Intelligent control system of deep well mining equipment

Abstract

The invention belongs to the technical field of equipment control systems, and discloses a deep well mining equipment intelligent control system which comprises a multi-source data sensing module, a decision module and an equipment cooperative control module. The multi-source data sensing module comprises a distributed Internet of Things sensing unit, an InSAR remote sensing detection unit and a data fusion gateway unit. The decision-making module receives the multi-source signal and comprises a prediction unit, an optimization unit and a regulation and control unit, the prediction unit processes the multi-source signal and outputs a height prediction signal and an aquifer penetration risk signal, and the optimization unit processes the multi-source signal and the height prediction signal to obtain an optimal grouting signal; and the regulation and control unit defines a water retention constraint condition according to the multi-source signal and outputs a safe mining threshold value. According to the intelligent control system of the deep well mining equipment, the problems of water resource loss and pollution caused by traditional deep well mining operation can be solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of equipment control systems, and specifically relates to an intelligent control system for deep-well mining equipment, which is particularly suitable for achieving safe and efficient mineral resource development under the technical requirements of water-conservation mining. Background Art

[0002] In the field of deep-well mineral resource mining, especially in underground mining operations such as coal and metal mines, achieving safe and efficient mining and protecting groundwater resources has always been a core challenge facing the industry. Traditional mining equipment control systems are mainly designed around production efficiency and personnel safety, lacking active protection mechanisms for the hydrogeological environment of the mining area. This has led to long-standing problems such as uncontrollable hydrogeological damage, weak data perception and fusion capabilities, extensive and inefficient grouting processes, low levels of equipment collaborative intelligence, and sharp conflicts between ecological protection and mining. Existing technologies make it difficult to balance mining intensity with the need to protect water resources. Conventional control systems target fixed production capacity and lack dynamic constraints on ecological indicators such as groundwater levels and water quality.

[0003] Existing technologies rely on localized point monitoring and lack the ability to sense the multi-physics coupling of rock stress, fracture development, and hydrological changes. The data update frequency and spatial resolution cannot meet the requirements of deep mining. Furthermore, control strategies are based on a pre-set rule base and cannot adapt to changing geological conditions. Summary of the Invention

[0004] In light of the shortcomings of the aforementioned existing technologies, the present invention aims to provide an intelligent control system for deep-well mining equipment, addressing the problems of water loss and pollution caused by deep-well mining operations. By developing an intelligent control system for deep-well mining equipment that deeply integrates IoT perception, intelligent decision-making, and precise execution, this system addresses the shortcomings of traditional mining technologies in terms of water conservation, equipment coordination, and risk prevention and control.

[0005] The intelligent control system for deep well mining equipment provided by the present invention includes: A multi-source data perception module includes a data fusion gateway unit, which forms a multi-source signal; The decision module receives multi-source signals and includes a prediction unit, an optimization unit, and a control unit. The prediction unit processes the multi-source signals and outputs a height prediction signal and an aquifer breakthrough risk signal. The optimization unit processes the multi-source signals and the height prediction signal to obtain the optimal grouting signal. The control unit defines water conservation constraints based on the multi-source signals and outputs a safe mining threshold. Equipment collaborative control module, the equipment collaborative control module includes an adaptive grouting execution unit, a mining equipment group control unit and a human-computer interaction and early warning module. The adaptive grouting execution unit adjusts the grouting parameters according to the optimal grouting signal. The mining equipment group control unit adjusts the mining equipment parameters according to the safe mining threshold and multi-source signals, and collects real-time detection signals. The human-computer interaction and early warning module processes the aquifer breakthrough risk signal and height prediction signal and outputs a comprehensive risk index.

[0006] Furthermore, the multi-source data perception module includes a distributed Internet of Things sensing unit and an InSAR (interferometric synthetic aperture radar) remote sensing detection unit. The multi-source data perception module is deployed in underground mining faces, tunnels and the surface to collect sensing signals. The InSAR remote sensing detection unit obtains surface deformation data of the mining area through satellite interferometry technology and forms surface signals. The data fusion gateway unit receives sensing signals and surface signals to realize heterogeneous data fusion and form multi-source signals. The multi-source signals include predicted multi-source signals, optimized multi-source signals and controlled multi-source signals.

[0007] Furthermore, the prediction unit includes a time series prediction model based on an LSTM (long short-term memory artificial neural network) neural network and a DS evidence theory model, and predicts multi-source signals including rock stress data, microseismic event energy distribution data, InSAR surface deformation gradient data, historical mining parameters, and historical water-conducting fracture zone height data. The rock stress data, historical mining parameters, microseismic event energy distribution data, InSAR surface deformation gradient data, and historical water-conducting fracture zone height data are input into the time series prediction model and output a height prediction signal. The rock stress data, microseismic event energy distribution data, and InSAR surface deformation gradient data are fused through the improved DS evidence theory model and an aquifer breakthrough risk signal is output.

