Autonomous planning mining control system based on big data

Abstract

The invention relates to the technical field of coal mine fully mechanized coal mining face automation control, in particular to an autonomous planning mining control system based on big data. Technical problems are that an automatic control system model of a fully mechanized coal mining face in the prior art is insufficient in precision, low in intelligent degree and poor in collaboration; according to the technical scheme, the autonomous planning mining control system based on big data comprises a high-precision dynamic geological modeling module, a big data intelligent decision module, an autonomous planning control module and an edge-cloud collaborative architecture module; aiming at the problems of insufficient model precision, low intelligent degree and poor collaboration of an automatic control system of a fully mechanized coal mining face of a coal mine in the prior art, the invention provides an autonomous planning mining control system based on big data, which combines high-precision geological modeling, an intelligent decision algorithm, real-time closed-loop control and edge-cloud collaboration technologies. And autonomous planning mining of fully mechanized coal mining face equipment under complex geological conditions is realized.

Description

Technical Field

[0001] This invention relates to the field of automated control technology for fully mechanized coal mining faces, and in particular to an autonomous planning and mining control system based on big data. Background Technology

[0002] The fully mechanized coal mining face automation control system is a highly integrated system designed to improve the safety and efficiency of coal mining. It integrates advanced technologies such as mechanics, electronics, computers, communications, software, system control, and networks, and realizes the automation and intelligence of the coal mining process through real-time monitoring and automated control of various equipment in the fully mechanized coal mining face. However, the geological modeling of the existing fully mechanized coal mining face automation control system relies on borehole exploration and static geological data, and uses Kriging interpolation to generate coal seam models. The model update cycle is long (usually on a monthly basis), and the modeling accuracy is generally above 30cm, which is difficult to meet the requirements of high-precision mining. At the same time, the existing coal mining machine control system mostly uses preset parameters (such as memory cutting) or manual remote control mode to operate, lacking real-time dynamic adjustment capabilities. Moreover, processes such as hydraulic support following the machine and scraper conveyor speed adjustment rely on fixed timing control, and the equipment coordination response delay is more than 2 seconds. Furthermore, it is not deeply coupled with the geological model, resulting in a low degree of matching between the cutting trajectory and geological conditions. Therefore, to address the issues of insufficient model accuracy, low intelligence, and poor coordination in existing fully mechanized mining face automation control systems, this paper proposes an autonomous planning and mining control system based on big data. This system combines high-precision geological modeling, intelligent decision-making algorithms, real-time closed-loop control, and edge-cloud collaborative technology to enable fully mechanized mining face equipment to autonomously plan and mine under complex geological conditions. It is applicable to the intelligent collaborative control of coal mining machines, hydraulic supports, scraper conveyors, and other equipment in underground coal mines. Summary of the Invention

[0003] In order to overcome the problems of insufficient accuracy, low level of intelligence and poor coordination in the existing automated control system model for fully mechanized coal mining faces.

[0004] The technical solution of this invention is: an autonomous planning and mining control system based on big data, comprising: The high-precision dynamic geological modeling module is used to connect with the underground coal mining machine, inertial navigation system, and cutting motor through the working face ring network to collect geological exploration data, equipment attitude data, and working condition data in real time. The big data intelligent decision-making module is deployed at the ground dispatch center to integrate with the IMS-P big data platform, receive geological model data output by the high-precision dynamic geological modeling module, and generate mining process optimization strategies through the HIMS-CDM algorithm. The autonomous planning and control module, which includes a black box control system and a closed-loop feedback unit, communicates with the coal mining machine's electrical control system and is used to dynamically adjust the cutting trajectory. The edge-cloud collaborative architecture module consists of an underground explosion-proof host and a ground workstation. It synchronizes data through an industrial ring network and a 5G redundant channel, and supports the autonomous operation of the local cache model in a network outage environment.

