Coal-rock interface recognition model training method, shearer cutting control method and device

Abstract

The present disclosure provides a method for training a coal-rock interface recognition model, a method and a device for controlling the cutting of a coal shearer. The method for training the coal-rock interface recognition model includes: receiving sample coal-rock distributions and sample multimodal data sent by an edge processor; obtaining sample load state features according to sample cutting arm vibration data and sample cutting coal-rock noise data; obtaining sample cutting coal-rock interface features according to sample drum cutting video data; obtaining sample drum cutting load features according to sample cutting motor current data and sample lifting cylinder pressure data; calling the coal-rock interface recognition model to perform decision-level fusion to generate a sample predicted coal-rock distribution; and training and updating the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution. Thus, it is possible to combine real-time working conditions and, based on multimodal data, train the coal-rock interface recognition model in real time online, which can improve the training efficiency and accuracy of the coal-rock interface recognition model.

Description

Technical Field

[0001] The present disclosure relates to the technical field of intelligent control in coal mine production, and particularly to a method for training a coal-rock interface recognition model, a method and device for controlling the cutting of a coal shearer. Background Art

[0002] In the related art, an image of a coal mining face is acquired through an image acquisition sensor, and the acquired image is recognized by using a pre-trained image recognition model. According to the image recognition result, the height of the coal shearer is controlled to perform automatic cutting of the coal seam.

[0003] However, the recognition accuracy of a single image modality is not high, resulting in poor effect of automatic coal seam cutting. Summary of the Invention

[0004] The present disclosure aims to at least solve one of the technical problems in the related art to some extent.

[0005] Therefore, the present disclosure provides a method for training a coal-rock interface recognition model, a method and device for controlling the cutting of a coal shearer, which can combine real-time working conditions, and according to multi-modal data, train the coal-rock interface recognition model in real time online, and can improve the training efficiency and accuracy of the coal-rock interface recognition model.

[0006] In a first aspect, a method for training a coal-rock interface recognition model is provided. The method is executed by a cloud server, and the method includes: receiving sample coal-rock distribution and sample multi-modal data sent by an edge processor; wherein, the sample multi-modal data includes: sample cutting motor current data, sample lifting oil cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; obtaining sample load state features according to the sample cutting arm vibration data and the sample cutting coal-rock noise data; obtaining sample cutting coal-rock interface features according to the sample drum cutting video data; obtaining sample drum cutting load features according to the sample cutting motor current data and the sample lifting oil cylinder pressure data; calling the coal-rock interface recognition model, and performing decision-level fusion according to the sample load state features, the sample cutting coal-rock interface features, and the sample drum cutting load features to generate a sample predicted coal-rock distribution; and training and updating the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution to obtain a trained coal-rock interface recognition model.

[0007] In the second aspect of the present disclosure, a shearer cutting control method is proposed. The method is executed by an edge processor and includes: receiving a trained coal-rock interface recognition model sent by a cloud server; wherein the trained coal-rock interface recognition model is obtained by using the method described in the above embodiments; acquiring cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data; obtaining a load state feature according to the cutting arm vibration data and the cutting coal-rock noise data; obtaining a cutting coal-rock interface feature according to the drum cutting video data; obtaining a drum cutting load feature according to the cutting motor current data and the lifting cylinder pressure data; calling the trained coal-rock interface recognition model, and generating a predicted coal-rock distribution according to the load state feature, the cutting coal-rock interface feature, and the drum cutting load feature; determining a target drum height and a target traction speed according to the predicted coal-rock distribution; and sending the target drum height and the target traction speed to a shearer controller to control the shearer to cut the coal seam.

[0008] In the third aspect of the present disclosure, a device for training a coal-rock interface recognition model is proposed, including: a data receiving unit configured to receive a sample coal-rock distribution and sample multimodal data sent by an edge processor; wherein the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; a first processing unit configured to obtain a sample load state feature according to the sample cutting arm vibration data and the sample cutting coal-rock noise data; a second processing unit configured to obtain a sample cutting coal-rock interface feature according to the sample drum cutting video data; a third processing unit configured to obtain a sample drum cutting load feature according to the sample cutting motor current data and the sample lifting cylinder pressure data; a fourth processing unit configured to call a coal-rock interface recognition model and perform decision-level fusion according to the sample load state feature, the sample cutting coal-rock interface feature, and the sample drum cutting load feature to generate a sample predicted coal-rock distribution; and a training and updating unit configured to perform model training and updating on the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution to obtain a trained coal-rock interface recognition model.

[0009] In a fourth aspect of the present disclosure, a shearer cutting control device is proposed, including: a first model receiving unit configured to receive a trained coal-rock interface recognition model sent by a cloud server; wherein, the trained coal-rock interface recognition model is obtained by training using the methods described in some of the above embodiments; a data acquisition unit configured to acquire cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data; a first feature acquisition unit configured to acquire a load state feature according to the cutting arm vibration data and the cutting coal-rock noise data; a second feature acquisition unit configured to acquire a cutting coal-rock interface feature according to the drum cutting video data; a third feature acquisition unit configured to acquire a drum cutting load feature according to the cutting motor current data and the lifting cylinder pressure data; a prediction unit configured to call the trained coal-rock interface recognition model and generate a predicted coal-rock distribution according to the load state feature, the cutting coal-rock interface feature, and the drum cutting load feature; a data determination unit configured to determine a target drum height and a target traction speed according to the predicted coal-rock distribution; and a data sending unit configured to send the target drum height and the target traction speed to a shearer controller to control the shearer to cut the coal seam.

[0010] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects:

[0011] By implementing the embodiments of the present disclosure, receiving a sample coal-rock distribution and sample multimodal data sent by an edge processor; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; acquiring a sample load state feature according to the sample cutting arm vibration data and the sample cutting coal-rock noise data; acquiring a sample cutting coal-rock interface feature according to the sample drum cutting video data; acquiring a sample drum cutting load feature according to the sample cutting motor current data and the sample lifting cylinder pressure data; calling a coal-rock interface recognition model and performing decision-level fusion according to the sample load state feature, the sample cutting coal-rock interface feature, and the sample drum cutting load feature to generate a sample predicted coal-rock distribution; and training and updating the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution to obtain a trained coal-rock interface recognition model. Thus, it is possible to combine real-time working conditions, and according to multimodal data, train the coal-rock interface recognition model in real time online, which can improve the training efficiency and accuracy of the coal-rock interface recognition model. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background art, the following will describe the drawings required to be used in the embodiments of the present application or the background art.

[0013] Figure 1 Structural diagram of an intelligent control system for a shearer provided by an embodiment of the present disclosure;

[0014] Figure 2 Flowchart of a method for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0015] Figure 3 Schematic diagram of the distribution of coal and rock provided by an embodiment of the present disclosure;

[0016] Figure 4 Structural diagram of an intelligent control system for a shearer provided by an embodiment of the present disclosure;

[0017] Figure 5 Flowchart of S2 in the method for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0018] Figure 6 Flowchart of S3 in the method for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0019] Figure 7 Flowchart of S4 in the method for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0020] Figure 8 Flowchart of a method for controlling the cutting of a shearer provided by an embodiment of the present disclosure;

[0021] Figure 9 Flowchart of S20 in the method for controlling the cutting of a shearer provided by an embodiment of the present disclosure;

[0022] Figure 10 Flowchart of another method for controlling the cutting of a shearer provided by an embodiment of the present disclosure;

[0023] Figure 11 Structural diagram of a device for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0024] Figure 12 Structural diagram of the first processing unit in the device for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0025] Figure 13 Structural diagram of the second processing unit in the device for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0026] Figure 14 Structural diagram of the third processing unit in the device for training a coal-rock interface recognition model provided by an embodiment of the present disclosure;

[0027] Figure 15The structural diagram of a shearer cutting control device provided by an embodiment of the present disclosure;

[0028] Figure 16 The structural diagram of the data acquisition unit in the shearer cutting control device provided by an embodiment of the present disclosure;

[0029] Figure 17 The structural diagram of another shearer cutting control device provided by an embodiment of the present disclosure;

[0030] Figure 18 The structural diagram of the sample data acquisition unit in the shearer cutting control device provided by an embodiment of the present disclosure. Detailed implementation manners

[0031] The embodiments of the present disclosure will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present disclosure, but should not be construed as limiting the present disclosure.