[0008] Furthermore, the time series prediction model includes a time series feature extraction submodule, a multimodal fusion submodule, a risk quantification submodule and a prediction result visualization module. The time series feature extraction submodule performs wavelet packet decomposition on the microseismic event energy distribution data, extracts the energy mutation characteristics of the frequency band that meets the frequency range of the crack extension precursor as the crack extension precursor signal, and the multimodal fusion submodule uses a bidirectional LSTM network weighted by an attention mechanism to process the rock layer stress data, microseismic event energy distribution data and InSAR surface deformation gradient data and outputs the data to the risk quantification submodule. The risk quantification submodule defines the risk index algorithm as shown below: R= ; Among them, R is the risk index, is the coefficient, is the maximum principal stress, is the mean deformation gradient, The prediction result visualization module generates a dynamic risk cloud map of the risk index based on the three-color blocks, and uses blocks of different colors to represent different risk indices at different heights.

[0009] Furthermore, the optimization unit processes the optimized multi-source signal and the height prediction signal through the grouting-fracture nonlinear relationship model, outputs the grouting signal, and processes the grouting diffusion signal using a genetic algorithm to obtain the optimal grouting signal; the optimized multi-source signal includes the aquifer permeability coefficient and the grouting material rheological signal, and the grouting-fracture nonlinear relationship model is constructed through the aquifer permeability coefficient, the height prediction signal, and the grouting material rheological signal. The grouting-fracture nonlinear relationship model is as follows: ; Among them, r is the diffusion radius, Q is the grouting flow rate, t is the grouting time, is the slurry viscosity, is the grouting pressure, is the average width of the crack, and k is the correction coefficient.

[0010] Furthermore, the optimization unit also includes a slurry performance dynamic matching library, a fracture network connectivity analysis submodule and a grouting path planning submodule. The slurry performance dynamic matching library stores historical grouting rheological signals of historical grouting materials. The fracture network connectivity analysis submodule calculates the hydraulic conductivity of the fracture network according to the grouting pressure and diffusion radius. The grouting path planning submodule obtains the optimal grouting signal through a genetic algorithm based on the hydraulic conductivity, the grouting-fracture nonlinear relationship model and the rheological signal of the grouting material.

[0011] Furthermore, the control unit defines a water conservation constraint condition based on the control multi-source signal, outputs a safe mining threshold, and dynamically adjusts the safe mining threshold to be no less than the minimum safe mining value through MPC (model predictive control). The control multi-source signal includes a real-time groundwater level signal and aquiclude thickness signal. The real-time groundwater level signal and aquiclude thickness signal define the water conservation constraint condition as shown below: ; in, is the safe mining threshold, is the depth of the aquifer bottom, For mining depth, To adjust the height of the water-conducting fracture zone, MPC dynamically adjusts the shearer traction speed and support frame shifting step distance to ensure that the safe mining threshold is not lower than the minimum safe mining value.

[0012] Furthermore, the human-computer interaction and early warning module includes a three-dimensional visualization platform and a risk early warning engine. The three-dimensional visualization platform receives multi-source signals and real-time detection signals and dynamically renders the expansion process of the water-conducting fracture zone. The risk early warning engine processes the aquifer breakthrough risk signal and the water-conducting fracture zone height prediction signal according to the improved TOPSIS (Top-of-the-Best Solution Distance Method) algorithm and outputs a comprehensive risk index. When the comprehensive risk index exceeds the risk index threshold, the human-computer interaction and early warning module controls the adaptive grouting execution unit to perform emergency grouting. The three-dimensional visualization platform supports cross-sectional analysis at any angle, can display the spatial relationship between the water-conducting fracture zone and the aquifer in real time, and simulates rock movement and water level changes under different mining schemes based on the discrete element method.

[0013] Furthermore, the risk warning engine includes a multi-level warning mechanism color status, namely yellow warning status, orange warning status and red warning status. When the height of the water-conducting fracture zone in the height prediction signal drops to the first height threshold, the yellow warning status is activated and the operating speed of the mining equipment is reduced to the warning operating speed. When the height of the water-conducting fracture zone in the height prediction signal drops to the second height threshold, the operation of the mining equipment is suspended and the adaptive grouting execution unit is controlled to perform emergency grouting. When the height of the water-conducting fracture zone in the height prediction signal drops to the third height threshold, the risk warning engine sends an emergency mine disaster signal and controls the deep well mining equipment intelligent control system to shut down urgently.

