Intelligent heat dissipation cooperative control high-low voltage reactive power compensation device and control method

Abstract

This invention discloses a high- and low-voltage reactive power compensation device with intelligent heat dissipation collaborative control, comprising a reactive power compensation main circuit module, an intelligent heat dissipation collaborative control module, a distributed adaptive sensor monitoring module, and a redundant switching heat dissipation execution module, forming a closed-loop intelligent heat dissipation collaborative system through an internal bus. The intelligent heat dissipation collaborative control module integrates a fault self-diagnosis unit and a thermal status early warning unit, employing an improved deep learning adaptive algorithm combined with multi-dimensional monitoring data to achieve accurate thermal status prediction and dynamic optimization of heat dissipation strategies. The distributed adaptive sensor monitoring module can dynamically adjust the sampling frequency according to the operating status, improving monitoring accuracy and energy consumption balance. The redundant switching heat dissipation execution module optimizes the three-level heat dissipation collaborative mechanism and adds a rapid component fault switching function. This invention solves the problems of poor heat dissipation collaboration, delayed fault response, and high monitoring energy consumption in traditional devices, and is suitable for scenarios with stringent requirements for heat dissipation stability, such as new energy power plants and industrial production.

Description

Technical Field

[0001] This invention belongs to the field of power electronic equipment technology, specifically relating to a high and low voltage reactive power compensation device with intelligent heat dissipation and coordinated control. It is suitable for low-voltage complete reactive power compensation scenarios with rated AC voltage not exceeding 1000V and frequency not exceeding 1000Hz, as well as high-voltage reactive power compensation scenarios of 6~110KV. It is especially suitable for application environments with stringent requirements for heat dissipation stability and fault tolerance, such as new energy power plants, industrial production, and commercial buildings. Background Technology

[0002] Reactive power compensation devices, as core equipment for optimizing power quality in power systems, primarily function to offset reactive power generated by loads by connecting capacitors or reactors in parallel, thereby improving the power factor of the power grid and reducing line losses. With the large-scale integration of new energy sources into the grid, the operating conditions of reactive power compensation devices are becoming increasingly complex, and heat dissipation has become a key factor restricting their performance and lifespan.

[0003] In the prior art, Chinese patent CN103401254A discloses a monitoring device for an automatic reactive power compensation system for high and low voltage distribution networks. This device monitors parameters through a power grid operation data monitoring module and an SVG module operation data monitoring module. When the temperature detection unit detects that the temperature exceeds the standard, it activates the cooling unit. However, this device only uses a single air cooling method, resulting in poor heat dissipation coordination and a lack of fault diagnosis and redundant protection for the heat dissipation components. Under high load or harsh environments, it is prone to equipment failure due to heat dissipation failure.

[0004] Chinese patent CN103401255A discloses an automatic reactive power compensation system for high and low voltage distribution networks. It adopts a multi-point detection unit to improve the comprehensiveness of monitoring and the cooling unit is an air-cooled device. However, the temperature detection of the system can only trigger simple cooling or shutdown actions and cannot adjust the heat dissipation strategy in advance according to the trend of thermal state changes, resulting in obvious response lag.

[0005] Chinese patent application CN120414289A proposes a high and low voltage reactive power compensation device that uses intelligent heat dissipation. It solves the problem of filter clogging by using a ring filter self-cleaning structure. However, the device still relies on a single air-cooling architecture, which has limited heat dissipation efficiency. Furthermore, it does not involve thermal state prediction and component redundancy switching design, making it difficult to meet the heat dissipation requirements of high power density scenarios.

[0006] Furthermore, existing reactive power compensation devices often use fixed sampling frequencies for sensor monitoring, leading to energy waste during normal operation or insufficient monitoring accuracy under abnormal conditions. The three-stage cooling mechanisms of the heat dissipation system are mostly simple superpositions, lacking precise coordinated triggering logic. Simultaneously, their diagnostic and response capabilities for heat dissipation component failures are weak; a failure in one component can easily cause the entire heat dissipation system to fail. These problems collectively result in insufficient reliability and adaptability of traditional reactive power compensation devices under complex operating conditions, limiting their application in harsh environments such as renewable energy power plants.

[0007] Therefore, there is an urgent need for a high- and low-voltage reactive power compensation device with intelligent collaborative heat dissipation, adaptive and precise monitoring, and fault redundancy switching functions to overcome the shortcomings of existing technologies. Summary of the Invention

[0008] To address the technical problems of existing high and low voltage reactive power compensation devices, such as poor heat dissipation coordination, insufficient thermal state prediction capability, fixed sampling frequency, lack of heat dissipation component fault diagnosis and redundancy switching, and lack of closed-loop feedback mechanism for heat dissipation strategy optimization, the present invention aims to provide a high and low voltage reactive power compensation device with intelligent coordinated heat dissipation, adaptive and accurate monitoring, and fault redundancy switching functions.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A high- and low-voltage reactive power compensation device with intelligent heat dissipation and coordinated control includes a reactive power compensation main circuit module, an intelligent heat dissipation and coordinated control module, a distributed adaptive sensor monitoring module, and a redundant switching heat dissipation execution module, which are connected in sequence via communication.

