Integrated thermal management system for explosion-proof charger

Abstract

The application discloses an integrated thermal management system for an explosion-proof charging machine, and relates to the technical field of charging machines. The integrated thermal management system comprises the following steps: step S100 three-dimensional thermal field modeling: obtaining a heat source point cloud by using a laser radar and a temperature sensor array, and constructing a three-dimensional thermal management model in combination with finite element analysis. The three-dimensional thermal field modeling is used to optimize the layout of a liquid cooling pipe (a heat flow density uniformity error is less than or equal to 10%), an integrated design of a ZL114A aluminum alloy shell and a micro-channel liquid cooling plate is combined (a thermal resistance is reduced by 40%), a metal sealing structure with an IP68 protection level and an ExdIICT6 explosion-proof certification are matched, and in a wide temperature range of-20 DEG C to 60 DEG C and a gas dust explosion environment, the safety heat dissipation of a power device with a temperature less than or equal to 85 DEG C is realized, and the liquid cooling medium is intelligently managed by using an electric conductivity sensor and an ion exchange resin filtering device, so that equipment failure caused by pollution is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of charging machines, in particular to an integrated thermal management system for an explosion-proof charging machine. BACKGROUND

[0002] Under the background of rapid development of industrial automation and new energy field, as a key equipment, the performance of the thermal management system of the explosion-proof charging machine directly affects the safety, reliability and service life of the equipment. Especially in the explosive environment of gas, dust and the like, the traditional thermal management system faces technical bottlenecks such as insufficient heat dissipation efficiency and uniformity (fixed pipe layout leads to heat flow density non-uniformity error exceeding 20%, split structure high thermal resistance), weak multi-source data fusion and real-time monitoring capability (poor parameter acquisition coordination, large infrared thermal imaging deviation, low sensor calibration accuracy), lagging abnormal detection and early warning mechanism (depending on preset threshold, early warning delay exceeding 1 minute, lack of intelligent optimization of heat dissipation strategy), insufficient explosion-proof safety and environmental adaptability (low thermal conductivity of shell material, low sealing structure protection level, extensive medium management) and low system integration and intelligence level (module independent operation, wide temperature working condition processing delay exceeding 500ms, poor load adaptive ability), etc. Therefore, an integrated, high-precision and high-reliability thermal management system is needed to solve the above problems. SUMMARY

[0003] The present application relates to the technical field of charging machines, in particular to an integrated thermal management system for an explosion-proof charging machine.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical scheme: an integrated thermal management system for an explosion-proof charging machine, comprising the following steps:

[0005] Step S100 three-dimensional thermal field modeling: obtaining heat source point cloud by using laser radar and temperature sensor array, combining finite element analysis to construct three-dimensional thermal management model; setting heat flow density monitoring point at the connection between power module and liquid cooling pipe, optimizing pipe layout to make heat flow density uniformity error ≤10%;

[0006] Step S200 multi-source data fusion: collecting flow, pressure and temperature field data of liquid cooling circuit, after reconstruction by multi-view stereo vision algorithm, registering through iterative closest point algorithm; denoising and calibrating infrared thermal imaging data, combining ray tracing algorithm to correct deviation;

[0007] Step S300 abnormal behavior detection: based on target detection algorithm and binocular camera calibration parameters, constructing three-dimensional thermal trajectory by triangulation, combining reference feature space established by K-means clustering to identify heat dissipation abnormalities;

[0008] Step S400: Dynamic sensor calibration: Arrange sensors on the liquid cooling radiator and pipelines, perform spatiotemporal calibration via the PTP protocol, and use the KD tree algorithm to generate a real-time three-dimensional thermal management model containing multi-source data;

[0009] Step S500: abnormal fluctuation judgment: use LSTM-GARCH model to predict liquid cooling parameters, and issue early warning when the prediction confidence is greater than 90%, and judge abnormality based on preset thresholds;

[0010] Step S600 multi-modal warning: when triggered, the liquid cooling system is linked to adjust the flow rate, and the optimal heat dissipation strategy is generated through the model predictive control algorithm.

[0011] Preferably, in step S100, the infrared heat flux sensor serves as the core component for heat flux density monitoring. It adopts the thermopile principle based on the Seebeck effect to convert thermal radiation into an electrical signal output. The sensor has a built-in temperature compensation circuit, which can automatically correct the influence of ambient temperature changes on the measurement results, ensuring that the accuracy is stable at ≤±3% within the range of 0-500W / m². During the data acquisition process, the sensor obtains the heat flux density data at the connection between the charger power module and the liquid cooling pipeline in real time at a frequency of 10Hz. Through finite element analysis software, a multi-physics field coupling model including fluid flow, heat transfer and structural mechanics is constructed. The grid adaptive technology is used to dynamically encrypt the grid in the key area, and the pipeline layout is iteratively optimized. During the optimization process, a genetic algorithm is introduced for global search. By adjusting the pipeline bending angle, pipe diameter and distribution position, the heat flux density distribution uniformity error is stably controlled at ≤10%, thereby improving the heat dissipation efficiency.

