A mine water supply and drainage dynamic optimization control method and system based on digital twinning

Abstract

This invention relates to the field of mine automation control technology, and in particular to a dynamic optimization control method and system for mine water supply and drainage based on digital twins. The technical solution is as follows: A dynamic optimization control method for mine water supply and drainage based on digital twins includes a process framework for dynamic optimization control of mine water supply and drainage. This invention obtains the planned water usage for future periods from the production scheduling system as the demand-side input, presets multi-dimensional constraints including the high-efficiency operating range of pump sets, the number of motor starts, and peak-valley electricity price periods, and uses the minimum total system energy consumption and balanced pump wear as dual optimization objectives. An intelligent optimization algorithm is employed to solve for the optimal pump combination scheme and operating parameters, enabling the water supply system to dynamically adjust its operating mode according to changes in actual water demand, ensuring that the pumps always operate in the high-efficiency range, avoiding energy waste, and achieving load balancing control when multiple pumps are running in parallel, reducing energy loss caused by ineffective start-stop operations.

Description

Technical Field

[0001] This invention relates to the field of mine automation control technology, and in particular to a dynamic optimization control method and system for mine water supply and drainage based on digital twins. Background Technology

[0002] In mining production enterprises, the water supply and drainage system is an important part of the entire mining and beneficiation production process. In addition to doing a good job in mining production and tailings dam seepage recovery, it is also necessary to ensure the supply of fresh water during grinding and beneficiation production and acid water treatment.

[0003] Existing mine water supply and drainage control technologies generally adopt a quantitative water supply and fixed-frequency operation control mode, lacking the ability to dynamically adjust according to actual water demand. Operators usually rely on experience to set the number of pumps in operation and valve opening. In order to ensure the water demand of end users, a conservative strategy is often adopted to keep the pumps in a high-flow state for a long time. When the production water consumption decreases, this fixed operation mode causes the pumps to deviate from the high-efficiency operating range, and the power utilization efficiency drops significantly. In the scenario of multiple pumps operating in parallel, there is a lack of coordinated control mechanism between the pumps. Frequent start-stop not only causes additional power loss, but also impacts the power grid, causing the electricity consumption per ton of water in the mine water supply and drainage system to remain at a high level for a long time. Summary of the Invention

[0004] To overcome the problem that existing mine water supply and drainage control technologies generally adopt quantitative water supply and fixed-frequency operation control modes, which lack the ability to dynamically adjust according to actual water demand.

[0005] The technical solution of this invention is: a dynamic optimization control method for mine water supply and drainage based on digital twins, comprising the following steps: S1: Acquire status data of pump units and electric valves in multiple pumping stations distributed in the mining area, as well as flow data, pressure data, liquid level data and water quality data of pipeline monitoring points; status data includes start-stop status, operating frequency, current, voltage, power factor, vibration signal and temperature signal; S2: Construct a digital twin model containing the topology of the water supply and drainage network, 3D models of equipment, and topographic information, and drive the operation of the digital twin model based on the real-time data obtained in S1 to perform synchronous mapping between physical pump stations and virtual models. S3: Obtain the planned production water volume for a future preset period from the mine production scheduling system as the demand-side input; the planned production water volume includes the predicted water volume and water quality requirements for each water-using node; S4: Preset multi-dimensional constraints, including water quality index constraints for each water source, efficient operating range constraints for pump sets, motor start-up times constraints, peak and valley electricity price time constraints, and upper and lower limits constraints for pipeline pressure; S5: With the dual optimization objectives of minimizing total system energy consumption and balancing pump wear, a multi-objective optimization function is constructed. Under the premise of satisfying the constraints in S4, an intelligent optimization algorithm is used to solve for the optimal pump combination scheme, valve opening scheme and variable frequency operation parameters. S6: Based on the optimal solution obtained from S5, control commands are issued to the actuators of each pumping station to perform dynamic optimization control of the water supply and drainage system.