[0014] Furthermore, the optimal grouting signal includes but is not limited to grouting area coordinates, slurry type priority, grouting pressure range and diffusion path planning. The adaptive grouting execution unit includes an intelligent grouting pump and an integrated slurry viscosity online detection device. The intelligent grouting pump can identify the optimal grouting signal and adjust the grouting flow and grouting pressure as needed. The integrated slurry viscosity online detection device receives the optimal grouting signal and adjusts the water-cement ratio of the grouting slurry in real time based on the ultrasonic phase difference method. The real-time detection signals include but are not limited to equipment temperature, equipment humidity, equipment voltage, equipment depth, mining pressure and expansion radius.

[0015] The intelligent control system for deep-well mining equipment provided by this invention addresses the shortcomings of traditional mining technologies in water resource protection, equipment coordination, and risk prevention and control by building a system that deeply integrates IoT perception, intelligent decision-making, and precise execution. Specifically, through a distributed sensor network and InSAR remote sensing technology, real-time monitoring and spatiotemporal alignment of the multi-physical fields of "rock stress, fracture development, and hydrological changes" are achieved, improving the accuracy of risk identification. Based on LSTM neural networks and genetic algorithms, closed-loop decision-making is achieved for predicting the expansion of water-conducting fracture zones, optimizing grouting parameters, and regulating mining intensity, ensuring water-conserving mining objectives. Fuzzy PID control and pulse grouting technology are used to dynamically adjust grouting pressure and diffusion paths, significantly improving plugging efficiency and material utilization. Multi-objective optimization and model predictive control enable precise synchronization of equipment such as coal mining machines, hydraulic supports, and conveyors, reducing unplanned downtime rates. A three-level early warning mechanism and three-dimensional dynamic risk cloud map are established to support AR remote collaboration, improving risk management efficiency and the human-computer interaction experience. This reduces the problem of water loss and pollution in aquifers. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 This is a system architecture diagram of the intelligent control system for deep well mining equipment of the present invention; Figure 2 This is a schematic diagram of the working principle of the intelligent control system for deep well mining equipment of the present invention. DETAILED DESCRIPTION

[0018] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. The details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless they conflict.

[0019] The present invention relates to an intelligent control system for deep-well mining equipment, which is mainly used in deep-well mining devices. Traditional mining methods rely on manual experience to judge the development of water-conducting fracture zones, and lack real-time monitoring and prediction methods. The rock fractures caused by mining often extend to the aquifer, causing the groundwater level to continue to drop. For example, in the shallow coal seam area of ​​northern Shaanxi, a single coal mining operation can cause the water level of the aquifer to drop by 2-3 meters, and the shortage of irrigation water for surrounding farmland is frequent. Existing control systems mostly use independently operated sensor networks, and underground stress monitoring, microseismic detection and surface deformation observation data have been in a state of separation for a long time. For example, a certain type of mining monitoring system can only achieve minute-level data updates, and InSAR remote sensing data and underground sensor data are difficult to effectively associate due to the inconsistent time and space benchmarks.

[0020] See Figure 1The figure shows the intelligent control system for deep-well mining equipment according to the present invention. It comprises a multi-source data perception module, a decision-making module, and an equipment collaborative control module. The multi-source data perception module includes a distributed Internet of Things (IoT) sensor unit, an InSAR (Interferometric Synthetic Aperture Radar) remote sensing unit, and a data fusion gateway unit. The multi-source data perception module is deployed at the underground mining face, tunnels, and surface to collect sensor signals. The InSAR remote sensing unit uses satellite interferometry to acquire surface deformation data in the mining area and generate surface signals. The data fusion gateway unit receives sensor signals and surface signals to achieve heterogeneous data fusion and generate multi-source signals. The decision-making module receives multi-source signals and includes a prediction unit, an optimization unit, and a control unit. The prediction unit processes the multi-source signals using a time series prediction model based on an LSTM (Long Short-Term Memory) neural network, outputting a predicted signal for the height of the water-conducting fracture zone, and employs an improved DS evidence theory to fuse the multi-source signals to calculate the probability distribution of fracture extension and output a signal indicating the risk of aquifer breakthrough. The optimization unit processes the optimized multi-source signals and the water-conducting fracture zone height prediction signal through the grouting-fracture nonlinear relationship model, outputs the grouting signal, and uses a genetic algorithm to process the grouting diffusion signal to obtain the optimal grouting signal. The control unit defines the water conservation constraint conditions based on the multi-source control signals, outputs the safe mining threshold, and dynamically adjusts the safe mining threshold to no less than the minimum safe mining value through MPC. The equipment collaborative control module includes an adaptive grouting execution unit, a mining equipment group control unit, and a human-computer interaction and early warning module. The adaptive grouting execution unit adjusts the grouting parameters based on the optimal grouting signal. The mining equipment group control unit adjusts the mining equipment parameters based on the safe mining threshold and multi-source signals and collects real-time detection signals. The human-computer interaction and early warning module includes a three-dimensional visualization platform and a risk warning engine. The three-dimensional visualization platform receives multi-source signals and real-time detection signals and dynamically renders the expansion process of the water-conducting fracture zone. The risk warning engine processes the aquifer breakthrough risk signal and the water-conducting fracture zone height prediction signal based on the improved TOPSIS (Top-of-the-Best Solution Distance) algorithm and outputs a comprehensive risk index. When the comprehensive risk index exceeds the risk index threshold, the human-computer interaction and early warning module controls the adaptive grouting execution unit to perform emergency grouting.