[0012] Preferably, in the infrared thermal imaging data processing in step S200, a 5×5 median filtering algorithm is used to suppress salt and pepper noise. The principle is to replace the grayscale value of each pixel in the image with the median of the grayscale values ​​of 5×5 pixels in the neighborhood of the point, thereby removing noise interference while retaining edge information. The blackbody radiation calibration link uses a high-precision blackbody furnace as a reference source. The temperature control accuracy of the blackbody furnace is ±0.1°C. By comparing the infrared thermal imager measurement value with the blackbody furnace standard value, a temperature error correction curve is established to make the calibration error ≤±0.5°C. The ray tracing algorithm is based on the Monte Carlo method to simulate the propagation path of thermal radiation in a complex pipeline structure. By calculating the reflection, absorption and transmission of radiation energy on the pipeline surface, the temperature field point cloud deviation caused by pipeline obstruction is corrected, so that the reconstructed temperature field is closer to the actual working condition.

[0013] Preferably, in the S300 step, the historical data collection covers the running conditions of the explosion-proof charger under different loads, environmental temperatures and heat dissipation conditions, ensuring that the data can reflect ≥90% of the actual working scenarios. Through preprocessing of the collected data, key characteristic parameters such as flow, temperature and power are extracted, the data dimension is reduced using principal component analysis method, more than 95% of the information entropy is retained, the data is clustered into different working condition modes based on K-means clustering algorithm, and the reference characteristic space containing flow-temperature mapping curve is constructed. At the same time, combined with the thermal characteristics of power devices, the safety temperature zone is set to ≤85℃, and the threshold values of flow mutation >1L / min・s and temperature gradient >4℃ / cm in the abnormal triggering rule are determined through a large number of experimental data combined with the thermal safety standard of the charger, which can timely and accurately detect the heat dissipation abnormality.

[0014] Preferably, the sensor calibration in the S400 step is crucial. The flow sensor is calibrated by weighing method. Specifically, the liquid cooling medium is collected within a certain time, the medium mass is measured by high-precision electronic scale, the actual flow is calculated combined with the medium density, and the error is ≤±0.5%FS compared with the sensor measurement value for calibration. The temperature sensor is calibrated at a calibration point of 37℃ using a high-precision constant temperature tank. By comparing the sensor output value with the standard platinum resistance thermometer measurement value, the sensor compensation parameters are adjusted to ensure that the error is ≤±0.2℃. The three-dimensional laser tracker measures the coordinate position of the sensor in space by emitting a laser beam, with a measurement accuracy of ±2mm, ensuring the accuracy of the spatial layout of the sensor and providing a basis for constructing an accurate real-time three-dimensional thermal management model.

[0015] Preferably, in the S500 step, the LSTM-GARCH model captures the time series characteristics of the liquid cooling parameters through the long short-term memory network, and processes the volatility and heteroscedasticity of the data through the generalized autoregressive conditional heteroscedasticity model, realizing accurate prediction of the liquid cooling parameters. When the predicted value exceeds the normal range and the confidence is >90%, the system triggers an early warning ≥8min in advance, providing sufficient processing time for the operation and maintenance personnel. After the multi-modal early warning is triggered, the flow regulating device of the liquid cooling system uses a fast-response electric regulating valve with a response time <50ms, which cooperates with a high-performance water pump to complete flow regulation within 200ms. By optimizing the pipeline resistance and water pump characteristic curve, the temperature drop rate of the power device is ≥1.5℃ / s, avoiding equipment damage caused by overheating.

[0016] Preferably, the model predictive control algorithm in step S600 takes maximizing heat dissipation efficiency and minimizing energy consumption as the goal, and establishes an optimization problem including a dynamic model of the liquid cooling system, constraints and objective functions. In the objective function J=min(ΔT / ΔQ), ΔT is the temperature change of the power device, and ΔQ is the flow change of the liquid cooling medium. The heat dissipation efficiency is improved by minimizing the function. At the same time, the pump power ≤150W is used as a constraint to ensure that the system energy consumption is within a reasonable range. The MPC algorithm adopts a rolling time domain optimization strategy. In each control cycle, the optimization problem is solved online according to the current system state and the predicted future working conditions, and the optimal flow regulation strategy is generated to achieve a balance between heat dissipation performance and energy consumption.

[0017] Preferably, the explosion-proof housing is integrally formed of ZL114A aluminum alloy. The aluminum alloy undergoes a T6 heat treatment process to significantly improve its tensile strength and thermal conductivity. The microchannel liquid cooling plate embedded in the inner wall is manufactured using precision machining technology. The channel width is ≤0.5mm and the depth is ≤2mm, which greatly increases the heat dissipation area and improves the heat dissipation efficiency. The microchannel is connected to the external circuit through a precision welding process to form an integrated heat dissipation structure. The metal bellows sealing joint at the pipe penetration point has a protection level of IP68, which can prevent dust and liquid intrusion. The temperature range is -40℃~125℃, which is suitable for harsh environments. With the fluororubber O-ring, the compression volume is controlled at 15%-20% to ensure good sealing performance, so that the heat conduction thermal resistance is ≤0.1℃・m² / W. The explosion-proof housing has passed the ExdIICT6 explosion-proof certification and meets the use requirements of explosive hazardous environments.