[0032] The coal-rock interface recognition model training method, the shearer cutting control method and device of the embodiments of the present disclosure will be described below with reference to the accompanying drawings.

[0033] Before explaining the coal-rock interface recognition model training method and the shearer cutting control method provided by the embodiments of the present disclosure, the shearer intelligent control system applicable to the coal-rock interface recognition model training method and the shearer cutting control method provided by the embodiments of the present disclosure will be explained first.

[0034] Figure 1 The structural diagram of a shearer intelligent control system provided by an embodiment of the present disclosure.

[0035] As Figure 1 shown, the shearer intelligent control system provided by the embodiments of the present disclosure includes: a shearer 1, a shearer controller 2, an edge processor 3, a 5G network layer 4, a cloud server 5, a Hall high-current sensor 6, a pressure sensor 7, a vibration sensor 8, a sound sensor 9, a camera 10, an inclination sensor 11, and a position encoder 12.

[0036] Among them, the shearer controller 2, the camera 10, and the sound sensor 9 are installed in the middle of the shearer. The shearer controller 2 is responsible for controlling the traction motor and the height-adjusting hydraulic cylinder, and thus controlling the traction speed and the drum height during the cutting of the shearer. The camera 10 captures the coal-rock image during cutting, and the sound sensor 9 collects the sound when cutting the coal wall.

[0037] The Hall large current sensor 6 is installed on the cutting motor to collect the cutting current signal, the pressure sensor 7 is installed in the lifting cylinder of the cutting arm to collect the cylinder pressure signal, and the vibration sensor 8 is installed on the cutting arm to collect the vibration signal during cutting coal and rock. The inclination sensor 11 is installed inside the wiring cavity of the cutting motor to detect the angle of the cutting arm and thus obtain the drum height. The position encoder 12 is installed at the axial positions of the high and low speed shafts of the machine traction box and is used for measuring the position and traction speed of the shearer on the working face.

[0038] Among them, the edge processor 3 includes a server and a switch. The server and the switch are installed inside the shearer. The switch is connected to the server and the CPE device of the 5G network layer 4. The switch is also connected to the sensors (the aforementioned devices for measuring data, including: the camera 10, the sound sensor 9, the Hall large current sensor 6, the pressure sensor 7, the inclination sensor 11, the position encoder 12). The switch realizes the access, aggregation and transmission of sensor data and model parameters. The server realizes functions such as data preprocessing, status push, status identification, task distribution, and collaborative operation.

[0039] Among them, the 5G network layer 4 includes a switch, a base station controller, a 5G base station, and a CPE (Customer Premise Equipment) device.

[0040] In the embodiments of the present disclosure, the switch can be a 980C switch, and the 980C switch can be installed in the ground computer room. The base station controller can include a first base station controller BBU (Baseband Unit) and a second controller RHUB (remote radio unit hub). A mine explosion-proof and intrinsically safe base station controller (BBU) is installed at the head of the first underground section. A mine explosion-proof and intrinsically safe base station controller (RHUB) is installed on the electrical equipment cart beside the console. A 5G base station is installed on the emulsion pump return liquid filter cart; a 5G base station is installed on the eighth pipeline cart near the Horseshoe; a 5G base station is installed on the 18th frame of the working face; a 5G base station is installed on the 60th frame of the working face; a 5G base station is installed on the 103rd frame of the working face; a CPE device including a transmitting antenna is installed on the body of the shearer.

[0041] In the embodiments of the present disclosure, the 980C switch can realize the access, aggregation and transmission of a large-capacity 5G network.

[0042] The first base station controller BBU, also known as the baseband processing unit, centrally controls and manages the entire base station system. The first base station controller BBU is mainly responsible for baseband signal processing, including FFT / IFFT, modulation / demodulation, and channel encoding / decoding, etc. The first base station controller BBU supports a plug-in modular structure. Users can configure different numbers of baseband processing single boards according to different network capacity requirements and support baseband resource sharing.

[0043] The second base station controller RHUB is a radio frequency remote CPRI data aggregation unit, which realizes the access bridging between the first base station controller BBU and the 5G base station pRRU (pico Remote Radio Unit). It has the ability to cascade with the BBU at 25GE level and the ability to access 8 pRRUs.

[0044] In the embodiments of the present disclosure, the 5G base station can be the KT618(5G)-F mine explosion-proof and intrinsically safe base station pRRU, also known as the radio frequency remote processing unit, mainly including: a high-speed interface module, a signal processing unit, a power amplifier unit, a duplexer unit, an extended interface, and a power supply module. This device receives the downlink baseband data sent by the second base station controller RHUB and sends the uplink baseband data to the second base station controller RHUB to realize communication with the first base station controller BBU. In the process of sending signals: modulate the baseband signal to the transmission frequency band, filter and amplify it, and then transmit it through the antenna; in the process of receiving signals: receive the radio frequency signal from the antenna, down-convert the received signal to an intermediate frequency signal, perform amplification processing and analog-to-digital conversion (A / D conversion), and then send it to the first base station controller BBU for processing. It supports external antennas and flexible configuration of multiple frequencies and multiple modes. The CPE device is responsible for converting the high-speed 5G signal into a WiFi signal for communication with the 5G base station.

[0045] The switch is connected to the algorithm server of the base station controller and the central cloud node 5. The base station controller is connected to the 5G base station, and the 5G base station is connected to the CPE device. Utilizing the characteristics of low latency and large bandwidth of the 5G network and flexible slicing technology, real-time data interaction between the edge processor 3 and the cloud server 5 is realized, and the data upload of the edge processor 3 and the algorithm download of the cloud server 5 are completed.

[0046] The cloud server 5 includes a database, an algorithm server, and a client that are interconnected with each other. The database, the algorithm server, and the client are all installed on the ground. The database is responsible for functions such as data management and statistical analysis. The algorithm server is responsible for state recognition, model training, and task creation. The client realizes the function of state monitoring.

[0047] It should be noted that one or more of the devices listed in the above shearer intelligent control system can be set, and the setting position can be adjusted according to needs. The embodiments of the present disclosure do not make specific limitations on this.

[0048] It should also be noted that the devices listed in the above intelligent control system of the shearer are only for illustration and do not constitute specific limitations on the embodiments of the present disclosure. Any of the devices can be replaced with the update and upgrade of each device. Under the condition of the same function, they all fall within the protection scope of this solution.

[0049] Based on the above intelligent control system of the shearer, the embodiments of the present disclosure provide a method for training a coal-rock interface recognition model and a method for controlling the cutting of the shearer.

[0050] Figure 2 It is a flowchart of a method for training a coal-rock interface recognition model provided by the embodiments of the present disclosure.

[0051] As Figure 2 shown, the method for training a coal-rock interface recognition model provided by the embodiments of the present disclosure is executed by a cloud server and includes but is not limited to the following steps:

[0052] S1: Receive the sample coal-rock distribution and sample multi-modal data sent by the edge processor; wherein, the sample multi-modal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data.

[0053] In the embodiments of the present disclosure, multi-modal data and coal-rock distribution can be obtained in real time through measurement devices such as Hall large-current sensors, pressure sensors, vibration sensors, sound sensors, cameras, inclination sensors, and position encoders.

[0054] Among them, the cutting motor current data is obtained through a Hall large-current sensor, the lifting cylinder pressure data is obtained through a pressure sensor, the cutting arm vibration data is obtained through a vibration sensor, the cutting coal-rock noise data is obtained through a sound sensor, the drum cutting video data is obtained through a camera, the drum height is obtained through an inclination sensor, and the position and traction speed of the shearer on the working face are obtained through a position encoder.

[0055] In the embodiments of the present disclosure, the coal-rock distribution can be the drum height, the position and traction speed of the shearer on the working face, or the coal-rock distribution position and ratio determined according to the drum height, the position and traction speed of the shearer on the working face.

[0056] In an exemplary embodiment, as Figure 3 shown, when the shearer cuts the coal seam on the working face, the rock may be distributed in a blocky mixture in the coal seam (such as Figure 3 (a) in), in this case, the coal-rock distribution can be the position distribution and ratio of the coal seam and the rock; or the coal seam and the rock are distributed in layers without mixing, in this case, the coal-rock distribution can be all coal (such as Figure 3In (b), the whole rock (such as Figure 3 In (c), the regular roof strata (such as Figure 3 In (d), the irregular roof strata (such as Figure 3 In (e), the regular floor strata (such as Figure 3 In (f), the irregular floor strata (such as Figure 3 In (g), the upper parting (such as Figure 3 In (h), the lower parting (such as Figure 3 In (i), etc.