[0005] Preferably, the high-precision dynamic geological modeling module includes a multi-source data fusion unit, a dynamic correction unit, and a three-dimensional coordinate transformation unit. The multi-source data fusion unit is used to fuse geological realistic data, coal mining machine inertial navigation attitude data, and cutting motor current fluctuation characteristics, wherein the pitch angle accuracy of the coal mining machine inertial navigation attitude data is ±0.1°, and the sampling rate of the cutting motor current fluctuation characteristics is 1kHz. The dynamic correction unit is used to generate a geological slice model with an interval of 0.2m using the CT scanning principle, triggering a model update every 10m, and integrating an LSTM neural network to perform reverse error correction on historical cutting data, with a coal seam thickness prediction error ≤5cm. The three-dimensional coordinate transformation unit is used to establish a coal mining machine kinematic model containing 21 degrees of freedom, and to perform real-time mapping between the geological grid point coordinates and the absolute coordinates of the equipment, with a transformation error <3cm.

[0006] Preferably, the dynamic correction unit operates by including the following steps: S101: Receives real-time cutting data from the coal mining machine and extracts the cutter number, cutting height, and turning point features; S102: Compare the current slice data with the prediction model and calculate the height deviation Δh; S103: When Δh > 8cm, activate the LSTM network to adjust the weights and output the correction coefficient α∈[0.8,1.2]; S104: Update the predicted coal seam thickness in the unmined area according to the formula H_new=H_old×α; S105: Generate an updated geological model and synchronize it to the big data decision-making module.

[0007] Preferably, a coal seam thickness prediction model is constructed by integrating geological exploration data, coal mining machine inertial navigation attitude data, and cutting motor current characteristics through a multi-source data fusion unit; a dynamic correction unit is used to update the model by CT scanning-style slices, updating it every 10m, and combining it with an LSTM neural network to perform reverse correction on historical cutting data, thereby controlling the coal seam thickness prediction error within ±5cm; a 21-DOF coal mining machine kinematic model is established through a three-dimensional coordinate transformation unit to achieve real-time mapping between geological grid points and the absolute coordinates of the equipment.

[0008] Preferably, the big data intelligent decision-making module includes a multi-objective optimization engine, a process knowledge graph library, and a real-time collaborative algorithm unit. The multi-objective optimization engine is used to establish the objective function min F(x) = 0.4T + 0.3E + 0.3W, where T is the mining time cost, E is the energy consumption coefficient, and W is the equipment wear index. The process knowledge graph library is used to store mining case data and support process recommendations based on semantic matching. The real-time collaborative algorithm unit is used to optimize the three-machine collaborative parameters using the DDPG reinforcement learning model. The big data intelligent decision-making module generates control parameters for the triangular coal collaborative process section through the HIMS-APD system and outputs a mining strategy instruction set including velocity gradient and turning point offset.

[0009] Preferably, the real-time collaborative algorithm unit performs the following steps during operation: S201: Acquire hydraulic support pressure sensor data, scraper conveyor load data, and coal mining machine positioning information; S202: Construct the state space S={s1,s2,s3}, where s1 is the relative position matrix of the equipment, s2 is the environmental parameter vector, and s3 is the geological model features; S203: Calculate the optimal strategy for motion space A={a1,a2,a3} using the Q-Learning algorithm, where a1 is the traction speed adjustment, a2 is the roller height compensation value, and a3 is the support pushing sequence. S204: Output control commands to the autonomous planning control module and record the execution results to the sample library.

[0010] Preferably, the big data intelligent decision-making module establishes an objective function min F(x) = 0.4T + 0.3E + 0.3W (T is time cost, E is energy consumption, and W is equipment wear) through a multi-objective optimization engine, generates the Pareto optimal solution set through an improved NSGA-II algorithm, integrates mining cases through a process knowledge graph, constructs a semantic matching rule base, and supports automatic recommendation of collaborative process parameters for triangular coal mining, and optimizes the timing of the three machines' actions based on DDPG reinforcement learning through a real-time collaborative algorithm, with a response latency of <500ms.

[0011] Preferably, the autonomous planning and control module includes a dynamic trajectory planning unit, a prediction and compensation unit, and a closed-loop verification unit. The dynamic trajectory planning unit is used to generate cutting trajectories with 20cm intervals using an improved NURBS curve algorithm, and supports real-time insertion of correction points. The prediction and compensation unit predicts the cutting resistance 3 seconds in advance based on the ARIMA model and adjusts the traction speed according to the formula v_new=v_base×(1-0.05ΔP), where ΔP is the pressure change rate. The closed-loop verification unit integrates a laser scanner to measure the cutting contour in real time and generates a PID compensation signal to be fed back to the electro-hydraulic servo system.