[0018] Preferably, the non-flammable heat-conducting medium circulation system in the liquid cooling medium management system adopts a high-purity synthetic heat-conducting fluid with low viscosity, high boiling point and good chemical stability. The conductivity sensor adopts a four-electrode design and uses AC excitation technology to measure the conductivity of the medium with an accuracy of ≤±1%. When the conductivity is greater than 50μS / cm, the ion exchange resin filter device is automatically started. The device adopts a multi-stage series structure to remove ionic impurities in the medium and restore the purity of the medium. The liquid level sensor adopts the magnetostrictive principle with an accuracy of ≤±1mm. It monitors the liquid level of the liquid cooling medium in real time. When the liquid level drops, the rehydration device is automatically triggered to achieve medium loss compensation and ensure the stable operation of the liquid cooling system.

[0019] Preferably, the integrated thermal management system is specially designed for explosive environments with gas concentration ≤1.5% and dust concentration ≤5mg / m³. All electrical components are explosion-proof and certified. The liquid cooling system operates under a wide temperature range of -20℃~60℃. It uses thermal conductive media with excellent low-temperature performance and low-temperature resistant sealing materials to ensure that the system works normally in a low-temperature environment. The edge server is equipped with a high-performance processor and an optimized anomaly detection algorithm with a processing delay of ≤100ms, realizing real-time monitoring and rapid response of the system status. The period of low system load is defined as the period when the average CPU load is <25% for 10 consecutive minutes and the memory occupancy rate is <30%. During this period, the system can automatically adjust the operating parameters to reduce energy consumption and improve operating efficiency.

[0020] Compared with related technologies, the integrated thermal management system for explosion-proof chargers provided by the present invention has the following beneficial effects:

[0021] 1. This invention provides an integrated thermal management system for explosion-proof chargers. This system optimizes the liquid cooling pipeline layout through three-dimensional thermal field modeling (heat flux uniformity error ≤ 10%). It integrates a ZL114A aluminum alloy housing and a microchannel liquid cooling plate (thermal resistance reduced by 40%). Combined with an IP68 metal sealing structure and ExdIICT6 explosion-proof certification, it achieves safe heat dissipation of power device temperatures ≤ 85°C in a wide temperature range of -20°C to 60°C and in gas and dust explosive environments. Furthermore, a conductivity sensor and ion exchange resin filter intelligently manage the liquid cooling medium to prevent equipment failures caused by contamination.

[0022] 2. The present invention provides an integrated thermal management system for explosion-proof chargers. It uses multimodal data collection such as lidar, temperature sensor arrays, and infrared thermal imaging, and calibrates through a multi-view stereo vision algorithm and ray tracing technology (temperature deviation ≤±0.5°C). It combines PTP protocol spatiotemporal calibration (sensor coordinate error ≤±2mm) with a KD tree algorithm to build a real-time three-dimensional thermal field model. Based on the LSTM-GARCH model, it provides an early warning of ≥8 minutes (traditional systems have a delay of more than 1 minute), achieving early and accurate identification of faults such as sudden flow rate changes and abnormal temperature gradients.

[0023] 3. The present invention provides an integrated thermal management system for explosion-proof chargers. This system dynamically adjusts flow (adjustment is completed within 200ms, and the temperature drop rate is ≥1.5°C / s) using a model predictive control (MPC) algorithm. This system maximizes heat dissipation efficiency (objective function J = min(ΔT / ΔQ)) while controlling pump power to ≤150W (reducing energy consumption by 30%). Furthermore, the system is integrated through edge computing (with anomaly detection and processing delay ≤100ms). Intelligent operation and maintenance is automatically performed during low-load periods (CPU load <25% and memory usage <30%), balancing heat dissipation efficiency and adaptive management. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a flow chart of the present invention;

[0025] Figure 2 An expanded flow chart for three-dimensional thermal field modeling of the present invention;

[0026] Figure 3 This is an extended flow chart of multi-source data fusion of the present invention;

[0027] Figure 4 This is an extended flow chart of abnormal behavior detection of the present invention;

[0028] Figure 5 An expanded flow chart of the dynamic calibration of the sensor of the present invention;

[0029] Figure 6 This is an expanded flow chart of abnormal fluctuation judgment of the present invention;

[0030] Figure 7 This is an extended flow chart of the multimodal early warning system of the present invention;

[0031] Figure 8 This is an expanded flow chart of the system integration features of the present invention. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0033] Example 1:

[0034] See also Figures 1-8 The present invention provides a technical solution: an integrated thermal management system for an explosion-proof charger, comprising the following steps:

[0035] Step S100: 3D thermal field modeling: Use laser radar and temperature sensor array to obtain heat source point cloud, and combine finite element analysis to build a 3D thermal management model; set heat flux density monitoring points at the connection between the power module and the liquid cooling pipeline, and optimize the pipeline layout to make the heat flux density uniformity error ≤10%;

