Mining flame-proof and intrinsic safety type lithium ion storage battery power supply monitoring management system

Abstract

The invention, which relates to the technical field of the mining emergency power supply equipment, discloses a mining explosion-proof and intrinsic safety type lithium ion storage battery power supply monitoring management system comprising a battery management unit, a thermal management unit, a load interface unit and a data acquisition unit used for acquiring system operation parameters. The system further comprises a central collaborative management unit, a system-level digital twin model and a multi-level verification module. According to the mining flame-proof and intrinsic safety type lithium ion storage battery power supply monitoring and management system, by introducing the central collaborative management unit and the system-level digital twinborn model, fusion perception and dynamic deduction of the multi-dimensional state of a power supply system are achieved, a collaborative control strategy can be generated based on multi-target optimization, and the power supply monitoring and management system is high in practicability. And the absolute safety and reliability of the strategy are ensured through a multi-stage verification mechanism, and the overall energy efficiency and decision robustness of the system in a complex mine environment are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of emergency power supply equipment technology for mines, specifically to a monitoring and management system for explosion-proof and intrinsically safe lithium-ion battery power supplies for mines. Background Technology

[0002] In the technological development of monitoring and management systems for intrinsically safe and explosion-proof lithium-ion batteries used in mining, existing solutions can achieve basic charge and discharge control, status monitoring, and fault protection functions. However, these functional modules typically operate independently, lacking deep integration and system-level collaboration. The actual operating environment in mines is complex and variable, with equipment often facing fluctuating load demands and stringent energy constraints. Current energy management strategies are mostly static or based on simple threshold responses, making it difficult to cope with dynamic operating conditions. For example, the lack of effective linkage between battery thermal management, power distribution, and load demand leads to problems such as low energy utilization, localized overheating, or unnecessary auxiliary energy consumption in scenarios like sudden high-power requests or prolonged light-load operation. Information isolation between subsystems prevents energy efficiency optimization from a global perspective, limiting further improvements in the overall performance and safety of the power system. Therefore, the urgent technical challenge is how to achieve dynamic collaboration and global energy efficiency optimization across multiple subsystems under complex and variable operating conditions. Summary of the Invention

[0003] The purpose of this invention is to provide a monitoring and management system for the power supply of explosion-proof and intrinsically safe lithium-ion batteries used in mining, so as to solve the problems mentioned in the background art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system, comprising: the battery management unit, thermal management unit, load interface unit, and data acquisition unit for collecting system operating parameters, and further comprising a central collaborative management unit, a system-level digital twin model, and a multi-level verification module; The central collaborative management unit is communicatively connected to the battery management unit, thermal management unit, load interface unit and data acquisition unit, respectively, and is used to receive and integrate data from the above units; The system-level digital twin model is integrated within the central collaborative management unit and is used to model and dynamically simulate the overall operating status of the power system. The central collaborative management unit is configured to: perform optimization calculations based on the deduction results of the system-level digital twin model, with the goal of improving the overall energy efficiency of the system, and generate an initial collaborative control strategy; The multi-level verification module interacts with the central collaborative management unit to sequentially perform security compliance verification and consistency verification on the initial collaborative control strategy. The central collaborative management unit is also configured to decompose the final collaborative control strategy verified by the multi-level verification module into specific control commands and send them to the battery management unit, thermal management unit and load interface unit respectively.

[0005] Preferably, the multi-level verification module includes a first-level safety constraint filter; the first-level safety constraint filter has a built-in quantified intrinsic safety and explosion-proof rule database, which is used to compare and screen the parameters in the initial cooperative control strategy with the rigid safety boundaries in the database.

[0006] Preferably, the multi-level verification module further includes a second-level cross-validation unit; the second-level cross-validation unit is configured to: use a fast verification model that is different from the optimization calculation algorithm principle in the central collaborative management unit to verify the key output parameters in the strategy after it has passed the screening of the first-level security constraint filter.

[0007] Preferably, the central collaborative management unit is further configured to: trigger the system-level digital twin model and optimization calculation process to perform iterative calculations when the deviation between the verification result of the second-level cross-validation unit and the corresponding parameter in the initial collaborative control strategy exceeds a preset tolerance range.

[0008] Preferably, the system-level digital twin model includes at least a battery electrochemical state deduction sub-model, a system heat conduction deduction sub-model, and a load power demand prediction sub-model.

[0009] Preferably, the system further includes a closed-loop learning module; the closed-loop learning module is connected to the central collaborative management unit and is used to receive the effect feedback data after the actual execution of the strategy, and feed back the comparison error between this data and the predicted data to the system-level digital twin model for adjusting the internal parameters of the model.

[0010] Preferably, the operating parameters collected by the data acquisition unit include the total battery pack voltage, total battery pack current, individual battery cell voltage, temperature sensor data, instantaneous load power, and historical load operating data.

[0011] Preferably, the control command sent to the battery management unit is a charge / discharge power limit command, the control command sent to the thermal management unit is a cooling fan speed adjustment command, and the control command sent to the load interface unit is an output power priority management command.