[0012] Preferably, the dynamic trajectory planning unit includes the following steps during operation: S301: Extract the coordinate sequence of the top and bottom plates of the current slice {P1,P2,…,Pn} from the geological model; S302: The initial trajectory is fitted using a cubic B-spline curve, and the node vectors are set to uniform parameterization; S303: When inserting a manual intervention point Q, perform weighted smoothing according to the formula Q_new=0.7Q_manual+0.3Q_auto; S304: Outputs a trajectory instruction set containing position-velocity-time parameters to the coal mining machine controller.

[0013] Preferably, the autonomous planning control module generates cutting trajectories with 20cm intervals by using an improved NURBS curve algorithm, supports weighted smoothing of manual intervention points (weight coefficient 0.7:0.3); predicts cutting resistance 3 seconds in advance based on the ARIMA model and dynamically adjusts the traction speed (formula: v_new=v_base×(1-0.05ΔP)); monitors the cutting contour in real time through a laser scanner, generates PID compensation signals to adjust the hydraulic servo system, and performs closed-loop feedback verification.

[0014] Preferably, the implementation steps of the edge-cloud collaborative architecture module specifically include: S401: Deploy edge computing nodes on the explosion-proof host to execute 50ms-level real-time control algorithms and cache the geological model and control parameters of the most recent 10m. S402: The ground workstation communicates with the downhole equipment via the OPC UA protocol, employs a dual-queue mechanism to ensure command transmission, and automatically switches to local control mode when the network delay is greater than 1 second. S403: In the offline state, the explosion-proof host will maintain autonomous operation for 30 minutes based on cached data, and attempt to reconnect once every 5 minutes.

[0015] Preferably, a safety control subsystem is also included, comprising a five-level safety response mechanism, a UWB personnel positioning linkage unit, and an emergency self-locking device. The five-level safety response mechanism specifically includes setting the cutting motor temperature thresholds to T1=85℃, T2=95℃, and vibration amplitude thresholds to A1=0.5g, A2=1.2g, triggering speed reduction and emergency stop. The UWB personnel positioning linkage unit automatically locks the traction system when personnel are detected entering an area 5m in front of the coal mining machine. The emergency self-locking device uses dual redundant solenoid valves for control, cutting off hydraulic power within 0.5s under abnormal operating conditions.

[0016] As a preferred option, the system operation steps specifically include: S1: Initialization phase: Load geological real data, construct the initial geological model and verify the accuracy ≥15cm; S2: Decision-making stage: Calculate the Pareto optimal solution set using the NSGA-II algorithm and select the mining strategy with the highest comprehensive score; S3: Control phase: The cutting trajectory is sent to the coal mining machine, and the laser scanner is started simultaneously for contour monitoring; S4: Feedback phase: Collect actual cutting data and equipment operating conditions, and reverse-correct the geological model and control parameters; S5: Iteration phase: After every 10m of progress, a new round of model optimization and strategy generation is automatically started.

