A double circulation self-adaptive liquid cooling system for charging piles

Abstract

This invention relates to the field of liquid cooling technology for charging piles, and discloses a dual-cycle adaptive liquid cooling system for charging piles. The system includes a main circulation cooling loop, an auxiliary circulation regulation loop, an operating condition database, a feature prediction module, a primary regulation module, a secondary regulation module, and a collaborative decision-making module. The main circulation loop provides basic heat dissipation for the charging pile power module; the auxiliary circulation loop dynamically adjusts the heat exchange rate. The feature prediction module outputs predicted temperature features based on the operating condition database and a deep learning model; the primary regulation module determines whether to initiate primary flow regulation and calculates the initial flow rate adjustment value accordingly; the secondary regulation module determines whether to initiate secondary dynamic flow rate regulation based on actual outlet temperature changes; and the collaborative decision-making module performs collaborative decision-making based on the status of multiple charging piles within the region to generate an optimal temperature balancing strategy. Through a dual-layer regulation combining prediction and feedback, and regional collaborative decision-making, precise and adaptive control of the charging pile's heat dissipation capacity is achieved.

Description

Technical Field

[0001] This invention relates to the field of liquid cooling technology for charging piles, specifically a dual-cycle adaptive liquid cooling system for charging piles. Background Technology

[0002] During operation, high-power charging piles generate significant heat from their power modules, making effective thermal management crucial for safe and stable operation. Existing liquid cooling systems for charging piles generally employ a single-cycle cooling structure, using a pump to drive coolant through a radiator for basic heat dissipation from the power modules. While this single-cycle system is relatively simple, its fixed heat dissipation capacity is insufficient to handle rapidly changing charging power and complex environmental conditions. At the initial stage of charging or during sudden increases in power demand, the module's heat generation increases dramatically, and the fixed heat dissipation capacity may lead to insufficient cooling, causing a rapid rise in module temperature and posing an overheating risk. Conversely, at the end of charging or during low-power operation, the fixed heat dissipation capacity can result in overcooling, leading to energy waste. Furthermore, changes in ambient temperature significantly affect the effectiveness of fixed-parameter cooling systems, which lack adaptive adjustment capabilities.

[0003] Conventional improvement solutions involve adding temperature sensors to the system and employing simple feedback control logic, such as increasing the speed of the circulating pump or activating the auxiliary fan when the outlet temperature exceeds a threshold. This reactive, lagging response strategy always adjusts based on already occurring temperature changes, resulting in a delayed response and difficulty in accurately controlling temperature, particularly in scenarios with rapidly fluctuating charging power. When multiple charging piles are deployed centrally, independently operating thermal management systems cannot optimize energy efficiency from a holistic perspective. This can lead to situations where some charging piles are operating at full capacity while adjacent charging piles are idle, resulting in excessively high localized heat loads or increased overall energy consumption. Current technologies lack the ability to predict the charging process in advance, cannot adjust cooling intensity based on upcoming charging demands, and cannot coordinate the operation of multiple charging piles at the regional level to achieve optimal global energy efficiency. The intelligence, adaptability, and collaborative capabilities of charging pile thermal management systems need further improvement. Summary of the Invention

[0004] The purpose of this invention is to provide a dual-cycle adaptive liquid cooling system for charging piles to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a dual-cycle adaptive liquid cooling system for charging piles, the system comprising:

[0006] The main cooling loop is used for basic heat dissipation of the charging pile power module, and includes a coolant circulation pump and a radiator.

[0007] The auxiliary circulation regulating loop is used to dynamically regulate the electrolyte heat exchange rate and includes an electrolyte storage tank, a flow regulating valve, and a heat exchanger.

[0008] The operating condition database stores historical charging process records, including peak charging demand, ambient temperature and humidity, main circulation inlet temperature, main circulation outlet temperature, and auxiliary circulation flow rate.

[0009] The feature prediction module is used to collect the charging demand characteristics and ambient temperature and humidity of the current charging pile, and combine the operating condition database and deep learning model to output the predicted temperature characteristics of the current charging process.

[0010] The primary control module is used to determine whether to initiate a flow regulation based on the predicted temperature characteristics, and to calculate the initial flow rate adjustment value of the auxiliary circulation regulation loop;

[0011] The secondary control module is used to calculate the actual temperature change characteristics based on the actual outlet temperature change of the main circulation cooling loop, and compare them with the predicted temperature characteristics to determine whether to start the secondary dynamic flow rate regulation.