[0012] Preferably, the closed-loop learning module adaptively adjusts the parameters of the system-level digital twin model.

[0013] Preferably, each hardware unit of the system is encapsulated in an explosion-proof enclosure, and all electrical connections and circuit designs meet the intrinsic safety standards for mining applications.

[0014] This invention provides a power monitoring and management system for explosion-proof and intrinsically safe lithium-ion batteries used in mining. It has the following beneficial effects: This mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system, by introducing a central collaborative management unit and a system-level digital twin model, achieves fusion perception and dynamic inference of the power system's multi-dimensional state. It can generate collaborative control strategies based on multi-objective optimization and ensure the absolute safety and reliability of the strategies through a multi-level verification mechanism, effectively improving the system's overall energy efficiency and decision robustness in complex mining environments.

[0015] This mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system achieves adaptive correction of digital twin model parameters through a closed-loop learning module. This enables the system to continuously track changes in equipment status and optimize control accuracy. At the same time, its hardware design strictly adheres to the dual safety standards of explosion-proof and intrinsically safe operation, which fundamentally ensures safe operation in explosive environments, extends equipment lifespan, and reduces maintenance requirements. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the module interaction of the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system of the present invention. Figure 2 This is the control logic timing diagram of the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 and Figure 2 The present invention provides a technical solution: a monitoring and management system for explosion-proof and intrinsically safe lithium-ion batteries used in mining, including: a battery management unit, a thermal management unit, a load interface unit, and a data acquisition unit for collecting system operating parameters, and also includes a central collaborative management unit, a system-level digital twin model, and a multi-level verification module; The central coordination management unit is communicatively connected to the battery management unit, thermal management unit, load interface unit, and data acquisition unit, respectively, and is used to receive and integrate data from the above units. A system-level digital twin model, integrated within the central collaborative management unit, is used to model and dynamically simulate the overall operating status of the power system. The central collaborative management unit is configured to perform optimization calculations based on the deduction results of the system-level digital twin model, with the goal of improving the overall energy efficiency of the system, and generate an initial collaborative control strategy. The multi-level verification module interacts with the central collaborative management unit to sequentially perform security compliance verification and consistency verification on the initial collaborative control strategy; The central collaborative management unit is also configured to decompose the final collaborative control strategy verified by the multi-level verification module into specific control commands and send them to the battery management unit, thermal management unit and load interface unit respectively.

[0019] It should be further explained that, in the specific implementation process, after the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system is started, its core central collaborative management unit immediately enters the working state. The system first continuously acquires the total voltage and current of the battery pack, the voltage of each individual cell, readings from multiple temperature sensors, real-time power of the load, and recorded historical load operating data through data acquisition units distributed throughout the key nodes of the power system. This multi-source heterogeneous data is transmitted in real time to the central collaborative management unit, where data fusion and preprocessing are performed, providing a consistent and reliable data foundation for subsequent in-depth analysis.

[0020] Subsequently, the central collaborative management unit invokes its internally integrated system-level digital twin model. This model is a multi-physics, multi-scale virtual mirror system that runs in parallel with three sub-models: a battery electrochemical state simulation sub-model, a system heat conduction simulation sub-model, and a load power demand prediction sub-model. The electrochemical state simulation sub-model estimates the progress and state change trends of the battery's internal chemical reactions based on current voltage, current, and historical data. The heat conduction simulation sub-model, considering ambient temperature, component heat generation power, and heat dissipation conditions, predicts the temperature field distribution and changes of the system over a future period. The load power demand prediction sub-model, by analyzing historical operating patterns and current operating conditions, makes forward-looking predictions of short-term load power demand. Based on these simulation and prediction results, the central collaborative management unit initiates its optimization decision-making logic, performing optimization calculations using a multi-objective function that integrates overall system energy efficiency, safe operating boundaries, and long-term equipment lifespan, thereby generating a preliminary collaborative control strategy. This strategy may include suggestions for adjusting battery charging and discharging power, setting the workload of the heat dissipation system, and re-planning the power supply priority for different loads.

[0021] Following this, the initial strategy is not implemented directly but must be rigorously verified through a multi-level validation module. The validation process begins with a first-level safety constraint filter, which pre-stores all rigid safety boundary parameters of the quantified intrinsically safe and explosion-proof rules for mining applications. The filter compares each parameter in the strategy, such as the proposed maximum discharge current or the highest permissible surface temperature, against the safety thresholds in the database. If any parameter exceeds the safety boundary, the strategy is immediately rejected, and the central collaborative management unit is required to recalculate and optimize it.

[0022] Only strategies that pass the aforementioned security screening will proceed to the second-level cross-validation unit. This unit employs a rapid validation model that is fundamentally different from the optimization algorithm within the central collaborative management unit, such as a rule-based expert system. It independently calculates and verifies key outputs of the strategy, such as predicted total energy consumption or the highest temperature point. If the deviation between the rapid validation model's calculation results and the initial strategy's predicted values ​​is within a preset reasonable tolerance range, the strategy is deemed to have good consistency and high reliability, and is approved for execution. If the deviation exceeds the tolerance range, it indicates potential uncertainty in the prediction. In this case, the deviation data will be fed back to the central collaborative management unit, triggering it to use new input information to drive iterative calculations between the digital twin model and the optimization algorithm until a new strategy that both passes the security screening and meets the consistency requirements is generated.

