A hydraulic support prediction cooperative control method and device

Abstract

The application provides a hydraulic support prediction cooperative control method and device, and relates to the technical field of mining equipment control. The method comprises the following steps: processing time-space structure field time sequence characteristics of a current hydraulic support based on a pre-trained time-space structure field evolution model to obtain a time-space structure field prediction result of the hydraulic support; performing risk trend prediction and state identification on a fully mechanized working face according to the time-space structure field prediction result to obtain a processing result; and performing hydraulic support prediction cooperative control according to the processing result. The hydraulic support prediction cooperative control method and device provided in the application embodiment can change the control basis from single stress feedback to structure evolution prediction through the time-space structure field evolution model, and can improve the prediction cooperative control precision and timeliness of the hydraulic support through active adjustment of the stability of the fully mechanized working face.

Description

Technical Field

[0001] This invention relates to the field of mining equipment control technology, specifically to a predictive and coordinated control method and device for hydraulic supports. Background Technology

[0002] The fully mechanized mining face consists of a coal mining machine, a scraper conveyor, and hydraulic supports, forming a "three-machine coordinated" mining system. Its safe and efficient operation depends on the stable spatial geometric relationship between the three. The hydraulic supports not only provide roof support but also determine the straightness of the scraper conveyor, the operating posture of the coal mining machine, and the exposed shape of the coal face. They are key equipment for maintaining the stability of the working face's spatial structure.

[0003] Current hydraulic support control primarily relies on threshold or rule-based control based on sensors such as column pressure, stroke, and tilt angle. Its control objectives focus on the quality of support and the accuracy of the shifting distance, essentially representing a passive control method based on mine pressure and stroke response. Under this method, the control system adjusts according to pressure, stroke, and tilt angle values, without constraining or optimizing the spatial relationship between the coal mining machine, scraper conveyor, and support.

[0004] However, in actual mining operations, anomalies at the working face often stem from instability in spatial relationships. Examples include poor fit between the sidewalls and the coal face leading to spalling, insufficient roof support causing roof falls, and low straightness of the scraper conveyor resulting in inadequate straightness of the hydraulic supports at the working face. These phenomena typically evolve first in terms of geometric shape, making it difficult to identify and intervene in a timely manner relying solely on pressure and stroke signals. Summary of the Invention

[0005] To address the problems in the prior art, embodiments of the present invention provide a predictive collaborative control method and apparatus for hydraulic supports, which can at least partially solve the problems existing in the prior art.

[0006] On one hand, this invention proposes a predictive collaborative control method for hydraulic supports, comprising: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0007] The step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: If the predicted result of the distance field between the coal wall and the support is greater than the preset distance threshold, and the predicted result of the local spatial structural stability index of the support is less than the preset stability threshold, and the increment of the predicted coal wall profile curvature is greater than the preset increment threshold, then the processing result is determined to be a sidewall trend. The predicted coal wall profile curvature is determined based on the prediction results of the predicted features of the coal wall profile.

[0008] The step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: Based on the prediction results of the predicted characteristics of the attitude relationship between the support frame and the scraper conveyor, the rate of change of the predicted attitude deviation between adjacent frames is determined. If the rate of change of the predicted attitude deviation between adjacent frames is determined to be greater than the preset bite risk threshold, then the processing result is determined to be a bite trend.

[0009] The prediction results of the predicted characteristics of the attitude relationship between the support and the scraper conveyor include the prediction results of the vertical attitude relationship between the support push rod and the scraper conveyor; correspondingly, the process of predicting the risk trend and identifying the state of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtaining the processing results, includes: If the predicted result of the vertical attitude relationship between the support push rod and the scraper conveyor is less than the preset tilt angle abnormality threshold, then the processing result is determined to be an abnormal tilt angle trend between the support and the conveyor.

[0010] The step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: The straightness curve of the working face scraper is calculated based on the prediction results of the predicted characteristics of the attitude relationship between the support and the scraper conveyor. The relative horizontal offset of each scraper is determined based on the straightness curve of the working face scraper, and the trend of the straightness change of the working face is predicted based on the relative horizontal offset of each scraper, thus obtaining the processing result.

[0011] The hydraulic support predictive collaborative control method further includes: Based on the spatiotemporal structure field prediction results and the operational parameter prediction results of the hydraulic support-related equipment, risk trend prediction and state identification are performed to obtain the processing results, including: If the predicted result of the distance field between the coal wall and the support is greater than the preset distance threshold, and the predicted pressure of the support column is less than the preset pressure threshold, then the processing result is determined to be an empty roof / virtual contact trend.

[0012] The acquisition of the local spatiotemporal structure field time series data includes: Features are extracted from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support to obtain the time series data of the local spatiotemporal structural field; The spatiotemporal structural field includes the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support.

[0013] The operating parameters include column pressure; correspondingly, the extraction of features from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support includes: Calculate the average value of the distance field between the coal wall and the support to obtain the statistical characteristics of the distance field between the hydraulic support and the coal wall; The coal wall profile is fitted and calculated, and the mean curvature value obtained is used as the coal wall profile feature. The attitude relationship between the support and the scraper conveyor is taken as the attitude characteristics of the hydraulic support and the scraper conveyor. The column pressure data per unit time is used as the column pressure characteristic. The average value of the local spatial structural stability index of the support is calculated to obtain the spatiotemporal field stability characteristics.

[0014] Obtaining the coal wall profile includes: Spatiotemporal registration was performed on the temporal visual images of the longwall mining face, and image preprocessing was performed on the spatiotemporally registered images. The coal wall target region is segmented from the preprocessed image to obtain the coal wall target region; The coal wall target area is subjected to coal wall edge detection and contour fitting to obtain the coal wall contour.

[0015] The acquisition of the distance field between the coal wall and the support includes: Extract the edge contour point set of the side panel and map it to physical coordinates; Calculate the distance set of the edge contour points of the support plate in the direction perpendicular to the coal wall based on the physical coordinates of the edge contour point set, and then interpolate and complete the distance set to obtain the distance field between the coal wall and the support.

[0016] The process of obtaining the attitude relationship between the support and the scraper conveyor includes: The verticality index is calculated based on the angle between the physical spatial direction vector of the scraper conveyor chute centerline and the physical spatial direction vector of the push rod axis. The consistency index of the current movement is calculated based on the standard movement distance of the current mining section and the actual movement distance of the current support. The attitude relationship between the support and the scraper conveyor is obtained by fusing the verticality index and the push consistency index.