[0021] like Figure 1As shown in Figure 1, the entire system consists of core components such as a multi-source data fusion perception module, a decision-making module, an adaptive grouting execution unit, and a mining equipment group controller. These modules interact with each other via an industrial bus and the Internet of Things protocol. During implementation, a distributed sensor network deployed in underground tunnels and on the surface begins operations. These sensors include fiber-optic water pressure sensors embedded in the rock formation, three-dimensional stress detectors mounted on hydraulic supports, and microseismic monitoring nodes deployed on the roof of the transport tunnel. The fiber-optic water pressure sensors collect pressure fluctuation data from the underground aquifer every second, covering a measurement range of 0 to 10 MPa. This data is transmitted to the data fusion gateway aboveground via explosion-proof optical cables. Simultaneously, vibration sensors installed in the shearer's cutting section monitor the equipment's operating status in real time, sampling vibration spectrum data 1000 times per second for analysis of pick wear. On the surface, the InSAR remote sensing monitoring subsystem acquires surface deformation data from satellites every seven days, generating a subsidence gradient map with millimeter-level accuracy. This data is then aligned with the underground sensor data in both temporal and spatial dimensions after coordinate conversion.

[0022] After the multi-source data fusion perception module completes initial data acquisition, the data fusion gateway initiates the heterogeneous data processing process. First, it interpolates data from different sampling frequencies, for example, dynamically time-warping sensor data collected once a second with satellite data collected once a week to generate a data sequence with a unified time base. It then uses the Kalman filter algorithm to eliminate sensor noise. For abnormal mutation values ​​detected by the fiber-optic water pressure sensor (such as an instantaneous pressure drop of more than 20%), the data verification program is automatically triggered, and data from adjacent sensors is retrieved for cross-validation. After data cleaning, the system standardizes the format of structured sensor data, semi-structured device log data, and unstructured InSAR image data, converts them into the Apache Avro intermediate format, and stores them in the corresponding partitions of the distributed database.

[0023] like Figure 2As shown, the decision-making module consists of a prediction unit, an optimization unit, and a control unit. The prediction unit processes multi-source signals using a time series prediction model based on an LSTM neural network, outputting a prediction signal for the height of the water-conducting fracture zone. It then fuses these multi-source signals using an improved DS evidence theory to calculate the probability distribution of fracture extension and output an aquifer breakthrough risk signal. These multi-source signals include rock stress data, microseismic event energy distribution data, InSAR surface deformation gradient data, historical mining parameters, and historical water-conducting fracture zone height data. These data are fed into the time series prediction model, which then outputs a prediction signal for the height of the water-conducting fracture zone. The improved DS evidence theory then fuses these data, which in turn outputs an aquifer breakthrough risk signal. The time series prediction model includes a time series feature extraction submodule, a multimodal fusion submodule, a risk quantification submodule, and a prediction result visualization module. The time series feature extraction submodule performs wavelet packet decomposition on the microseismic event energy distribution data and extracts the energy mutation characteristics of the frequency band that meets the frequency range of the crack extension precursor as the crack extension precursor signal. The multimodal fusion submodule uses a bidirectional LSTM network with attention mechanism weighting to process the rock layer stress data, microseismic event energy distribution data, and InSAR surface deformation gradient data and outputs the data to the risk quantification submodule. The risk quantification submodule defines the risk index algorithm as shown below: R= ; Among them, R is the risk index, is the coefficient, is the maximum principal stress, is the mean deformation gradient, The prediction result visualization module generates a dynamic risk cloud map of the risk index based on the three-color blocks for the microseismic energy peak, and uses blocks of different colors to represent different risk indices in the region. The optimization unit processes the multi-source signals and the water-conducting fracture zone height prediction signal through the grouting-fracture nonlinear relationship model, outputs the grouting signal, and uses the genetic algorithm to process the grouting diffusion signal to obtain the optimal grouting signal. The control unit defines the water conservation constraint condition based on the multi-source signal, outputs the safe mining threshold, and dynamically adjusts the safe mining threshold to not less than the minimum value of safe mining through MPC. The multi-source signals include the permeability coefficient of the aquifer and the rheological signal of the grouting material. The grouting-fracture nonlinear relationship model is constructed through the permeability coefficient of the aquifer, the prediction signal of the water-conducting fracture zone height and the rheological signal of the grouting material. The grouting-fracture nonlinear relationship model is as follows: ; Among them, r is the diffusion radius, Q is the grouting flow rate, t is the grouting time, is the slurry viscosity, is the grouting pressure, is the average width of the crack, and k is the correction coefficient.