[0036] In step S100, the infrared heat flow sensor is used as the core element of heat flow density monitoring. It converts thermal radiation into electrical signals based on the Seebeck effect. The sensor has a built-in temperature compensation circuit that can automatically correct the influence of environmental temperature changes on measurement results, ensuring that the accuracy is stable within ≤±3% in the 0-500W / m² range. During data collection, the sensor acquires heat flow density data at the connection between the power module and the liquid cooling pipeline at a frequency of 10Hz. Through finite element analysis software, a multi-physical field coupling model is constructed, including fluid flow, heat transfer, and structural mechanics. The grid adaptive technology is used to dynamically encrypt the key area grid. The pipeline layout is iteratively optimized. In the optimization process, the genetic algorithm is introduced for global search. By adjusting the pipeline bending angle, pipe diameter size, and distribution position, the heat flow density distribution uniformity error is stably controlled within ≤10%, and the heat dissipation efficiency is improved.

[0037] In this embodiment, a 16-line laser radar (model: Velodyne VLP-16, scanning frequency 10Hz, measurement accuracy ±2cm) and a temperature sensor array (model: PT1000, accuracy ±0.1℃) are used to scan the charger power module and the liquid cooling pipeline. The laser radar acquires heat source point cloud data within a horizontal 0°-360° and vertical -60°-+60° angle range. ANSYS finite element analysis software is used to simulate the liquid cooling pipeline layout. Infrared heat flow sensors (model: HFM-7, range 0-500W / m², accuracy ±3%) are deployed at the connection between the power module and the pipeline. By optimizing the pipeline bending radius and flow velocity distribution, the heat flow density uniformity error is controlled within 10%.

[0038] In step S200, multi-source data fusion: collect the flow, pressure, and temperature field data of the liquid cooling circuit. After reconstruction by the multi-view stereo vision algorithm, the data is registered by the iterative closest point algorithm. The infrared thermal imaging data is denoised and calibrated, and the deviation is corrected by the ray tracing algorithm.

[0039] In the infrared thermal imaging data processing in step S200, the 5×5 median filter algorithm suppresses salt and pepper noise. The principle is to replace the gray value of each pixel in the image with the median value of the 5×5 pixel gray values in the neighborhood. This preserves the edge information while removing noise interference. A high-precision blackbody furnace is used as the reference source in the blackbody radiation calibration process. The temperature control accuracy of the blackbody furnace is ±0.1℃. By comparing the measured value of the infrared thermal imager with the standard value of the blackbody furnace, a temperature error correction curve is established, and the calibration error is ≤±0.5℃. The ray tracing algorithm is based on the Monte Carlo method, which simulates the propagation path of thermal radiation in complex pipeline structures. By calculating the reflection, absorption, and transmission of radiant energy on the pipeline surface, the temperature field point cloud deviation caused by pipeline shielding is corrected, making the reconstructed temperature field closer to the actual working condition.

[0040] In this implementation, flow rate (turbine flowmeter, accuracy ±0.5%FS), pressure (piezoresistive sensor, accuracy ±0.2kPa), and temperature field data from the liquid cooling circuit are collected via an RS485 bus. Three-dimensional reconstruction of the data is performed using the CMVS / PMVS multi-view stereo vision algorithm. Registration is performed using the iterative closest point (ICP) algorithm with a convergence threshold of 1e-6m. Infrared thermal imaging data (FLIRT1040, resolution 640×512) is denoised using a 5×5 median filter. Temperature deviation is calibrated using a blackbody radiation source (model: HT-300, accuracy ±0.5°C). A ray tracing algorithm is used to simulate pipe occlusion of thermal radiation, correcting the spatial deviation of the temperature field point cloud to ±2mm.

[0041] Step S300: Abnormal behavior detection: Based on the target detection algorithm and the calibration parameters of the binocular camera, a three-dimensional thermal trajectory is constructed through triangulation, and the baseline feature space established by K-means clustering is combined to identify heat dissipation anomalies;

[0042] In step S300, historical data collection covers the operating conditions of explosion-proof chargers under different loads, ambient temperatures, and heat dissipation conditions, ensuring that the data reflects ≥90% of actual operating scenarios. The collected data is preprocessed to extract key characteristic parameters of flow, temperature, and power. Principal component analysis is used to reduce data dimensionality, retaining more than 95% of the information entropy. Based on the K-means clustering algorithm, the data is clustered into different operating modes, and a baseline feature space containing flow-temperature mapping curves is constructed. Furthermore, a safe temperature zone of ≤85°C is set based on the thermal characteristics of power devices. The thresholds for flow rate mutations >1L / min·s and temperature gradients >4°C / cm in the anomaly triggering rules are determined by combining extensive experimental data with charger thermal safety standards, enabling timely and accurate detection of heat dissipation anomalies.