[0012] The collaborative decision-making module is used to receive status data of multiple charging piles in the area, make collaborative decisions based on the operating condition database, generate the optimal temperature equalization strategy, and distribute it to the primary control module and secondary control module of each charging pile.

[0013] Preferably, the data acquisition unit of the operating condition database performs the following operations:

[0014] Collect complete records for each charging process, including charging start time, peak charging demand, ambient temperature and humidity, main circulation inlet temperature timing, main circulation outlet temperature timing, and auxiliary circulation set flow rate;

[0015] Records whose maximum and minimum values ​​of the main circulation outlet temperature time series are within the safe temperature range are selected to form a standard record set;

[0016] Calculate the time interval from the charging start time to the temperature stabilization time for each record in the standard record set, and use it as the first time reference parameter;

[0017] The average increment of the main cycle outlet temperature in the standard record set from the start of charging to the time when the temperature stabilizes is extracted and used as the first temperature reference threshold.

[0018] Calculate the average rate of change of the main circulation outlet temperature per unit time interval in each record, and use it as a characteristic of historical temperature changes.

[0019] Preferably, the feature extraction unit of the feature prediction module performs the following operations:

[0020] Real-time monitoring of the charging start time and main cycle outlet temperature sequence during the current charging process;

[0021] When the main circulation outlet temperature exceeds the first temperature reference threshold, mark that moment as the temperature monitoring start point;

[0022] The difference between the main circulation outlet temperature corresponding to the temperature monitoring start point and the charging start point temperature is calculated and used as the current temperature increment.

[0023] Based on the current peak charging demand and ambient temperature and humidity, extract matching historical temperature change features from the operating condition database;

[0024] The current peak charging demand, ambient temperature and humidity, current temperature increment, and the initial set flow rate of the auxiliary circulation regulation loop are input into the deep learning model, which outputs the predicted temperature change characteristics.

[0025] Preferably, the control determination unit of the primary control module performs the following operations:

[0026] The product of the predicted temperature change feature and the first time reference parameter is calculated as the predicted temperature offset.

[0027] The first temperature reference threshold is added to the predicted temperature offset to obtain the control determination parameter;

[0028] When the control judgment parameter exceeds the safe temperature range, a flow control command is triggered.

[0029] The difference between the lower limit of the safe temperature range and the first temperature reference threshold is calculated and used as the demand adjustment amount;

[0030] The ratio of the demand adjustment amount to the first time reference parameter is calculated and used as the target adjustment feature.

[0031] Preferably, the flow rate calculation unit of the primary control module performs the following operations:

[0032] Input the current peak charging demand, ambient temperature and humidity, and current temperature increment into the deep learning model;

[0033] Using the target regulation features as output constraints, the target flow velocity value of the auxiliary circulation regulation loop is obtained by inversion calculation through the deep learning model.

[0034] The initial set flow rate of the auxiliary circulation regulating loop is updated to the target flow rate value.

[0035] Preferably, the error analysis unit of the secondary control module performs the following operations:

[0036] Starting from the temperature monitoring start point, the main circulation outlet temperature is collected at fixed time intervals;

[0037] Calculate the rate of change of the main circulation outlet temperature between adjacent time intervals as a characteristic of actual temperature change;

[0038] The absolute difference between the actual temperature change characteristics and the predicted temperature change characteristics is calculated and used as a dynamic adjustment judgment quantity.

[0039] When the dynamic adjustment judgment value exceeds the preset error threshold, a secondary dynamic flow rate adjustment command is triggered.

[0040] Preferably, the dynamic adjustment unit of the secondary control module performs the following operations:

[0041] When the actual temperature change characteristic is lower than the predicted temperature change characteristic, the current flow rate of the auxiliary circulation regulation loop is reduced by a preset ratio.

[0042] When the actual temperature change characteristic is higher than the predicted temperature change characteristic, the current flow rate of the auxiliary circulation regulation loop is increased by a preset ratio.

[0043] The preset ratio is equal to the product of the preset weight coefficient and the dynamic adjustment judgment quantity.