[0023] Ultimately, the fully validated collaborative control strategy is broken down into specific execution instructions by the central collaborative management unit: adjusted charge / discharge power limit instructions are sent to the battery management unit, new cooling fan speed control instructions are sent to the thermal management unit, and updated output power priority management instructions are sent to the load interface unit. Simultaneously, the closed-loop learning module begins operation, continuously monitoring the actual execution effect of the strategy, such as actual temperature rise and energy consumption. This measured data is compared with previous prediction data, and the resulting errors are used to adaptively and progressively correct the internal parameters of the digital twin model. This allows the model's prediction accuracy to continuously evolve and improve as the system operates, forming a virtuous cycle of increasing intelligence with use. All hardware units of the entire system are enclosed in an explosion-proof enclosure, and all internal circuit designs and electrical connections strictly adhere to intrinsically safe mining standards, ensuring absolute safety during operation in potentially explosive environments.

[0024] The multi-level verification module includes a first-level safety constraint filter. This filter contains a quantified intrinsically safe and explosion-proof rule database, used to compare and screen parameters in the initial collaborative control strategy against rigid safety boundaries in the database. It should be further noted that, in practical implementation, the first-level safety constraint filter in the multi-level verification module of the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system bears the rigid responsibility of ensuring the absolute safety of all control decisions. The core of this filter is a pre-built, quantified database of intrinsically safe and explosion-proof rules. This database is not a simple list of parameters, but rather transforms complex mining safety regulations and product explosion-proof standards into a series of explicit mathematical boundary conditions that can be directly compared by a computer. These boundary conditions cover multiple dimensions, including electrical, thermal, and time parameters, such as: the maximum permissible instantaneous voltage and continuous current of the battery pack or any circuit branch; the maximum permissible operating temperature and short-time withstand temperature of any battery cell or component surface; the maximum energy that can be safely released under a specific resistance; and the minimum permissible switching time interval at different power levels.

[0025] Once the central collaborative management unit generates the initial collaborative control strategy, all proposed operational parameters contained within this strategy are extracted and input into the first-level safety constraint filter for automated screening. This process is a rigorous, rule-based logical judgment process. The filter compares each key parameter proposed in the strategy, such as the planned peak discharge current, the upper limit of charging voltage, the cooling fan start threshold, or the power value allocated to a certain load, with the corresponding rigid safety boundary in the database. This comparison is not a rough estimate, but a precise numerical comparison.

[0026] Its judgment logic is clear and direct: if all proposed parameters are completely within their respective security boundaries, the screening result is "passed," and the strategy is allowed to proceed to the next level of verification unit. Conversely, if any parameter even slightly exceeds the limit value specified in the database, no matter how superior its optimization objective, the filter will immediately trigger a "rejection" instruction and return the strategy to the central collaborative management unit, along with information identifying the specific non-compliant parameter. Upon receiving the rejection information, the central collaborative management unit will adjust the constraints of its optimization algorithm based on this feedback, recalculate to generate a new strategy that conforms to all security boundaries, thereby ensuring from the source that any operation that might endanger explosion-proof safety is absolutely impossible to execute.

[0027] The multi-level verification module also includes a second-level cross-validation unit. This second-level cross-validation unit is configured to use a fast verification model, different in principle from the optimization algorithm within the central collaborative management unit, to verify the key output parameters of the strategy after passing the first-level safety constraint filter. It should be further explained that, in specific implementation, the core function of the second-level cross-validation unit in the multi-level verification module is to perform an independent technical credibility assessment of the initial strategy that has passed the safety screening. This unit uses a fast verification model that is fundamentally different from the main optimization algorithm within the central collaborative management unit. For example, if the main optimization algorithm uses model predictive control based on numerical iteration, the cross-validation unit can use an expert system based on historical data and a rule base. When the initial collaborative control strategy enters this unit, the cross-validation unit extracts the key output parameters from the strategy, such as the predicted total power consumption, the predicted maximum temperature of the battery pack, or the predicted maximum current of the main circuit within a future operating cycle. Subsequently, this unit uses its own fast verification model, based on the same input conditions—the current system state and load prediction—to perform an independent and rapid calculation of these key parameters.

[0028] After the calculation is completed, the system quickly compares the calculation results of the verification model with the corresponding predicted values ​​in the initial strategy. This comparison process makes logical judgments based on a preset tolerance range: if the difference between each key parameter is within its corresponding tolerance range, the strategy is deemed to have passed the consistency verification, its reliability is confirmed, and it is allowed to proceed to the subsequent execution stage. If the difference of any key parameter exceeds its set tolerance range, a verification failure signal is immediately triggered. This signal, along with specific deviation information, is fed back to the central collaborative management unit, indicating that its current optimization strategy has uncertainties and that an iterative optimization process needs to be initiated. The central collaborative management unit then uses this deviation information as new constraints or input conditions to drive the system-level digital twin model and optimization algorithm to recalculate and adjust, aiming to generate a new strategy that meets the requirements in terms of both security and consistency. Thus, through this cross-comparison mechanism of heterogeneous algorithms, the robustness and credibility of the decision results are greatly improved.