[0017] The process of obtaining the local spatial structural stability index of the scaffold includes: The distance distribution information between the coal wall and the support is calculated based on the distance field between the coal wall and the support. Based on the pre-established correspondence, determine the rationality index of the distance between the coal wall and the support corresponding to the distance distribution information between the coal wall and the support; The correspondence is the relationship between the preset coal wall and support distance distribution information and the preset coal wall and support distance rationality index. The attitude relationship between the bracket and the scraper conveyor is used as an index of the attitude matching degree between the bracket and the conveyor. The support quality index of the support column is calculated based on the column pressure and the preset pressure threshold. The local spatial structural stability index of the support is obtained by weighting the rationality index of the distance between the coal wall and the support, the attitude matching index between the support and the conveyor, and the support column support quality index.

[0018] On one hand, the present invention proposes a hydraulic support predictive collaborative control device comprising: The first processing unit is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support based on the pre-trained spatiotemporal structure field evolution model, and obtain the spatiotemporal structure field prediction result of the hydraulic support. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. The second processing unit is used to predict the risk trend and identify the status of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtain the processing results. The control unit is used to perform predictive and coordinated control of the hydraulic support based on the processing results.

[0019] In another aspect, embodiments of the present invention provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0020] This invention provides a computer-readable storage medium, comprising: The computer-readable storage medium stores a computer program that, when executed by a processor, implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0021] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0022] The hydraulic support predictive collaborative control method and apparatus provided in this invention process the temporal characteristics of the current spatiotemporal structural field of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support. The spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the local spatiotemporal structural field temporal data of the hydraulic support. The spatiotemporal structural field prediction result includes prediction results corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structural field prediction result, risk trend prediction and state identification are performed on the fully mechanized mining face to obtain the processing result. Based on the processing result, predictive collaborative control of the hydraulic support is performed. By using the spatiotemporal structural field evolution model, the control basis is transformed from single force feedback to structural evolution prediction. Through active adjustment of the stability of the fully mechanized mining face, the accuracy and timeliness of the hydraulic support predictive collaborative control can be improved. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating a predictive collaborative control method for hydraulic supports provided in an embodiment of the present invention.

[0024] Figure 2 This is a flowchart illustrating a predictive collaborative control method for hydraulic supports provided in another embodiment of the present invention.

[0025] Figure 3 This is a schematic diagram of the structure of a hydraulic support predictive collaborative control device provided in an embodiment of the present invention.

[0026] Figure 4 This is a schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.

[0028] Figure 1 This is a flowchart illustrating a predictive collaborative control method for hydraulic supports according to an embodiment of the present invention, as shown below. Figure 1 As shown, the hydraulic support predictive collaborative control method provided in this embodiment of the invention includes: Step S1: Process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support based on the pre-trained spatiotemporal structure field evolution model to obtain the spatiotemporal structure field prediction results of the hydraulic support; The spatiotemporal structure field evolution model is obtained by training a GRU neural network using a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of the prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively.

[0029] Step S2: Based on the spatiotemporal structure field prediction results, perform risk trend prediction and status identification on the fully mechanized mining face to obtain the processing results.

[0030] Step S3: Perform predictive collaborative control of the hydraulic support based on the processing results.

[0031] In step S1 above, the device processes the temporal characteristics of the current spatiotemporal structure field of the hydraulic support based on a pre-trained spatiotemporal structure field evolution model to obtain the spatiotemporal structure field prediction result of the hydraulic support. The spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the time-series data of the local spatiotemporal structural field of the hydraulic support. The spatiotemporal structural field prediction results include prediction results corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. The device can be a computer device that executes this method. The acquisition, storage, use, and processing of data in this application's technical solution all comply with relevant regulations.

[0032] In step S2 above, the device performs risk trend prediction and status identification on the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtains the processing results.

[0033] In step S3 above, the device performs predictive collaborative control of the hydraulic support based on the processing results.

[0034] like Figure 2 As shown, the method of the present invention can be implemented based on modularity, as detailed below: The system includes an image acquisition module, a working condition parameter acquisition module, a spatiotemporal structure field construction module, a state recognition and prediction module, a collaborative control decision-making module, and an electro-hydraulic actuation control module. The image acquisition module continuously acquires images of the working face through fixed-focus cameras installed on each hydraulic support. The images include at least the coal face outline, side guards, telescopic beams, top beams, scraper conveyor trough, and local areas of the coal mining machine, forming time-series visual data Ii(t).

[0035] The operating parameter acquisition module collects the operating parameters of the hydraulic support and related equipment, including column pressure Pi (t), pushing stroke Li (t), telescopic beam stroke Ei (t), and coal mining machine position S (t).

[0036] The spatiotemporal structure field construction module is the core modeling step in the transformation of hydraulic supports from passive control to active predictive and collaborative control. This module takes the time-series visual data Ii(t) from the image acquisition module and the operating parameters from the working condition parameter acquisition module as dual inputs. Through a four-stage algorithm process of visual spatial field transformation, core geometric feature extraction, spatial relationship quantification calculation, and multi-dimensional index fusion, it realizes a continuous and quantitative description of the spatial geometric relationship between the coal wall, the support, and the scraper conveyor, and finally constructs the local spatiotemporal structure field of the support. .

[0037] Where i is the hydraulic support number and t is the timestamp, the algorithm takes into account the characteristics of the industrial scene of strong dust and high vibration in the fully mechanized mining face, and meets the computational efficiency requirements of real-time control while ensuring the accuracy of geometric modeling. The specific process is as follows: Step 1: Visual spatial field transformation: The transformation of spatial physical coordinates based on visual images is fundamental to subsequent analyses of coal face contour changes, the distance between the coal face and the sidewalls, the contact state of the roof beam, and the relative positions of the supports and the scraper conveyor. The core process involves establishing a mapping relationship between image pixel coordinates and actual spatial physical coordinates through on-site calibration of feature points and the collected image data. This achieves the transformation from two-dimensional images to three-dimensional actual spatial coordinates, providing a spatial reference for subsequent quantitative analysis. The specific core steps are as follows: (1) Establishment of spatial coordinate system: Physical coordinate system ( ): Construct a spatial physical coordinate system with the fixed-focus camera on the hydraulic support as the reference, with the camera installation position as the origin O, the extension direction of the scraper conveyor as the X-axis (horizontal forward), the vertical upward as the Y-axis (perpendicular to the ground), and the direction perpendicular to the coal wall as the Z-axis (horizontal to the coal wall).