[0057] It should be noted that in the embodiments of the present disclosure, the edge processor needs to determine the sample multi-modal data and the sample coal-rock distribution according to the real-time acquired multi-modal data and the coal-rock distribution, so as to send them to the cloud server, and train the coal-rock interface recognition model according to the sample multi-modal data and the sample coal-rock distribution in the cloud server, so as to obtain a trained coal-rock interface recognition model.

[0058] Based on this, in the embodiments of the present disclosure, the edge processor acquires measurement devices such as Hall large current sensors, pressure sensors, vibration sensors, sound sensors, cameras, inclination sensors, and position encoders, and can preprocess the multi-modal data acquired in real time and the coal-rock distribution. For example: the video of the camera is enhanced and denoised by using the Retinex image enhancement algorithm for the images in the drum cutting video data.

[0059] In an exemplary embodiment, the memory cutting parameters and the memory cutting template are pre-stored in the edge processor. The memory cutting parameters are the drum height and the traction speed when the shearer cuts at different positions on the working face, and the memory cutting template is the curve of the roof and floor of the working face.

[0060] Among them, the shearer can independently cut the coal seam according to the memory cutting parameters and the memory cutting template set by the memory cutting function. At this time, measurement devices such as Hall large current sensors, pressure sensors, vibration sensors, sound sensors, cameras, inclination sensors, and position encoders can acquire the multi-modal data and the coal-rock distribution in real time.

[0061] Among them, in the process of the shearer independently cutting the coal seam according to the memory cutting parameters and the memory cutting template set by the memory cutting function, the drum height and the traction speed during the cutting of the shearer can be controlled based on manual intervention. For example: the electro-hydraulic control and the frequency converter are adjusted by manual remote control, and then the drum height and the traction speed are adjusted.

[0062] It is understandable that during the process of the shearer autonomously cutting the coal seam according to the memory cutting parameters and memory cutting templates set by the memory cutting function, if the coal seam is cut autonomously according to the memory cutting parameters and memory cutting templates, and the shearer traction speed is normal, at this time, the operator can judge that the working condition is normal and no manual intervention is required. When the shearer autonomously cuts the coal seam according to the memory cutting parameters and memory cutting templates, if it is found that the shearer cuts to the roof or floor, or the shearer traction speed is abnormal, manual intervention can be carried out at this time to adjust the electro-hydraulic control and frequency converter, and adjust the drum height and traction speed to enable the shearer to cut the coal seam normally.

[0063] Based on this, on the basis of the memory cutting parameters and memory cutting templates, combined with manual intervention, the real-time measured multi-modal data and coal-rock distribution are obtained, which can ensure that the obtained multi-modal data and coal-rock distribution are data under normal working conditions.

[0064] Furthermore, the edge processor determines the sample multi-modal data and sample coal-rock distribution according to the real-time obtained multi-modal data and coal-rock distribution. Among them, the real-time obtained multi-modal data and coal-rock distribution can be directly determined as the sample multi-modal data and sample coal-rock distribution, or the real-time obtained multi-modal data and coal-rock distribution can be preprocessed to obtain the sample multi-modal data and sample coal-rock distribution.

[0065] Thus, the edge processor sends the sample multi-modal data and sample coal-rock distribution to the cloud server, and trains the coal-rock interface recognition model according to the sample multi-modal data and sample coal-rock distribution in the cloud service. When the accuracy of the training sample data is relatively high, a coal-rock interface recognition model that better conforms to the working condition can be obtained, so that more accurate prediction results can be obtained when the coal-rock interface recognition model is used for prediction in the subsequent process.

[0066] It should be noted that in the embodiments of the present disclosure, the sample multi-modal data may also include other data other than the above examples. For example, it also includes: the relevant cutting state information and warning information of the shearer: the power of the scraper conveyor, the dip length of the caving face, the average initial coal seam thickness, the initial coal seam dip angle, the average initial coal seam gangue rate, the initial coal seam recoverable index, and the gas concentration, etc.

[0067] Among them, the method of digital twin can be used for three-dimensional simulation modeling to obtain the relevant cutting state information and warning information of the shearer.

[0068] In the embodiments of the present disclosure, the edge processor sends the sample coal-rock distribution and sample multi-modal data to the cloud server. Among them, the edge processor can send the sample coal-rock distribution and sample multi-modal data to the cloud server through the 5G network layer. The 5G network has the characteristics of low latency, large bandwidth, and flexible slicing. Sending the sample coal-rock distribution and sample multi-modal data to the cloud server through the 5G network layer can realize the data interaction between the edge processor and the cloud server, complete the data upload of the edge processor and the algorithm distribution of the cloud server, and the latency is relatively small.

[0069] Among them, for the setting of the 5G network layer, reference can be made to the relevant descriptions in the above examples, which will not be elaborated here.

[0070] It should be noted that in the embodiments of the present disclosure, the edge processor can obtain the sample cutting motor current data, sample lifting oil cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data in real time and send them to the cloud server in real time.

[0071] S2: Obtain the sample load state characteristics according to the sample cutting arm vibration data and the sample cutting coal-rock noise data.

[0072] In the embodiments of the present disclosure, after the cloud server receives the sample cutting arm vibration data and the sample cutting coal-rock noise data sent by the edge processor, it can obtain the sample load state characteristics according to the sample cutting arm vibration data and the sample cutting coal-rock noise data.

[0073] Among them, call the combined GSV model of the pre-trained GMM-UBM improved Gaussian mixture model and the SVM model to obtain the sound characteristics according to the sample cutting coal-rock noise data, call the pre-trained neural network model to obtain the vibration characteristics according to the sample cutting arm vibration data, and perform feature fusion on the sound characteristics and the vibration characteristics to generate the sample load state characteristics. Thus, the sample load state characteristics can be obtained according to the sample cutting arm vibration data and the sample cutting coal-rock noise data.

[0074] S3: Obtain the sample cutting coal-rock interface characteristics according to the sample drum cutting video data.

[0075] In the embodiments of the present disclosure, after the cloud server receives the sample drum cutting video data sent by the edge processor, it can obtain the sample cutting coal-rock interface characteristics according to the sample drum cutting video data.

[0076] Among them, call the pre-trained deep adversarial network (Adversarial Learnin) algorithm model to obtain the sample cutting coal-rock interface characteristics according to the sample drum cutting video data.

[0077] S4: Obtain the sample drum cutting load characteristics based on the sample cutting motor current data and the sample lifting cylinder pressure data.

[0078] In the embodiments of the present disclosure, after the cloud server receives the sample cutting motor current data and the sample lifting cylinder pressure data sent by the edge processor, it can obtain the sample drum cutting load characteristics based on the sample cutting motor current data and the sample lifting cylinder pressure data.

[0079] Among them, call the pre-trained load characteristic model based on the Bayesian network model, and judge the relationship between the current change and the cut coal and rock according to the sample cutting motor current data. Call the pre-trained cylinder pressure model based on the principle of equivalent average load, and judge the height position of the rock formation relative to the drum according to the sample lifting cylinder pressure data.

[0080] S5: Call the coal and rock interface recognition model, and perform decision-level fusion based on the sample load state characteristics, the sample cutting coal and rock interface characteristics, and the sample drum cutting load characteristics to generate the sample predicted coal and rock distribution.

[0081] In the embodiments of the present disclosure, a collaborative representation method is adopted. Based on the similarity model, each modality in the multi-modal is respectively mapped to its own representation space to obtain the sample load state characteristics, the sample cutting coal and rock interface characteristics, and the sample drum cutting load characteristics. Call the coal and rock interface recognition model based on the generative adversarial network model, combine the sample load state characteristics, the sample cutting coal and rock interface characteristics, and the sample drum cutting load characteristics, perform decision-level fusion, perform target prediction, and generate the sample predicted coal and rock distribution.

[0082] Among them, according to the cutting motor current load characteristic model, predict whether to cut the rock formation, predict the relative position of the rock formation through the pressure model of the two chambers of the drum lifting hydraulic cylinder, and monitor the differences in vibration, noise, and video under different coal and rock occurrence conditions to perform decision-level fusion and generate the sample predicted coal and rock distribution.

[0083] S6: Update the model training of the coal and rock interface recognition model according to the sample predicted coal and rock distribution and the sample coal and rock distribution to obtain the trained coal and rock interface recognition model.