[0029] The central collaborative management unit is further configured to trigger iterative calculations in the system-level digital twin model and optimization calculation process when the deviation between the verification results of the second-level cross-validation unit and the corresponding parameters in the initial collaborative control strategy exceeds a preset tolerance range. It should be further explained that, in specific implementation, when the deviation between the verification results of the second-level cross-validation unit and the corresponding key parameters in the initial collaborative control strategy is confirmed to exceed the preset tolerance range, the central collaborative management unit immediately initiates a rigorous iterative optimization process. This process first analyzes and classifies the received deviation information, identifying which output parameters, such as the predicted maximum temperature or total power consumption, have become uncertain, and determining the direction and magnitude of the deviation. These specific deviation data are then transformed into new, quantified constraints or optimization target weight adjustment factors and input into the system-level digital twin model and dynamic optimization decision module.

[0030] Upon receiving this feedback, the digital twin model fine-tunes the parameters of one or more of its sub-models, such as the heat conduction simulation sub-model or the electrochemical state simulation sub-model. This includes adjusting the heat capacity coefficient or internal resistance estimate to better reflect the actual operating characteristics of the current system. Subsequently, the optimization decision module recalculates based on the updated digital twin model within the existing multi-objective optimization framework. This recalculation not only considers the initial energy efficiency, safety, and lifetime targets but also prioritizes reducing identified prediction biases as a key optimization objective.

[0031] This process may not be completed in one step, but rather forms an iterative sequence containing multiple calculation-verification loops. In each iteration, the newly generated strategy is submitted again to the multi-level verification module, especially the second-level cross-validation unit, for a new round of consistency verification. The iteration process only terminates when the differences between all key output parameters of the new strategy and the calculation results of the fast verification model fall back to within the preset tolerance range, and all requirements of the first-level safety constraint filter are met. Only then is the strategy finally confirmed as an executable solution. Through this feedback-based closed-loop iterative mechanism, the system can proactively correct errors in its decision-making model, gradually approaching an optimal control strategy that is both safe and reliable, as well as accurate and efficient. This effectively addresses the challenges posed by the slow drift of system characteristics or sudden operating conditions in the complex environment of a mine.

[0032] The system-level digital twin model includes at least a battery electrochemical state simulation sub-model, a system heat conduction simulation sub-model, and a load power demand prediction sub-model. It should be further explained that, in the specific implementation process, the system-level digital twin model, as the core decision support for the central collaborative management unit, is organically integrated from these three highly collaborative professional sub-models. The battery electrochemical state simulation sub-model continuously receives data on the total voltage and current of the battery pack and the voltage of each individual cell from the data acquisition unit. Through an embedded battery state estimation algorithm, it dynamically simulates the progress of internal chemical reactions within the battery, comprehensively assesses the changing trends of its state of charge and health, and estimates minute fluctuations in key internal parameters such as equivalent internal resistance, providing an electrochemical basis for overall power allocation.

[0033] The system heat conduction simulation sub-model runs concurrently. It integrates real-time readings from multiple temperature sensors, current heat generation calculated by the electrochemical model, and known performance parameters of the heat dissipation system. Dynamic thermal simulation is performed within a simplified thermal network model to predict temperature changes in key areas of the battery compartment, such as battery module surfaces and power device heat sinks, over a future period. This allows for proactive assessment of thermal risks and evaluation of the potential effectiveness of different heat dissipation strategies. The load power demand prediction sub-model focuses on analyzing real-time power and historical operating data reported by the load interface unit. By identifying patterns and regularities in load start-up and shutdown, and power fluctuations, it provides short-term forward-looking predictions of total power demand and potential peak power within a future decision-making cycle. These three sub-models do not operate in isolation but rather run in parallel under the scheduling of the central collaborative management unit, frequently exchanging data.

[0034] For example, the current internal resistance value derived by the electrochemical model directly affects the heat generation calculation in the thermal model, while the temperature predicted by the thermal model, in turn, affects the correction of battery performance parameters in the electrochemical model. Simultaneously, the future power demand output by the load prediction model is a crucial boundary condition necessary for state extrapolation between the electrochemical and thermal models. This deep inter-model coupling and data exchange enables the digital twin model to form a virtual mirror that highly approximates the actual operating state of the physical system, laying a solid and reliable model foundation for the subsequent generation of collaborative control strategies that satisfy multi-objective optimization requirements and closely match actual operating conditions.