[0038] Camera coordinate system ( ): Taking the camera installation position as the origin O, the extension direction of the scraper conveyor is... The axis (horizontal forward) and vertical upward are... Axis (perpendicular to the ground), perpendicular to the coal face direction is Axis (horizontal direction towards the coal face).

[0039] Image coordinate system ( The origin is the top left corner of the image, the u-axis is the horizontal pixel direction, and the v-axis is the vertical pixel direction.

[0040] (2) Feature point data acquisition and preprocessing: Feature point extraction: Visual detection algorithms (such as corner detection and edge detection) are used to extract s (s≥3) non-collinear feature points of the hydraulic support top beam, and the image pixel coordinates of each feature point are recorded. , (i=1,2,...,s).

[0041] True length measurement: Based on the dimensions of the hydraulic support top beam, obtain the length constraints of the feature points in the image corresponding to the real physical space, and the true Euclidean distance between adjacent feature points. (e.g., the actual distance between feature points A and B on the top beam); the geometric constraints of feature points in the physical coordinate system (e.g., the top beam is parallel to the Z-axis, and all feature points have the same Y-coordinate).

[0042] (3) Image pixel coordinate system - camera coordinate system mapping: Establish the mapping relationship between image pixel coordinates and camera coordinate system:

[0043] Where K is the camera intrinsic parameter matrix, These are the pixel coordinates of the image; These are coordinates in the camera coordinate system. This represents the depth value.

[0044] Since a fixed-focus camera cannot directly acquire depth information, intrinsic parameter calibration and scale restoration are performed using feature points with known true lengths. Let the normalized coordinates of two feature points in the image be... , The physical coordinates of the two feature points in physical space with a depth of Z are: , Then the true length L between the two points satisfies:

[0045] With the objective of minimizing the sum of squared errors between the image mapping length and the physical true length among all feature point pairs, an expectation function is established, and the optimal solution of the intrinsic parameter matrix K is obtained using the least squares method. .

[0046] (4) Camera coordinate system to physical coordinate system mapping: Since both the camera coordinate system and the physical coordinate system are constructed with the camera as the origin, the rotation matrix R = I (identity matrix), the translation vector t = 0, and the transformation matrix... At this point, the camera coordinate system-physical coordinate system mapping can be directly simplified to:

[0047] Finally, based on the intrinsic parameters K and transformation matrix T obtained from the calibration, arbitrary pixel points are realized. To physical coordinates The mapping.

[0048] Step 2: Coal Wall Profile Extraction and quantification: Based on the spatiotemporally registered visual image, the physical spatial contour of the coal face is extracted through a process of image preprocessing, target region segmentation, edge detection, and contour fitting, achieving a quantitative and structured representation of the coal face morphology. Specific details are as follows: Image preprocessing: To address image noise caused by dust, uneven lighting, and equipment vibration at the fully mechanized mining face, grayscale conversion, Gaussian filtering, and contrast adaptive histogram equalization are performed sequentially to suppress noise while enhancing the edge feature differences between the coal face and background targets such as the sidewalls and scraper conveyors.

[0049] Coal wall target region segmentation: A region growing method based on grayscale features is adopted. Taking the grayscale difference between the coal wall and the support equipment as the feature, an adaptive grayscale threshold is set, and the growth starts from the center area of ​​the coal wall. Interference areas such as side plates, telescopic beams, and scraper conveyors are removed to obtain the region of interest (ROI) containing only the coal wall, which is the coal wall target region.

[0050] Coal wall edge detection and contour fitting: The Canny edge detection algorithm is used for the ROI region of the coal wall, and a dual threshold is set to filter continuous and effective coal wall edge point sets. Step 1 maps the coal face edge points to the physical coordinate system of the support. The coal wall profile curve in the image coordinate system is obtained by performing least-squares polynomial fitting on the edge point set. Its mathematical expression is a three-dimensional point set in the physical coordinate system, which accurately reflects the geometric shape of the coal wall in physical space at time t.

[0051] Step 3: Distance between coal face and support Construction: Extract the edge contour point set of the side panel using visual algorithms. Based on step 1, these pixels are mapped to physical coordinates. Calculate the set of distances from the edge contour points of the protective panel at time t in the direction perpendicular to the coal face (Z-axis direction). Interpolation completion was performed using linear interpolation to obtain a continuous distance field between the coal wall and the support. .

[0052] Step 4: Attitude relationship between the support frame and the scraper conveyor Quantization calculation: By integrating visual features and operating parameters, the spatial attitude matching relationship between the support and the scraper conveyor at time t is quantitatively described from two core dimensions: verticality and consistency of movement. The normalized index is between 0 and 1. The closer the value is to 1, the higher the pose matching degree between the two and the more stable the spatial geometric relationship.

[0053] Verticality Calculation: The centerline of the scraper conveyor chute and the axis of the push rod are obtained through feature extraction using visual algorithms. Step 1 then yields the physical spatial direction vector of the scraper conveyor chute centerline. The physical spatial direction vector of the push rod axis Calculate the spatial angle between two vectors. And normalize it to obtain the verticality index, the formula is:

[0054] in, The closer it is to 1, the closer the support is to the scraper conveyor.

[0055] Progression Consistency Calculation: Based on the position of the coal mining machine Based on the benchmark and combined with the mining process parameters of the working face, determine the standard traverse distance of the current mining section. ; Calculate the current actual displacement of the stent The relative deviation rate from the standard shift stroke, after normalization, yields the shift consistency index, calculated using the following formula:

[0056] in, The closer it is to 1, the more consistent the movement of the support.

[0057] Attitude Relationship Fusion: A weighted fusion method is used to integrate verticality and lateral consistency into a unified attitude relationship index for the support and scraper conveyor. Weighting coefficient , The formula is determined by the on-site process calibration at the working face: .

[0058] Step 5: Structural stability indicators Fusion computing: Based on the distance between the coal face and the support The relationship between the support frame and the scraper conveyor's attitude and column pressure Based on this, a multi-dimensional feature fusion structural stability evaluation model is constructed. Through index normalization, weight allocation, and comprehensive calculation, a quantitative structural stability index of 0 to 1 is obtained, realizing a comprehensive assessment and classification of the local spatial structural stability of the support.