[0010] The reactive power compensation main circuit module includes capacitor banks, reactors, and a composite switching switch. The capacitor banks adopt a three-phase delta connection, with each phase consisting of 10 50kvar capacitor units connected in series, for a total capacity of 150kvar. The optimized connection structure reduces the risk of local overheating. The reactors are dry-type air-core reactors with a reactance of 7%, effectively suppressing harmonics and inrush current. The composite switching switch combines the advantages of contactors and thyristors to achieve zero-voltage switching and zero-current interruption, reducing inrush current and overvoltage during the switching process and minimizing additional heat generation.

[0011] The intelligent heat dissipation collaborative control module is the core innovation of this invention. It adopts a master-slave multi-processor hardware architecture. The master controller is an ARM Cortex-M7 series 32-bit microcontroller (STM32F767), which is responsible for overall system coordination and data processing. The coprocessor is a DSP digital signal processor (TMS320F28335), which focuses on real-time processing of sensor data and algorithm calculation. The dedicated heat dissipation control chip uses an FPGA programmable logic device (XilinxArtix-7) to achieve precise control of the heat dissipation execution module. A fault self-diagnosis unit and a thermal state early warning unit are added to improve fault response and risk avoidance capabilities.

[0012] In terms of software, an improved deep learning adaptive algorithm is adopted, which is divided into five layers: the data acquisition layer receives multi-dimensional monitoring data; the feature enhancement layer uses wavelet packet transform and improved principal component analysis algorithm to strengthen the correlation of key features such as temperature gradient and load change rate; the prediction and decision layer uses a combination of bidirectional long short-term memory network (Bi-LSTM) and deep reinforcement learning, inputs the feature data of the most recent 15 seconds, outputs the thermal state prediction value of the next 45 seconds, and formulates the optimal heat dissipation strategy; the execution control layer converts the strategy into control commands; and the feedback optimization layer dynamically adjusts the algorithm parameters according to the heat dissipation effect and fault information to form a closed-loop optimization.

[0013] The distributed adaptive sensor monitoring module adopts a distributed network architecture, including 24 sensor nodes and 1 central data fusion unit. The sensor nodes cover temperature, humidity, wind speed, current, and voltage sensors: the temperature sensors use a combination of a ±0.1℃ PT100 platinum resistance thermometer, an MLX90614 infrared sensor, and a DS18B20 digital sensor, distributed at key heat-generating areas and air inlets / outlets; the humidity sensor is a high-precision SHT31 sensor; the wind speed sensor is a thermal wind speed sensor; and the current and voltage sensors use Hall effect sensors to monitor the main circuit's operating status.

[0014] The central data fusion unit uses an STM32H7 series microcontroller and communicates with sensor nodes via a CAN-FD bus, with a data transmission rate of ≥5Mbps. The sensor nodes employ an adaptive sampling mechanism: the sampling frequency is 1Hz during normal operation; when temperature fluctuations exceed ±2℃ / min or load fluctuations exceed 10%, the sampling frequency automatically increases to 15Hz, achieving a balance between monitoring accuracy and energy consumption. The central data fusion unit improves data reliability through multi-source data calibration and fusion algorithms, supporting the storage of historical data for the past 60 days and anomaly backtracking.

[0015] The redundant switching heat dissipation execution module optimizes the three-level heat dissipation coordination mechanism, including the variable frequency fan subsystem, the micro-pump liquid cooling subsystem, and the phase change material heat dissipation subsystem: The variable frequency fan subsystem adopts a redundant design of two DeltaElectronics BFB0412VHD brushless DC fans with a speed adjustment range of 800~12000rpm. Stepless speed regulation is achieved through PWM control, and the noise is controlled below 32dB. The switching response time in the event of a single fan failure is ≤50ms. The micro-pump liquid cooling subsystem uses two Bartels Mikrotechnik mp6 micro-pumps (with redundancy), with a flow rate of 500 ml / min and a pressure of 30 kPa. The coolant is an environmentally friendly nanofluid composed of deionized water and 1% volume fraction Al2O3 nanoparticles, with a thermal conductivity that is more than 40% higher than that of traditional coolants. The liquid cooling system is a fully sealed structure and is equipped with dual monitoring of pressure and flow rate. The phase change material heat dissipation subsystem uses paraffin-based phase change material with a phase change temperature of 48℃~52℃, a latent heat value of ≥185J / g, a total mass of 2kg, and is encapsulated in an aluminum container. It is in close contact with the capacitor bank through thermally conductive silicone grease.

[0016] The three-level heat dissipation coordination triggering logic is as follows: When the heat flux density is below 100W / cm², only the phase change material heat dissipation subsystem operates; when it is between 100 and 500W / cm², the phase change material and variable frequency fan subsystems operate in coordination; when it exceeds 500W / cm², all three subsystems operate simultaneously. When the fault self-diagnosis unit detects a fault in a certain subsystem, it automatically increases the operating power of the remaining subsystems to ensure continuous heat dissipation.