[0044] Preferably, the state synchronization unit of the collaborative decision-making module performs the following operations:

[0045] The current peak charging demand, ambient temperature and humidity, main circulation outlet temperature, and auxiliary circulation real-time flow rate of each charging pile within the receiving area;

[0046] When the main circulation outlet temperature of any charging pile exceeds the first preset temperature difference threshold, the charging pile is marked as a high load node.

[0047] When the main circulation outlet temperature of any charging pile is lower than the second preset temperature difference threshold, the charging pile is marked as a low load node.

[0048] Preferably, the strategy generation unit of the collaborative decision-making module performs the following operations:

[0049] Based on the historical load correlation of each charging pile in the operating condition database, a dynamic weight vector is generated;

[0050] The global temperature compensation is obtained by weighting the temperature increment of all high-load nodes in the region based on the dynamic weight vector.

[0051] Calculate the flow rate compensation coefficient for each low-load node based on the global temperature compensation amount;

[0052] The flow rate compensation coefficient is distributed to the primary control module of the corresponding charging pile.

[0053] Preferably, the optimization execution unit of the collaborative decision-making module performs the following operations:

[0054] Real-time monitoring of the main circulation outlet temperature change rate after each charging pile applies the aforementioned flow rate compensation coefficient;

[0055] When the overall temperature change rate of the region does not reach the expected convergence rate, the dynamic weight vector is recalculated.

[0056] A new global temperature compensation amount is generated iteratively based on the updated dynamic weight vector.

[0057] The iteratively generated global temperature compensation is input into the secondary control module for dynamic flow rate fine-tuning.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] By setting up a main cooling loop responsible for basic heat dissipation and an independent auxiliary circulation regulating loop to dynamically adjust the heat exchange rate with the electrolyte, a dual-layer structure is formed, separating basic heat dissipation from fine-tuning temperature control. The main circulation ensures continuous and stable basic heat dissipation capacity, while the auxiliary circulation acts as a dynamic compensation unit. Primary initial regulation of the auxiliary circulation flow rate is performed based on predicted temperature characteristics, realizing a shift from passive response to active intervention. This structure allows the system to proactively adjust heat exchange conditions according to predicted heat dissipation demands, preparing for upcoming changes in heat load. Once the actual charging process begins, changes in the main circulation outlet temperature are monitored in real time. By calculating the actual temperature change characteristics and comparing them with the predicted value, secondary dynamic flow rate regulation is triggered. This process constitutes a dual-layer closed-loop control combining prediction and feedback. The prediction front end provides pilot adjustment commands, while the feedback back end performs real-time calibration and fine-tuning. This mechanism effectively overcomes the hysteresis of simple feedback control, enabling pre-adjustment of heat dissipation capacity before significant fluctuations in heat load occur, controlling the temperature fluctuations of the power module within a narrower range, and improving temperature stability. At the same time, it avoids energy loss from maintaining high heat dissipation intensity when the heat load is low, and precisely matches cooling resources according to actual needs, thereby reducing the system's operating energy consumption.

[0060] A collaborative decision-making mechanism is introduced to collect real-time status data and historical operating condition data from multiple charging piles within the region, and perform centralized analysis and collaborative calculations. This mechanism can perceive the thermal field distribution and grid load conditions of the entire region, generating a temperature balancing strategy aimed at optimizing overall operating efficiency. The collaborative decision-making module analyzes the current charging stage, power level, thermal state, and environmental conditions of each charging pile to formulate strategies. For example, when the total heat load is close to its limit, it can appropriately increase the allowable temperature limit of charging piles that are about to complete to reduce their cooling energy consumption, or prioritize the allocation of heat dissipation resources to charging piles with more urgent heat loads. This minimizes the total heat dissipation energy consumption of the region while ensuring the safe operation of each charging pile. This collective intelligent optimization breaks the limitations of the independent operation of a single charging pile thermal management system, coordinating resource allocation at the system level, avoiding local overheating or competition for cooling capacity, and improving the overall energy efficiency and operational economy of the charging station cluster. The strategies generated by the collaborative decision-making are distributed to the primary and secondary control modules of each charging pile, guiding their individual control behaviors and ensuring that individual goals are consistent with the overall goals. Attached Figure Description

[0061] Figure 1 This is a timing diagram of the dual-cycle adaptive liquid cooling system for charging piles described in this invention.