[0059] Evaluation index system construction: Three core evaluation dimensions were selected, and all indicators were normalized to the [0,1] interval. The larger the value, the better the state of that dimension and the higher the stability.

[0060] Reasonableness of coal face-support distance Based on the distance field The mean and variance are calculated; the smaller the distance distribution, the better the side protection plate fits the coal face. The closer it is to 1.

[0061] Support frame - transport aircraft attitude matching : Directly use the attitude relationship index obtained in step 4 This reflects the spatial compatibility between the support frame and the scraper conveyor.

[0062] scaffold column support quality : Reflects the stress on the support Deviation from rated value The degree of (i.e., the preset pressure threshold):

[0063] Weighting coefficients were determined by using the analytic hierarchy process (AHP) combined with on-site measurement data from the working face and expert experience to determine the weighting coefficients for each evaluation dimension. , and ,satisfy .

[0064] Comprehensive stability calculation: The structural stability index is calculated using the linear weighted sum method. The formula is: .

[0065] Step 6: Integration and continuous updating of the local spatiotemporal structural field of the support: Using timestamp t as the time axis, the coal wall profile obtained in the above steps is... Distance between coal face and support field The relationship between the support frame and the scraper conveyor's attitude Structural stability index Spatiotemporal integration is performed to achieve continuous and dynamic updates of the spatiotemporal structure field, ultimately outputting the local spatiotemporal structure field of the support:

[0066] The structural field is a two-dimensional continuous function in time and space, which can reflect the dynamic evolution of the spatial geometric relationship between the coal face, support, and scraper conveyor in real time and accurately. It is used as the core output to the state identification and prediction module, providing a complete, quantitative, and structured modeling foundation for predicting the instability trend of the working face structure, identifying risks, and making decisions on the coordinated control of hydraulic supports.

[0067] The state recognition and prediction module is the core decision-making pre-process of the active predictive and collaborative control of hydraulic supports. This module uses the spatiotemporal structural field time series data of the local spatiotemporal structural field of the support output by the spatiotemporal structural field construction module. Using time-series features of the temporal structure field as input, a three-stage algorithm flow is employed: construction of a spatiotemporal structure field evolution model, and identification of risk trend states. This process enables the identification of core risks such as splintering, frame entanglement, and open / virtual contact, as well as the state identification and short-term trend prediction of the support-transporter tilt angle and the straightness of the working face. An evolution model of the spatiotemporal structure field from the current time t to future times is established. .

[0068] Step 1: Construction of temporal features of spatiotemporal structure field: To capture the temporal evolution of the working face state, key statistical features are first extracted from the spatiotemporal structure field (k=1,2,…,n) to construct a temporal feature sequence.

[0069] Statistical characteristics of the hydraulic support-coal wall distance field: Calculate the average distance between the hydraulic support and the coal wall. Characterizes the overall horizontal distance between the coal face and the support: .

[0070] Coal wall profile features: The mean curvature is calculated by fitting the contour points. It reflects the flatness of the coal face; the greater the curvature, the more irregular the coal face, and the higher the risk of coal face spalling.

[0071] Hydraulic support-scraper conveyor attitude characteristics: Directly adopt the attitude relationship between the support frame and the transport aircraft As a temporal feature, among For verticality, To ensure consistency over time.

[0072] Column pressure characteristics: Calculate pressure data per unit time As a key indicator of stress changes in the surrounding rock of the roof.

[0073] Characteristics of spatiotemporal field stability: Take the mean of the stability field Characterizes the overall safety level of the stent in a localized area: .

[0074] The above features are combined into a time-series feature vector: .

[0075] Forming a time sequence for model training and inference. .

[0076] Step 2: Construction of the spatiotemporal structure field evolution model: With preprocessed time series dataset Using the input as input, a spatiotemporal fusion evolution model based on gated cyclic units (GRU) is constructed to realize the spatiotemporal structure field from the current time t. To the future spatiotemporal structure field Mapping: .

[0077] This model combines GRU's strong fitting ability to time series data with the mechanistic constraints of the working surface spatial structure, balancing prediction accuracy and physical interpretability. The specific construction process is as follows: Model input and output design: The standardized spatiotemporal structure field feature sequence (T represents the historical time window) is used as the model input, and future moments are used as input. spatiotemporal structure field characteristics (Δt is the prediction time window) is used as the model output.

[0078] GRU Network Structure Construction: Construct a GRU neural network containing an input layer, hidden layers, and an output layer. The number of neurons in the input layer is equal to the number of feature dimensions, and the number of neurons in the output layer is the same as that in the input layer. The ReLU activation function is used to solve the gradient vanishing problem, and a Dropout layer is used to improve the model's generalization ability and avoid overfitting.

[0079] Model Training and Optimization: Using time-series data of the spatiotemporal structural field collected at the working face as the training set, mean squared error (MSE) was used as the loss function. The Adam optimizer was used to train the model until the loss function converged and tended to stabilize. The loss function is as follows: .

[0080] Where N is the number of training samples. Let i be the predicted feature of the i-th sample. These are the corresponding real features.

[0081] During the inference phase, the spatiotemporal structure field at the latest m moments is... Convert it into a feature sequence and input it into the model to obtain... Predicted features The predicted spatiotemporal structure field is reconstructed: .

[0082] Step 3: Risk Trend Prediction and Status Identification: Based on the prediction of spatiotemporal structure field Construct risk trend status identification rules: Film gang trend: When there exists a region in the predicted scraper-coal wall distance field that satisfies the prediction characteristics of the coal wall-support distance field, the prediction results are as follows. ( The prediction results of the predicted characteristics of the local spatial structural stability index of the support in the corresponding region (based on a preset distance threshold). ( If the predicted increase in the curvature of the coal face profile is greater than the preset stability threshold (i.e., significantly increased), it is considered a spalling trend. This rule integrates three key precursors to spalling: the distance between the scraper and the coal face, low stability of the spatial field, and irregular coal face morphology.

[0083] Biting trend: Calculate the predicted attitude deviation between adjacent frames The rate of change, combined with the travel distance of the adjacent frame on the working face, if the rate of change of deviation ( If a pre-set threshold for fighting risk is set, then it is determined to be a fighting trend risk, and the adjacent frame combination with fighting risk is output.