[0017] The beneficial effects of this invention are: 1. This invention employs an LSTM neural network to iteratively correct historical cutting data from coal mining machines, combined with a CT scan-style slicing update mechanism, improving the accuracy of the geological model from 30cm in traditional static models to the 10cm level, with a coal seam thickness prediction error ≤5cm. This invention integrates inertial navigation attitude data, cutting motor current characteristics, and geological real-world data to achieve real-time dynamic modeling of the coal seam roof and floor dip angles, shortening the model update cycle to automatic triggering every 10m advance. A 21-DOF kinematic model enables precise conversion between geological grid points and equipment absolute coordinates with an error <3cm, avoiding cutting trajectory deviations caused by coordinate biases; thus improving the accuracy of geological modeling. 2. This invention generates the Pareto optimal solution set based on the improved NSGA-II algorithm, comprehensively optimizing mining time, energy consumption, and equipment wear. In practical applications, production efficiency is increased by 35% and energy consumption is reduced by 18%. Through the DDPG reinforcement learning model, millisecond-level response coordination of the coal mining machine, hydraulic support, and scraper conveyor is achieved, reducing equipment idle waiting time by 60% and extending equipment life by 20%. By using the ARIMA model to predict cutting resistance 3 seconds in advance and dynamically adjusting the traction speed (formula: v_new = v_base ×(1 - 0.05ΔP)), the overload failure rate of the cutting motor is reduced by 45%. Thus, mining efficiency and resource utilization are optimized. 3. This invention uses an improved NURBS curve algorithm to generate cutting trajectories with 20cm intervals. Combined with closed-loop feedback from a laser scanner, it achieves a cutting height tracking error of ±3cm, which is 50% higher than the traditional system. The electro-hydraulic control mining model uses pressure sensors to calibrate the support pushing stroke in real time, with a straightness control error of <2cm and a follow-up action synchronization rate of 98%. The PID compensation mechanism compensates for interference factors such as cutting vibration and sudden changes in coal seam hardness in real time, with a trajectory adjustment response time of <200ms; thereby improving control accuracy and stability. Attached Figure Description

[0018] Figure 1 The diagram illustrates the workflow of the autonomous planning and mining control system module based on big data of the present invention. Figure 2 The diagram shown is a schematic representation of the operation steps of the autonomous planning and mining control system based on big data according to the present invention. Detailed Implementation

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0020] Please see Figure 1 and Figure 2 This invention provides an embodiment of an autonomous planning and mining control system based on big data, comprising: The high-precision dynamic geological modeling module is used to connect with the underground coal mining machine, inertial navigation system, and cutting motor through the working face ring network to collect geological exploration data, equipment attitude data, and working condition data in real time. The big data intelligent decision-making module is deployed at the ground dispatch center to integrate with the IMS-P big data platform, receive geological model data output by the high-precision dynamic geological modeling module, and generate mining process optimization strategies through the HIMS-CDM algorithm. The autonomous planning and control module, which includes a black box control system and a closed-loop feedback unit, communicates with the coal mining machine's electrical control system and is used to dynamically adjust the cutting trajectory. The edge-cloud collaborative architecture module consists of an underground explosion-proof host and a ground workstation. It synchronizes data through an industrial ring network and a 5G redundant channel, and supports the autonomous operation of the local cache model in a network outage environment.

[0021] Preferably, the high-precision dynamic geological modeling module includes a multi-source data fusion unit, a dynamic correction unit, and a three-dimensional coordinate transformation unit. The multi-source data fusion unit is used to fuse geological realistic data, coal mining machine inertial navigation attitude data, and cutting motor current fluctuation characteristics, wherein the pitch angle accuracy of the coal mining machine inertial navigation attitude data is ±0.1°, and the sampling rate of the cutting motor current fluctuation characteristics is 1kHz. The dynamic correction unit is used to generate a geological slice model with an interval of 0.2m using the CT scanning principle, triggering a model update every 10m, and integrating an LSTM neural network to perform reverse error correction on historical cutting data, with a coal seam thickness prediction error ≤5cm. The three-dimensional coordinate transformation unit is used to establish a coal mining machine kinematic model containing 21 degrees of freedom, and to perform real-time mapping between the geological grid point coordinates and the absolute coordinates of the equipment, with a transformation error <3cm.

[0022] Preferably, the dynamic correction unit operates by including the following steps: S101: Receives real-time cutting data from the coal mining machine and extracts the cutter number, cutting height, and turning point features; S102: Compare the current slice data with the prediction model and calculate the height deviation Δh; S103: When Δh > 8cm, activate the LSTM network to adjust the weights and output the correction coefficient α∈[0.8,1.2]; S104: Update the predicted coal seam thickness in the unmined area according to the formula H_new=H_old×α; S105: Generate an updated geological model and synchronize it to the big data decision-making module.

[0023] Preferably, a coal seam thickness prediction model is constructed by integrating geological exploration data, coal mining machine inertial navigation attitude data, and cutting motor current characteristics through a multi-source data fusion unit; a dynamic correction unit is used to update the model by CT scanning-style slices, updating it every 10m, and combining it with an LSTM neural network to perform reverse correction on historical cutting data, thereby controlling the coal seam thickness prediction error within ±5cm; a 21-DOF coal mining machine kinematic model is established through a three-dimensional coordinate transformation unit to achieve real-time mapping between geological grid points and the absolute coordinates of the equipment.