[0017] Furthermore, the device employs a three-level collaborative control strategy: device level, system level, and cloud level. Equipment-level control is performed by the local controller of each module, realizing basic control and protection; System-level control is led by the intelligent heat dissipation collaborative control module, which uses model predictive control algorithms to optimize the heat dissipation collaborative logic; Cloud-level control is achieved through the ESP32-WROOM-32 IoT module, which supports dual-mode communication via Wi-Fi and Ethernet. It uses the MQTT protocol to exchange data with the cloud platform. Users can remotely monitor, configure parameters, receive fault alarms, and perform OTA firmware upgrades via a mobile APP or web interface.

[0018] Meanwhile, the present invention also provides a control method for the above-mentioned device, comprising the following steps: (1) The distributed adaptive sensor monitoring module collects multi-dimensional data according to the adaptive sampling mechanism, and sends the data to the intelligent heat dissipation collaborative control module after processing by the central data fusion unit. (2) The intelligent heat dissipation collaborative control module predicts the trend of thermal state change of the device and formulates the initial heat dissipation strategy through an improved deep learning adaptive algorithm; (3) The redundant switching heat dissipation execution module starts the corresponding level of heat dissipation subsystem according to the initial heat dissipation strategy to achieve coordinated heat dissipation; (4) The fault self-diagnosis unit monitors the operating status of the heat dissipation components in real time and triggers the redundancy switching mechanism when a fault occurs to ensure the continuity of heat dissipation. (5) The feedback optimization layer dynamically adjusts the algorithm parameters and heat dissipation strategy based on the heat dissipation effect data to form a closed-loop optimization.

[0019] Due to the adoption of the above solution, the beneficial effects of the present invention are as follows: 1. Significantly improved heat dissipation synergy: Through precise three-level heat dissipation triggering logic and improved deep learning algorithm, the heat flux density processing range is expanded to 30~1200W / cm², which is 20%~30% more efficient than traditional devices in terms of heat dissipation. Under high load conditions, the temperature of key components can be controlled below 60℃.

[0020] 2. It has outstanding thermal state prediction and early warning capabilities. The bidirectional long short-term memory network enables accurate thermal state prediction within 45 seconds. The thermal state early warning unit avoids overheating risks in advance, reducing the occurrence rate of device overheating failure by more than 80%.

[0021] 3. Significant optimization of monitoring energy consumption: The distributed adaptive sampling mechanism reduces sensor monitoring energy consumption by 35% to 45% while ensuring monitoring accuracy. The central data fusion unit improves data reliability and provides accurate basis for the formulation of heat dissipation strategies.

[0022] 4. The fault tolerance capability is greatly enhanced. The redundant switching design makes the response time of the heat dissipation component fault switching ≤50ms, the fault self-diagnosis accuracy rate ≥98%, and the mean time between failures (MTBF) of the device is improved by more than 300% compared with traditional devices.

[0023] 5. The closed-loop optimization mechanism improves adaptability. By dynamically adjusting the heat dissipation strategy through the feedback optimization layer, the device can operate stably in a temperature range of -40℃ to +55℃, in environments with 95% high humidity, and in areas with strong winds and sandstorms. The maintenance cycle is extended to 2 to 3 times that of traditional devices.

[0024] 6. Power quality is continuously optimized. With stable temperature control and intelligent reactive power compensation strategies, the power factor of the power grid can be improved to above 0.97 and the total harmonic distortion (THD) can be reduced to below 3%, further improving the stability and reliability of the power grid. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall module architecture and communication connection of the intelligent heat dissipation and collaborative control high and low voltage reactive power compensation device of the present invention. Figure 2 This is a schematic diagram of the hardware and algorithm hierarchy of the intelligent heat dissipation collaborative control module of the present invention; Figure 3 This is a block diagram illustrating the three-level heat dissipation coordinated triggering and fault redundancy switching principle of the redundancy switching heat dissipation execution module of the present invention; Figure 4 This is a schematic diagram of the closed-loop process of the intelligent heat dissipation collaborative control method of the present invention; Figure 5 This is a schematic diagram of the three-level collaborative control architecture of the present invention: device level, system level, and terminal level.

[0026] In the diagram, 1-Reactive power compensation main circuit module; 11-Capacitor bank; 12-Dry-type air-core reactor; 13-Composite switching switch; 2-Distributed adaptive sensor monitoring module; 21-Distributed sensor node; 22-Central data fusion unit; 3-Intelligent heat dissipation collaborative control module; 31-Main controller; 32-Coprocessor; 33-Dedicated heat dissipation control chip; 34-Fault self-diagnosis unit; 35-Thermal status early warning unit; 4-Redundant switching heat dissipation execution module; 41-Variable frequency fan subsystem; 42-Micro-pump liquid cooling subsystem; 43-Phase change material heat dissipation subsystem; 51-Data acquisition layer; 52-Feature enhancement layer; 53-Predictive decision layer; 54-Execution control layer; 55-Feedback optimization layer; 61-Equipment level; 62-System level; 63-Cloud level. Detailed Implementation

[0027] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. The embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0028] Example 1: Overall Assembly and Collaborative Operation of the Device In this embodiment, the intelligent heat dissipation and collaborative control high and low voltage reactive power compensation device is integrated into a standard 19-inch industrial cabinet (600mm×800mm×2000mm). The cabinet adopts a modular design specifically for power equipment. Each functional module is relatively independent but collaborates with each other through an industrial-grade internal bus. The overall modular architecture and communication connection relationship of the device are shown in Figure 1.