[0084] In the embodiments of the present disclosure, the cloud server updates the model training of the coal and rock interface recognition model according to the sample predicted coal and rock distribution and the sample coal and rock distribution, where the sample predicted coal and rock distribution is predicted based on the sample multi-modal data.

[0085] Among them, according to the predicted coal and rock distribution and the sample coal and rock distribution, the coal and rock interface recognition model is trained and updated. The loss can be calculated based on the predicted coal and rock distribution and the sample coal and rock distribution, and then the model parameters of the coal and rock interface recognition model are updated according to the loss. When the calculated loss is less than a certain value, it indicates that the recognition accuracy of the coal and rock interface recognition model is high and stable at this time, so as to obtain a trained coal and rock interface recognition model.

[0086] For ease of understanding, an exemplary embodiment is provided in this disclosure.

[0087] As Figure 4 shown, the edge processor acquires real-time multimodal data sent by the multimodal sensor device. For example: the cutting motor current data is acquired through a Hall large-current sensor, the lifting cylinder pressure data is acquired through a pressure sensor, the cutting arm vibration data is acquired through a vibration sensor, the cutting coal and rock noise data is acquired through a sound sensor, the drum cutting video data is acquired through a camera, the drum height is acquired through an inclination sensor, and the position and traction speed of the shearer on the working face are acquired through a position encoder.

[0088] The edge processor can preprocess the real-time acquired multimodal data. Among them, the preprocessing can be to combine manual intervention with the memory cutting template and memory cutting parameters to obtain sample multimodal data and sample coal and rock distribution, and then upload the data and send it to the cloud server through the 5G network layer.

[0089] The cloud server receives the sample coal and rock distribution and sample multimodal data sent by the edge processor; among them, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal and rock noise data, and sample drum cutting video data.

[0090] After that, according to the sample cutting arm vibration data and the sample cutting coal and rock noise data, the sample load state characteristics are obtained; according to the sample drum cutting video data, the sample cutting coal and rock interface characteristics are obtained; according to the sample cutting motor current data and the sample lifting cylinder pressure data, the sample drum cutting load characteristics are obtained; the coal and rock interface recognition model is called, and decision-level fusion is performed according to the sample load state characteristics, the sample cutting coal and rock interface characteristics, and the sample drum cutting load characteristics to generate a sample predicted coal and rock distribution; according to the sample predicted coal and rock distribution and the sample coal and rock distribution, the coal and rock interface recognition model is trained and updated to obtain a trained coal and rock interface recognition model.

[0091] When the cloud server obtains the trained coal and rock interface recognition model, it can perform model deployment and send the trained coal and rock interface recognition model to the edge processor through the 5G network layer.

[0092] The edge processor can use the trained coal-rock interface recognition model to predict the coal-rock distribution based on the multi-modal data obtained in real time. Further, it calibrates the memory cutting template and memory cutting parameters to obtain the drum height and traction speed when the shearer cuts the coal seam. Further, the edge processor can send the drum height and traction speed to the shearer controller to control the shearer to cut the coal seam according to the drum height and traction speed. Thus, the application based on 5G industrial Internet and cloud-edge integration can greatly improve the training efficiency and accuracy of the coal-rock interface recognition model and enhance the stability of intelligent cutting control.

[0093] In the embodiments of the present disclosure, the training of the coal-rock interface recognition model is carried out in a closed-loop iterative and cyclic convergence manner. Combining manual intervention to control parameters and based on the cloud-edge collaboration technology, it realizes efficient model training, deployment and control, and adaptively extracts the optimal control parameters for the cutting state to correct the memory cutting parameters and memory cutting template.

[0094] By implementing the embodiments of the present disclosure, it receives the sample coal-rock distribution and sample multi-modal data sent by the edge processor; wherein, the sample multi-modal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; according to the sample cutting arm vibration data and sample cutting coal-rock noise data, it obtains the sample load state characteristics; according to the sample drum cutting video data, it obtains the sample cutting coal-rock interface characteristics; according to the sample cutting motor current data and sample lifting cylinder pressure data, it obtains the sample drum cutting load characteristics; it calls the coal-rock interface recognition model, and based on the sample load state characteristics, sample cutting coal-rock interface characteristics and sample drum cutting load characteristics, conducts decision-level fusion to generate the sample predicted coal-rock distribution; according to the sample predicted coal-rock distribution and sample coal-rock distribution, it conducts model training and updating on the coal-rock interface recognition model to obtain the trained coal-rock interface recognition model. Thus, it can combine the real-time working conditions and, according to the multi-modal data, train the coal-rock interface recognition model in real time online, and can improve the training efficiency and accuracy of the coal-rock interface recognition model.

[0095] As Figure 5 shown, in some embodiments, S2: obtaining the sample load state characteristics according to the sample cutting arm vibration data and sample cutting coal-rock noise data includes:

[0096] S21: obtaining the trained vibration frequency spectrum model and trained sound recognition model.

[0097] In the embodiments of the present disclosure, the trained vibration spectrum model may be a GSV model combined with a Gaussian mixture model improved by a trained GMM-UBM and an SVM model; the trained voice recognition model may be a trained neural network model.

[0098] Among them, to obtain the trained vibration spectrum model, vibration data and vibration features corresponding to the vibration data may be obtained in advance. The vibration data is input into the vibration spectrum model to obtain predicted vibration features. According to the vibration features and the predicted vibration features, the parameters of the vibration spectrum model are updated. Among them, the vibration loss between the vibration features and the predicted vibration features may be calculated, and the parameters of the vibration spectrum model are updated. When the vibration loss meets the vibration optimization condition, the trained vibration spectrum model is determined. Based on this, the trained vibration spectrum model is obtained.

[0099] In the embodiments of the present disclosure, for the method of obtaining the trained vibration spectrum model, reference may also be made to the methods in the related art, which are not limited to the methods provided in the embodiments of the present disclosure, and the embodiments of the present disclosure do not make specific limitations in this regard.

[0100] Among them, to obtain the trained voice recognition model, audio data and audio features corresponding to the audio data may be obtained in advance. The audio data is input into the voice recognition model to obtain predicted audio features. According to the audio features and the predicted audio features, the parameters of the voice recognition model are updated. Among them, the audio loss between the audio features and the predicted audio features may be calculated, and the parameters of the voice recognition model are updated. When the audio loss meets the audio optimization condition, the trained voice recognition model is determined. Based on this, the trained voice recognition model is obtained.

[0101] In the embodiments of the present disclosure, for the method of obtaining the trained voice recognition model, reference may also be made to the methods in the related art, which are not limited to the methods provided in the embodiments of the present disclosure, and the embodiments of the present disclosure do not make specific limitations in this regard.

[0102] S22: Call the trained vibration spectrum model, and obtain sample vibration features according to the sample cutting rocker arm vibration data.

[0103] S23: Call the trained voice recognition model, and obtain sample voice features according to the sample cutting coal and rock noise data.

[0104] In the embodiments of the present disclosure, in the case of obtaining the trained vibration spectrum model and the trained voice recognition model, further, call the trained vibration spectrum model, and obtain sample vibration features according to the sample cutting rocker arm vibration data. Call the trained voice recognition model, and obtain sample voice features according to the sample cutting coal and rock noise data.

[0105] S24: Feature - level fusion of the sample vibration characteristics and the sample sound characteristics is performed to obtain the sample load - state characteristics.

[0106] In the embodiments of the present disclosure, in the case of obtaining the sample vibration characteristics and the sample sound characteristics, feature - level fusion of the sample vibration characteristics and the sample sound characteristics is performed to obtain the sample load - state characteristics. Thus, the sample load - state characteristics can be obtained based on the sample cutting rocker - arm vibration data and the sample cutting coal - rock noise data.

[0107] As Figure 6 shown, in some embodiments, S3: Based on the sample drum - cutting video data, sample cutting coal - rock interface characteristics are obtained, including:

[0108] S31: Obtain a trained deep adversarial network model.

[0109] Among them, to obtain a trained deep adversarial network model, video data and the corresponding cutting coal - rock interface characteristics of the video data can be obtained in advance. The video data is input into the deep adversarial network model to obtain the predicted cutting coal - rock interface characteristics. Based on the cutting coal - rock interface characteristics and the predicted cutting coal - rock interface characteristics, the parameters of the deep adversarial network model are updated. Among them, the video loss between the cutting coal - rock interface characteristics and the predicted cutting coal - rock interface characteristics can be calculated to update the parameters of the deep adversarial network model. When the video loss meets the video optimization condition, a trained deep adversarial network model is determined. Based on this, a trained deep adversarial network model is obtained.