[0024] Specifically, in one embodiment of the present invention, the decision module extracts pre-processed data from the database and starts the fracture development prediction algorithm; the algorithm first performs wavelet packet decomposition on the acoustic emission signal captured by the microseismic monitoring node to extract the energy mutation characteristics of the 6-8 Hz frequency band, which have been proven to be strongly correlated with the expansion of rock fractures; at the same time, the stress tensor data collected by the three-dimensional stress sensor array at a grid density of 5 meters is reduced to three principal stress components after principal component analysis and input into the trained LSTM neural network model; the neural network model is pre-trained based on historical mining data (including the measured values ​​of the height of the water-conducting fracture zone of 200 working faces in the past three years) and can predict the expansion trend of the water-conducting fracture zone within the next 72 hours; the surface deformation gradient data obtained by InSAR is associated with the downhole stress field through a spatial interpolation algorithm. When the deformation gradient is detected to exceed 5 mm per meter, the algorithm automatically increases the risk level weight. Figure 2 As shown, the equipment collaborative control module includes an adaptive grouting execution unit, a mining equipment group control unit, and a human-computer interaction and early warning module. The optimization unit also includes a slurry performance dynamic matching library, a fracture network connectivity analysis submodule, and a grouting path planning submodule. The slurry performance dynamic matching library stores historical grouting rheological signals of historical grouting materials. The fracture network connectivity analysis submodule calculates the water conductivity of the fracture network based on the grouting pressure and diffusion radius. The grouting path planning submodule obtains the optimal grouting signal through a genetic algorithm based on the water conductivity, the grouting-fracture nonlinear relationship model, and the rheological signal of the grouting material. The multi-source signals here include real-time groundwater level signals and aquiclude thickness signals. The real-time groundwater level signals and aquiclude thickness signals define the water conservation constraints as follows: ; in, is the safe mining threshold, is the depth of the aquifer bottom, For mining depth, To address the height of the water-conducting fracture zone, MPC dynamically adjusts the shearer traction speed and support step distance to ensure that the safe mining threshold does not fall below the minimum safe mining value. The optimal grouting signal includes but is not limited to the grouting area coordinates, slurry type priority, grouting pressure range, and diffusion path planning. The adaptive grouting execution unit includes an intelligent grouting pump and an integrated slurry viscosity online detection device. The intelligent grouting pump can identify the optimal grouting signal and adjust the grouting flow and grouting pressure as needed. The integrated slurry viscosity online detection device receives the optimal grouting signal and adjusts the water-cement ratio of the grouting slurry in real time based on the ultrasonic phase difference method.