[0084] Top / Virtual Contact Trend: When the predicted scraper-coal wall distance is abnormal, ( (based on a preset distance threshold), and the predicted pressure of the support column. If this occurs, the hydraulic support is determined to have a tendency for empty top / virtual contact.

[0085] Trend of tilt angle variation of bracket-scraper conveyor: When the predicted vertical attitude relationship between the support push rod and the scraper conveyor is obtained... ( If the preset tilt angle abnormality threshold is used, it is determined to be an abnormal tilt angle trend of the support-transporter.

[0086] Trend of straightness variation of working face: Using the predicted attitude relationship of the support-scraper conveyor Calculate the straightness curve of the working face scraper, determine the relative horizontal offset of each scraper based on the straightness curve of the working face scraper, and judge the trend of the straightness change of the working face based on the relative horizontal offset of each scraper.

[0087] Step 4: Prediction Result Fusion and Output: The multi-dimensional state prediction and identification results are integrated and standardized prediction reports are generated in the format of risk type, prediction time, location of occurrence, risk level, and magnitude of change. Simultaneously, the predicted spatiotemporal structural field and state identification results are transmitted in real time to the collaborative control decision module, providing accurate prediction basis for subsequent support, relocation, sidewall attitude adjustment, and collaborative control with adjacent supports. Furthermore, the algorithm incorporates a self-updating mechanism to update the data collected on-site. Using spatiotemporal structural field data as new samples, the evolution model is continuously fine-tuned online to adapt it to the dynamic changes in geological conditions and mining technology at the working face, ensuring long-term prediction accuracy.

[0088] The collaborative control decision-making module outputs the future from the state recognition and prediction module. Local spatiotemporal structure field of the support at any moment Using the instability risk prediction results as input, and through the core process of prediction state matching—control strategy generation—command quantization encoding, the system achieves the mapping from structural evolution prediction to precise control commands for hydraulic supports, generating commands that meet the requirements of... The quantitative control commands enable proactive intervention in risks such as roof collapse and open roof, as well as coordinated regulation of the working face spatial structure, as follows: Precise matching of risk status: extraction The core characteristics of the working face, such as the risk of the working face being exposed, the tendency of the roof to bite, and the false contact of the roof beam, are reflected in the data. These characteristics are then accurately matched with the preset instability state to clarify the core control objects of the working face in the present and future.

[0089] Differentiated strategy generation: Based on the matched instability state, the preset targeted control strategy is invoked. For side slope risk, the support plate is extended in advance and the support pressure is increased. For open roof, the shifting rhythm is adjusted adaptively. For biting frame trend, the single frame shifting restriction and adjacent frame coordination are implemented. For top beam virtual contact, the top beam spatial attitude is corrected, taking into account both single frame control and the overall spatial structural stability of the working face.

[0090] Control command quantization encoding: converting qualitative control strategies into quantized control commands that the hydraulic support actuator can recognize. The instructions contain core information such as action type, execution parameters, and coordination logic, enabling... arrive Precise function mapping.

[0091] The electro-hydraulic actuator control module coordinates with the quantitative control commands output by the control decision module. The input stage is the execution phase where the control strategy is translated into the actual actions of the hydraulic support. The core process is instruction parsing—action conversion—execution feedback—coordinated control, ultimately achieving stable control of the overall spatial structure of the working face. Specifically: Control command parsing: for input Real-time analysis is performed to extract the action commands and quantitative parameters of each actuator, clarifying key information such as action type, adjustment amplitude, and execution sequence.

[0092] Electro-hydraulic signal conversion: The parsed digital control commands are converted into analog drive signals for the electro-hydraulic control system of the hydraulic support, which correspond to the action drive of actuators such as the column, the push cylinder, the side guard plate, and the telescopic beam.

[0093] Precise execution by sub-mechanisms: Based on the drive signals, each mechanism completes independent action control, namely adjusting the column pressure to regulate the support strength, controlling the pushing cylinder to achieve the moving / pushing action, and controlling the side plate / telescopic beam to complete the reinforcement support of the coal wall and roof.

[0094] Synchronous and coordinated control of adjacent supports: For commands involving prevention of support biting and maintenance of working face straightness that require linkage between adjacent supports, synchronous / differential action control of the actuators of multiple supports is realized to ensure the coordination of adjacent support actions.

[0095] Action status feedback: The actual action status of each actuator is collected in real time and fed back to the collaborative control decision module to form a control closed loop, ensuring the consistency between the support action and the control command, and ultimately achieving stable control of the overall spatial structure of the working face.

[0096] The support control principle of this invention, which takes the spatial relationship of the three machines as the control target, is as follows: Instead of using column pressure, pushing stroke, and support tilt angle as the main adjustment basis, the control target is to stabilize the spatial geometric relationship between the coal mining machine, scraper conveyor, and hydraulic support, thus transforming the hydraulic support from a force response device into a spatial structure adjustment actuator.

[0097] The method for modeling the local spatiotemporal structural field of the support in this invention is as follows: By fusing fixed-view sensor information from the support frame with working parameters, a four-dimensional spatiotemporal structural model is established to describe the relative positions of the coal face, roof, sidewalls, and scraper conveyor over time, thus enabling continuous representation of the working face structure.

[0098] The mapping mechanism from spatial structural state to risk in this invention is as follows: Risks such as roof collapse, roof fall, support biting, and attitude instability are identified by geometric quantities such as changes in coal wall distance, roof beam contact status, and relative attitude of supports and scraper conveyors, rather than relying on sudden pressure changes for judgment.

[0099] The predictive control mechanism based on structural evolution of this invention is as follows: By utilizing temporal variation patterns to predict structural instability trends, and generating support actions before pressure changes occur, proactive control with early intervention can be achieved.

[0100] The collaborative adjustment mechanism of this invention for the overall stability of the working face is as follows: The single-frame control is extended to a group coordinated control that is topologically associated along the working face, and the overall geometric continuity of the working face is maintained by adjusting the frame movement distance and attitude adjustment.

[0101] This invention does not use column pressure, pushing stroke, and support tilt angle as the main control basis. Instead, it takes the spatial constraint relationship between the three machines in the fully mechanized mining face—the coal mining machine, the scraper conveyor, and the hydraulic support—as the control target. By establishing a four-dimensional spatiotemporal model of the working face structure changing over time, the coal wall morphology, support posture, and scraper conveyor position are continuously modeled. This enables the hydraulic support to transform from a pressure, stroke, and tilt angle response actuator to a spatial structure adjustment actuator.