[0029] In this embodiment, the reactive power compensation main circuit module 1 is the core power module of the device. Its internal capacitor bank 11 has a total capacity of 150kvar and adopts a three-phase delta connection structure. The reactor 12 is a dry-type air-core reactor with a reactance rate of 7%. The composite switching switch 13 realizes zero-voltage switching and zero-current disconnection, effectively reducing the impact heat generation during the switching process. The intelligent heat dissipation and collaborative control module 3 serves as the control core of the device, using an STM32F767 main controller 31, a TMS320F28335 coprocessor 32, and a Xilinx controller. The Artix-7 FPGA dedicated heat dissipation control chip 33 features a master-slave hardware architecture. The three components communicate with shared memory via a high-speed SPI bus. The fault self-diagnosis unit 34 and the thermal status early warning unit 35 are integrated into this module to achieve fault diagnosis and thermal risk early warning. The distributed adaptive sensor monitoring module 2 is equipped with 24 distributed sensor nodes 21. The central data fusion unit 22 uses an STM32H7 microcontroller and communicates with the sensor nodes 21 via a CAN-FD bus. The redundant switching heat dissipation execution module 4 is equipped with a variable frequency fan subsystem 41 consisting of two variable frequency fans, a micro-pump liquid cooling subsystem 42 consisting of two micro-pumps, and a phase change material heat dissipation subsystem 43 consisting of a 2kg phase change material heat sink, forming a three-level heat dissipation collaborative system.

[0030] The device as a whole adopts a three-level collaborative control strategy of device level, system level, and cloud level, and its architecture is as follows: Figure 5 As shown: Device level 61 is completed by the local controller of four core modules for basic control and protection; system level 62 is dominated by intelligent heat dissipation collaborative control module 3, which uses model predictive control algorithm to optimize heat dissipation collaborative logic; cloud level 63 realizes dual-mode communication of Wi-Fi and Ethernet through ESP32-WROOM-32 IoT module, and uses MQTT protocol to complete remote monitoring and parameter configuration.

[0031] During normal operation, the distributed adaptive sensor monitoring module 2 collects multi-dimensional monitoring data according to the adaptive sampling mechanism. After processing by the central data fusion unit 22, the data is sent to the intelligent heat dissipation collaborative control module 3 via the internal bus. The intelligent heat dissipation collaborative control module 3 uses an improved deep learning adaptive algorithm to predict the thermal state of the device and formulate the optimal heat dissipation strategy. After receiving the control command, the redundant switching heat dissipation execution module 4 triggers the corresponding level of the heat dissipation subsystem to achieve collaborative heat dissipation according to the heat flux density threshold. The fault self-diagnosis unit 34 monitors the operating status of each heat dissipation component in real time and immediately triggers the redundant switching mechanism when a fault is detected. The feedback optimization layer 55 of the intelligent heat dissipation collaborative control module 3 dynamically adjusts the algorithm parameters according to the heat dissipation effect and fault information to form a closed-loop collaborative operation of the device, combining the overall operating logic. Figure 1 , Figure 2 , Figure 3 , Figure 4 accomplish.

[0032] Example 2: Execution of the intelligent heat dissipation collaborative control module algorithm This embodiment focuses on the algorithm operation logic and hardware working mechanism of the intelligent heat dissipation collaborative control module 3. The hardware and algorithm hierarchical architecture of the intelligent heat dissipation collaborative control module 3 is as follows: Figure 2 As shown, its software system is developed based on the FreeRTOS real-time operating system. It is divided into six independent tasks according to the algorithm level. Each task has a fixed execution cycle to ensure the real-time performance of data processing and control command output.