[0035] The system also includes a closed-loop learning module. This module connects to the central collaborative management unit and receives feedback data on the actual effects of the strategy execution. It then feeds back the comparison error between this data and the predicted data to the system-level digital twin model to adjust the model's internal parameters. It's important to note that in practice, the closed-loop learning module is a crucial component for the system's continuous self-optimization and performance improvement. This module begins operation after the central collaborative management unit issues a control command and executes a complete operating cycle. It collects feedback data on the actual effects of the strategy execution again through the data acquisition unit. This data includes, but is not limited to: the actual average temperature and hot spot temperature of the battery pack, the actual operating power of the cooling fan, the actual energy consumption of the load, and the response curves of the battery voltage and current. Subsequently, the module accurately compares this collected measured data with the corresponding predicted data made by the system-level digital twin model during optimization decisions. For example, it compares the actual measured maximum temperature with the model's predicted maximum temperature, and the actual total power consumption with the predicted total power consumption, thereby calculating the prediction error for each key parameter.

[0036] The calculated error data are not simply recorded, but systematically input into a parameter adjustment algorithm. This algorithm analyzes the pattern and magnitude of the error, and generates corrections for specific parameters within the digital twin model accordingly. For example, if the predicted temperature of the heat conduction simulation sub-model is consistently higher than the actual measured value, the algorithm may determine that the estimated heat capacity coefficient or convective heat transfer coefficient in the model is too high, and make corresponding fine adjustments to reduce that value; conversely, if the electrochemical state simulation sub-model deviates from its prediction of the battery terminal voltage, the algorithm may calibrate the parameters characterizing internal resistance or polarization voltage in the model.

[0037] This parameter adjustment is a continuous, adaptive process. New error data generated after each strategy execution is used to perform a new round of gradual corrections to the model. This allows the digital twin model to progressively compensate for errors arising from its own simplification assumptions and slowly track characteristic drift caused by system aging, environmental changes, or the gradual degradation of component performance. Through this cyclical "prediction-execution-comparison-learning-correction" closed loop, the inference accuracy of the digital twin model is continuously optimized over time and with data accumulation. This leads to increasingly precise and realistic collaborative control decisions made by the entire monitoring and management system, ultimately resulting in a steady improvement in the overall system energy efficiency and reliability.

[0038] The data acquisition unit collects operating parameters including total battery pack voltage, total battery pack current, individual battery cell voltage, temperature sensor data, instantaneous load power, and historical load operating data. It should be further noted that, in practice, the data acquisition unit provides a comprehensive and multi-dimensional time-series data foundation for the entire monitoring and management system. This unit continuously collects and uploads key parameters necessary for system operation through a high-precision, intrinsically safe sensor network. These parameters primarily include the total battery pack voltage and current, reflecting the overall power supply status; the voltage of each individual battery cell, used for detailed assessment of battery consistency and health; readings from multiple temperature sensors distributed in key thermal management areas within the battery compartment; instantaneous load power, used for real-time sensing of energy consumption demands; and historical load operating data records, used for analyzing load behavior patterns and making trend predictions. The key thermal management areas include the areas between battery modules, the heat dissipation substrates of power devices, and air inlets and outlets.

[0039] After undergoing necessary signal conditioning and analog-to-digital conversion, this data is transmitted to the central collaborative management unit via isolated intrinsically safe communication lines. Upon receiving this raw data, the central collaborative management unit first performs a series of preprocessing operations, including but not limited to data validity verification, such as removing outliers that significantly exceed the physical range, data filtering and smoothing to suppress random interference, and synchronizing and aligning the timestamps of multi-source data. The clean and spatiotemporally consistent data after processing is distributed to the corresponding sub-models and optimization decision-making modules of the system-level digital twin model, serving as the fundamental data input source for their state extrapolation, trend prediction, and multi-objective optimization calculations. The smooth and accurate flow of the entire data is a prerequisite and cornerstone for the reliable implementation of all subsequent advanced intelligent decision-making functions.

[0040] The control commands sent to the battery management unit are charge / discharge power limit commands, the control commands sent to the thermal management unit are cooling fan speed adjustment commands, and the control commands sent to the load interface unit are output power priority management commands. It should be further explained that, in the specific implementation process, the central collaborative management unit decomposes the finally verified collaborative control strategy into three specific execution commands and sends them to the corresponding execution units. The charge / discharge power limit command sent to the battery management unit explicitly sets the maximum allowable charging current and maximum discharging current values ​​for the battery pack in the current and subsequent time periods. The battery management unit strictly limits the inflow and outflow of energy according to this command, ensuring that the battery operates within its electrochemical and thermal safety boundaries. The cooling fan speed adjustment command sent to the thermal management unit specifies the specific speed setting or duty cycle that the cooling fan should operate at based on the predicted heat load and optimization objectives. After receiving the command, the thermal management unit drives the fan to operate at the corresponding speed to achieve the best balance between heat dissipation efficiency and its own energy consumption.

[0041] The load interface unit receives output power priority management commands. These commands define the power supply priority and possible power reduction ratios for different load circuits when the system's total power is limited or needs adjustment. Based on these commands, the load interface unit can, when necessary, reduce power to low-priority loads in an orderly manner or temporarily interrupt power supply, thereby ensuring uninterrupted power supply to critical loads. The issuance and execution of all these commands are under continuous system monitoring. Any abnormal status reported by any execution unit will trigger a re-decision and command adjustment by the central coordination management unit, thus forming a dynamic, closed-loop control circuit.