[0024] Preferably, the big data intelligent decision-making module includes a multi-objective optimization engine, a process knowledge graph library, and a real-time collaborative algorithm unit. The multi-objective optimization engine is used to establish the objective function min F(x) = 0.4T + 0.3E + 0.3W, where T is the mining time cost, E is the energy consumption coefficient, and W is the equipment wear index. The process knowledge graph library is used to store mining case data and support process recommendations based on semantic matching. The real-time collaborative algorithm unit is used to optimize the three-machine collaborative parameters using the DDPG reinforcement learning model. The big data intelligent decision-making module generates control parameters for the triangular coal collaborative process section through the HIMS-APD system and outputs a mining strategy instruction set including velocity gradient and turning point offset.

[0025] Preferably, the real-time collaborative algorithm unit performs the following steps during operation: S201: Acquire hydraulic support pressure sensor data, scraper conveyor load data, and coal mining machine positioning information; S202: Construct the state space S={s1,s2,s3}, where s1 is the relative position matrix of the equipment, s2 is the environmental parameter vector, and s3 is the geological model features; S203: Calculate the optimal strategy for motion space A={a1,a2,a3} using the Q-Learning algorithm, where a1 is the traction speed adjustment, a2 is the roller height compensation value, and a3 is the support pushing sequence. S204: Output control commands to the autonomous planning control module and record the execution results to the sample library.

[0026] Preferably, the big data intelligent decision-making module establishes an objective function min F(x) = 0.4T + 0.3E + 0.3W (T is time cost, E is energy consumption, and W is equipment wear) through a multi-objective optimization engine, generates the Pareto optimal solution set through an improved NSGA-II algorithm, integrates mining cases through a process knowledge graph, constructs a semantic matching rule base, and supports automatic recommendation of collaborative process parameters for triangular coal mining, and optimizes the timing of the three machines' actions based on DDPG reinforcement learning through a real-time collaborative algorithm, with a response latency of <500ms.

[0027] Preferably, the autonomous planning and control module includes a dynamic trajectory planning unit, a prediction and compensation unit, and a closed-loop verification unit. The dynamic trajectory planning unit is used to generate cutting trajectories with 20cm intervals using an improved NURBS curve algorithm, and supports real-time insertion of correction points. The prediction and compensation unit predicts the cutting resistance 3 seconds in advance based on the ARIMA model and adjusts the traction speed according to the formula v_new=v_base×(1-0.05ΔP), where ΔP is the pressure change rate. The closed-loop verification unit integrates a laser scanner to measure the cutting contour in real time and generates a PID compensation signal to be fed back to the electro-hydraulic servo system.

[0028] Preferably, the dynamic trajectory planning unit includes the following steps during operation: S301: Extract the coordinate sequence of the top and bottom plates of the current slice {P1,P2,…,Pn} from the geological model; S302: The initial trajectory is fitted using a cubic B-spline curve, and the node vectors are set to uniform parameterization; S303: When inserting a manual intervention point Q, perform weighted smoothing according to the formula Q_new=0.7Q_manual+0.3Q_auto; S304: Outputs a trajectory instruction set containing position-velocity-time parameters to the coal mining machine controller.

[0029] Preferably, the autonomous planning control module generates cutting trajectories with 20cm intervals by using an improved NURBS curve algorithm, supports weighted smoothing of manual intervention points (weight coefficient 0.7:0.3); predicts cutting resistance 3 seconds in advance based on the ARIMA model and dynamically adjusts the traction speed (formula: v_new=v_base×(1-0.05ΔP)); monitors the cutting contour in real time through a laser scanner, generates PID compensation signals to adjust the hydraulic servo system, and performs closed-loop feedback verification.

