, and the environment is harsh, wired communication cannot be used normally underground. Therefore, many underground coal mine equipment are still monitored and diagnosed manually. This method has serious subjective influence and the accuracy rate is not high. To solve the problem of fault diagnosis in this offline mode, CN117826751A discloses a remote fault diagnosis method and system for underground equipment in coal mines. By integrating a storage device into the underground coal mine equipment, when a fault occurs in the equipment, the data is manually transferred to the ground, and the equipment manufacturer diagnoses the fault according to the data.

[0004] However, the above patent is only applicable after a fault occurs in the equipment. At this time, the coal mine equipment has stopped running, resulting in the interruption of coal mine production. Therefore, it is necessary to give early warnings for faults of coal mine equipment when the equipment is running, so that the vast majority of faults can be solved in advance before they occur. Summary of the Invention

[0005] The embodiments of the present application provide a remote intelligent operation and maintenance system and method for coal mine equipment to solve the problem of diagnosing only after a fault occurs in the prior art.

[0006] On the one hand, the embodiments of the present application provide a remote intelligent operation and maintenance system for coal mine equipment, including:

[0007] A local client, configured to obtain fault sample data of coal mine equipment. The fault sample data includes state sample data of coal mine equipment and operation sample data input by an operator; extract feature information of the fault sample data, and train a local model based on the feature information;

[0008] A cloud server, configured to receive parameters of the local model uploaded by the local client. The cloud server aggregates the parameters of the local model to obtain a diagnosis model;

[0009] A controller is set in coal mine equipment. During the operation of the coal mine equipment, the controller collects real-time status data and real-time operation data, inputs the real-time status data and real-time operation data into a status prediction model to obtain predicted status data after a period of time in the future. The controller inputs the predicted status data into an operation prediction model to obtain predicted operation data corresponding to the time of the predicted status data, and inputs the predicted status data and predicted operation data into a diagnosis model to obtain a warning result for the faults of the coal mine equipment.

[0010] On the other hand, an embodiment of the present application also provides a remote intelligent operation and maintenance method for coal mine equipment, including:

[0011] Obtain fault sample data of coal mine equipment, where the fault sample data includes status sample data of the coal mine equipment and operation sample data input by the operator;

[0012] Extract the feature information of the fault sample data and train a local model based on the feature information;

[0013] Aggregate the parameters of the local model to obtain a diagnosis model;

[0014] Download the diagnosis model to the controller of the coal mine equipment. During the operation of the coal mine equipment, the controller collects real-time status data and real-time operation data, and inputs the real-time status data and real-time operation data into a status prediction model to obtain predicted status data after a period of time in the future;

[0015] The controller inputs the predicted status data into an operation prediction model to obtain predicted operation data corresponding to the time of the predicted status data;

[0016] The controller inputs the predicted status data and predicted operation data into the diagnosis model to obtain a warning result for the faults of the coal mine equipment.

[0017] A remote intelligent operation and maintenance system and method for coal mine equipment in the present application has the following advantages:

[0018] By using the method of federated learning to construct a diagnosis model, not only the accuracy of the diagnosis model is improved, but also the security of the privacy data in each mine is improved. Writing the diagnosis model into the controller of the coal mine equipment can carry out fault warning during the operation of the coal mine equipment and generate reminder information before the fault occurs, without the need for the equipment manufacturer to diagnose every time a fault occurs, which greatly shortens the downtime of the equipment and improves the efficiency of coal mine production. Description of the Drawings

[0019] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0020] Figure 1 It is a schematic diagram of the functional module composition of a remote intelligent operation and maintenance system for coal mine equipment provided by an embodiment of the present application.

[0021] Figure 2 It is a flowchart of a remote intelligent operation and maintenance method for coal mine equipment provided by an embodiment of the present application. Detailed implementation manners

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0023] Figure 1 It is a schematic diagram of the functional modules of a remote intelligent operation and maintenance system for coal mine equipment provided by an embodiment of the present application. An embodiment of the present application provides a remote intelligent operation and maintenance system for coal mine equipment, including the following modules:

[0024] The local client is used to obtain the fault sample data of the coal mine equipment. The fault sample data includes the status sample data of the coal mine equipment and the operation sample data input by the operator; extract the feature information of the fault sample data and train the local model based on the feature information.