[0043] In this implementation, a three-dimensional thermal trajectory (with an error of ≤5mm) is constructed using triangulation, based on the YOLOv8 object detection algorithm and a binocular camera (focal length f = 12mm, baseline distance 80mm) with calibration parameters. 30 days of historical operating data (covering 90% of operating conditions, including full load and no-load scenarios) is collected. A baseline feature space is established using the K-means clustering algorithm. The power device safety temperature range is set to ≤85°C. Abnormal warnings are triggered when the flow rate changes by more than 1L / min·s or the temperature gradient exceeds 4°C / cm.

[0044] Step S400: Dynamic sensor calibration: Arrange sensors on the liquid cooling radiator and pipelines, perform spatiotemporal calibration via the PTP protocol, and use the KD tree algorithm to generate a real-time three-dimensional thermal management model containing multi-source data;

[0045] The sensor calibration work in step S400 is crucial. The flow sensor is calibrated using the weighing method. The specific operation is to collect liquid cooling medium over a certain period of time, measure the medium mass using a high-precision electronic scale, calculate the actual flow rate based on the medium density, and compare it with the sensor measurement value for calibration to ensure that the error is ≤±0.5%FS. The temperature sensor is calibrated at the calibration point of 37°C using a high-precision constant temperature bath. By comparing the sensor output value with the measurement value of a standard platinum resistance thermometer, the sensor compensation parameters are adjusted to ensure that the error is ≤±0.2°C. The 3D laser tracker measures the coordinate position of the sensor in space by emitting a laser beam. Its measurement accuracy reaches ±2mm, ensuring the accuracy of the sensor spatial layout and providing a foundation for building an accurate real-time 3D thermal management model.

[0046] In this implementation, sensors were placed on a liquid-cooled heat sink (thermal conductivity 150 W / m·K) and in the piping, and spatial and temporal calibration was performed using the PTP protocol (IEEE1588, time synchronization accuracy ±5 ms). The flow sensor was calibrated gravimetrically (with a standard weight error of ±0.5% FS), and the temperature sensor was compared in a 37°C constant temperature bath (with an error of ±0.2°C). The sensor's spatial coordinates were calibrated using a Leica AT960 laser tracker (with an accuracy of ±15 μm), maintaining an error within ±2 mm. A real-time 3D thermal management model incorporating multi-source data was generated using a KD tree algorithm (with a spatial threshold of 0.3 m).

[0047] Step S500: abnormal fluctuation judgment: use LSTM-GARCH model to predict liquid cooling parameters, and issue early warning when the prediction confidence is greater than 90%, and judge abnormality based on preset thresholds;

[0048] In step S500, the LSTM-GARCH model captures the time series characteristics of the liquid cooling parameters through a long-short-term memory network. Combined with the generalized autoregressive conditional heteroscedasticity model, it processes the volatility and heteroscedasticity of the data to achieve accurate predictions of the liquid cooling parameters. When the predicted value exceeds the normal range and the confidence level is greater than 90%, the system triggers an early warning at least 8 minutes in advance, reserving sufficient processing time for operation and maintenance personnel. After the multimodal warning is triggered, the flow control device of the liquid cooling system uses a fast-response electric control valve with a response time of less than 50ms. Combined with a high-performance water pump, flow control is completed within 200ms. By optimizing the pipeline resistance and the water pump characteristic curve, the temperature drop rate of the power device is kept at least 1.5°C / s, preventing equipment damage caused by overheating.

[0049] In this implementation, an LSTM-GARCH model (with 128 hidden layer neurons and a 10-minute time step) was used to predict liquid cooling parameters. When the prediction confidence level exceeded 90%, an early warning was triggered 8 minutes in advance. Once the multimodal warning was triggered, the liquid cooling system, using a proportional valve (with a 20-ms response time), adjusted the flow rate within 200 ms, reducing the power device temperature by a rate of 1.5°C / s (monitored in real time by thermocouples).

[0050] Step S600 multi-modal warning: when triggered, the liquid cooling system is linked to adjust the flow rate and an optimal cooling strategy is generated through a model predictive control algorithm;

[0051] The model predictive control algorithm in step S600 aims to maximize heat dissipation efficiency and minimize energy consumption. It establishes an optimization problem that includes a dynamic model of the liquid cooling system, constraints, and an objective function. In the objective function J=min(ΔT / ΔQ), ΔT is the temperature change of the power device, and ΔQ is the flow change of the liquid cooling medium. By minimizing this function, heat dissipation efficiency is improved. At the same time, the pump power is set to ≤150W as a constraint to ensure that the system energy consumption is within a reasonable range. The MPC algorithm adopts a rolling time domain optimization strategy. Within each control cycle, the optimization problem is solved online based on the current system state and predicted future operating conditions to generate the optimal flow regulation strategy to achieve a balance between heat dissipation performance and energy consumption.

[0052] In this implementation plan, the model predictive control (MPC) algorithm is adopted with the goals of maximizing heat dissipation efficiency (objective function J=min(ΔT / ΔQ), ΔT is the temperature change, ΔQ is the flow change) and minimizing energy consumption (pump power ≤150W). The optimal flow regulation strategy is generated through rolling optimization, with an adjustment step of 0.1L / min and a response delay of ≤50ms.