[0102] To achieve the above objectives, this invention installs a fixed-focus camera facing the coal face on each hydraulic support to acquire a continuous sequence of images including the coal face, side guards, telescopic beams, top beam, and scraper conveyor. These images are then fused with operating parameters such as column pressure, pushing stroke, telescopic beam stroke, support inclination angle, and coal mining machine position to construct a local spatiotemporal structural field for the support. By analyzing changes in the coal face contour, the distance between the coal face and side guards, the contact state of the top beam, the relative positions of the support and scraper conveyor, and their temporal evolution, a spatial stability assessment model and a risk prediction model are formed.

[0103] Based on this, a support control strategy is generated according to the prediction results. The support, movement, side plate posture and adjacent support coordination are actively adjusted to maintain a stable spatial geometric relationship of the working face, realize side plate prediction, support bite prevention and posture adaptive adjustment, thereby improving support safety and automation.

[0104] This invention couples visual perception with working condition parameters into a unified spatiotemporal structural model, transforming the control basis from single force feedback to structural evolution prediction, thereby achieving proactive adjustment of the working face stability.

[0105] The hydraulic support predictive collaborative control method provided in this invention processes the temporal characteristics of the current spatiotemporal structural field of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support. The spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the local spatiotemporal structural field temporal data of the hydraulic support. The spatiotemporal structural field prediction result includes prediction results corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structural field prediction result, risk trend prediction and state identification are performed on the fully mechanized mining face to obtain the processing result. Based on the processing result, predictive collaborative control of the hydraulic support is performed. By using the spatiotemporal structural field evolution model, the control basis is transformed from single force feedback to structural evolution prediction. Active adjustment of the stability of the fully mechanized mining face can improve the accuracy and timeliness of the hydraulic support predictive collaborative control.

[0106] In the above optional embodiments, the step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: If the predicted result of the distance field between the coal wall and the support is greater than a preset distance threshold, and the predicted result of the local spatial structural stability index of the support is less than a preset stability threshold, and the increment of the predicted coal wall profile curvature is greater than a preset increment threshold, then the processing result is determined to be a sidewall trend; the above embodiments can be referred to for explanation, and will not be repeated here.

[0107] The predicted curvature of the coal face profile is determined based on the prediction results of the predicted features of the coal face profile. This can be referred to the above embodiments for further explanation, and will not be repeated here.

[0108] In the above optional embodiments, the step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: Based on the prediction results of the predicted characteristics of the attitude relationship between the support frame and the scraper conveyor, the rate of change of the predicted attitude deviation between adjacent frames is determined; this can be referred to the above embodiment for explanation, and will not be repeated here.

[0109] If the rate of change of the predicted attitude deviation between adjacent frames is determined to be greater than the preset bite risk threshold, then the processing result is determined to be a bite trend. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0110] In the above optional embodiments, the prediction results of the predicted characteristics of the attitude relationship between the support and the scraper conveyor include the prediction results of the vertical attitude relationship between the support push rod and the scraper conveyor; correspondingly, the step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: If the predicted vertical attitude relationship between the support push rod and the scraper conveyor is less than a preset tilt angle anomaly threshold, then the processing result is determined to be an abnormal tilt angle trend between the support and the conveyor. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0111] In the above optional embodiments, the step of performing risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structure field prediction results to obtain the processing results includes: The straightness curve of the working face scraper is calculated based on the prediction results of the predicted characteristics of the attitude relationship between the support and the scraper conveyor. The relative horizontal offset of each scraper is determined based on the straightness curve, and the trend of straightness change of the working face is predicted based on the relative horizontal offset of each scraper, thus obtaining the processing result. This can be referred to the above embodiment for explanation, and will not be repeated here.

[0112] In the above optional embodiments, the hydraulic support predictive collaborative control method further includes: Based on the spatiotemporal structure field prediction results and the operational parameter prediction results of the hydraulic support-related equipment, risk trend prediction and state identification are performed to obtain the processing results, including: If the predicted characteristic of the distance field between the coal wall and the support is greater than a preset distance threshold, and the predicted pressure on the support column is less than a preset pressure threshold, then the processing result is determined to be a false roof / virtual contact trend. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0113] In the above optional embodiments, obtaining the local spatiotemporal structure field time series data includes: Features are extracted from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support to obtain the time series data of the local spatiotemporal structural field; the above embodiments can be referred to for explanation, and will not be repeated here.

[0114] The spatiotemporal structural field includes the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support. These can be referred to the above embodiments for further explanation and will not be repeated here.

[0115] In the above optional embodiments, the operating parameters include column pressure; correspondingly, the extraction of features from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support includes: The average value of the distance field between the coal wall and the support is calculated to obtain the statistical characteristics of the distance field between the hydraulic support and the coal wall; this can be referred to the above embodiment for explanation, and will not be repeated here.

[0116] The coal wall profile is fitted and calculated, and the mean curvature value obtained is used as the coal wall profile feature; the above embodiments can be referred to for explanation, and will not be repeated here.

[0117] The attitude relationship between the support and the scraper conveyor is taken as the attitude characteristics of the hydraulic support and the scraper conveyor; this can be described with reference to the above embodiments, and will not be repeated here.

[0118] The column pressure data per unit time is used as the column pressure characteristic; the above embodiments can be referred to for explanation, and will not be repeated here.

[0119] The average value of the local spatial structural stability index of the support is calculated to obtain the spatiotemporal field stability characteristics. This can be referred to the above embodiment for further explanation, and will not be repeated here.

[0120] In the above optional embodiments, obtaining the coal wall profile includes: Spatiotemporal registration is performed on the temporal visual images of the longwall mining face, and image preprocessing is performed on the spatiotemporally registered images; the above embodiments can be referred to for explanation, and will not be repeated here.

[0121] The coal wall target region is segmented from the preprocessed image to obtain the coal wall target region; the above embodiments can be referred to for explanation, and will not be repeated here.

[0122] The coal wall target area is subjected to coal wall edge detection and contour fitting to obtain the coal wall contour. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0123] In the above optional embodiments, obtaining the distance field between the coal wall and the support includes: Extract the edge contour point set of the side panel and map it to physical coordinates; this can be described with reference to the above embodiments and will not be repeated here.

[0124] The distance set between the edge contour points of the support plate and the coal wall in the direction perpendicular to the coal wall is calculated based on the physical coordinates of the edge contour point set, and the distance set is interpolated to obtain the distance field between the coal wall and the support. This can be referred to the above embodiment for explanation, and will not be repeated here.