[0042] The closed-loop learning module adaptively adjusts the parameters of the system-level digital twin model. It's important to further clarify that, in practice, the parameter adjustment process of the closed-loop learning module is an autonomous and continuously optimizing adaptive process. This module does not use fixed, preset correction coefficients, but rather dynamically adjusts the intensity and direction of its learning and correction based on the correlation between historical error data and model output characteristics. Its core lies in a built-in adaptive algorithm that continuously analyzes a series of predicted and measured error data. It evaluates the statistical characteristics of these errors, such as their stability, fluctuation direction, and trends over time.

[0043] Based on this analysis, the algorithm intelligently determines whether the current error is primarily caused by initial parameter deviations in the model, accidental environmental disturbances, or long-term, slow changes in system characteristics. For different types of error causes, the algorithm adopts corresponding adjustment strategies. For example, for persistent errors in a consistent direction, the algorithm generates a relatively stable parameter correction amount with an appropriate magnitude to gradually eliminate systematic model bias. For errors with large fluctuations or no clear trend, the algorithm adopts a more conservative correction strategy, potentially making only minor adjustments or temporarily keeping the parameters unchanged to avoid excessively disturbing the model and introducing new instabilities.

[0044] This adjustment is gradual; each parameter update is a fine-tuning based on the previous model state, ensuring that the digital twin model's performance evolves smoothly and stably towards higher accuracy, avoiding drastic performance fluctuations. Through this adaptive learning mechanism, the closed-loop learning module enables the digital twin model not only to compensate for the uncertainties of the initial modeling but also to continuously track the dynamic characteristics changes caused by component aging, changes in environmental conditions, or adjustments to operating strategies, ultimately achieving a long-term, reliable match between the model's predictive capabilities and the actual system's operating state.

[0045] All hardware units of the system are enclosed in an explosion-proof enclosure, and all electrical connections and circuit designs meet intrinsically safe standards for mining applications. It should be further noted that, in practical implementation, all hardware units of the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system, including the central coordination management unit, battery management unit, data acquisition unit, thermal management unit, and load interface unit, are integrated and installed within a specially designed explosion-proof enclosure. This enclosure is made of high-strength mechanical materials, and its structural design ensures that in the event of a potential battery failure leading to an explosion, it can withstand the internal explosion pressure with sufficient structural strength, and effectively cool and extinguish the explosion flame through a specific explosion-proof joint surface structure, preventing its propagation to the external hazardous environment. All electrical connections and circuit designs within the system strictly adhere to intrinsically safe standards for mining applications, achieved through multi-layered safety measures.

[0046] Intrinsically safe protection circuits are installed at the power input and interfaces connecting to external sensors and loads. These circuits typically consist of fast-response current-limiting and voltage-limiting components, as well as isolation barriers, ensuring that under any single fault condition—whether a short circuit, open circuit, or component failure—the energy transmitted to hazardous locations is strictly limited to a safe level that cannot ignite explosive gases. Internal wiring is neat and uses high-temperature resistant and flame-retardant materials, and critical circuit boards undergo three-proof treatment to improve reliability in harsh environments. The entire system, from physical structure to electrical design, constitutes a dual safety guarantee system. Its explosion-proof characteristics address potential ignition sources within the equipment, while the intrinsically safe design fundamentally limits the energy that may be generated. Together, they ensure the absolute safety and reliability of this power monitoring and management system operating in explosive gas environments underground in mines.

[0047] It should be further explained that, in the specific implementation process, the operation of the mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system begins with the comprehensive acquisition of multi-source data. The system continuously acquires the total voltage and current values ​​of the battery pack through a distributed sensor network, monitors the voltage of each individual battery cell to assess consistency, collects temperature readings from multiple key temperature measurement points to construct the system's thermal state, and simultaneously records the real-time power consumption of the load and stores its historical operating data. After necessary filtering and synchronization processing, this data is transmitted to the system's core decision-making unit, namely the central collaborative management unit.

[0048] The core of the central collaborative management unit is the construction and operation of a system-level digital twin model. This model is a multi-physics coupled virtual system containing three sub-models that operate in parallel and exchange data with each other. The battery electrochemical state simulation sub-model simulates the internal chemical reaction process of the battery based on input voltage and current data, estimating the changing trends of its state of charge and health. The system heat conduction simulation sub-model integrates the current temperature distribution, estimated heat generation power, and heat dissipation conditions to predict temperature changes in key components over a future period. The load power demand prediction sub-model analyzes historical and real-time load data to predict the short-term power demand profile. Based on these simulation results, the central collaborative management unit executes a multi-objective optimization calculation process. Its goal is to optimize the overall system energy efficiency while strictly meeting safe operating boundaries and considering equipment lifespan degradation. This generates a preliminary collaborative control strategy, which includes adjustment suggestions for charging and discharging power, heat dissipation intensity, and load priority.