[0025] The cloud server is used to receive the parameters of the local model uploaded by the local client. The cloud server aggregates the parameters of the local model to obtain the diagnostic model.

[0026] The controller is set in the coal mine equipment. During the operation of the coal mine equipment, the controller collects the real-time status data and real-time operation data, inputs the real-time status data and real-time operation data into the status prediction model to obtain the predicted status data after a period of time in the future. The controller inputs the predicted status data into the operation prediction model to obtain the predicted operation data corresponding to the time of the predicted status data, and inputs the predicted status data and predicted operation data into the diagnostic model to obtain the early warning result of the fault of the coal mine equipment.

[0027] Exemplarily, the local client is deployed in each mine. During operation, the controller can not only convert various operation data input by the operator into corresponding control instructions and send them to the corresponding actuators, but also record these operation data. At the same time, it will collect status data including vibration, oil temperature, oil pressure, rotational speed, and current through sensors installed at various key positions in the coal mine equipment, and these status data will also be saved by the controller. At regular intervals, for example, during the routine maintenance of the coal mine equipment, the operator needs to remove the controller from the coal mine equipment and upload the saved operation data and status data to the local client after bringing it to the ground.

[0028] After experienced operators view the operation data and status data, they will set labels for each set of operation data and status data recorded at the same time. This label indicates whether the coal mine equipment is normal or faulty under the corresponding status data and operation data, and if there is a fault, what the specific fault type is. These operation data and status data will respectively become operation sample data and status sample data, and the status sample data, operation sample data, and labels will form the fault sample data.

[0029] After obtaining the fault sample data, it needs to be preprocessed first. The preprocessing can include data cleaning, outlier removal, and normalization, etc. The preprocessed fault sample data will be divided into a training set and a test set according to a ratio of 7:3. The local model will be trained using the training set, and the trained local model will then be tested using the test set.

[0030] In the embodiment of this application, the local model adopts a Transformer model, and during the training process using the training set, the local client will use a generative adversarial network (GAN) to train the local model.

[0031] The generative adversarial network consists of two core parts: a generator and a discriminator. In the scenario of coal mine equipment fault diagnosis, the task of the generator is to learn the potential distribution rules of various real operating states and fault modes of the coal mine equipment. Based on these learning results, it can generate seemingly real fault sample data. For example, for the operation of a shearer, the generator can simulate the current, rotational speed fluctuation data when the motor has different degrees of faults, as well as the corresponding equipment vibration data, etc. It starts from a random noise input and gradually constructs complex fault scenario simulation data through a multi-layer neural network.

[0032] The discriminator receives real coal mine equipment operation data and simulated data generated by the generator. Through comparative learning of a large amount of real and simulated data, the discriminator has a powerful discrimination ability and can accurately determine whether the input data comes from real equipment or is fabricated by the generator. The internal neural network structure deeply extracts and analyzes the features of the data, such as discriminating features like vibration spectra and temperature change curves.

[0033] The generator and the discriminator confront and play against each other. In this process, the generator continuously optimizes the generated data to make it more and more realistic, and the discriminator also continuously improves its discrimination ability. Eventually, a balanced state is reached, at which time the generator can generate high-quality simulated fault scenario data, and these data are highly similar to real fault data in terms of feature distribution and so on.

[0034] Using a generative adversarial network can generate a large number of high-quality fault sample data, thereby improving the accuracy of the local model.

[0035] Furthermore, the local model is the result of aggregating multiple different types of computer models. After training each computer model using feature information, the parameters of all the trained computer models are aggregated to establish the local model.