[0125] In the above optional embodiments, obtaining the attitude relationship between the support and the scraper conveyor includes: The verticality index is calculated based on the angle between the physical spatial direction vector of the scraper conveyor chute centerline and the physical spatial direction vector of the push rod axis; the above embodiments can be referred to for explanation, and will not be repeated here.

[0126] The consistency index of the current movement is calculated based on the standard movement distance of the current mining section and the actual movement distance of the current support; the above embodiments can be referred to for explanation, and will not be repeated here.

[0127] The verticality index and the push consistency index are fused to obtain the attitude relationship between the support and the scraper conveyor. This can be referred to the above embodiments for explanation, and will not be repeated here.

[0128] In the above optional embodiments, obtaining the local spatial structural stability index of the stent includes: The distance distribution information between the coal wall and the support is calculated based on the distance field between the coal wall and the support; this can be referred to the above embodiment for explanation, and will not be repeated here.

[0129] The reasonableness index of the distance between the coal wall and the support, corresponding to the distance distribution information between the coal wall and the support, is determined based on the pre-established correspondence; the above embodiments can be referred to for explanation, and will not be repeated here.

[0130] The correspondence is the relationship between the preset coal wall and support distance distribution information and the preset coal wall and support distance rationality index; it can be referred to the above embodiments for explanation, and will not be repeated here.

[0131] The attitude relationship between the bracket and the scraper conveyor is used as an indicator of the attitude matching degree between the bracket and the conveyor; this can be referred to the above embodiments for explanation, and will not be repeated here.

[0132] The support quality index of the support column is calculated based on the column pressure and the preset pressure threshold; the above embodiments can be referred to for explanation, and will not be repeated here.

[0133] The local spatial structural stability index of the support is obtained by weighting the rationality index of the distance between the coal face and the support, the attitude matching index between the support and the conveyor, and the support column support quality index. This can be referred to the above embodiment for further explanation and will not be repeated here.

[0134] Figure 3 This is a schematic diagram of the structure of a hydraulic support predictive collaborative control device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the hydraulic support predictive collaborative control device provided in this embodiment of the invention includes a first processing unit 301, a second processing unit 302, and a control unit 303, wherein: The first processing unit 301 is used to process the temporal characteristics of the current spatiotemporal structural field of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support; wherein, the spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the local spatiotemporal structural field temporal data of the hydraulic support; the spatiotemporal structural field prediction result includes prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively; the second processing unit 302 is used to perform risk trend prediction and state identification of the fully mechanized mining face based on the spatiotemporal structural field prediction result to obtain the processing result; the control unit 303 is used to perform predictive collaborative control of the hydraulic support based on the processing result.

[0135] Specifically, the first processing unit 301 in the device is used to process the current spatiotemporal structural field time-series features of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support; wherein, the spatiotemporal structural field evolution model is obtained by training a GRU neural network with a training set constructed from the local spatiotemporal structural field time-series data of the hydraulic support; the spatiotemporal structural field prediction result includes the prediction results of the prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively; the second processing unit 302 is used to perform risk trend prediction and state identification of the fully mechanized mining face according to the spatiotemporal structural field prediction result to obtain the processing result; the control unit 303 is used to perform predictive collaborative control of the hydraulic support according to the processing result.

[0136] The hydraulic support predictive collaborative control device provided in this embodiment of the invention processes the temporal characteristics of the current spatiotemporal structural field of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support. The spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the local spatiotemporal structural field temporal data of the hydraulic support. The spatiotemporal structural field prediction result includes prediction results corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structural field prediction result, risk trend prediction and state identification are performed on the fully mechanized mining face to obtain the processing result. Based on the processing result, predictive collaborative control of the hydraulic support is performed. By using the spatiotemporal structural field evolution model, the control basis is transformed from single force feedback to structural evolution prediction. Through active adjustment of the stability of the fully mechanized mining face, the accuracy and timeliness of the hydraulic support predictive collaborative control can be improved.

[0137] The embodiments of the present invention provide a hydraulic support predictive collaborative control device that can be used to execute the processing flow of the above-described method embodiments. Its functions will not be repeated here, but can be referred to the detailed description of the above-described method embodiments.

[0138] Figure 4 This is a schematic diagram of the physical structure of a computer device provided in an embodiment of the present invention, such as... Figure 4 As shown, the computer device includes: a memory 401, a processor 402, and a computer program stored in the memory 401 and executable on the processor 402. When the processor 402 executes the computer program, it implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0139] This embodiment discloses a computer program product, which includes a computer program that, when executed by a processor, implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0140] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the following method: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

[0141] Compared with existing technologies, the hydraulic support predictive collaborative control method provided in this invention processes the temporal characteristics of the current spatiotemporal structural field of the hydraulic support based on a pre-trained spatiotemporal structural field evolution model to obtain the spatiotemporal structural field prediction result of the hydraulic support. The spatiotemporal structural field evolution model is obtained by training a GRU neural network using a training set constructed from the local spatiotemporal structural field temporal data of the hydraulic support. The spatiotemporal structural field prediction result includes prediction results corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support. Based on the spatiotemporal structural field prediction result, risk trend prediction and state identification are performed on the fully mechanized mining face to obtain the processing result. Based on the processing result, predictive collaborative control of the hydraulic support is performed. By using the spatiotemporal structural field evolution model, the control basis is transformed from single force feedback to structural evolution prediction. Active adjustment of the stability of the fully mechanized mining face can improve the accuracy and timeliness of the hydraulic support predictive collaborative control.

[0142] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0143] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0144] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0145] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0146] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0147] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A predictive collaborative control method for hydraulic supports, characterized in that, include: The spatiotemporal structure field evolution model is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support, and the spatiotemporal structure field prediction results of the hydraulic support are obtained. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. Based on the spatiotemporal structure field prediction results, risk trend prediction and status identification are performed on the fully mechanized mining face to obtain the processing results; Based on the processing results, predictive and coordinated control of hydraulic supports is performed.

2. The hydraulic support predictive collaborative control method according to claim 1, characterized in that, The process of predicting risk trends and identifying the state of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtaining the processing results, includes: If the predicted result of the distance field between the coal wall and the support is greater than the preset distance threshold, and the predicted result of the local spatial structural stability index of the support is less than the preset stability threshold, and the increment of the predicted coal wall profile curvature is greater than the preset increment threshold, then the processing result is determined to be a sidewall trend. The predicted coal wall profile curvature is determined based on the prediction results of the predicted features of the coal wall profile.