[0006] As a preferred embodiment, the pump set wear equalization target in S5 is achieved through the following method: The running time and number of starts of each water pump are accumulated, and then normalized to obtain the normalized value of the running time. and the normalized value of the number of starts By introducing a first weighting coefficient α and a second weighting coefficient β, a wear leveling index function is constructed. = Σ(α· + β· ), where α+β=1; The addition of a penalty term to the multi-objective optimization function makes the optimization algorithm tend to select a pump combination scheme with a balanced distribution of runtime and number of starts during the solution process.

[0007] As a preferred embodiment, a dynamic optimization control method for mine water supply and drainage based on digital twins also includes equipment health assessment and proactive maintenance, with specific steps including: The system collects real-time vibration and temperature signals from the water pump unit, extracts features, and inputs them into a pre-trained fault diagnosis model to output a health score. When the health score is lower than a preset first threshold, an early warning message is generated. When it is predicted that the equipment will reach a second threshold within a preset time period in the future, the system automatically performs load transfer and generates a maintenance work order.

[0008] As a preferred option, the training and application of the fault diagnosis model include: Wavelet packet decomposition or empirical mode decomposition methods are used to extract time-frequency domain features from vibration signals. The extracted features include energy, peak value, peak-to-peak value, kurtosis index, and waveform index for each frequency band. A fault classification model is constructed using a deep belief network or convolutional neural network, and supervised training is performed using historical fault data. The input of the trained model is a feature vector, and the output is the equipment health score and fault mode recognition result.

[0009] Preferably, the water quality constraints in S4 include: For production water nodes with different water quality requirements, allowable ranges for pH value, turbidity, and suspended solids concentration are set. During the optimization process, priority is given to scheduling seepage recovery pump stations or pit water accumulation pump stations that meet the water quality requirements of the water nodes. New water supply pump stations are only started when the amount of recovered water is insufficient or the water quality does not meet the standards, so as to carry out cascade utilization of multiple water sources.

[0010] As a preferred embodiment, the digital twin model constructed in S2 includes hydraulic gradient line analysis functionality, specifically: Based on the pipeline network topology and real-time pressure data, the pressure distribution of each node in the pipeline network is calculated through hydraulic modeling, and the most unfavorable point and its pressure demand are identified. The pressure at the most unfavorable point is used as one of the core constraints to ensure that the water supply system can meet the end-user water demand during the optimization scheduling process.

[0011] As a preferred approach, the intelligent optimization algorithm in S5 adopts the improved non-dominated sorting genetic algorithm NSGA-II or the multi-objective particle swarm optimization algorithm MOPSO. During the solution process, a heuristic strategy is used to initialize the population, prioritizing the selection of pump combinations currently in the high-efficiency operating range as initial individuals. During the iteration process, an elite retention strategy is introduced to ensure the diversity of the Pareto front. Finally, the Pareto optimal solution set is output for operators to choose from or the solution with the highest overall satisfaction can be automatically selected according to preset rules.

[0012] As a preferred embodiment, a dynamic optimization control method for mine water supply and drainage based on digital twins also includes a network interruption resumption mechanism at the edge computing layer. A time-series database is built into the edge computing gateway deployed at each pumping station. When the network is normal, data is uploaded to the digital twin platform layer in real time. When the network is interrupted, the collected data is stored in a local cache and the data timestamp is recorded. When the network is restored, the cached data is automatically retransmitted to the digital twin platform layer.

[0013] Preferably, load transfer includes: When the health score of a water pump falls below the warning threshold, the system automatically adjusts the optimization solution process of S5, reducing the call priority of that water pump and increasing the call priority of other standby or healthy water pumps. When load transfer is required, the system first starts the alternative water pump and confirms that it is running normally before stopping the water pump to be maintained, thus achieving a seamless switch and avoiding impact on the pipeline pressure.