[0053] The explosion-proof enclosure is integrally molded from ZL114A aluminum alloy, which undergoes a T6 heat treatment process to significantly improve its tensile strength and thermal conductivity. The microchannel liquid cooling plate embedded in the inner wall is manufactured using precision machining technology. The channel width is ≤0.5mm and the depth is ≤2mm, greatly increasing the heat dissipation area and improving heat dissipation efficiency. The microchannel is connected to the external circuit through precision welding to form an integrated heat dissipation structure. The metal bellows sealing joints at the pipe penetrations have an IP68 protection rating, protecting against dust and liquid intrusion. The temperature range is -40°C to 125°C, making it suitable for harsh environments. The fluororubber O-ring is used, and the compression is controlled at 15%-20%, ensuring excellent sealing performance and a heat conduction resistance of ≤0.1°C / m² / W. The explosion-proof enclosure is ExdIICT6 certified and meets the requirements for use in explosive environments.

[0054] In this implementation, the explosion-proof housing is die-cast from ZL114A aluminum alloy (tensile strength 380 MPa, thermal conductivity 180 W / m・K). A microchannel liquid cooling plate (channel width 0.5 mm, depth 2 mm) is embedded within the inner wall and integrated with the housing via vacuum brazing, resulting in a 40% reduction in thermal resistance compared to traditional split-type designs. Pipeline penetrations utilize metal bellows sealing joints (IP68 protection, temperature resistance -40°C to 125°C) with fluororubber O-rings (18% compression), resulting in a thermal resistance of 0.08°C・m² / W. The system is ExdIICT6 certified.

[0055] The pipe penetrations use metal bellows sealing joints (IP68 protection, temperature resistance -40°C to 125°C) with fluororubber O-rings (compression volume 18%), with a heat conduction resistance of 0.08°C・m² / W to prevent the intrusion of explosive gases.

[0056] 3M™ Novec™ 7500 non-flammable fluorinated fluid is used, and its purity is monitored by a conductivity sensor (accuracy ±1%). When the conductivity is greater than 50μS / cm, the 0.1μm precision ion exchange resin filter device is automatically activated. A liquid level sensor (accuracy ±1mm) is used for real-time monitoring. When the liquid level drops by more than 5mm, the 5L / min rehydration pump is triggered to automatically compensate.

[0057] The non-flammable heat-conducting medium circulation system in the liquid cooling medium management system uses a high-purity synthetic heat-conducting fluid with low viscosity, high boiling point, and good chemical stability. The conductivity sensor adopts a four-electrode design and uses AC excitation technology to measure the medium conductivity with an accuracy of ≤±1%. When the conductivity is greater than 50μS / cm, the ion exchange resin filter device is automatically activated. This device adopts a multi-stage series structure to effectively remove ionic impurities in the medium and restore the medium purity. The liquid level sensor adopts the magnetostrictive principle with an accuracy of ≤±1mm. It monitors the liquid level of the liquid cooling medium in real time. When the liquid level drops, the refill device is automatically triggered to compensate for the medium loss and ensure the stable operation of the liquid cooling system.

[0058] In this implementation, the cooling medium is a non-flammable fluorinated liquid (model: 3M™ Novec™ 7500). A conductivity sensor (accuracy ±1%) monitors the cooling medium in real time. When the conductivity exceeds 50 μS / cm, an ion exchange resin filter (filtration accuracy 0.1 μm) is automatically activated. A liquid level sensor (accuracy ±1 mm) monitors the liquid level in real time. When the level drops by more than 5 mm, a refill pump (flow rate 5 L / min) is triggered to automatically compensate.

[0059] This integrated thermal management system is specially designed for explosive environments with gas concentrations ≤1.5% and dust concentrations ≤5mg / m³. All electrical components are explosion-proof and certified. The liquid cooling system operates in a wide operating temperature range of -20°C to 60°C. It uses a thermally conductive medium with excellent low-temperature performance and low-temperature-resistant sealing materials to ensure normal operation of the system in low-temperature environments. The edge server is equipped with a high-performance processor and an optimized anomaly detection algorithm, with a processing delay of ≤100ms, enabling real-time monitoring and rapid response of system status. Periods of low system load are defined as periods when the average CPU load is less than 25% for 10 consecutive minutes and the memory utilization is less than 30%. During this period, the system automatically adjusts operating parameters to reduce energy consumption and improve operational efficiency.