[0033] 1. Data acquisition task: The execution cycle is 1ms. It receives multi-dimensional monitoring data transmitted by the distributed adaptive sensor monitoring module 2 through the CAN-FD bus. The CRC check and range check mechanism is used to ensure data integrity. After the acquisition is completed, the data is transmitted to the data acquisition layer 51 of the algorithm architecture to provide a basis for subsequent feature extraction. 2. Feature Enhancement Task: The execution cycle is 10ms, corresponding to feature enhancement layer 52 of the algorithm architecture. It uses wavelet packet transform with 4-layer decomposition and improved principal component analysis algorithm to extract features from the collected raw data, strengthen the correlation between temperature gradient and load change features, remove invalid and redundant data, and improve the effectiveness of feature data. 3. Prediction and Decision Task: The execution cycle is 150ms, corresponding to prediction and decision layer 53 in the algorithm architecture. It adopts a Bi-LSTM network with 4 hidden layers and 80 neurons in each layer. It takes the feature data of the most recent 15 seconds as input and outputs the predicted value of the device thermal state for the next 45 seconds. At the same time, it combines the double Q-learning deep reinforcement learning algorithm, with temperature, humidity, load, etc. as the state space and fan speed level, micro-pump flow level, etc. as the action space, to formulate the optimal heat dissipation strategy that matches the thermal state. 4. Execution control task: The execution cycle is 10ms. The corresponding execution control layer 54 of the algorithm architecture uses the PID adaptive control algorithm to convert the heat dissipation strategy formulated by the prediction decision layer 53 into precise electrical control commands. These commands are sent to the redundant switching heat dissipation execution module 4 through the Xilinx Artix-7FPGA dedicated heat dissipation control chip 33. The FPGA realizes the microsecond-level precise output of control commands to meet the real-time requirements of industrial heat dissipation control. 5. Feedback optimization task: The execution cycle is 200ms. It corresponds to the feedback optimization layer 55 of the algorithm architecture. It collects the heat dissipation effect data of the redundant switching heat dissipation execution module 4 and the fault information of the fault self-diagnosis unit 34 in real time. It dynamically adjusts the core parameters of the Bi-LSTM network and the dual Q-learning algorithm, and optimizes the logic of heat dissipation strategy formulation to form a closed-loop optimization of the algorithm. The parameter adjustment signal of the feedback optimization layer 55 is simultaneously transmitted back to the data acquisition layer 51 and the prediction decision layer 53 to achieve optimization and adaptation at the whole algorithm level. 6. Communication Task: With an execution cycle of 1000ms, the Modbus TCP / IP protocol is used to realize the communication between the intelligent heat dissipation collaborative control module 3 and the external system and the cloud-based 63 IoT platform, and to complete data uploading and remote command reception.

[0034] In this embodiment, both the fault self-diagnosis unit 34 and the thermal state early warning unit 35 communicate with the main controller 31 in real time. The fault self-diagnosis unit 34 transmits the diagnosed sensor data anomalies, heat dissipation component failures, and other information to the main controller 31, and the main controller 31 simultaneously sends the information to the feedback optimization layer 55. The thermal state early warning unit 35, based on the thermal state prediction value output by the prediction decision layer 53, triggers an early warning through both the local indicator light and the cloud platform when the predicted temperature approaches the safety threshold of 80%. The early warning signal also serves as a reference for algorithm optimization.

[0035] Example 3: Data Processing of Distributed Adaptive Sensor Monitoring Module This embodiment focuses on the sensor node distribution, data acquisition, and processing mechanism of the distributed adaptive sensor monitoring module 2. This module provides accurate multi-dimensional monitoring data for the intelligent heat dissipation collaborative control module 3. Its module composition is as follows: Figure 1 As shown, the collected data is processed and then transmitted to... Figure 2 The data acquisition layer 51 of the intelligent heat dissipation collaborative control module 3 shown.

[0036] In this embodiment, the distributed adaptive sensor monitoring module 2 has a total of 24 sensor nodes 21, including 10 temperature sensor nodes, 4 humidity sensor nodes, 4 wind speed sensor nodes, and 6 current / voltage sensor nodes. The temperature sensor nodes adopt a combination of PT100 platinum resistance thermometer, MLX90614 infrared sensor and DS18B20 digital sensor, and are distributed in key heat-generating parts such as capacitor bank 11 and reactor 12, as well as the air inlet and outlet of the cabinet. The humidity sensor nodes are SHT31 high-precision sensors and are installed in the four corners inside the device. The wind speed sensor nodes are thermal wind speed sensors and are installed at the fan outlet of the variable frequency fan subsystem 41. The current and voltage sensor nodes are Hall effect sensors and are installed at the input and output terminals of the reactive power compensation main circuit module 1.

[0037] The central data fusion unit 22 uses an STM32H7 series microcontroller and receives data collected from 24 sensor nodes 21 via a 4-channel CAN-FD bus with a data transmission rate of ≥5Mbps. The central data fusion unit 22 first uses a Kalman filter algorithm to filter the raw data, eliminating abnormal data caused by environmental interference. Then, it improves the accuracy of the data through multi-source data calibration and fusion algorithms. The processed effective data is sent to the intelligent heat dissipation collaborative control module 3 in real time and stored on a local SD card, supporting the storage of historical data for the past 60 days and the function of data anomaly backtracking.

[0038] In this embodiment, sensor node 21 adopts an adaptive sampling mechanism: under normal operating conditions, the sampling frequency of all sensor nodes 21 is 1Hz, taking into account both monitoring requirements and energy consumption control; when a temperature fluctuation exceeding ±2℃ / min or a load change exceeding 10% is detected, the central data fusion unit 22 sends a control command, and the sampling frequency of all sensor nodes 21 is automatically increased to 15Hz, realizing high-precision monitoring under abnormal conditions and ensuring that the intelligent heat dissipation collaborative control module 3 can capture changes in the thermal state of the device in a timely manner.