[0110] In the embodiments of the present disclosure, for the method of obtaining a trained deep adversarial network model, reference can also be made to the methods in the related art, which is not limited to the method provided in the embodiments of the present disclosure, and the present disclosure does not make specific limitations in this regard.

[0111] S32: Invoke the trained deep adversarial network model, and based on the sample drum - cutting video data, obtain the sample cutting coal - rock interface characteristics.

[0112] In the embodiments of the present disclosure, in the case of obtaining a trained deep adversarial network model, further, the trained deep adversarial network model is invoked, and based on the sample drum - cutting video data, the sample cutting coal - rock interface characteristics are obtained.

[0113] As Figure 7 shown, in some embodiments, S4: Based on the sample cutting - motor current data and the sample lifting - oil - cylinder pressure data, sample drum - cutting load characteristics are obtained, including:

[0114] S41: Obtain a trained load - characteristic model.

[0115] In the embodiments of the present disclosure, the trained load - characteristic model can be a trained Bayesian network model.

[0116] Among them, to obtain the trained load feature model, the cutting motor current data and the lifting cylinder pressure data can be obtained in advance, as well as the drum cutting load features corresponding to the cutting motor current data and the lifting cylinder pressure data. The cutting motor current data and the lifting cylinder pressure data are input into the load feature model to obtain the predicted drum cutting load features. According to the drum cutting load features and the predicted drum cutting load features, the parameters of the load feature model are updated. Among them, the load loss between the drum cutting load features and the predicted drum cutting load features can be calculated, and the parameters of the load feature model are updated. When the load loss meets the vibration optimization condition, the trained load feature model is determined. Based on this, the trained load feature model is obtained.

[0117] In the embodiments of the present disclosure, for the method of obtaining the trained load feature model, reference can also be made to the methods in the related art, which are not limited to the methods provided in the embodiments of the present disclosure. The embodiments of the present disclosure do not make specific limitations in this regard.

[0118] S42: Call the trained load feature model, and obtain the sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data.

[0119] In the embodiments of the present disclosure, in the case of obtaining the trained load feature model, further, call the trained load feature model, and obtain the sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data.

[0120] Figure 8 It is a flowchart of a shearer cutting control method provided by the embodiments of the present disclosure.

[0121] As Figure 8 shown, the shearer cutting control method provided by the embodiments of the present disclosure is executed by an edge processor, including but not limited to the following steps:

[0122] S10: Receive the trained coal-rock interface recognition model sent by the cloud server; among them, the trained coal-rock interface recognition model is trained by using the method in the above embodiments.

[0123] In the embodiments of the present disclosure, the edge processor receives the trained coal-rock interface recognition model sent by the cloud server. The edge processor can receive the trained coal-rock interface recognition model sent by the cloud server through the 5G network layer. The 5G network has the characteristics of low latency, large bandwidth, and flexible slicing. By receiving the trained coal-rock interface recognition model sent by the cloud server through the 5G network layer, the model deployment can be quickly completed for subsequent shearer cutting prediction and controlling the shearer to cut the coal seam.

[0124] Among them, the trained coal-rock interface recognition model is obtained by using the method in the above-mentioned embodiment. For the method in the above-mentioned embodiment, reference can be made to the relevant description of the above-mentioned embodiment, which will not be elaborated here.

[0125] S20: Obtain cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data.

[0126] In the embodiment of the present disclosure, the cutting motor current data, the lifting cylinder pressure data, the cutting arm vibration data, the cutting coal-rock noise data, and the drum cutting video data can be obtained in real time.

[0127] Such as Figure 9 shown, in some embodiments, S20: Obtain cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data, including:

[0128] S201: Obtain the cutting motor current data through the current sensor set on the cutting motor.

[0129] S202: Obtain the lifting cylinder pressure data through the pressure sensor set in the lifting cylinder of the cutting arm.

[0130] S203: Obtain the cutting arm vibration data through the vibration sensor set on the cutting arm.

[0131] S204: Collect the sound signal of the cutting coal wall through the sound sensor set at the bottom of the rocker arm, and obtain the cutting coal-rock noise data.

[0132] S205: Collect the coal-rock image of the cutting through the video acquisition device set at the bottom of the rocker arm, and obtain the drum cutting video data.

[0133] In the embodiment of the present disclosure, the cutting motor current data can be obtained through the current sensor set on the cutting motor. Among them, the current sensor can be a Hall large current sensor. The lifting cylinder pressure data can be obtained through the pressure sensor set in the lifting cylinder of the cutting arm. The cutting arm vibration data can be obtained through the vibration sensor set on the cutting arm. The cutting coal-rock noise data can be obtained by collecting the sound signal of the cutting coal wall through the sound sensor set at the bottom of the rocker arm. The drum cutting video data can be obtained by collecting the coal-rock image of the cutting through the video acquisition device set at the bottom of the rocker arm.

[0134] It can be understood that in the embodiment of the present disclosure, by obtaining multi-modal data through the above-mentioned multimedia sensor device, it can be obtained in real time. Thus, the real-time working conditions can be obtained, so that when making subsequent predictions, the obtained prediction results can better meet the requirements of the actual working conditions, and the accuracy of the prediction can be improved.

[0135] In some embodiments, the shearer cutting control method provided by the embodiments of the present disclosure further includes: receiving the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server.

[0136] In the embodiments of the present disclosure, the edge processor receives the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server. The edge processor can receive the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server through the 5G network layer. The 5G network has the characteristics of low latency, large bandwidth, and flexible slicing. Receiving the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server through the 5G network layer can quickly complete model deployment for subsequent shearer cutting prediction and control the shearer to cut the coal seam.

[0137] It can be understood that the cloud server sends the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model to the edge processor. The vibration spectrum model, the sound recognition model, the deep adversarial network model, and the load feature model can complete the training process in the cloud server.

[0138] S30: Obtain the load state characteristics according to the cutting arm vibration data and the cutting coal and rock noise data.

[0139] In the embodiments of the present disclosure, the trained vibration spectrum model is called to generate vibration characteristics according to the cutting arm vibration data, and the trained sound recognition model is called to generate audio characteristics according to the cutting coal and rock noise data. Then, the vibration characteristics and the sound characteristics are fused at the feature level to obtain the load state characteristics.

[0140] S40: Obtain the cutting coal and rock interface characteristics according to the drum cutting video data.

[0141] In the embodiments of the present disclosure, the trained deep adversarial network model is called to obtain the cutting coal and rock interface characteristics according to the drum cutting video data.

[0142] S50: Obtain the drum cutting load characteristics according to the cutting motor current data and the lifting cylinder pressure data.

[0143] In the embodiments of the present disclosure, the trained load feature model is called to obtain the drum cutting load characteristics according to the cutting motor current data and the lifting cylinder pressure data.

[0144] S60: Invoke the trained coal-rock interface recognition model, and generate a predicted coal-rock distribution based on the load state characteristics, the characteristics of the cutting coal-rock interface, and the characteristics of the drum cutting load.

[0145] In the embodiments of the present disclosure, in the case of obtaining the load state characteristics, the characteristics of the cutting coal-rock interface, and the characteristics of the drum cutting load, the trained coal-rock interface recognition model is invoked, and a predicted coal-rock distribution is generated based on the load state characteristics, the characteristics of the cutting coal-rock interface, and the characteristics of the drum cutting load.

[0146] S70: Determine the target drum height and the target traction speed according to the predicted coal-rock distribution.

[0147] In the embodiments of the present disclosure, the target drum height and the target traction speed are determined according to the predicted coal-rock distribution. It can be understood that in the case of different predicted coal-rock distributions, the corresponding drum heights and traction speeds are different. The target drum height and the target traction speed are determined according to the Deep Deterministic Policy Gradient algorithm DDPG and the adaptive cutting control strategy.

[0148] Among them, the Deep Deterministic Policy Gradient algorithm DDPG and the adaptive cutting control strategy can be preset and set according to needs, and can be pre-trained or formulated. The embodiments of the present disclosure do not make specific limitations on this.

[0149] In the embodiments of the present disclosure, the edge processor can also determine the target roof and floor cutting curves according to the predicted coal-rock distribution.