[0030] Preferably, the implementation steps of the edge-cloud collaborative architecture module specifically include: S401: Deploy edge computing nodes on the explosion-proof host to execute 50ms-level real-time control algorithms and cache the geological model and control parameters of the most recent 10m. S402: The ground workstation communicates with the downhole equipment via the OPC UA protocol, employs a dual-queue mechanism to ensure command transmission, and automatically switches to local control mode when the network delay is greater than 1 second. S403: In the offline state, the explosion-proof host will maintain autonomous operation for 30 minutes based on cached data, and attempt to reconnect once every 5 minutes.

[0031] Preferably, a safety control subsystem is also included, comprising a five-level safety response mechanism, a UWB personnel positioning linkage unit, and an emergency self-locking device. The five-level safety response mechanism specifically includes setting the cutting motor temperature thresholds to T1=85℃, T2=95℃, and vibration amplitude thresholds to A1=0.5g, A2=1.2g, triggering speed reduction and emergency stop. The UWB personnel positioning linkage unit automatically locks the traction system when personnel are detected entering an area 5m in front of the coal mining machine. The emergency self-locking device uses dual redundant solenoid valves for control, cutting off hydraulic power within 0.5s under abnormal operating conditions.

[0032] As a preferred option, the system operation steps specifically include: S1: Initialization phase: Load geological real data, construct the initial geological model and verify the accuracy ≥15cm; S2: Decision-making stage: Calculate the Pareto optimal solution set using the NSGA-II algorithm and select the mining strategy with the highest comprehensive score; S3: Control phase: The cutting trajectory is sent to the coal mining machine, and the laser scanner is started simultaneously for contour monitoring; S4: Feedback phase: Collect actual cutting data and equipment operating conditions, and reverse-correct the geological model and control parameters; S5: Iteration phase: After every 10m of progress, a new round of model optimization and strategy generation is automatically started.

[0033] Example 1 Optionally, in a coal mine working face, the coal seam thickness varies drastically (1.5-3.2m), and the roof and floor dip angles fluctuate by ±15°. Traditional systems have cutting trajectory deviations of ±20cm, requiring manual adjustments more than 15 times per shift. The present invention employs a big data-based autonomous planning and mining control system for dynamic modeling and high-precision cutting under complex coal seam conditions. The specific implementation process is as follows: A1: Load geological realism data (drill hole spacing 50m) and coal mining machine inertial navigation data (pitch angle accuracy ±0.1°) to construct an initial geological model (accuracy 25cm). A2: Enable LSTM dynamic correction unit, set the model update to be triggered by advancing 10m, and the historical truncation data feedback cycle to 2 minutes; A13: The coal mining machine runs along the NURBS trajectory and automatically inserts correction points in the coal seam thinning area (2.1m→1.8m), with the drum height adjustment range Δh=0.3m; A4: Laser scanner detects cutting contours in real time (sampling rate 100Hz), PID compensation hydraulic system adjustment ΔP=1.2MPa; The results are compared below:

[0034] Among them, the model deviation is corrected in real time by the dynamic correction unit to ensure the cutting accuracy; the closed-loop feedback mechanism eliminates mechanical errors through laser scanning and PID compensation.

[0035] Example 2 Optionally, in a certain mine's fully mechanized mining face, the coal mining machine and hydraulic support are not synchronized, with a delay of up to 2.5 seconds, resulting in a cumulative downtime of 4 hours per day. The big data-based autonomous planning and mining control system of this invention is used to optimize the collaborative mining efficiency of multiple equipment. The specific implementation process is as follows: B1: In the DDPG algorithm, define the state space S={support pressure matrix, scraper conveyor load curve, coal mining machine position}, and the action space A={pushing time sequence Δt, traction speed v}; B2: Set optimization goals: support follow-up delay < 0.5 seconds, scraper conveyor full load rate > 85%; B3: When the traction speed of the coal mining machine increases from 3m / min to 4.2m / min, the DDPG model automatically adjusts the support pushing interval from 5 supports / minute to 7 supports / minute; B4: When the scraper conveyor load exceeds the limit (>90%), the speed coordination adjustment is triggered: the coal mining machine speed is reduced to 3.5m / min, and the scraper conveyor speed is increased by 10%; The results are compared below:

[0036] Among them, a real-time collaborative algorithm achieves millisecond-level response, and a multi-objective optimization engine balances efficiency and equipment wear.