[0025] In one embodiment of the present invention, after the fracture development prediction results are generated, the grouting parameter optimization algorithm begins. The algorithm first retrieves available material data from a dynamic grouting property matching library, including rheological curves for over 20 grouting materials, such as cement-based grouts and polymer chemical grouts. Based on real-time water quality parameters (such as pH and conductivity), the system automatically excludes grout types that could cause secondary contamination. For example, if excessive chloride ion concentration is detected in the water, chemical grouts containing metal components are prohibited. Next, a topological analysis of the fracture network is performed using a Delaunay triangulation algorithm, calculating the hydraulic conductivity of key seepage paths and generating a three-dimensional grouting priority map. Finally, a genetic algorithm is used to determine the optimal grouting parameter combination, minimizing material consumption while ensuring a grouting curtain coverage exceeding 95%. The calculated results are then transmitted to the adaptive grouting execution unit via the OPC UA protocol. Upon receiving the grouting instruction, the intelligent grouting pump initiates the workflow. First, the water-cement ratio is automatically adjusted according to the slurry type. The slurry rheological properties are monitored in real time using an ultrasonic viscosity tester, and the mixer speed is dynamically adjusted to maintain optimal mixing uniformity. The grouting pump's pressure control system uses a fuzzy PID algorithm to stabilize the grouting pressure within ±0.2 MPa of the set value. When abnormal pressure fluctuations are detected (such as a momentary increase of more than 0.5 MPa / s), it immediately switches to pulse grouting mode, performing intermittent high-pressure shocks at a frequency of 2-5 Hz to prevent crack blockage. During the grouting process, distributed resistivity imagers arranged around the borehole continuously monitor the slurry diffusion range, generating a resistivity distribution map every second. The three-dimensional distribution of the slurry is reconstructed using an inversion algorithm, and the results are fed back to the decision-making module in real time for dynamic optimization of grouting parameters. The mining equipment group controller synchronously receives mining intensity control instructions from the decision-making module; the coal mining machine navigation system automatically adjusts the cutting trajectory according to the real-time updated water-conducting fracture zone prediction model. When the prediction shows that there is a high risk of water inrush in the area 5 meters ahead, the control system will reduce the traction speed from 4 meters per minute to 2 meters and trigger the roof support reinforcement mode; the electro-hydraulic control system of the hydraulic support dynamically adjusts the initial support force according to the three-dimensional stress sensor data, and increases the support resistance to more than 35 MPa in the stress concentration area. At the same time, the misalignment of adjacent supports is monitored through the infrared ranging sensor. When a posture deviation of more than 50 mm is detected, the correction action is automatically performed; the scraper conveyor's variable frequency drive adjusts the chain speed according to the load current prediction model, reducing energy consumption by 15%-20% while ensuring transportation efficiency. When the coal quantity sensor detects that the instantaneous load exceeds 120% of the rated value, the overload protection program is automatically triggered.The intelligent storage and regulation unit of the underground reservoir forms a linkage mechanism with the grouting execution system; when the grouting operation consumes a large amount of water resources, the storage and regulation unit starts the ecological water replenishment mode, and adjusts the opening of the water release valve through the fuzzy PID controller to maintain the reservoir water level within the set range; the water purification device adopts a combination of electrochemical oxidation and membrane filtration to deeply treat the reinjection water. When the online monitoring instrument detects that the COD value exceeds 100 mg / L, it automatically increases the electrolysis current intensity and extends the backwash frequency of the membrane component; when the rainy season comes, the system frees up storage space in advance according to the weather forecast data, and uses the goaf to store flood water, which can reduce the peak flood flow by more than 30%; during the winter freezing period, the insulation mode is activated, and the mine water waste heat recovery device is used to maintain the reservoir temperature above 5°C to prevent the water supply pipeline from freezing and cracking.

[0026] In one embodiment of the present invention, the risk warning engine plays a central role in the entire system. When a multi-indicator fusion analysis indicates that the water-conducting fracture zone is close to the aquifer floor, the warning engine initiates a three-level response mechanism: first, a flashing red icon is marked on the 3D visualization platform, and an alert message is sent to all mobile devices via the industrial ring network; second, mining intensity in the affected area is automatically reduced, limiting the speed of coal mining machines to below the safety threshold; and finally, the grouting system is activated to seal key channels, forming a multi-level protection system. The warning system's built-in case-based reasoning library stores over 300 sets of historical accident data. When a pattern similar to a precursor to a water inrush accident is detected, the emergency response plan at that time is automatically retrieved and recommended to the operator. The 3D visualization platform integrates the operating status data of all modules to provide a panoramic monitoring interface. The platform constructs a geological model of the mining area based on BIM technology, displaying information such as InSAR deformation data, microseismic event distribution, and grouting diffusion range as overlays in different color layers. Operators can use gestures to rotate and slice the model at any angle to view the stress distribution within the rock formation. When selecting a hydraulic support, the interface automatically displays the device's real-time pressure data, historical maintenance records, and the status of adjacent equipment. The AR remote collaboration feature allows experts to overlay the virtual model with the real equipment through mixed reality glasses and annotate abnormal points using voice commands. This information is instantly synchronized to the on-site personnel's tablet computer. The energy collaborative optimization system collects real-time energy consumption data from each device through smart meters and dynamically adjusts power usage strategies based on photovoltaic power generation forecasts and the SOC status of energy storage batteries. During peak electricity price periods, the system prioritizes energy storage batteries to power control devices, adjusting the operating hours of high-power equipment such as coal shearers to off-peak electricity price periods. A mine water waste heat recovery system extracts heat from wastewater and uses plate heat exchangers to heat above-ground buildings. This system can replace coal-fired boilers in winter, reducing CO2 emissions by over 40%. If the energy efficiency of a device is detected to be below a set threshold during continuous operation, the system automatically generates a maintenance work order and pushes it to the equipment management system.