[0150] S80: Send the target drum height and the target traction speed to the shearer controller to control the shearer to cut the coal seam.

[0151] In the embodiments of the present disclosure, in the case of obtaining the target drum height and the target traction speed, the target drum height and the target traction speed are further sent to the shearer controller to control the shearer to cut the coal seam. Thus, it is possible to predict the coal-rock distribution of the shearer, and perform intelligent cutting control on the shearer according to the prediction result, which conforms to the real-time working conditions and has a high accuracy.

[0152] In some embodiments, the shearer cutting control method provided by the embodiments of the present disclosure further includes: obtaining a sample coal-rock distribution and sample multimodal data, and sending them to the cloud server to obtain the trained coal-rock interface recognition model; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting oil cylinder pressure data, sample cutting boom vibration data, sample cutting coal-rock noise data, and sample drum cutting video data.

[0153] In the embodiments of the present disclosure, the edge processor also obtains the sample coal-rock distribution and sample multi-modal data, and sends them to the cloud server for model training in the cloud server to obtain a trained coal-rock interface recognition model.

[0154] In the embodiments of the present disclosure, the coal-rock distribution may be the drum height, the position of the shearer on the working face, and the traction speed, or the coal-rock distribution position and ratio determined according to the drum height, the position of the shearer on the working face, and the traction speed.

[0155] Among them, to obtain the sample coal-rock distribution, the drum height can be obtained through an inclination sensor, and the position and traction speed of the shearer on the working face can be obtained through a position encoder, and then the sample coal-rock distribution can be further determined.

[0156] Among them, for the method of obtaining the sample multi-modal data, reference can be made to the relevant descriptions in the above embodiments, which will not be elaborated here.

[0157] As Figure 10 shown, in some embodiments, obtaining the sample coal-rock distribution and sample multi-modal data includes:

[0158] S100: Determine the memory top and bottom plate cutting curve, memory drum height, and memory traction speed.

[0159] In the embodiments of the present disclosure, the memory top and bottom plate cutting curve, memory drum height, and memory traction speed can be stored in the edge processor. Among them, the memory top and bottom plate cutting curve, memory drum height, and memory traction speed can be obtained by using the methods in related technologies and pre-stored in the edge processor.

[0160] S200: Obtain the real-time cutting motor current data, real-time lifting cylinder pressure data, real-time cutting arm vibration data, real-time cutting coal-rock noise data, real-time drum cutting video data, real-time top and bottom plate cutting curve, real-time drum height, and real-time traction speed when the shearer cuts the coal seam under preset conditions; where the preset conditions are the memory top and bottom plate cutting curve, memory drum height, and memory traction speed based on manual intervention.

[0161] In the embodiments of the present disclosure, the shearer can independently cut the coal seam according to the memory top and bottom plate cutting curve, memory drum height, and memory traction speed set by the memory cutting function. At this time, measuring devices such as Hall large current sensors, pressure sensors, vibration sensors, sound sensors, cameras, inclination sensors, and position encoders can obtain multi-modal data and coal-rock distribution in real time.

[0162] Among them, during the process of the shearer independently cutting the coal seam according to the memory cutting parameters and memory cutting templates set by the memory cutting function, based on manual intervention, the drum height and traction speed during the shearer cutting can be controlled. For example, by manually remote controlling the electro-hydraulic control and frequency converter, the drum height and traction speed can be adjusted accordingly.

[0163] It can be understood that during the process of the shearer independently cutting the coal seam according to the memory cutting parameters and memory cutting templates set by the memory cutting function, if the coal seam is cut according to the memory cutting parameters and memory cutting templates, and the shearer traction speed is normal, at this time, the operator can judge that the working condition is normal and no manual intervention is required. When the shearer independently cuts the coal seam according to the memory cutting parameters and memory cutting templates, if it is found that the shearer cuts to the roof or floor, or the shearer traction speed is abnormal, at this time, manual intervention can be carried out to adjust the electro-hydraulic control and frequency converter, and adjust the drum height and traction speed, so that the shearer can cut the coal seam normally.

[0164] Based on this, on the basis of the memory cutting parameters and memory cutting templates, combined with manual intervention, the multi-modal data and coal-rock distribution measured in real time are obtained, which can ensure that the obtained multi-modal data and coal-rock distribution are data under normal working conditions.

[0165] S300: Determine the sample coal-rock distribution according to the real-time roof and floor cutting curve, real-time drum height and real-time traction speed.

[0166] In the embodiments of the present disclosure, the sample coal-rock distribution is determined according to the real-time roof and floor cutting curve, real-time drum height and real-time traction speed.

[0167] It can be understood that in different cases of predicting the coal-rock distribution, the corresponding drum height and traction speed are different. According to the real-time roof and floor cutting curve, real-time drum height and real-time traction speed, the sample coal-rock distribution can be determined.

[0168] S400: Determine the real-time cutting motor current data as the sample cutting motor current data, the real-time lifting cylinder pressure data as the sample lifting cylinder pressure data, the real-time cutting arm vibration data as the sample cutting arm vibration data, the real-time cutting coal-rock noise data as the sample cutting coal-rock noise data, and the real-time drum cutting video data as the sample drum cutting video data.

[0169] In the embodiments of the present disclosure, the edge processor determines the sample multi-modal data and the sample coal-rock distribution according to the multi-modal data and the coal-rock distribution obtained in real time. Among them, the multi-modal data and the coal-rock distribution obtained in real time can be directly determined as the sample multi-modal data and the sample coal-rock distribution, or the multi-modal data and the coal-rock distribution obtained in real time can be preprocessed to obtain the sample multi-modal data and the sample coal-rock distribution.

[0170] The edge processor sends the sample multimodal data and the sample coal-rock distribution to the cloud server, where the coal-rock interface recognition model is trained based on the sample multimodal data and the sample coal-rock distribution. When the accuracy of the training sample data is relatively high, a coal-rock interface recognition model that better conforms to the working conditions can be obtained, so that more accurate prediction results can be obtained when the coal-rock interface recognition model is used for prediction subsequently.

[0171] Figure 11 It is a structural diagram of a device for training a coal-rock interface recognition model provided by an embodiment of the present disclosure.

[0172] As Figure 11 shown, the coal-rock interface recognition model training device 100 includes: a data receiving unit 101, a first processing unit 102, a second processing unit 103, a third processing unit 104, a fourth processing unit 105, and a training and updating unit 106.

[0173] The data receiving unit 101 is configured to receive the sample coal-rock distribution and the sample multimodal data sent by the edge processor; among them, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data.

[0174] The first processing unit 102 is configured to obtain sample load state features according to the sample cutting arm vibration data and the sample cutting coal-rock noise data.

[0175] The second processing unit 103 is configured to obtain sample cutting coal-rock interface features according to the sample drum cutting video data.

[0176] The third processing unit 104 is configured to obtain sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data.

[0177] The fourth processing unit 105 is configured to call the coal-rock interface recognition model, and perform decision-level fusion according to the sample load state features, the sample cutting coal-rock interface features, and the sample drum cutting load features to generate a sample predicted coal-rock distribution.

[0178] The training and updating unit 106 is configured to perform model training and updating on the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution, so as to obtain a trained coal-rock interface recognition model.

[0179] As Figure 12As shown, in some embodiments, the first processing unit 102 includes: a first model acquisition module 1021, a sample vibration feature acquisition module 1022, a sample sound feature acquisition module 1023, and a load feature fusion module 1024.

[0180] The first model acquisition module 1021 is configured to acquire a trained vibration spectrum model and a trained sound recognition model.

[0181] The sample vibration feature acquisition module 1022 is configured to call the trained vibration spectrum model and acquire sample vibration features according to the sample cutting rocker arm vibration data.

[0182] The sample sound feature acquisition module 1023 is configured to call the trained sound recognition model and acquire sample sound features according to the sample cutting coal and rock noise data.

[0183] The load feature fusion module 1024 is configured to perform feature-level fusion on the sample vibration features and the sample sound features to obtain sample load state features.

[0184] As Figure 13 shown, in some embodiments, the second processing unit 103 includes: a second model acquisition module 1031 and an interface feature fusion module 1032.

[0185] The second model acquisition module 1031 is configured to acquire a trained deep adversarial network model.

[0186] The interface feature fusion module 1032 is configured to call the trained deep adversarial network model and acquire sample cutting coal and rock interface features according to the sample drum cutting video data.