[0037] Example 3 Optionally, a mine encounters a 1.2m vertical fault. Traditional systems require a 4-hour shutdown for manual surveying, impacting production progress. The present invention employs a big data-based autonomous planning and mining control system for emergency handling of fault geological anomalies. The specific implementation process is as follows: C1: The cutting motor current suddenly increased by 120% (threshold 100A→225A), the vibration sensor detected an amplitude of 1.5g (exceeding threshold A2=1.2g), triggering a level three safety response; C2: The system automatically switches to fault mode: the coal mining machine slows down to 1.5m / min, and the drum is raised 0.4m to avoid the rock strata; C3: Initiate geological CT scan, generate a 0.1m interval slice model of the fault area, and plan the detour trajectory; C4: Production Resumption: It took 8 minutes from anomaly identification to generating a new trajectory. Mining resumed after manual confirmation, with a resource recovery rate of 82% in the fault area. The results are compared below:

[0038] Among these features, a five-level safety response system enables rapid emergency stops, and dynamic geological modeling supports rapid replanning.

[0039] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A big data-based autonomous planning and mining control system, characterized by: Including: The high-precision dynamic geological modeling module is used to connect with the underground coal mining machine, inertial navigation system, and cutting motor through the working face ring network to collect geological exploration data, equipment attitude data, and working condition data in real time. The big data intelligent decision-making module is deployed at the ground dispatch center to integrate with the IMS-P big data platform, receive geological model data output by the high-precision dynamic geological modeling module, and generate mining process optimization strategies through the HIMS-CDM algorithm. The autonomous planning and control module, which includes a black box control system and a closed-loop feedback unit, communicates with the coal mining machine's electrical control system and is used to dynamically adjust the cutting trajectory. The edge-cloud collaborative architecture module consists of an underground explosion-proof host and a ground workstation. It synchronizes data through an industrial ring network and a 5G redundant channel, and supports the autonomous operation of the local cache model in a network outage environment.

2. The autonomous planning and mining control system based on big data according to claim 1, characterized in that: The high-precision dynamic geological modeling module includes a multi-source data fusion unit, a dynamic correction unit, and a three-dimensional coordinate transformation unit. The multi-source data fusion unit is used to fuse geological realistic data, coal mining machine inertial navigation attitude data, and cutting motor current fluctuation characteristics. The pitch angle accuracy of the coal mining machine inertial navigation attitude data is ±0.1°, and the sampling rate of the cutting motor current fluctuation characteristics is 1kHz. The dynamic correction unit is used to generate a geological slice model with an interval of 0.2m using the CT scanning principle. The model is updated every 10m and an LSTM neural network is integrated to perform reverse error correction on historical cutting data, with a coal seam thickness prediction error ≤5cm. The three-dimensional coordinate transformation unit is used to establish a coal mining machine kinematic model with 21 degrees of freedom and to map the geological grid point coordinates to the absolute coordinates of the equipment in real time, with a transformation error <3cm.

3. The autonomous planning and mining control system based on big data according to claim 2, characterized in that: When the dynamic correction unit is in operation, the specific working steps include: S101: Receives real-time cutting data from the coal mining machine and extracts the cutter number, cutting height, and turning point features; S102: Compare the current slice data with the prediction model and calculate the height deviation Δh; S103: When Δh > 8cm, activate the LSTM network to adjust the weights and output the correction coefficient α∈[0.8,1.2]; S104: Update the predicted coal seam thickness in the unmined area according to the formula H_new=H_old×α; S105: Generate an updated geological model and synchronize it to the big data decision-making module.

4. The autonomous planning and mining control system based on big data according to claim 1, characterized in that: The big data intelligent decision-making module includes a multi-objective optimization engine, a process knowledge graph library, and a real-time collaborative algorithm unit. The multi-objective optimization engine is used to establish the objective function min F(x)=0.4T+0.3E+0.3W, where T is the mining time cost, E is the energy consumption coefficient, and W is the equipment wear index. The process knowledge graph library is used to store mining case data and support process recommendations based on semantic matching. The real-time collaborative algorithm unit is used to optimize the three-machine collaborative parameters using the DDPG reinforcement learning model; the big data intelligent decision-making module generates control parameters for the triangular coal collaborative process section through the HIMS-APD system, and outputs a mining strategy instruction set including speed gradient and turning point offset.