[0039] Example 4: Redundancy switching and heat dissipation execution modules work together This embodiment focuses on the three-level heat dissipation collaborative triggering logic, fault redundancy switching, and heat dissipation capacity compensation mechanism of the redundancy switching heat dissipation execution module 4. The principle of the three-level heat dissipation collaborative triggering and fault redundancy switching of this module is as follows: Figure 3 As shown, its module composition is as follows Figure 1 As shown, receiving Figure 2 The intelligent heat dissipation collaborative control module 3 shown executes the control commands issued by the control layer 54.

[0040] In this embodiment, the variable frequency fan subsystem 41 of the redundant switching heat dissipation execution module 4 adopts a redundant design of two DeltaElectronics BFB0412VHD brushless DC fans with a speed adjustment range of 800~12000rpm. Stepless speed regulation is achieved through PWM control, and the noise is controlled below 32dB. The micro-pump liquid cooling subsystem 42 adopts a redundant configuration of two BartelsMikrotechnik mp6 micro-pumps. The coolant is an environmentally friendly nanofluid (deionized water and 1% volume fraction Al2O3 nanoparticles), with a thermal conductivity that is more than 40% higher than that of traditional coolants. The liquid cooling system is a fully sealed structure and is equipped with dual pressure and flow monitoring. The phase change material heat dissipation subsystem 43 uses paraffin-based phase change materials with a phase change temperature of 48~52℃, a latent heat value ≥185J / g, and a total mass of 2kg. It is encapsulated in an aluminum container and is in close contact with the capacitor bank 11 through thermally conductive silicone grease.

[0041] (I) Three-level heat dissipation coordinated triggering logic The redundant switching heat dissipation execution module 4 triggers the coordinated operation of different levels of heat dissipation subsystems according to the real-time heat flux density detection value of the device and a preset threshold. Specifically: 1. When the heat flux density of the device is less than 100W / cm², only the phase change material heat dissipation subsystem 43 works independently, and heat dissipation without energy consumption is achieved by absorbing the heat generated instantaneously by the phase change latent heat absorption device of the phase change material. 2. When the heat flux density of the device is between 100 and 500 W / cm², the phase change material heat dissipation subsystem 43 and the variable frequency fan subsystem 41 work together. The variable frequency fan subsystem 41 dynamically adjusts the fan speed according to the real-time temperature of the device to achieve a combination of forced convection heat dissipation and phase change heat dissipation. 3. When the heat flux density of the device exceeds 500W / cm², the three-stage heat dissipation subsystems of phase change material heat dissipation subsystem 43, variable frequency fan subsystem 41 and micro pump liquid cooling subsystem 42 operate simultaneously. The micro pump liquid cooling subsystem 42 quickly removes the high heat of the core heat-generating components of the device through closed-loop circulation of coolant. Combined with phase change heat dissipation and forced convection heat dissipation, efficient heat dissipation is achieved.

[0042] (II) Fault redundancy switching and heat dissipation compensation mechanism The fault self-diagnosis unit 34 monitors the operating status of the variable frequency fan subsystem 41, the micro-pump liquid cooling subsystem 42, and the phase change material heat dissipation subsystem 43 in real time, with a diagnostic accuracy of ≥98%. When a fault is detected, it immediately triggers a redundancy switching and heat dissipation capacity compensation mechanism. 1. When a single fan failure is detected in the variable frequency fan subsystem 41, the standby fan will start and switch within 50ms to maintain more than 75% of the heat dissipation capacity of the variable frequency fan subsystem 41. 2. When a failure of the main micropump in the micropump liquid cooling subsystem 42 is detected, the backup micropump will immediately switch over. The flow rate and pressure will be monitored to ensure the stability of the liquid cooling system flow and maintain the heat dissipation capacity of the micropump liquid cooling subsystem 42 by more than 75%. 3. When a fault is detected in the phase change material heat dissipation subsystem 43, since there are no spare redundant components, the intelligent heat dissipation coordination control module 3 immediately issues an instruction to increase the fan speed of the variable frequency fan subsystem 41 and the micro pump flow rate of the micro pump liquid cooling subsystem 42. By actively increasing the operating power of the other two heat dissipation subsystems, the heat dissipation capacity loss of the phase change material heat dissipation subsystem 43 is compensated, ensuring that the overall heat dissipation effect of the device meets the operating requirements.

[0043] In this embodiment, the real-time heat dissipation effect data of the redundant switching heat dissipation execution module 4 will be continuously fed back to the feedback optimization layer 55 of the fault self-diagnosis unit 34 and the intelligent heat dissipation collaborative control module 3, providing a basis for fault diagnosis and algorithm parameter optimization.