[0014] A dynamic optimization control system for mine water supply and drainage based on digital twins, comprising: Physical pump station layer: This includes multiple pump rooms distributed throughout the mining area. Each pump room is equipped with a water pump unit, electric valves, a frequency converter control cabinet, and integrates a multi-dimensional sensing terminal. The multi-dimensional sensing terminal includes vibration sensors, temperature sensors, flow meters, pressure transmitters, level gauges, pH meters, and electrical parameter acquisition modules, which are used to collect vibration signals, temperature signals, pipeline flow, pipeline pressure, water tank level, water quality pH value, and motor current, voltage, and power factor parameters, respectively. Edge control layer: The edge computing gateways deployed at each pump station site communicate with the PLC control cabinet of the physical pump station layer via industrial Ethernet or industrial fieldbus. They are responsible for the preprocessing, cleaning, timestamp alignment and local caching of real-time data, and execute control commands from the digital twin platform layer. The edge computing gateways have a built-in mechanism for resuming data transmission after network interruption. They automatically store data when the network is interrupted and automatically retransmit it after the network is restored. Digital twin platform layer: Deployed on the server cluster of the mine control center, it includes a virtual mine water supply and drainage network constructed by BIM+GIS 3D model, as well as an embedded dynamic hydraulic balance model, equipment health assessment model and multi-objective optimization engine; the digital twin platform layer interacts bidirectionally with the edge control layer through the industrial ring network; Remote monitoring layer: including the large screen display system and operator station in the central control center, providing a human-machine interface to display the overall system status, equipment operating parameters, early warning information and optimization scheduling schemes in a three-dimensional visualization form.

[0015] The beneficial effects of this invention are: 1. This solution obtains the planned water consumption for future periods from the production scheduling system as the demand-side input. It presets multi-dimensional constraints, including the high-efficiency operating range of pump sets, the number of motor starts, and peak and off-peak electricity price periods. With the dual optimization objectives of minimizing total system energy consumption and balancing pump wear, it uses an intelligent optimization algorithm to solve for the optimal pump combination scheme and operating parameters. This allows the water supply system to dynamically adjust its operation mode according to changes in actual water demand, ensuring that the pumps always operate in the high-efficiency range, avoiding energy waste. When multiple pumps are running in parallel, it achieves load balancing control, reduces energy loss caused by ineffective start-stop, and effectively extends the overall service life of the equipment units. 2. By setting water quality index constraints, this solution first assesses whether the water quality parameters of each water source meet the requirements of different water use nodes during the optimization process. For seepage and recycled water and pit water that meet the standards for pH value, turbidity and other indicators, they are prioritized for use in production links with relatively relaxed water quality requirements. Only when the amount of recycled water is insufficient or the water quality does not meet the standards will the new water supply pump station be activated. This enables the originally independent multi-water source system to form a coordinated linkage, the high-quality water resources are rationally allocated, the utilization rate of recycled water is significantly improved, the amount of new water taken out is reduced, and the water resource procurement cost is reduced. 3. This invention constructs a complete system architecture consisting of a physical pumping station layer, an edge control layer, a digital twin platform layer, and a remote monitoring layer. By deploying multi-dimensional sensing terminals at each pumping station, it collects pump operating parameters and pipeline hydraulic data in real time. After data preprocessing using an edge computing gateway, the data is uploaded to the digital twin platform. In the digital twin platform, a virtual model containing the topology of the water supply and drainage pipeline network is constructed based on BIM and GIS technologies, realizing synchronous mapping between physical pumping stations and the virtual model. Dispatch and management personnel can monitor the operating status of the entire mine's water supply and drainage system in real time through a three-dimensional visualization interface at the central control center. Any equipment abnormalities or pipeline pressure fluctuations can be detected and dealt with immediately, significantly improving the system's safe operation level. Attached Figure Description