[0060] In this implementation, the integrated thermal management system is suitable for high-risk explosive environments with gas concentrations ≤ 1.5% by volume (reaching the safety threshold for underground coal mines) and dust concentrations ≤ 5 mg / m³ (meeting the standard for Zone 21 explosive dust environments). It operates stably in a wide operating temperature range of -20°C to 60°C. The system is ExdIICT6 certified, protecting against the ignition risk of Class IIC explosive gases (such as hydrogen and acetylene). The explosion-proof housing is die-cast from ZL114A aluminum alloy and features an integrated microchannel liquid cooling plate. This allows the system to maintain liquid cooling flow through a preheat cycle even at -20°C. At 60°C, the liquid cooling system dynamically adjusts the flow rate to 3 L / min using an MPC algorithm, ensuring power device temperatures remain within the safety threshold of 85°C. The anomaly detection algorithm on the edge server achieves a processing latency of only 85ms under typical operating conditions with a gas concentration of 1.2% and a dust concentration of 4 mg / m³, meeting the requirements for real-time safety monitoring in explosive environments.

[0061] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that modifications or equivalent replacements of the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.

Claims

1. An integrated thermal management system for an explosion-proof charger, characterized in that: The following steps are involved: Step S100: 3D thermal field modeling: Use laser radar and temperature sensor array to obtain heat source point cloud, and combine finite element analysis to build a 3D thermal management model; set heat flux density monitoring points at the connection between the power module and the liquid cooling pipeline, and optimize the pipeline layout to make the heat flux density uniformity error ≤10%; Step S200: Multi-source data fusion: Flow, pressure, and temperature field data of the liquid cooling circuit are collected, reconstructed using a multi-view stereo vision algorithm, and then aligned using an iterative closest point algorithm; infrared thermal imaging data are denoised and calibrated, and deviations are corrected using a ray tracing algorithm; Step S300: Abnormal behavior detection: Based on the target detection algorithm and the calibration parameters of the binocular camera, a three-dimensional thermal trajectory is constructed through triangulation, and the baseline feature space established by K-means clustering is combined to identify heat dissipation anomalies; Step S400: Dynamic sensor calibration: Arrange sensors on the liquid cooling radiator and pipelines, perform spatiotemporal calibration via the PTP protocol, and use the KD tree algorithm to generate a real-time three-dimensional thermal management model containing multi-source data; Step S500: abnormal fluctuation judgment: use LSTM-GARCH model to predict liquid cooling parameters, and issue early warning when the prediction confidence is greater than 90%, and judge abnormality based on preset thresholds; Step S600 multi-modal warning: when triggered, the liquid cooling system is linked to adjust the flow rate, and the optimal heat dissipation strategy is generated through the model predictive control algorithm.

2. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: In step S100, the infrared heat flux sensor serves as the core component for heat flux density monitoring. It adopts the thermopile principle based on the Seebeck effect to convert thermal radiation into electrical signal output. The sensor has a built-in temperature compensation circuit that can automatically correct the impact of ambient temperature changes on the measurement results, ensuring that the accuracy is stable at ≤±3% within the range of 0-500W / m². During the data acquisition process, the sensor obtains the heat flux density data at the connection between the charger power module and the liquid cooling pipeline in real time at a frequency of 10Hz. Through finite element analysis software, a multi-physics field coupling model including fluid flow, heat transfer and structural mechanics is constructed. The grid adaptive technology is used to dynamically encrypt the grid in key areas, and the pipeline layout is iteratively optimized. During the optimization process, a genetic algorithm is introduced for global search. By adjusting the pipeline bending angle, pipe diameter and distribution position, the heat flux density distribution uniformity error is stably controlled at ≤10%, thereby improving the heat dissipation efficiency.

3. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: In the infrared thermal imaging data processing in step S200, the 5×5 median filtering algorithm is used to suppress salt and pepper noise. Its principle is to replace the grayscale value of each pixel in the image with the median of the grayscale values ​​of 5×5 pixels in the neighborhood of the point, removing noise interference while retaining edge information. The blackbody radiation calibration link uses a high-precision blackbody furnace as a reference source. The temperature control accuracy of the blackbody furnace reaches ±0.1°C. By comparing the infrared thermal imager measurement value with the blackbody furnace standard value, a temperature error correction curve is established to make the calibration error ≤±0.5°C. The ray tracing algorithm is based on the Monte Carlo method to simulate the propagation path of thermal radiation in complex pipeline structures. By calculating the reflection, absorption and transmission of radiation energy on the pipeline surface, the temperature field point cloud deviation caused by pipeline obstruction is corrected, making the reconstructed temperature field closer to the actual working conditions.

4. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: In step S300, historical data collection covers the operating conditions of explosion-proof chargers under different loads, ambient temperatures, and heat dissipation conditions, ensuring that the data can reflect ≥90% of actual working scenarios. By preprocessing the collected data, key characteristic parameters of flow, temperature, and power are extracted, and the principal component analysis method is used to reduce the data dimension, retaining more than 95% of the information entropy. Based on the K-means clustering algorithm, the data is clustered into different operating modes, and a baseline feature space containing flow-temperature mapping curves is constructed. At the same time, combined with the thermal characteristics of power devices, a safe temperature zone is set to ≤85°C. In the abnormal trigger rules, the thresholds of flow mutation >1L / min・s and temperature gradient >4°C / cm are determined by combining a large amount of experimental data with the charger thermal safety standards, and can detect heat dissipation anomalies in a timely and accurate manner.

5. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: The sensor calibration work in step S400 is crucial. The flow sensor is calibrated using the weighing method. The specific operation is to collect liquid cooling medium within a certain period of time, measure the medium mass using a high-precision electronic scale, calculate the actual flow rate based on the medium density, and compare it with the sensor measurement value for calibration to ensure that the error is ≤±0.5%FS. The temperature sensor is calibrated at the calibration point of 37°C using a high-precision constant temperature bath. By comparing the sensor output value with the measurement value of a standard platinum resistance thermometer, the sensor compensation parameters are adjusted to ensure that the error is ≤±0.2°C. The three-dimensional laser tracker measures the coordinate position of the sensor in space by emitting a laser beam. Its measurement accuracy reaches ±2mm, ensuring the accuracy of the sensor spatial layout and providing a basis for building an accurate real-time three-dimensional thermal management model.

6. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: In step S500, the LSTM-GARCH model captures the time series characteristics of the liquid cooling parameters through a long short-term memory network, and combines it with a generalized autoregressive conditional heteroscedasticity model to process the volatility and heteroscedasticity of the data, achieving accurate prediction of the liquid cooling parameters. When the predicted value exceeds the normal range and the confidence level is greater than 90%, the system triggers an early warning at least 8 minutes in advance, reserving sufficient processing time for operation and maintenance personnel. After the multimodal warning is triggered, the flow control device of the liquid cooling system adopts a fast-response electric control valve with a response time of less than 50ms. In conjunction with a high-performance water pump, flow regulation is completed within 200ms. By optimizing the pipeline resistance and the water pump characteristic curve, the temperature drop rate of the power device is made ≥1.5℃ / s, avoiding equipment damage caused by overheating.

7. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: The model predictive control algorithm in step S600 aims to maximize heat dissipation efficiency and minimize energy consumption. It establishes an optimization problem that includes a dynamic model of the liquid cooling system, constraints, and an objective function. In the objective function J=min(ΔT / ΔQ), ΔT is the temperature change of the power device, and ΔQ is the flow change of the liquid cooling medium. By minimizing this function, the heat dissipation efficiency is improved. At the same time, the pump power ≤150W is set as a constraint to ensure that the system energy consumption is within a reasonable range. The MPC algorithm adopts a rolling time domain optimization strategy. In each control cycle, based on the current system status and predicted future operating conditions, the optimization problem is solved online to generate the optimal flow regulation strategy to achieve a balance between heat dissipation performance and energy consumption.

8. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: The explosion-proof enclosure is integrally molded from ZL114A aluminum alloy, which undergoes a T6 heat treatment process to significantly improve its tensile strength and thermal conductivity. The microchannel liquid cooling plate embedded in the inner wall is manufactured using precision machining techniques. The channel width is ≤0.5mm and the depth is ≤2mm, greatly increasing the heat dissipation area and improving heat dissipation efficiency. The microchannels are connected to the external circuit through precision welding to form an integrated heat dissipation structure. The metal bellows sealing joints at the pipe penetrations have an IP68 protection rating, protecting against dust and liquid intrusion. The temperature range is -40°C to 125°C, making it suitable for harsh environments. Fluororubber O-rings with compression controlled at 15%-20% ensure excellent sealing performance, with a thermal resistance of ≤0.1°C / m² / W. The explosion-proof enclosure is ExdIICT6 certified and meets the requirements for use in explosive environments.

9. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: The non-flammable heat-conducting medium circulation system in the liquid cooling medium management system adopts high-purity synthetic heat-conducting fluid with low viscosity, high boiling point and good chemical stability. The conductivity sensor adopts a four-electrode design and uses AC excitation technology to measure the conductivity of the medium with an accuracy of ≤±1%. When the conductivity is greater than 50μS / cm, the ion exchange resin filter device is automatically started. The device adopts a multi-stage series structure to remove ionic impurities in the medium and restore the purity of the medium. The liquid level sensor adopts the magnetostrictive principle with an accuracy of ≤±1mm. It monitors the liquid level of the liquid cooling medium in real time. When the liquid level drops, the rehydration device is automatically triggered to compensate for the medium loss and ensure the stable operation of the liquid cooling system.

10. The integrated thermal management system for an explosion-proof charger according to claim 1, characterized in that: This integrated thermal management system is specially designed for explosive environments with gas concentration ≤1.5% and dust concentration ≤5mg / m³. All electrical components are explosion-proof and certified. The liquid cooling system operates under a wide operating temperature range of -20℃~60℃. It uses a thermal conductive medium with excellent low-temperature performance and low-temperature resistant sealing materials to ensure that the system operates normally in a low-temperature environment. The edge server is equipped with a high-performance processor and an optimized anomaly detection algorithm with a processing delay of ≤100ms, realizing real-time monitoring and rapid response of the system status. The period of low system load is defined as the period when the average CPU load is less than 25% for 10 consecutive minutes and the memory usage is less than 30%. During this period, the system can automatically adjust the operating parameters to reduce energy consumption and improve operating efficiency.

Images