[0027] In one embodiment of the present invention, the operational data of the entire system is continuously optimized via a digital twin training platform. The platform constructs a high-precision virtual mine model and uses the discrete element method to simulate rock formation movement patterns under different mining scenarios. After each simulation iteration, the platform automatically compares the predicted results with actual monitoring data and reversely corrects the model parameters. During this process, a deep reinforcement learning algorithm autonomously explores optimal control strategies. For example, it discovered that shortening the grouting interval under specific geological conditions can improve the sealing effect by more than 15%. This empirical knowledge is automatically encoded and stored in a decision rule library. Operations and maintenance personnel can use the historical data playback function to trace the system status at any point in time, analyze the causes of failures, and improve operating procedures.

[0028] In one embodiment of the invention, the system was continuously tested in a mining area for 12 months; during the test period, it successfully predicted three water inrush risk events, with an average warning lead time of 56 hours, which is more than 20 times more efficient than manual monitoring methods; the height control accuracy of the water-conducting fracture zone was improved from ±2.5 meters of the traditional method to ±0.8 meters, and the water consumption per ton of coal was reduced to 0.13 cubic meters; the unplanned downtime of underground equipment was reduced by 68%, reducing a large amount of maintenance costs; these empirical data fully verify the technical advantages and practical value of this system in deep well water conservation mining scenarios.

[0029] The intelligent control system for deep-well mining equipment of the present invention focuses on solving the shortcomings of traditional mining technology in water resource protection, equipment coordination and risk prevention and control by constructing a deep-well mining equipment intelligent control system that deeply integrates Internet of Things perception, intelligent decision-making and precise execution.

[0030] Therefore, the deep well mining equipment intelligent control system of the present invention can solve the problems of water resource loss and pollution caused by deep well mining operations.

[0031] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. Intelligent control system for deep well mining equipment, characterized by: include: A multi-source data perception module, the multi-source data perception module includes a data fusion gateway unit, the data fusion gateway unit forms a multi-source signal; A decision module, wherein the decision module receives the multi-source signal, and the decision module includes a prediction unit, an optimization unit, and a control unit. The prediction unit processes the multi-source signal and outputs a height prediction signal and an aquifer breakthrough risk signal. The optimization unit processes the multi-source signal and the height prediction signal to obtain an optimal grouting signal. The control unit defines a water conservation constraint condition according to the multi-source signal and outputs a safe mining threshold. An equipment collaborative control module, the equipment collaborative control module includes an adaptive grouting execution unit, a mining equipment group control unit and a human-computer interaction and early warning module, the adaptive grouting execution unit adjusts the grouting parameters according to the optimal grouting signal, the mining equipment group control unit adjusts the mining equipment parameters according to the safe mining threshold and the multi-source signal, and collects real-time detection signals, and the human-computer interaction and early warning module processes the aquifer breakthrough risk signal and the height prediction signal and outputs a comprehensive risk index.

2. The intelligent control system for deep well mining equipment according to claim 1 is characterized in that: The multi-source data perception module includes a distributed Internet of Things sensor unit and an InSAR remote sensing detection unit. The multi-source data perception module is deployed in the underground mining face, tunnel and surface, and collects sensor signals. The InSAR remote sensing detection unit obtains the surface deformation data of the mining area through satellite interferometry measurement technology and forms a surface signal. The data fusion gateway unit receives the sensor signal and the surface signal to realize heterogeneous data fusion and form the multi-source signal. The multi-source signal includes predicted multi-source signal, optimized multi-source signal and regulated multi-source signal.

3. The deep well mining equipment intelligent control system according to claim 2 is characterized in that: The prediction unit includes a time series prediction model based on an LSTM neural network and a DS evidence theory model. The predicted multi-source signal includes rock formation stress data, microseismic event energy distribution data, InSAR surface deformation gradient data, historical mining parameters and historical water-conducting fracture zone height data. The rock formation stress data, the historical mining parameters, the microseismic event energy distribution data, the InSAR surface deformation gradient data and the historical water-conducting fracture zone height data are input into the time series prediction model and the height prediction signal is output. The rock formation stress data, the microseismic event energy distribution data and the InSAR surface deformation gradient data are fused through the improved DS evidence theory model and the aquifer breakthrough risk signal is output.

4. The deep well mining equipment intelligent control system according to claim 3 is characterized in that: The time series prediction model includes a time series feature extraction submodule, a multimodal fusion submodule, a risk quantification submodule and a prediction result visualization module. The time series feature extraction submodule performs wavelet packet decomposition on the microseismic event energy distribution data, extracts the energy mutation characteristics of the frequency band that meets the crack extension precursor frequency band range as the crack extension precursor signal, and the multimodal fusion submodule uses a bidirectional LSTM network weighted by an attention mechanism to process the rock layer stress data, the microseismic event energy distribution data and the InSAR surface deformation gradient data and outputs the data to the risk quantification submodule. The risk quantification submodule defines a risk index algorithm as shown below: R= ； Among them, R is the risk index, is the coefficient, is the maximum principal stress, is the mean deformation gradient, The prediction result visualization module generates a dynamic risk cloud map of the risk index based on the three-color blocks, and uses blocks of different colors to represent different risk indexes at different heights.