5. The autonomous planning and mining control system based on big data according to claim 4, characterized in that: When the real-time collaborative algorithm unit is working, it performs the following steps: S201: Acquire hydraulic support pressure sensor data, scraper conveyor load data, and coal mining machine positioning information; S202: Construct the state space S={s1,s2,s3}, where s1 is the relative position matrix of the equipment, s2 is the environmental parameter vector, and s3 is the geological model features; S203: Calculate the optimal strategy for motion space A={a1,a2,a3} using the Q-Learning algorithm, where a1 is the traction speed adjustment, a2 is the roller height compensation value, and a3 is the support pushing sequence. S204: Output control commands to the autonomous planning control module and record the execution results to the sample library.

6. The autonomous planning and mining control system based on big data according to claim 1, characterized in that: The autonomous planning and control module includes a dynamic trajectory planning unit, a prediction and compensation unit, and a closed-loop verification unit. The dynamic trajectory planning unit uses an improved NURBS curve algorithm to generate cutting trajectories with 20cm intervals and supports real-time insertion of correction points. The prediction and compensation unit predicts the cutting resistance 3 seconds in advance based on the ARIMA model and adjusts the traction speed according to the formula v_new=v_base×(1-0.05ΔP), where ΔP is the pressure change rate. The closed-loop verification unit integrates a laser scanner to measure the cutting contour in real time and generates a PID compensation signal to be fed back to the electro-hydraulic servo system.

7. The autonomous planning and mining control system based on big data according to claim 6, characterized in that: The dynamic trajectory planning unit, when in operation, includes the following steps: S301: Extract the coordinate sequence of the top and bottom plates of the current slice {P1,P2,…,Pn} from the geological model; S302: The initial trajectory is fitted using a cubic B-spline curve, and the node vectors are set to uniform parameterization; S303: When inserting a manual intervention point Q, perform weighted smoothing according to the formula Q_new=0.7Q_manual+0.3Q_auto; S304: Outputs a trajectory instruction set containing position-velocity-time parameters to the coal mining machine controller.

8. The autonomous planning and mining control system based on big data according to claim 1, characterized in that: The specific implementation steps of the edge-cloud collaborative architecture module include: S401: Deploy edge computing nodes on the explosion-proof host to execute 50ms-level real-time control algorithms and cache the geological model and control parameters of the most recent 10m. S402: The ground workstation communicates with the downhole equipment via the OPC UA protocol, employs a dual-queue mechanism to ensure command transmission, and automatically switches to local control mode when the network delay is greater than 1 second. S403: In the offline state, the explosion-proof host will maintain autonomous operation for 30 minutes based on cached data, and attempt to reconnect once every 5 minutes.

9. The autonomous planning and mining control system based on big data according to claim 1, characterized in that: It also includes a safety control subsystem, which includes a five-level safety response mechanism, a UWB personnel positioning linkage unit, and an emergency self-locking device. The five-level safety response mechanism specifically includes setting the cutting motor temperature thresholds to T1=85℃, T2=95℃, and vibration amplitude thresholds to A1=0.5g, A2=1.2g, triggering deceleration and emergency stop. The UWB personnel positioning linkage unit is used to automatically lock the traction system when personnel are detected entering the area 5m in front of the coal mining machine; the emergency self-locking device is used to cut off the hydraulic power within 0.5s under abnormal working conditions by using dual redundant solenoid valve control.

10. The autonomous planning and mining control system based on big data according to any one of claims 1-9, characterized in that: The specific steps of system operation include: S1: Initialization phase: Load geological real data, construct the initial geological model and verify the accuracy ≥15cm; S2: Decision-making stage: Calculate the Pareto optimal solution set using the NSGA-II algorithm and select the mining strategy with the highest comprehensive score; S3: Control phase: The cutting trajectory is sent to the coal mining machine, and the laser scanner is started simultaneously for contour monitoring; S4: Feedback phase: Collect actual cutting data and equipment operating conditions, and reverse-correct the geological model and control parameters; S5: Iteration phase: After every 10m of progress, a new round of model optimization and strategy generation is automatically started.

Images