[0016] Figure 1 The diagram shown is a schematic of a dynamic optimization control process for mine water supply and drainage based on digital twins according to the present invention. Figure 2 The diagram illustrates the process of equipment health assessment, proactive maintenance, and load transfer in a dynamic optimization control method for mine water supply and drainage based on digital twins, as described in this invention. Figure 3 The diagram shown is a schematic of the multi-source cascade utilization and water quality constraint optimization process of a dynamic optimization control method for mine water supply and drainage based on digital twins according to the present invention. Detailed Implementation

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Please see Figure 1-3 This invention provides an embodiment: a dynamic optimization control method for mine water supply and drainage based on digital twins, comprising the following steps: S1: Acquire status data of pump units and electric valves in multiple pumping stations distributed in the mining area, as well as flow data, pressure data, liquid level data and water quality data of pipeline monitoring points; status data includes start-stop status, operating frequency, current, voltage, power factor, vibration signal and temperature signal; S2: Construct a digital twin model containing the topology of the water supply and drainage network, 3D models of equipment, and topographic information, and drive the operation of the digital twin model based on the real-time data obtained in S1 to perform synchronous mapping between physical pump stations and virtual models. S3: Obtain the planned production water volume for a future preset period from the mine production scheduling system as the demand-side input; the planned production water volume includes the predicted water volume and water quality requirements for each water-using node; S4: Preset multi-dimensional constraints, including water quality index constraints for each water source, efficient operating range constraints for pump sets, motor start-up times constraints, peak and valley electricity price time constraints, and upper and lower limits constraints for pipeline pressure; S5: With the dual optimization objectives of minimizing total system energy consumption and balancing pump wear, a multi-objective optimization function is constructed. Under the premise of satisfying the constraints in S4, an intelligent optimization algorithm is used to solve for the optimal pump combination scheme, valve opening scheme and variable frequency operation parameters. S6: Based on the optimal solution obtained from S5, control commands are issued to the actuators of each pumping station to perform dynamic optimization control of the water supply and drainage system.

[0019] As a preferred embodiment, the pump set wear equalization target in S5 is achieved through the following method: The running time and number of starts of each water pump are accumulated, and then normalized to obtain the normalized value of the running time. and the normalized value of the number of starts By introducing a first weighting coefficient α and a second weighting coefficient β, a wear leveling index function is constructed. = Σ(α· + β· ), where α+β=1; The addition of a penalty term to the multi-objective optimization function makes the optimization algorithm tend to select a pump combination scheme with a balanced distribution of runtime and number of starts during the solution process.

[0020] As a preferred embodiment, a dynamic optimization control method for mine water supply and drainage based on digital twins also includes equipment health assessment and proactive maintenance, with specific steps including: The system collects real-time vibration and temperature signals from the water pump unit, extracts features, and inputs them into a pre-trained fault diagnosis model to output a health score. When the health score is lower than a preset first threshold, an early warning message is generated. When it is predicted that the equipment will reach a second threshold within a preset time period in the future, the system automatically performs load transfer and generates a maintenance work order.

[0021] As a preferred option, the training and application of the fault diagnosis model include: Wavelet packet decomposition or empirical mode decomposition methods are used to extract time-frequency domain features from vibration signals. The extracted features include energy, peak value, peak-to-peak value, kurtosis index, and waveform index for each frequency band. A fault classification model is constructed using a deep belief network or convolutional neural network, and supervised training is performed using historical fault data. The input of the trained model is a feature vector, and the output is the equipment health score and fault mode recognition result.

[0022] Preferably, the water quality constraints in S4 include: For production water nodes with different water quality requirements, allowable ranges for pH value, turbidity, and suspended solids concentration are set. During the optimization process, priority is given to scheduling seepage recovery pump stations or pit water accumulation pump stations that meet the water quality requirements of the water nodes. New water supply pump stations are only started when the amount of recovered water is insufficient or the water quality does not meet the standards, so as to carry out cascade utilization of multiple water sources.