[0187] As Figure 14 shown, in some embodiments, the third processing unit 104 includes: a third model acquisition module 1041 and a cutting feature fusion module 1042.

[0188] The third model acquisition module 1041 is configured to acquire a trained load feature model.

[0189] The cutting feature fusion module 1042 is configured to call the trained load feature model and acquire sample drum cutting load features according to the sample cutting motor current data and the sample lifting oil cylinder pressure data.

[0190] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0191] The beneficial effects that can be achieved by the coal-rock interface recognition model training device in the embodiments of the present disclosure are the same as those that can be achieved by the above-described coal-rock interface recognition model training method, and will not be elaborated here.

[0192] Figure 15 It is a structural diagram of a shearer cutting control device provided by an embodiment of the present disclosure.

[0193] As Figure 15 shown, the shearer cutting control device 1000 includes: a first model receiving unit 1001, a data acquisition unit 1002, a first feature acquisition unit 1003, a second feature acquisition unit 1004, a third feature acquisition unit 1005, a prediction unit 1006, a data determination unit 1007, and a data sending unit 1008.

[0194] The first model receiving unit 1001 is configured to receive the trained coal-rock interface recognition model sent by the cloud server; wherein, the trained coal-rock interface recognition model is trained by using the method according to any one of claims 1 to 4.

[0195] The data acquisition unit 1002 is configured to acquire cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data.

[0196] The first feature acquisition unit 1003 is configured to acquire a load state feature according to the cutting arm vibration data and the cutting coal-rock noise data.

[0197] The second feature acquisition unit 1004 is configured to acquire a cutting coal-rock interface feature according to the drum cutting video data.

[0198] The third feature acquisition unit 1005 is configured to acquire a drum cutting load feature according to the cutting motor current data and the lifting cylinder pressure data.

[0199] The prediction unit 1006 is configured to call the trained coal-rock interface recognition model and generate a predicted coal-rock distribution according to the load state feature, the cutting coal-rock interface feature, and the drum cutting load feature.

[0200] The data determination unit 1007 is configured to determine a target drum height and a target traction speed according to the predicted coal-rock distribution.

[0201] The data sending unit 1008 is configured to send the target drum height and the target traction speed to the shearer controller to control the shearer to cut the coal seam.

[0202] As Figure 16As shown, in some embodiments, the data acquisition unit 1002 includes: a current acquisition module 10021, a pressure acquisition module 10022, a vibration acquisition module 10023, a noise acquisition module 10024, and an image acquisition module 10025.

[0203] The current acquisition module 10021 is configured to acquire the cutting motor current data through a current sensor disposed on the cutting motor.

[0204] The pressure acquisition module 10022 is configured to acquire the lifting cylinder pressure data through a pressure sensor disposed in the cutting arm lifting cylinder.

[0205] The vibration acquisition module 10023 is configured to acquire the cutting rocker arm vibration data through a vibration sensor disposed on the cutting arm.

[0206] The noise acquisition module 10024 is configured to collect the sound signal of the cutting coal wall through a sound sensor disposed at the bottom of the rocker arm to acquire the cutting coal and rock noise data.

[0207] The image acquisition module 10025 is configured to collect the cutting coal and rock images through a video acquisition device disposed at the bottom of the rocker arm to acquire the drum cutting video data.

[0208] As Figure 17 shown, in some embodiments, the shearer cutting control device 1000 further includes: a second model receiving unit 1009.

[0209] The second model receiving unit 1009 is configured to receive the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server.

[0210] Please continue to refer to Figure 17 , in some embodiments, the shearer cutting control device 1000 further includes: a sample data acquisition unit 1010.

[0211] The sample data acquisition unit 1010 is configured to acquire the sample coal and rock distribution and the sample multimodal data to send to the cloud server to obtain the trained coal and rock interface recognition model; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting rocker arm vibration data, sample cutting coal and rock noise data, and sample drum cutting video data.

[0212] As Figure 18 shown, in some embodiments, the sample data acquisition unit 1010 includes: a memory data determination module 10101, a real-time data acquisition module 10102, a coal and rock distribution determination module 10103, and a sample data determination module 10104.

[0213] The memory data determination module 10101 is configured to determine the memory cutting curves of the top and bottom coal seams, the memory drum height, and the memory traction speed.

[0214] The real-time data acquisition module 10102 is configured to acquire real-time cutting motor current data, real-time lifting cylinder pressure data, real-time cutting arm vibration data, real-time cutting coal and rock noise data, real-time drum cutting video data, real-time cutting curves of the top and bottom coal seams, real-time drum height, and real-time traction speed during the coal seam cutting of the shearer under preset conditions; wherein, the preset conditions are the memory cutting curves of the top and bottom coal seams, the memory drum height, and the memory traction speed based on manual intervention.

[0215] The coal and rock distribution determination module 10103 is configured to determine the sample coal and rock distribution according to the real-time cutting curves of the top and bottom coal seams, the real-time drum height, and the real-time traction speed.

[0216] The sample data determination module 10104 is configured to determine the real-time cutting motor current data as the sample cutting motor current data, the real-time lifting cylinder pressure data as the sample lifting cylinder pressure data, the real-time cutting arm vibration data as the sample cutting arm vibration data, the real-time cutting coal and rock noise data as the sample cutting coal and rock noise data, and the real-time drum cutting video data as the sample drum cutting video data.

[0217] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0218] The beneficial effects that can be achieved by the shearer cutting control device in the embodiments of the present disclosure are the same as those of the above-described shearer cutting control method, and will not be repeated here.

[0219] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present disclosure. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common general knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are merely exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0220] It should be understood that the various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and no limitations are imposed herein.

[0221] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. 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 principle of the present disclosure shall be included within the protection scope of the present disclosure.

[0222] In the description of this specification, the descriptions referring to the terms "one embodiment", "some embodiments", "example", "exemplary embodiment", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0223] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present disclosure, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0224] Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment or part of code including one or more executable instructions for implementing a customized logic function or process, and the scope of the preferred embodiments of the present disclosure includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art of the embodiments of the present disclosure.

[0225] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definable sequence list of executable instructions for implementing logical functions, which can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in conjunction with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then storing it in a computer memory.

[0226] It should be understood that various parts of the present disclosure can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0227] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by a program instructing relevant hardware, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0228] In addition, in various embodiments of the present disclosure, each functional unit can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above-mentioned integrated module can be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0229] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A method for training a coal-rock interface recognition model, characterized in that, The method is executed by a cloud server and includes: Receiving the sample coal-rock distribution and sample multi-modal data sent by an edge processor; wherein, the sample multi-modal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting arm vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; Obtaining sample load state features according to the sample cutting arm vibration data and the sample cutting coal-rock noise data; Obtaining sample cutting coal-rock interface features according to the sample drum cutting video data; Obtaining sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data; Invoking a coal-rock interface recognition model, and performing decision-level fusion according to the sample load state features, the sample cutting coal-rock interface features, and the sample drum cutting load features to generate a sample predicted coal-rock distribution; Performing model training and updating on the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution to obtain a trained coal-rock interface recognition model; The obtaining of the sample load state features according to the sample cutting arm vibration data and the sample cutting coal-rock noise data includes: Obtaining a trained vibration spectrum model and a trained sound recognition model; Invoking the trained vibration spectrum model and obtaining sample vibration features according to the sample cutting arm vibration data; Invoking the trained sound recognition model and obtaining sample sound features according to the sample cutting coal-rock noise data; Performing feature-level fusion on the sample vibration features and the sample sound features to obtain the sample load state features.

2. The method according to claim 1, wherein The obtaining of the sample cutting coal-rock interface features according to the sample drum cutting video data includes: Obtaining a trained deep adversarial network model; Invoking the trained deep adversarial network model and obtaining the sample cutting coal-rock interface features according to the sample drum cutting video data.

3. The method according to claim 1, wherein The obtaining of the sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data includes: Obtaining a trained load feature model; Invoking the trained load feature model and obtaining the sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data.