[0049] To ensure absolute safety, this initial strategy must undergo rigorous verification by multiple levels of validation modules. The first-level safety constraint filter compares each parameter in the strategy, such as the proposed current value or temperature threshold, against a database pre-stored with quantified safety boundaries for all intrinsically safe and explosion-proof rules in mining. Any parameter exceeding the rigid safety boundaries will result in the immediate rejection of the strategy and a recalculation. Strategies that pass the safety screening then proceed to the second-level cross-validation unit. This unit employs a rapid validation model, distinct from the main optimization algorithm, to independently calculate the key outputs of the strategy. If the deviation between the rapid validation result and the predicted value of the initial strategy exceeds a preset tolerance range, it indicates uncertainty in the prediction. The deviation data is fed back to the central collaborative management unit to trigger iterative calculations between the digital twin model and the optimization algorithm until a new strategy that is both safe and consistent is generated.

[0050] The final approved strategy is broken down into specific instructions: sending new charge and discharge power limit instructions to the battery management unit to ensure the battery operates within a safe range; sending instructions to the thermal management unit to adjust the cooling fan speed for precise temperature control; and sending output power priority management instructions to the load interface unit to ensure power supply to critical loads when necessary. After the strategy is executed, its actual effect is monitored and compared with the predicted values. The generated error data is fed into a closed-loop learning module, where an adaptive algorithm analyzes error patterns and accordingly performs progressive and targeted fine-tuning of the relevant internal parameters of the digital twin model. This allows the model's predictive ability to continuously track the slow changes in system characteristics, thereby achieving continuous self-optimization of system performance.

[0051] All hardware in the entire system is enclosed in a high-strength explosion-proof enclosure. All internal circuit designs strictly adhere to intrinsic safety standards for mining applications. Through meticulous protection circuit design, it is ensured that under any single fault condition, the energy transmitted to the hazardous environment is limited to an absolutely safe level, thus forming a dual safety guarantee of explosion-proof and intrinsic safety.

[0052] The implementation process of the monitoring and management system for explosion-proof and intrinsically safe lithium-ion batteries used in mining is as follows, and the key models and algorithms include the following: The system operates based on a system-level digital twin model for multi-objective collaborative optimization. This digital twin model consists of three coupled sub-models: a. Battery Electrochemical State Deduction Sub-model: An equivalent circuit model is used to describe the dynamic characteristics of the battery, and its terminal voltage prediction formula is: U terminal =OCV(SOC)+I×R internal +U polarization Among them, U terminal The battery terminal voltage is the predicted value; OCV(SOC) is the open-circuit voltage under the current state of charge, obtained by looking up a table from a pre-measured OCV-SOC curve; I is the operating current, positive for charging and negative for discharging; R internal U is the ohmic internal resistance of the battery. polarization The polarization voltage is used to characterize the polarization phenomenon.

[0053] The state of charge (SOC) is estimated using a combination of ampere-hour integral method and model calibration: Where SOC(t) is the state of charge at time t, SOC(t0) is the initial state of charge, and Q... nominal Where η is the battery's rated capacity, K is the coulombic efficiency, and U is the correction factor. measured This is the actual measured terminal voltage.

[0054] b. System heat conduction derivation sub-model: Temperature is predicted using a lumped-parameter thermal model or a simplified finite-difference thermal network model. A simplified lumped-parameter model formula is: C th dT / dt=I 2 R internal +Q other -hA(TT ambient ); where C th Let Q be the system's heat capacity, T be the predicted temperature, t be time, and Q be the temperature. other The heat dissipation power of components other than the battery is given by h, the heat dissipation coefficient is given by A, and the effective heat dissipation area is given by T. ambient The ambient temperature.

[0055] c. Load power demand forecasting sub-model: Short-term forecasting is performed using a sliding window weighted average or autoregressive method based on historical data, with the formula: P load_predict (t+1)=α1P load (t)+α2P load (t-1)+...+α n P load (t-n+1); where P load_predict (t+1) represents the predicted load power at the next time step, P load (t) represents the actual load power at the current moment, α1, α2, ..., α3. n These are the weighting coefficients that are fitted or adaptively adjusted based on historical data.

[0056] The multi-objective optimization process of the central collaborative management unit is expressed by the objective function: Minimize: λ 1· Energy_Cost+λ 2· Aging_Cost+λ 3· Thermal Risk; The constraints are as follows: ; Among them, Energy_Cost represents the energy efficiency target, such as total loss; Aging_Cost represents the lifetime decay cost, which is related to SOC, current, and temperature; Thermal_Risk represents thermal risk; λ1, λ2, and λ3 are weighting coefficients that are dynamically adjusted according to system preferences. The optimization solution process must meet constraints such as charging and discharging current, SOC, temperature, and power supply and demand balance.

[0057] The operation of the multi-level verification module involves explicit logical judgments, including the following: First-level security constraint filter: Execution logic judgment is as follows: IF(Proposed_Current>I_safe_max)ORProposed_Temperature>T_safe_max)...THENPolicy_Rejected.