[0023] As a preferred embodiment, the digital twin model constructed in S2 includes hydraulic gradient line analysis functionality, specifically: Based on the pipeline network topology and real-time pressure data, the pressure distribution of each node in the pipeline network is calculated through hydraulic modeling, and the most unfavorable point and its pressure demand are identified. The pressure at the most unfavorable point is used as one of the core constraints to ensure that the water supply system can meet the end-user water demand during the optimization scheduling process.

[0024] As a preferred approach, the intelligent optimization algorithm in S5 adopts the improved non-dominated sorting genetic algorithm NSGA-II or the multi-objective particle swarm optimization algorithm MOPSO. During the solution process, a heuristic strategy is used to initialize the population, prioritizing the selection of pump combinations currently in the high-efficiency operating range as initial individuals. During the iteration process, an elite retention strategy is introduced to ensure the diversity of the Pareto front. Finally, the Pareto optimal solution set is output for operators to choose from or the solution with the highest overall satisfaction can be automatically selected according to preset rules.

[0025] As a preferred embodiment, a dynamic optimization control method for mine water supply and drainage based on digital twins also includes a network interruption resumption mechanism at the edge computing layer. A time-series database is built into the edge computing gateway deployed at each pumping station. When the network is normal, data is uploaded to the digital twin platform layer in real time. When the network is interrupted, the collected data is stored in a local cache and the data timestamp is recorded. When the network is restored, the cached data is automatically retransmitted to the digital twin platform layer.

[0026] Preferably, load transfer includes: When the health score of a water pump falls below the warning threshold, the system automatically adjusts the optimization solution process of S5, reducing the call priority of that water pump and increasing the call priority of other standby or healthy water pumps. When load transfer is required, the system first starts the alternative water pump and confirms that it is running normally before stopping the water pump to be maintained, thus achieving a seamless switch and avoiding impact on the pipeline pressure.

[0027] A dynamic optimization control system for mine water supply and drainage based on digital twins, comprising: Physical pump station layer: This includes multiple pump rooms distributed throughout the mining area. Each pump room is equipped with a water pump unit, electric valves, a frequency converter control cabinet, and integrates a multi-dimensional sensing terminal. The multi-dimensional sensing terminal includes vibration sensors, temperature sensors, flow meters, pressure transmitters, level gauges, pH meters, and electrical parameter acquisition modules, which are used to collect vibration signals, temperature signals, pipeline flow, pipeline pressure, water tank level, water quality pH value, and motor current, voltage, and power factor parameters, respectively. Edge control layer: The edge computing gateways deployed at each pump station site communicate with the PLC control cabinet of the physical pump station layer via industrial Ethernet or industrial fieldbus. They are responsible for the preprocessing, cleaning, timestamp alignment and local caching of real-time data, and execute control commands from the digital twin platform layer. The edge computing gateways have a built-in mechanism for resuming data transmission after network interruption. They automatically store data when the network is interrupted and automatically retransmit it after the network is restored. Digital twin platform layer: Deployed on the server cluster of the mine control center, it includes a virtual mine water supply and drainage network constructed by BIM+GIS 3D model, as well as an embedded dynamic hydraulic balance model, equipment health assessment model and multi-objective optimization engine; the digital twin platform layer interacts bidirectionally with the edge control layer through the industrial ring network; Remote monitoring layer: including the large screen display system and operator station in the central control center, providing a human-machine interface to display the overall system status, equipment operating parameters, early warning information and optimization scheduling schemes in a three-dimensional visualization form.