[0044] Example 5: Verification of Practical Application Effect To verify the practical application effect of the intelligent heat dissipation and coordinated control high and low voltage reactive power compensation device of this invention, the device was applied to a reactive power compensation retrofit project of a 100MW wind farm. This wind farm is located in a high-altitude area, where winter temperatures can reach -35℃ and summer temperatures can reach 42℃, with large wind speed fluctuations. This places stringent requirements on the heat dissipation stability and fault tolerance of the reactive power compensation device. The original reactive power compensation device at this wind farm used a traditional single air-cooling method, resulting in 8 failures per year, 6 of which were due to heat dissipation system failures. The grid power factor could only be maintained at around 0.90, and the total harmonic distortion (THD) was higher than 5%.

[0045] The modified high and low voltage reactive power compensation device using the intelligent heat dissipation collaborative control of this invention is constructed as described in Example 1. The intelligent heat dissipation collaborative control module algorithm is set as described in Example 2. The distributed adaptive sensor monitoring module data processing is configured as described in Example 3. The redundant switching heat dissipation execution module collaborative operation is debugged as described in Example 4. The overall closed-loop control logic of the device is combined. Figure 1-5 Implementation. After 18 months of continuous operation and testing, the following significant application results were achieved: 1. Significantly improved heat dissipation: The maximum temperature of the core heat-generating components (capacitor bank, reactor) is controlled below 58℃, the temperature fluctuation range is reduced by 70% compared to the original device, the heat flux density processing range is expanded to 30~1200W / cm², and the heat dissipation efficiency is improved by 25% compared to the original device; 2. Significantly reduced failure rate: The average number of failures per year of the device has been reduced to 1, and all of them are grid-side line failures. There are no failures related to the heat dissipation system. The over-temperature failure rate has been reduced by more than 80% compared to the original device. The failure switching response time of the heat dissipation components is ≤50ms. The mean time between failures (MTBF) of the device has been increased by more than 300% compared to the traditional device. 3. Energy consumption and maintenance cost optimization: The distributed adaptive sampling mechanism reduces sensor monitoring energy consumption by 45%, and the on-demand tiered operation of the heat dissipation system reduces heat dissipation energy consumption by 40%. The overall annual power consumption of the device is reduced from the original 7200kWh to 3960kWh. The device maintenance cycle is extended to 2 years / time, which is 3 times higher than the original 6 months / time, and the maintenance cost is reduced by 70%. 4. Continuous optimization of power quality: Stable temperature control keeps the reactive power compensation strategy of the device in the optimal state, the power factor of the grid is stable above 0.97, and the total harmonic distortion (THD) is reduced to below 3%, which effectively improves the grid stability and reliability of the wind farm. 5. Strong adaptability to complex working conditions: The device can operate stably in a wide temperature range of -35℃ to 42℃, a high humidity environment of 95%, and a strong wind and sand area in the wind farm, which verifies the high reliability and strong adaptability of the device under complex and harsh working conditions.

[0046] The practical application results of this embodiment show that the intelligent heat dissipation coordinated control high and low voltage reactive power compensation device of the present invention effectively solves the problems of poor heat dissipation coordination, delayed fault response, and high monitoring energy consumption of traditional reactive power compensation devices, significantly improves the operational reliability of the device under complex working conditions, and can continuously optimize the power quality of the power grid, thus having significant economic and social benefits.

Claims

1. A high- and low-voltage reactive power compensation device with intelligent heat dissipation and coordinated control, characterized in that, It includes a reactive power compensation main circuit module, an intelligent heat dissipation collaborative control module, a distributed adaptive sensor monitoring module, and a redundancy switching heat dissipation execution module that are connected in sequence via communication. The reactive power compensation main circuit module includes a capacitor bank, a reactor, and a composite switching switch. The capacitor bank adopts a three-phase delta connection structure, the reactor is a dry-type air-core reactor, and the composite switching switch realizes zero-voltage switching and zero-current disconnection. The intelligent heat dissipation collaborative control module includes a main controller, a coprocessor, a dedicated heat dissipation control chip, a fault self-diagnosis unit, and a thermal status early warning unit. The distributed adaptive sensor monitoring module includes multiple distributed sensor nodes and a central data fusion unit. The sensor nodes cover temperature, humidity, wind speed, current, and voltage sensors. The redundant switching heat dissipation execution module includes a variable frequency fan subsystem, a micro-pump liquid cooling subsystem, and a phase change material heat dissipation subsystem, which together constitute a three-level heat dissipation mechanism triggered in a coordinated manner.

2. The high and low voltage reactive power compensation device with intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The intelligent heat dissipation collaborative control module adopts an improved deep learning adaptive algorithm, which includes a data acquisition layer, a feature enhancement layer, a prediction decision layer, an execution control layer, and a feedback optimization layer. The data acquisition layer receives multi-dimensional data from the distributed adaptive sensor monitoring module; The feature enhancement layer uses wavelet packet transform and improved principal component analysis algorithm to extract key features and enhance the correlation between temperature gradient and load change. The prediction decision layer uses a combination of a bidirectional long short-term memory network (Bi-LSTM) and a deep reinforcement learning algorithm. It takes the feature data of the most recent 15 seconds as input and outputs the thermal state prediction value for the next 45 seconds to formulate the optimal heat dissipation strategy. The execution control layer converts the heat dissipation strategy into precise control commands and sends them to the redundant switching heat dissipation execution module. The feedback optimization layer dynamically adjusts the algorithm parameters based on the heat dissipation effect and fault information.

3. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 2, characterized in that, The bidirectional long short-term memory network has four hidden layers, each with 80 neurons, and introduces an attention mechanism to strengthen the weight of hot features during critical periods. The deep reinforcement learning adopts the dual Q-learning algorithm. The state space includes temperature, humidity, load, wind speed and component operating status parameters, the action space includes fan speed level, micro-pump flow level, phase change heat dissipation trigger threshold, and the reward function integrates heat dissipation efficiency, energy consumption level and component life protection.

4. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The sensor nodes of the distributed adaptive sensor monitoring module adopt an adaptive sampling mechanism: Under normal operating conditions, the sampling frequency is 1Hz. When the temperature fluctuation exceeds ±2℃ / min or the load change exceeds 10%, the sampling frequency is automatically increased to 15Hz. The temperature sensor is a combination of PT100 platinum resistance, MLX90614 infrared sensor and DS18B20 digital sensor; the humidity sensor is SHT31 high-precision sensor; the wind speed sensor is thermal wind speed sensor; and the current and voltage sensors are Hall effect sensors. The central data fusion unit uses an STM32H7 series microcontroller and improves the reliability of monitoring data through multi-source data calibration and fusion algorithms.

5. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The three-level heat dissipation coordination mechanism of the redundant switching heat dissipation execution module satisfies: When the heat flux density of the device is below 100W / cm², only the phase change material heat dissipation subsystem is working. When the heat flux density is between 100 and 500 W / cm², the phase change material heat dissipation subsystem and the variable frequency fan subsystem work together. When the heat flux density exceeds 500 W / cm², the three-stage heat dissipation subsystem operates simultaneously; The variable frequency fan subsystem adopts a dual-fan redundancy design, and the switching response time is ≤50ms when a single fan fails. The micropump liquid cooling subsystem is equipped with a backup micropump, which enables rapid fault switching through flow monitoring.

6. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The fault self-diagnosis unit of the intelligent heat dissipation collaborative control module achieves accurate diagnosis of fan failure, micro-pump malfunction, filter blockage, and coolant leakage by monitoring abnormal sensor data, deviation of component operating parameters from thresholds, and attenuation of heat dissipation effect, with a diagnostic accuracy rate of ≥98%. The thermal state early warning unit triggers an early warning when the predicted temperature approaches 80% of the safety threshold, based on the output of the prediction decision layer, and provides dual reminders through local indicator lights and the cloud platform.

7. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The device employs a three-level collaborative control strategy: device level, system level, and cloud level. Device-level control enables local basic control and protection for each module; System-level control is led by the intelligent heat dissipation collaborative control module, which uses model predictive control algorithms to optimize the heat dissipation collaborative logic; The cloud-based control supports dual-mode communication via Wi-Fi and Ethernet, and uses the MQTT protocol to achieve remote monitoring, parameter configuration, fault alarms, and OTA firmware upgrades.

8. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The variable frequency fan subsystem has a fan speed adjustment range of 800~12000rpm, and stepless speed regulation is achieved through PWM control, with noise controlled below 32dB; The micro-pump liquid cooling subsystem uses environmentally friendly nano-cooling fluid, which has a thermal conductivity that is more than 40% higher than that of traditional coolants. The liquid cooling system is a fully sealed structure and is equipped with dual monitoring of pressure and flow. The phase change material of the phase change material heat dissipation subsystem has a phase change temperature of 48℃~52℃ and a latent heat value of ≥185J / g.

9. The high and low voltage reactive power compensation device for intelligent heat dissipation and coordinated control according to claim 1, characterized in that, The sensor nodes of the distributed adaptive sensor monitoring module communicate with the central data fusion unit via a CAN-FD bus, with a data transmission rate of ≥5Mbps; The central data fusion unit supports the storage of historical data for the most recent 60 days and has the function of backtracking data anomalies.

10. A smart heat dissipation control method based on the device according to any one of claims 1-9, characterized in that, Includes the following steps: (1) The distributed adaptive sensor monitoring module collects multi-dimensional data according to the adaptive sampling mechanism, and sends the data to the intelligent heat dissipation collaborative control module after processing by the central data fusion unit. (2) The intelligent heat dissipation collaborative control module predicts the trend of thermal state change of the device and formulates the initial heat dissipation strategy through an improved deep learning adaptive algorithm; (3) The redundant switching heat dissipation execution module starts the corresponding level of heat dissipation subsystem according to the initial heat dissipation strategy to achieve coordinated heat dissipation; (4) The fault self-diagnosis unit monitors the operating status of the heat dissipation components in real time and triggers the redundancy switching mechanism when a fault occurs to ensure the continuity of heat dissipation. (5) The feedback optimization layer dynamically adjusts the algorithm parameters and heat dissipation strategy based on the heat dissipation effect data to form a closed-loop optimization.

Images