4. A cutting control method for a coal shearer, characterized in that, The method is executed by an edge processor and includes: Receiving the trained coal-rock interface recognition model sent by the cloud server; wherein, the trained coal-rock interface recognition model is trained by using the method described in any one of claims 1 to 3; Obtaining cutting motor current data, lifting cylinder pressure data, cutting arm vibration data, cutting coal-rock noise data, and drum cutting video data; Obtaining load state features according to the cutting arm vibration data and the cutting coal-rock noise data; Obtaining cutting coal-rock interface features according to the drum cutting video data; Obtaining drum cutting load features according to the cutting motor current data and the lifting cylinder pressure data; Call the trained coal-rock interface recognition model, and generate a predicted coal-rock distribution according to the load state characteristics, the cutting coal-rock interface characteristics, and the drum cutting load characteristics; Determine the target drum height and the target traction speed according to the predicted coal-rock distribution; Send the target drum height and the target traction speed to the shearer controller to control the shearer to cut the coal seam.

5. The method according to claim 4, wherein The obtaining the cutting motor current data, the lifting cylinder pressure data, the cutting boom vibration data, the cutting coal-rock noise data, and the drum cutting video data includes: Obtain the cutting motor current data through a current sensor arranged on the cutting motor; Obtain the lifting cylinder pressure data through a pressure sensor arranged in the cutting arm lifting cylinder; Obtain the cutting boom vibration data through a vibration sensor arranged on the cutting boom; Collect the sound signal of the cutting coal wall through a sound sensor arranged at the bottom of the boom, and obtain the cutting coal-rock noise data; Collect the coal-rock image of the cutting through a video acquisition device arranged at the bottom of the boom, and obtain the drum cutting video data.

6. The method according to claim 4, wherein The method further includes: Receive the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load characteristic model sent by the cloud server.

7. The method according to claim 4, characterized in that The method further includes: Obtain a sample coal-rock distribution and sample multimodal data to send to the cloud server to obtain the trained coal-rock interface recognition model; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting boom vibration data, sample cutting coal-rock noise data, and sample drum cutting video data.

8. The method according to claim 7, wherein The obtaining the sample coal-rock distribution and the sample multimodal data includes: Determine the memory top and bottom coal cutting curves, the memory drum height, and the memory traction speed; Obtain the real-time cutting motor current data, the real-time lifting cylinder pressure data, the real-time cutting boom vibration data, the real-time cutting coal-rock noise data, the real-time drum cutting video data, the real-time top and bottom coal cutting curves, the real-time drum height, and the real-time traction speed of the shearer cutting the coal seam under preset conditions; wherein, the preset conditions are the memory top and bottom coal cutting curves, the memory drum height, and the memory traction speed based on manual intervention; Determine the sample coal-rock distribution according to the real-time top and bottom coal cutting curves, the real-time drum height, and the real-time traction speed; Determine the real-time cutting motor current data as the sample cutting motor current data, the real-time lifting cylinder pressure data as the sample lifting cylinder pressure data, the real-time cutting boom vibration data as the sample cutting boom vibration data, the real-time cutting coal-rock noise data as the sample cutting coal-rock noise data, and the real-time drum cutting video data as the sample drum cutting video data.

9. A training device for a coal-rock interface recognition model, characterized in that, including: A data receiving unit, configured to receive the sample coal-rock distribution and sample multimodal data sent by an edge processor; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting boom vibration data, sample cutting coal-rock noise data, and sample drum cutting video data; A first processing unit, configured to obtain sample load state features according to the sample cutting boom vibration data and the sample cutting coal-rock noise data; A second processing unit, configured to obtain sample cutting coal-rock interface features according to the sample drum cutting video data; A third processing unit, configured to obtain sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data; A fourth processing unit, configured to call a coal-rock interface recognition model, and perform decision-level fusion according to the sample load state features, the sample cutting coal-rock interface features, and the sample drum cutting load features, to generate a sample predicted coal-rock distribution; A training and updating unit, configured to perform model training and updating on the coal-rock interface recognition model according to the sample predicted coal-rock distribution and the sample coal-rock distribution, so as to obtain a trained coal-rock interface recognition model; The first processing unit includes: A first model acquisition module, configured to acquire a trained vibration spectrum model and a trained voice recognition model; A sample vibration feature acquisition module, configured to call the trained vibration spectrum model, and obtain sample vibration features according to the sample cutting boom vibration data; A sample voice feature acquisition module, configured to call the trained voice recognition model, and obtain sample voice features according to the sample cutting coal-rock noise data; A load feature fusion module, configured to perform feature-level fusion on the sample vibration features and the sample voice features to obtain the sample load state features.

10. The device according to claim 9, characterized in that, The second processing unit includes: A second model acquisition module, configured to acquire a trained deep adversarial network model; An interface feature fusion module, configured to call the trained deep adversarial network model, and obtain the sample cutting coal-rock interface features according to the sample drum cutting video data.

11. The device according to claim 9, characterized in that, The third processing unit includes: A third model acquisition module, configured to acquire a trained load feature model; A cutting feature fusion module, configured to call the trained load feature model, and obtain the sample drum cutting load features according to the sample cutting motor current data and the sample lifting cylinder pressure data.

12. A shearer cutting control device, characterized in that It includes: A first model receiving unit, configured to receive a trained coal-rock interface recognition model sent by a cloud server; wherein, the trained coal-rock interface recognition model is trained by using the method described in any one of claims 1 to 3; A data acquisition unit, configured to acquire cutting motor current data, lifting cylinder pressure data, cutting boom vibration data, cutting coal-rock noise data, and drum cutting video data; The first feature acquisition unit is configured to acquire a load state feature according to the cutting boom vibration data and the cutting coal and rock noise data; The second feature acquisition unit is configured to acquire a cutting coal and rock interface feature according to the drum cutting video data; The third feature acquisition unit is configured to acquire a drum cutting load feature according to the cutting motor current data and the lifting cylinder pressure data; The prediction unit is configured to call the trained coal and rock interface recognition model and generate a predicted coal and rock distribution according to the load state feature, the cutting coal and rock interface feature, and the drum cutting load feature; The data determination unit is configured to determine a target drum height and a target traction speed according to the predicted coal and rock distribution; The data sending unit is configured to send the target drum height and the target traction speed to the shearer controller to control the shearer to cut the coal seam.

13. The device according to claim 12, wherein, The data acquisition unit includes: The current acquisition module is configured to acquire the cutting motor current data through a current sensor arranged on the cutting motor; The pressure acquisition module is configured to acquire the lifting cylinder pressure data through a pressure sensor arranged in the cutting arm lifting cylinder; The vibration acquisition module is configured to acquire the cutting boom vibration data through a vibration sensor arranged on the cutting boom; The noise acquisition module is configured to acquire the cutting coal and rock noise data by collecting a sound signal of the cutting coal wall through a sound sensor arranged at the bottom of the boom; The image acquisition module is configured to acquire the drum cutting video data by collecting a coal and rock image of the cutting through a video acquisition device arranged at the bottom of the boom.

14. The device according to claim 12, characterized in that The device further includes: The second model receiving unit is configured to receive the trained vibration spectrum model, the trained sound recognition model, the trained deep adversarial network model, and the trained load feature model sent by the cloud server.

15. The device according to claim 14, characterized in that The device further includes: The sample data acquisition unit is configured to acquire a sample coal and rock distribution and sample multimodal data to send to the cloud server to obtain the trained coal and rock interface recognition model; wherein, the sample multimodal data includes: sample cutting motor current data, sample lifting cylinder pressure data, sample cutting boom vibration data, sample cutting coal and rock noise data, and sample drum cutting video data.

16. The device according to claim 15, characterized in that, The sample data acquisition unit includes: The memory data determination module is configured to determine a memory top and bottom coal cutting curve, a memory drum height, and a memory traction speed; The real-time data acquisition module is configured to acquire real-time cutting motor current data, real-time lifting cylinder pressure data, real-time cutting boom vibration data, real-time cutting coal and rock noise data, real-time drum cutting video data, real-time top and bottom coal cutting curve, real-time drum height, and real-time traction speed of the shearer for coal seam cutting under preset conditions; wherein, the preset conditions are the memory top and bottom coal cutting curve, the memory drum height, and the memory traction speed based on manual intervention. A coal and rock distribution determination module, configured to determine the sample coal and rock distribution according to the real-time top and bottom floor cutting curves, the real-time drum height, and the real-time traction speed; A sample data determination module, configured to determine the real-time cutting motor current data as the sample cutting motor current data, the real-time lifting cylinder pressure data as the sample lifting cylinder pressure data, the real-time cutting arm vibration data as the sample cutting arm vibration data, the real-time cutting coal and rock noise data as the sample cutting coal and rock noise data, and the real-time drum cutting video data as the sample drum cutting video data.

Images