[0028] Example 1 In a large mining enterprise, the water supply and drainage system undertakes multiple tasks, including mining drainage water recovery, mineral processing water supply, tailings dam seepage recovery, and fresh water replenishment. The numerous pump stations are scattered throughout the mining area. Before the upgrade, most pump stations relied on manual on-site operation, with dispatchers obtaining information via telephone, making it impossible to monitor equipment operating status and pipeline hydraulic parameters in real time, leading to delayed emergency response. Furthermore, to ensure end-user water demand, operators often adopted a conservative quantitative water supply model, causing pumps to deviate from their efficient operating range for extended periods, resulting in wasted energy. Equipment maintenance relied on periodic repairs, lacking online monitoring of equipment health, leading to frequent unplanned shutdowns. Multiple water sources operated independently, resulting in low utilization rates of recycled water and significant waste of high-quality water resources. To address the aforementioned issues, the mine introduced the dynamic optimization control system for water supply and drainage based on digital twins proposed in this invention. First, the pump stations were automated at the physical pump station level. Each pump room was equipped with a PLC control cabinet, and manual valves were replaced with electric regulating valves. Vibration sensors, temperature sensors, flow meters, pressure transmitters, level gauges, pH meters, and electrical parameter acquisition modules were integrated to achieve comprehensive perception of the operating status of the pump units and the hydraulic parameters of the pipeline network. All sensing terminals were connected to the PLC through an industrial fieldbus to complete real-time data acquisition and local control. At the edge control layer, each pump station is equipped with an edge computing gateway, which establishes a communication connection with the PLC via industrial Ethernet. The edge gateway is responsible for preprocessing, cleaning and timestamp alignment of the raw data, and has a built-in time-series database for local caching. When the network is normal, the data is uploaded to the digital twin platform layer in real time; when the network is interrupted, the data is automatically stored in the local cache and re-uploaded after the network is restored, ensuring the continuity and integrity of the data. The digital twin platform layer is deployed on the server cluster of the mine's centralized control center. Based on BIM and GIS technologies, it constructs a virtual mine water supply and drainage digital twin model that includes the topology of the water supply and drainage network, 3D models of equipment, and topographic information. This model interacts bidirectionally with the edge control layer through the industrial ring network, receives field data in real time to drive the model's operation, and realizes the synchronous mapping between physical pump stations and the virtual model. At the same time, the model has an embedded hydraulic gradient line analysis function, which calculates the pressure distribution of each node based on the network topology and real-time pressure data, identifies the most unfavorable point in the network and its pressure demand, and uses it as one of the core constraints. The remote monitoring layer consists of a large screen display system in the central control center and an operator station. It displays the overall status of the mine's water supply and drainage system, equipment operating parameters, early warning information, and optimized scheduling schemes in a three-dimensional visualization form. The scheduling and management personnel can monitor the system operation in real time through the human-machine interface and make manual interventions when necessary. In daily operation, the system obtains the planned water consumption for the next hour from the mine production scheduling system, including the predicted water consumption and water quality requirements of each water node. Based on this demand-side input, the system presets multi-dimensional constraints: water quality index constraints of each water source, efficient operating range constraints of pump groups, motor start-up times constraints, peak and valley electricity price constraints, and upper and lower limits of pipeline pressure constraints. With the dual optimization objectives of minimizing total system energy consumption and balancing pump wear, an improved non-dominated sorting genetic algorithm is used to solve the problem. During the solution process, a heuristic strategy is adopted for population initialization, prioritizing the selection of pump combinations currently in the efficient operating range as initial individuals. An elite retention strategy is introduced in the iteration to ensure the diversity of the Pareto front. Finally, the optimal pump combination scheme, valve opening scheme, and variable frequency operation parameters are output and distributed to the actuators of each pump station through the edge gateway to realize dynamic optimization control of the water supply and drainage system. In terms of equipment health management, the system continuously collects vibration and temperature signals from the pump units, extracts time-frequency domain features using wavelet packet decomposition, inputs them into a pre-trained deep belief network fault diagnosis model, and outputs equipment health scores and fault mode recognition results. When the health score of a pump is lower than the preset warning threshold, the system automatically generates a warning message and highlights it on the monitoring interface. When it is predicted that the equipment will reach the fault threshold in the future, the system automatically adjusts and optimizes the equipment call priority in the solution process, reducing the call weight of the pump to be maintained and increasing the priority of other pumps with good health status. When performing load transfer, the system first starts the replacement pump and confirms that it is running normally before stopping the pump to be maintained, achieving seamless switching and avoiding impact on the pipeline pressure. In terms of multi-source coordinated scheduling, the system prioritizes the scheduling of seepage recovery pumping stations and pit water accumulation pumping stations based on water quality index constraints. When the amount of recovered water is sufficient and the water quality meets the requirements of the water use node, the recovered water is used first. Only when the amount of recovered water is insufficient or the water quality does not meet the standards will the new water supply pumping station be activated, thereby realizing the cascade utilization of multiple water sources and improving the comprehensive utilization rate of water resources.