5. The deep well mining equipment intelligent control system according to claim 2 is characterized in that: The optimization unit processes the optimized multi-source signal and the height prediction signal through the grouting-crack nonlinear relationship model, outputs the grouting signal, and processes the grouting diffusion signal by using a genetic algorithm to obtain the best grouting signal; the optimized multi-source signal includes the aquifer permeability coefficient and the grouting material rheological signal, and the grouting-crack nonlinear relationship model is constructed through the aquifer permeability coefficient, the height prediction signal and the grouting material rheological signal. The grouting-crack nonlinear relationship model is as follows: ； Where r is the diffusion radius, Q is the grouting flow rate, and t is the grouting time. is the slurry viscosity, is the grouting pressure, is the average width of the crack, and k is the correction coefficient.

6. The deep well mining equipment intelligent control system according to claim 5 is characterized in that: The optimization unit also includes a slurry performance dynamic matching library, a fracture network connectivity analysis submodule and a grouting path planning submodule. The slurry performance dynamic matching library stores historical grouting rheological signals of historical grouting materials. The fracture network connectivity analysis submodule calculates the water conductivity of the fracture network according to the grouting pressure and the diffusion radius. The grouting path planning submodule obtains the optimal grouting signal through a genetic algorithm based on the water conductivity, the grouting-fracture nonlinear relationship model and the grouting material rheological signal.

7. The intelligent control system for deep well mining equipment according to claim 2 is characterized in that: The control unit defines a water conservation constraint condition according to the control multi-source signal, outputs a safe mining threshold, and dynamically adjusts the safe mining threshold to be not less than the minimum safe mining value through MPC. The control multi-source signal includes a real-time groundwater level signal and a water-repellent layer thickness signal. The real-time groundwater level signal and the water-repellent layer thickness signal define the water conservation constraint condition as shown below: ； in, is the safe mining threshold, is the depth of the aquifer bottom, For mining depth, The MPC dynamically adjusts the shearer traction speed and the support step distance to ensure that the safe mining threshold is not lower than the minimum safe mining value.

8. The intelligent control system for deep well mining equipment according to claim 1 is characterized in that: The human-computer interaction and early warning module includes a three-dimensional visualization platform and a risk early warning engine. The three-dimensional visualization platform receives the multi-source signal and the real-time detection signal and dynamically renders the expansion process of the water-conducting fracture zone. The risk early warning engine processes the aquifer breakthrough risk signal and the water-conducting fracture zone height prediction signal according to the improved TOPSIS algorithm and outputs a comprehensive risk index. When the comprehensive risk index exceeds the risk index threshold, the human-computer interaction and early warning module controls the adaptive grouting execution unit to perform emergency grouting. The three-dimensional visualization platform supports cross-sectional analysis at any angle, can display the spatial relationship between the water-conducting fracture zone and the aquifer in real time, and simulates the rock movement and water level changes under different mining schemes based on the discrete element method.

9. The deep well mining equipment intelligent control system according to claim 8 is characterized in that: The risk warning engine includes color states of a multi-level warning mechanism, namely yellow warning state, orange warning state and red warning state. When the height of the water-conducting fracture zone in the height prediction signal drops to a first height threshold, the yellow warning state is activated, and the operating speed of the mining equipment is reduced to a warning operating speed. When the height of the water-conducting fracture zone in the height prediction signal drops to a second height threshold, the operation of the mining equipment is suspended and the adaptive grouting execution unit is controlled to perform emergency grouting. When the height of the water-conducting fracture zone in the height prediction signal drops to a third height threshold, the risk warning engine sends out an emergency mine disaster signal and controls the deep well mining equipment intelligent control system to shut down urgently.

10. The intelligent control system for deep well mining equipment according to claim 1, characterized in that: The optimal grouting signal includes but is not limited to grouting area coordinates, slurry type priority, grouting pressure range and diffusion path planning. The adaptive grouting execution unit includes an intelligent grouting pump and an integrated slurry viscosity online detection device. The intelligent grouting pump can identify the optimal grouting signal and adjust the grouting flow and grouting pressure as needed. The integrated slurry viscosity online detection device receives the optimal grouting signal and adjusts the water-cement ratio of the grouting slurry in real time based on the ultrasonic phase difference method. The real-time detection signal includes but is not limited to equipment temperature, equipment humidity, equipment voltage, equipment depth, mining pressure and expansion radius.

Images