[0058] Second-level cross-validation unit: Execution logic judgment is as follows: IF|Value_primary-Value_cross_validation|>ToleranceTHENPolicy_Rejected_With_Feedback.

[0059] The parameters of the closed-loop learning module are adjusted using gradient descent or similar optimization algorithms to minimize the prediction error, as shown in the following formula: θ new=θ old -γ θ Loss(M θ (X), Y); where θ represents the parameter vector to be adjusted in the digital twin model, such as internal resistance, thermal resistance, prediction model coefficients, etc.; γ is the learning rate, θ Loss is the gradient of the loss function with respect to the parameters, M. θ (X) is the model's predicted output for input X under parameter θ, and Y is the actual observed value.

[0060] All the algorithms and models described above are implemented on a hardware platform that meets the requirements for explosion-proof and intrinsically safe operation in mining. Intrinsic safety protection is achieved through a designed current-limiting and voltage-limiting protection circuit, ensuring that the output energy under any single fault condition is below the minimum ignition energy threshold. The explosion-proof enclosure is designed and manufactured in accordance with relevant standards, capable of withstanding possible internal explosions and preventing flame propagation.

[0061] By introducing a central collaborative management unit and a system-level digital twin model, the system achieves fusion perception and dynamic simulation of the multi-dimensional state of the power system. It can generate collaborative control strategies based on multi-objective optimization and ensure the absolute security and reliability of the strategies through a multi-level verification mechanism, effectively improving the overall energy efficiency and decision robustness of the system in complex mining environments.

[0062] The closed-loop learning module enables adaptive correction of the parameters of the digital twin model, allowing the system to continuously track changes in equipment status and optimize control accuracy. Meanwhile, its hardware design strictly adheres to the dual safety standards of explosion-proof and intrinsically safe operation, which fundamentally ensures safe operation in explosive environments, extends equipment lifespan, and reduces maintenance requirements.

[0063] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0064] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A monitoring and management system for explosion-proof and intrinsically safe lithium-ion batteries used in mining, including: The battery management unit, thermal management unit, load interface unit, and data acquisition unit for collecting system operating parameters are characterized in that they also include a central collaborative management unit, a system-level digital twin model, and a multi-level verification module. The central collaborative management unit is communicatively connected to the battery management unit, thermal management unit, load interface unit and data acquisition unit, respectively, and is used to receive and integrate data from the above units; The system-level digital twin model is integrated within the central collaborative management unit and is used to model and dynamically simulate the overall operating status of the power system. The central collaborative management unit is configured to: perform optimization calculations based on the deduction results of the system-level digital twin model, with the goal of improving the overall energy efficiency of the system, and generate an initial collaborative control strategy; The multi-level verification module interacts with the central collaborative management unit to sequentially perform security compliance verification and consistency verification on the initial collaborative control strategy. The central collaborative management unit is also configured to decompose the final collaborative control strategy verified by the multi-level verification module into specific control commands and send them to the battery management unit, thermal management unit and load interface unit respectively.

2. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: The multi-level verification module includes a first-level safety constraint filter; the first-level safety constraint filter has a built-in quantified intrinsic safety and explosion-proof rule database, which is used to compare and screen the parameters in the initial cooperative control strategy with the rigid safety boundaries in the database.

3. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 2, characterized in that: The multi-level verification module also includes a second-level cross-verification unit; the second-level cross-verification unit is configured to use a fast verification model that is different from the optimization calculation algorithm in the central collaborative management unit to verify the key output parameters in the strategy after it has passed the screening of the first-level security constraint filter.

4. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 3, characterized in that: The central collaborative management unit is further configured to: when the deviation between the verification result of the second-level cross-validation unit and the corresponding parameter in the initial collaborative control strategy exceeds the preset tolerance range, trigger the system-level digital twin model and optimization calculation process to perform iterative calculations.

5. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: The system-level digital twin model includes at least a battery electrochemical state deduction sub-model, a system heat conduction deduction sub-model, and a load power demand prediction sub-model.

6. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: The system also includes a closed-loop learning module; the closed-loop learning module is connected to the central collaborative management unit and is used to receive the effect feedback data after the actual execution of the strategy, and feed back the comparison error between this data and the predicted data to the system-level digital twin model for adjusting the internal parameters of the model.

7. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: The data acquisition unit collects operating parameters including total battery pack voltage, total battery pack current, individual battery cell voltage, temperature sensor data, instantaneous load power, and historical load operating data.

8. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: The control command sent to the battery management unit is a charge / discharge power limit command, the control command sent to the thermal management unit is a cooling fan speed adjustment command, and the control command sent to the load interface unit is an output power priority management command.

9. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 6, characterized in that: The closed-loop learning module adaptively adjusts the parameters of the system-level digital twin model.

10. The mining explosion-proof and intrinsically safe lithium-ion battery power monitoring and management system according to claim 1, characterized in that: Each hardware unit of the system is encapsulated in an explosion-proof enclosure, and all electrical connections and circuit designs meet the intrinsic safety standards for mining applications.

Images