[0029] 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 mine water supply and drainage dynamic optimization control method based on digital twinning; characterized in that: It includes the following steps: S1: Acquire status data of pump units and electric valves in multiple pumping stations distributed in the mining area, as well as flow data, pressure data, liquid level data and water quality data of pipeline monitoring points; status data includes start-stop status, operating frequency, current, voltage, power factor, vibration signal and temperature signal; S2: Construct a digital twin model containing the topology of the water supply and drainage network, 3D models of equipment, and topographic information, and drive the operation of the digital twin model based on the real-time data obtained in S1 to perform synchronous mapping between physical pump stations and virtual models. S3: Obtain the planned production water volume for a future preset period from the mine production scheduling system as the demand-side input; the planned production water volume includes the predicted water volume and water quality requirements for each water-using node; S4: Preset multi-dimensional constraints, including water quality index constraints for each water source, efficient operating range constraints for pump sets, motor start-up times constraints, peak and valley electricity price time constraints, and upper and lower limits constraints for pipeline pressure; S5: With the dual optimization objectives of minimizing total system energy consumption and balancing pump wear, a multi-objective optimization function is constructed. Under the premise of satisfying the constraints in S4, an intelligent optimization algorithm is used to solve for the optimal pump combination scheme, valve opening scheme and variable frequency operation parameters. S6: Based on the optimal solution obtained from S5, control commands are issued to the actuators of each pumping station to perform dynamic optimization control of the water supply and drainage system.

2. The mine water supply and drainage dynamic optimization control method based on digital twinning according to claim 1, characterized in that: The pump set wear equalization target in S5 is achieved through the following methods: The running time and number of starts of each water pump are accumulated, and then normalized to obtain the normalized value of the running time. and the normalized value of the number of starts By introducing a first weighting coefficient α and a second weighting coefficient β, a wear leveling index function is constructed. ;Will The addition of a penalty term to the multi-objective optimization function makes the optimization algorithm tend to select a pump combination scheme with a balanced distribution of runtime and number of starts during the solution process.

3. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 1, characterized in that: A dynamic optimization control method for mine water supply and drainage based on digital twins also includes equipment health assessment and proactive maintenance. Specific steps include: The system collects real-time vibration and temperature signals from the water pump unit, extracts features, and inputs them into a pre-trained fault diagnosis model to output a health score. When the health score is lower than a preset first threshold, an early warning message is generated. When it is predicted that the equipment will reach a second threshold within a preset time period in the future, the system automatically performs load transfer and generates a maintenance work order.

4. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 3, characterized in that: Training and application of fault diagnosis models, including: Wavelet packet decomposition or empirical mode decomposition methods are used to extract time-frequency domain features from vibration signals. The extracted features include energy, peak value, peak-to-peak value, kurtosis index, and waveform index for each frequency band. A fault classification model is constructed using a deep belief network or convolutional neural network, and supervised training is performed using historical fault data. The input of the trained model is a feature vector, and the output is the equipment health score and fault mode recognition result.

5. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 1, characterized in that: The water quality constraints in S4 include: For production water nodes with different water quality requirements, allowable ranges for pH value, turbidity, and suspended solids concentration are set. During the optimization process, priority is given to scheduling seepage recovery pump stations or pit water accumulation pump stations that meet the water quality requirements of the water nodes. New water supply pump stations are only started when the amount of recovered water is insufficient or the water quality does not meet the standards, so as to carry out cascade utilization of multiple water sources.

6. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 1, characterized in that: The digital twin model built in S2 includes hydraulic gradient line analysis capabilities, specifically: Based on the pipeline network topology and real-time pressure data, the pressure distribution of each node in the pipeline network is calculated through hydraulic modeling, and the most unfavorable point and its pressure demand are identified. The pressure at the most unfavorable point is used as one of the core constraints to ensure that the water supply system can meet the end-user water demand during the optimization scheduling process.

7. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 1, characterized in that: The intelligent optimization algorithm in S5 adopts the improved non-dominated sorting genetic algorithm NSGA-II or the multi-objective particle swarm optimization algorithm MOPSO. During the solution process, a heuristic strategy is used to initialize the population, giving priority to selecting the pump combination that is currently in the high-efficiency operating range as the initial individual. During the iteration process, an elite retention strategy is introduced to ensure the diversity of the Pareto frontier; finally, the Pareto optimal solution set is output for operators to choose from or the solution with the highest overall satisfaction is automatically selected according to preset rules.

8. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 1, characterized in that: A dynamic optimization control method for mine water supply and drainage based on digital twins also includes a network interruption resume mechanism in the edge computing layer. A time-series database is built into the edge computing gateway deployed in each pumping station. When the network is normal, the data is uploaded to the digital twin platform layer in real time. When the network is interrupted, the collected data is stored in the local cache and the data timestamp is recorded. Once the network is restored, the cached data will be automatically retransmitted to the digital twin platform layer.

9. The dynamic optimization control method for mine water supply and drainage based on digital twin as described in claim 3, characterized in that: Load shifting includes: When the health score of a water pump falls below the warning threshold, the system automatically adjusts the optimization solution process of S5, reducing the call priority of that water pump and increasing the call priority of other standby or healthy water pumps. When load transfer is required, the system first starts the alternative water pump and confirms that it is running normally before stopping the water pump to be maintained, thus achieving a seamless switch and avoiding impact on the pipeline pressure.

10. A dynamic optimization control method for mine water supply and drainage based on digital twins according to any one of claims 1-9, characterized in that: A dynamic optimization control system for mine water supply and drainage based on digital twins, comprising: Physical pump station layer: This includes multiple pump rooms distributed throughout the mining area. Each pump room is equipped with a water pump unit, electric valves, a frequency converter control cabinet, and integrates a multi-dimensional sensing terminal. The multi-dimensional sensing terminal includes vibration sensors, temperature sensors, flow meters, pressure transmitters, level gauges, pH meters, and electrical parameter acquisition modules, which are used to collect vibration signals, temperature signals, pipeline flow, pipeline pressure, water tank level, water quality pH value, and motor current, voltage, and power factor parameters, respectively. Edge control layer: The edge computing gateways deployed at each pump station site communicate with the PLC control cabinet of the physical pump station layer via industrial Ethernet or industrial fieldbus. They are responsible for the preprocessing, cleaning, timestamp alignment and local caching of real-time data, and execute control commands from the digital twin platform layer. The edge computing gateways have a built-in mechanism for resuming data transmission after network interruption. They automatically store data when the network is interrupted and automatically retransmit it after the network is restored. Digital twin platform layer: Deployed on the server cluster of the mine control center, it includes a virtual mine water supply and drainage network constructed by BIM+GIS 3D model, as well as an embedded dynamic hydraulic balance model, equipment health assessment model and multi-objective optimization engine; the digital twin platform layer interacts bidirectionally with the edge control layer through the industrial ring network; Remote monitoring layer: including the large screen display system and operator station in the central control center, providing a human-machine interface to display the overall system status, equipment operating parameters, early warning information and optimization scheduling schemes in a three-dimensional visualization form.

Images