[0036] Specifically, these computer models include decision tree models, neural network models, and support vector machine models. The characteristics of different mine sites or equipment areas vary significantly, and a single type of model may not be able to fully adapt to all situations. For example, in some coal mines, the underground environment has high humidity and a lot of dust, which has a unique impact on equipment erosion; in some coal mines with a large mining depth, the pressure and stress conditions on the equipment are complex. Aggregating multiple different computer models, such as decision trees, neural networks, and support vector machines, can cover different modeling ideas and advantages. Among these computer models, the decision tree model can intuitively display the hierarchical relationship of fault features and is easy to interpret; the neural network model has strong ability to mine complex non-linear equipment operation data associations; the support vector machine model performs well in scenarios with small sample fault data. Combining their advantages can more accurately capture the subtle features of equipment faults under different working conditions.

[0037] By training different computer models using the same training set and then testing them using the same test set, multiple trained computer models can be obtained. These computer models have the same type of parameters. Therefore, after aggregating the corresponding parameters, they can be assigned to the local model in the initial state, and then a local model with good adaptability and accuracy can be obtained.

[0038] After the processing of federated learning, it is possible to desensitize the private data of the mine to a certain extent, improving the security of the private data. On the basis of federated learning, this application also aggregates multiple computer models to form a local model, further enhancing the security of the private data in the mine.

[0039] The cloud server is set outside each mine. The parameters of the local models trained by the local clients in multiple mines will be gathered at the cloud server. After the cloud server sets weights according to the accuracy of each local model, it performs weighted summation on the parameters of the corresponding local models according to the weights to obtain the diagnostic model. In the embodiments of this application, both the diagnostic model and the local model are Transformer models and they have the same structure.

[0040] The state prediction model is used to predict the future state of the coal mine equipment, while the operation prediction model is used to predict the operations that the operators may perform on the coal mine equipment. In the embodiments of this application, both the state prediction model and the operation prediction model are established based on LSTM (Long Short-Term Memory network). Moreover, the state prediction model is trained using the historical state data and historical operation data stored in the local client, and the operation prediction model is also trained using the historical operation data and historical state data stored in the local client. The difference is that the state prediction model is obtained by analyzing the influence of historical operation data on historical state data, while the operation prediction model is obtained by analyzing the influence of historical state data on historical operation data.

[0041] Since the future state of the coal mine equipment is the result of the combined action of the current state and the control operations made by the operators in the current state, historical state data and historical operation data are required to participate in the training process. Moreover, in the actual use process, real-time state data and real-time operation data also need to be input into the state prediction model simultaneously to obtain accurate predicted state data. Although the control operations made by the operators have a certain degree of subjectivity and randomness, there are still certain rules, and these rules are related to the current state of the coal mine equipment. The operators need to understand the current state of the coal mine equipment before making any control operations, and the state of the coal mine equipment plays a decisive role in the control operations of the operators. Therefore, historical state data and historical operation data are also required to participate in the training process. In the actual use process, by inputting the predicted state data into the operation prediction model, the predicted results of the operations that the operators may perform in the future can be obtained.

[0042] After obtaining the predicted status data and predicted operation data at the same future time, they can be input into the diagnostic model. The diagnostic model is used to determine whether the coal mine equipment will fail after a period of time in the future, and if a failure occurs, what the specific type of the failure is. Since the data input into the diagnostic model includes the status data of the coal mine equipment and the operation data of the operators, the judgment results output by the diagnostic model may be failures caused by the poor status of the coal mine equipment itself, or failures caused by incorrect control operations made by the operators. When obtaining the detailed fault warning results, the controller on the coal mine equipment will display reminder information in the form of voice or text, so that the on-site operators can handle it in the first time.

[0043] Although the diagnostic model can warn of possible failures, for some complex or rare failures, simply relying on the diagnostic model cannot obtain accurate warning results. When the coal mine equipment fails but no reminder information is generated in advance, the controller will record the detailed data in this case. After the operator takes the controller to the ground, the local client will save the data in the controller, and the experienced operator will set labels for these data again, update the diagnostic model and write it into the controller again to improve the controller's warning ability for such failures.