3. The hydraulic support predictive collaborative control method according to claim 1, characterized in that, The process of predicting risk trends and identifying the state of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtaining the processing results, includes: Based on the prediction results of the predicted characteristics of the attitude relationship between the support frame and the scraper conveyor, the rate of change of the predicted attitude deviation between adjacent frames is determined. If the rate of change of the predicted attitude deviation between adjacent frames is determined to be greater than the preset bite risk threshold, then the processing result is determined to be a bite trend.

4. The hydraulic support predictive collaborative control method according to claim 1, characterized in that, The prediction results of the attitude relationship between the support and the scraper conveyor include the prediction results of the vertical attitude relationship between the support push rod and the scraper conveyor; correspondingly, the process of predicting the risk trend and identifying the state of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtaining the processing results, includes: If the predicted result of the vertical attitude relationship between the support push rod and the scraper conveyor is less than the preset tilt angle abnormality threshold, then the processing result is determined to be an abnormal tilt angle trend between the support and the conveyor.

5. The hydraulic support predictive collaborative control method according to claim 1, characterized in that, The process of predicting risk trends and identifying the state of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtaining the processing results, includes: The straightness curve of the working face scraper is calculated based on the prediction results of the predicted characteristics of the attitude relationship between the support and the scraper conveyor. The relative horizontal offset of each scraper is determined based on the straightness curve of the working face scraper, and the trend of the straightness change of the working face is predicted based on the relative horizontal offset of each scraper, thus obtaining the processing result.

6. The hydraulic support predictive collaborative control method according to claim 1, characterized in that, The hydraulic support predictive collaborative control method further includes: Based on the spatiotemporal structure field prediction results and the operational parameter prediction results of the hydraulic support-related equipment, risk trend prediction and state identification are performed to obtain the processing results, including: If the predicted result of the distance field between the coal wall and the support is greater than the preset distance threshold, and the predicted pressure of the support column is less than the preset pressure threshold, then the processing result is determined to be an empty roof / virtual contact trend.

7. The hydraulic support predictive collaborative control method according to any one of claims 1 to 6, characterized in that, Acquiring the time series data of the local spatiotemporal structure field includes: Features are extracted from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support to obtain the time series data of the local spatiotemporal structural field; The spatiotemporal structural field includes the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support.

8. The hydraulic support predictive collaborative control method according to claim 7, characterized in that, The operating parameters include column pressure; correspondingly, the extraction of features from the operating parameters of the hydraulic support-related equipment and the pre-acquired spatiotemporal structural field of the hydraulic support includes: Calculate the average value of the distance field between the coal wall and the support to obtain the statistical characteristics of the distance field between the hydraulic support and the coal wall; The coal wall profile is fitted and calculated, and the mean curvature value obtained is used as the coal wall profile feature. The attitude relationship between the support and the scraper conveyor is taken as the attitude characteristics of the hydraulic support and the scraper conveyor. The column pressure data per unit time is used as the column pressure characteristic. The average value of the local spatial structural stability index of the support is calculated to obtain the spatiotemporal field stability characteristics.

9. The predictive and coordinated control method for hydraulic supports according to claim 7, characterized in that, Obtaining the coal face profile includes: Spatiotemporal registration was performed on the temporal visual images of the longwall mining face, and image preprocessing was performed on the spatiotemporally registered images. The coal wall target region is segmented from the preprocessed image to obtain the coal wall target region; The coal wall target area is subjected to coal wall edge detection and contour fitting to obtain the coal wall contour.

10. The hydraulic support predictive collaborative control method according to claim 7, characterized in that, Obtaining the distance field between the coal face and the support includes: Extract the edge contour point set of the side panel and map it to physical coordinates; Calculate the distance set of the edge contour points of the support plate in the direction perpendicular to the coal wall based on the physical coordinates of the edge contour point set, and then interpolate and complete the distance set to obtain the distance field between the coal wall and the support.

11. The hydraulic support predictive collaborative control method according to claim 7, characterized in that, Obtaining the attitude relationship between the support and the scraper conveyor includes: The verticality index is calculated based on the angle between the physical spatial direction vector of the scraper conveyor chute centerline and the physical spatial direction vector of the push rod axis. The consistency index of the current movement is calculated based on the standard movement distance of the current mining section and the actual movement distance of the current support. The attitude relationship between the support and the scraper conveyor is obtained by fusing the verticality index and the push consistency index.

12. The hydraulic support predictive collaborative control method according to claim 7, characterized in that, Obtaining the local spatial structural stability index of the scaffold includes: The distance distribution information between the coal wall and the support is calculated based on the distance field between the coal wall and the support. Based on the pre-established correspondence, determine the rationality index of the distance between the coal wall and the support corresponding to the distance distribution information between the coal wall and the support; The correspondence is the relationship between the preset coal wall and support distance distribution information and the preset coal wall and support distance rationality index. The attitude relationship between the bracket and the scraper conveyor is used as an index of the attitude matching degree between the bracket and the conveyor. The support quality index of the support column is calculated based on the column pressure and the preset pressure threshold. The local spatial structural stability index of the support is obtained by weighting the rationality index of the distance between the coal wall and the support, the attitude matching index between the support and the conveyor, and the support column support quality index.

13. A predictive and coordinated control device for hydraulic supports, characterized in that, include: The first processing unit is used to process the temporal characteristics of the current spatiotemporal structure field of the hydraulic support based on the pre-trained spatiotemporal structure field evolution model, and obtain the spatiotemporal structure field prediction result of the hydraulic support. The spatiotemporal structure field evolution model is obtained by training a GRU neural network with a training set constructed from the time series data of the local spatiotemporal structure field of the hydraulic support; the spatiotemporal structure field prediction results include prediction results of prediction features corresponding to the coal wall profile, the distance field between the coal wall and the support, the attitude relationship between the support and the scraper conveyor, and the local spatial structural stability index of the support, respectively. The second processing unit is used to predict the risk trend and identify the status of the fully mechanized mining face based on the spatiotemporal structure field prediction results, and obtain the processing results. The control unit is used to perform predictive and coordinated control of the hydraulic support based on the processing results.

14. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 12.

16. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 12.

Images