[0044] The embodiment of the present application also provides a remote intelligent operation and maintenance method for coal mine equipment, as Figure 2 shown, the method includes the following steps:

[0045] S200, obtaining the fault sample data of the coal mine equipment, where the fault sample data includes the status sample data of the coal mine equipment and the operation sample data input by the operator;

[0046] S210, extracting the feature information of the fault sample data, and training the local model based on the feature information;

[0047] S220, aggregating the parameters of the local model to obtain the diagnostic model;

[0048] S230, downloading the diagnostic model to the controller of the coal mine equipment. During the operation of the coal mine equipment, the controller collects the real-time status data and real-time operation data, and inputs the real-time status data and real-time operation data into the status prediction model to obtain the predicted status data after a period of time in the future;

[0049] S240, the controller inputs the predicted status data into the operation prediction model to obtain the predicted operation data corresponding to the time of the predicted status data;

[0050] S250, the controller inputs the predicted status data and the predicted operation data into the diagnostic model to obtain the warning result of the coal mine equipment failure.

[0051] Although the preferred embodiments of the present application have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present application.

[0052] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. A remote intelligent operation and maintenance system for coal mine equipment, characterized in that, Including: A local client for obtaining fault sample data of coal mine equipment, where the fault sample data includes status sample data of the coal mine equipment and operation sample data input by an operator; Extracting feature information of the fault sample data and training a local model based on the feature information; A cloud server for receiving parameters of the local model uploaded by the local client, and the cloud server aggregating the parameters of the local model to obtain a diagnostic model; A controller is arranged in the coal mine equipment. During the operation of the coal mine equipment, the controller collects real-time status data and real-time operation data, inputs the real-time status data and the real-time operation data into a status prediction model to obtain predicted status data after a period of time in the future. The controller inputs the predicted status data into an operation prediction model to obtain predicted operation data corresponding to the time of the predicted status data, and inputs the predicted status data and the predicted operation data into the diagnostic model to obtain a warning result of the fault of the coal mine equipment.

2. The remote intelligent operation and maintenance system for a coal mine device according to claim 1, characterized in that, The local client trains the local model using a generative adversarial network.

3. The remote intelligent operation and maintenance system for a coal mine device according to claim 1, characterized in that, The local model is the result of aggregating multiple different types of computer models. After training each computer model using the feature information, the parameters of all the trained computer models are aggregated to establish the local model.

4. The remote intelligent operation and maintenance system for a coal mine device according to claim 3, characterized in that The computer models include a decision tree model, a neural network model, and a support vector machine model.

5. The remote intelligent operation and maintenance system for a coal mine device according to claim 1, characterized in that The status prediction model is obtained by analyzing the influence of historical operation data stored in the local client on historical status data.

6. The remote intelligent operation and maintenance system for a coal mine device according to claim 1, characterized in that, The operation prediction model is obtained by analyzing the influence of historical status data stored in the local client on historical operation data.

7. The remote intelligent operation and maintenance system for a coal mine device according to claim 1, characterized in that Both the local model and the diagnostic model adopt a Transformer model, and both the status prediction model and the operation prediction model adopt an LSTM model.

8. A method for a remote intelligent operation and maintenance system of coal mine equipment according to any one of claims 1-7, characterized in that, Including the following steps: Obtaining fault sample data of coal mine equipment, where the fault sample data includes status sample data of the coal mine equipment and operation sample data input by an operator; Extracting feature information of the fault sample data and training a local model based on the feature information; Aggregating the parameters of the local model to obtain a diagnostic model; Downloading the diagnostic model to the controller of the coal mine equipment. During the operation of the coal mine equipment, the controller collects real-time status data and real-time operation data, and inputs the real-time status data and the real-time operation data into a status prediction model to obtain predicted status data after a period of time in the future; The controller inputs the predicted status data into an operation prediction model to obtain predicted operation data corresponding to the time of the predicted status data; The controller inputs the predicted status data and the predicted operation data into the diagnostic model to obtain a warning result of the fault of the coal mine equipment.

Images