[0062] Figure 2 This is a flowchart of the data acquisition and processing for the operating condition database;

[0063] Figure 3 The flowchart for the control determination of the primary control module;

[0064] Figure 4 The flowchart for error analysis of the secondary control module;

[0065] Figure 5 A flowchart for generating strategies for the collaborative decision-making module. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] Please see Figure 1 The present invention provides a dual-cycle adaptive liquid cooling system for charging piles, the system comprising: a main circulation cooling loop, an auxiliary circulation adjustment loop, an operating condition database, a feature prediction module, a primary control module, a secondary control module, and a collaborative decision-making module.

[0068] The main circulation cooling loop uses a coolant circulation pump to drive coolant flow through the radiator, providing basic cooling for the charging pile's power modules. The auxiliary circulation regulating loop stores electrolyte in an electrolyte tank, controls its flow rate via a flow regulating valve, and exchanges heat with the main circulation loop through a heat exchanger, achieving dynamic adjustment of the electrolyte heat exchange rate. The operating condition database stores historical charging process records, including peak charging demand, ambient temperature and humidity, main circulation inlet temperature, main circulation outlet temperature, and auxiliary circulation flow rate. The feature prediction module collects real-time charging demand characteristics and ambient temperature and humidity of the current charging pile, combines this data with historical data from the operating condition database and a deep learning model, and outputs predicted temperature characteristics for the current charging process. The first-level control module determines whether to initiate primary flow regulation based on the predicted temperature characteristics and calculates the initial flow rate adjustment value for the auxiliary circulation regulating loop. The second-level control module calculates the actual temperature change characteristics based on the actual outlet temperature change of the main circulation cooling loop, compares it with the predicted temperature characteristics, and determines whether to initiate secondary dynamic flow rate regulation. The collaborative decision-making module receives status data from multiple charging piles within the area, performs collaborative decision-making based on the operating condition database, generates the optimal temperature balancing strategy, and distributes it to the primary and secondary control modules of each charging pile.

[0069] Example 1: See Figure 2 In the actual operation of the dual-cycle adaptive liquid cooling system for charging piles, data acquisition and feature prediction constitute the starting point for intelligent system control. The data acquisition unit of the operating condition database continuously collects information through a sensor network deployed at key nodes of the charging pile. These sensors record detailed parameters of each charging process at high frequency, including the charging start time accurate to milliseconds, real-time fluctuating peak charging demand, changing ambient temperature and humidity data, continuous temperature time-series readings at the inlet and outlet of the main circulation loop, and the set flow rate of the electrolyte in the auxiliary circulation loop. All this data is timestamped and transmitted to the central storage system, forming the raw data pool. The data cleaning process begins, its primary task being to filter out records from the massive amount of data where the main circulation outlet temperature consistently remains within a safe operating range. This filtering process does not simply eliminate outliers, but rather identifies complete charging cycle data that meets heat dissipation requirements without overcooling by establishing a safe boundary model for temperature changes. This high-quality data is archived as a standard record set.

[0070] After forming a standard record set, the system begins extracting representative feature parameters from it. For each qualified charging record, the algorithm precisely calculates the time elapsed from the start of charging to the temperature reaching a steady state. This time interval reflects the thermal inertia characteristics of the system under that specific operating condition and is stored as the first time reference parameter. Simultaneously, the system analyzes the change in the main circulation outlet temperature throughout the entire temperature rise process, calculating the average temperature increment from the starting point to the steady point. This value reflects the average load level of the cooling system and is defined as the first temperature reference threshold. Furthermore, the system performs differential analysis on the temperature change curve, calculating the rate of temperature change per unit time, thereby obtaining historical temperature change characteristics describing the system's dynamic response. All these feature parameters are categorized and stored in an index table of the operating condition database for rapid retrieval and matching.

[0071] When a new charging process begins, the feature prediction module immediately activates. This module continuously monitors the charging pile's current operating parameters via a real-time data interface, including the precise start-up time and the real-time data stream from the main circulation outlet temperature sensor. The system compares the current reading with a previously stored first temperature reference threshold. Once the outlet temperature exceeds this threshold, this moment is immediately marked as the temperature monitoring start point, indicating that the system has confirmed the heat load has reached a level requiring close monitoring. From this starting point, the system calculates the difference between the current temperature and the temperature at the start of charging, obtaining the current temperature increment. This value reflects the actual heat load generated so far.

[0072] Based on the current peak charging demand and ambient temperature and humidity conditions, the feature prediction module begins pattern matching in the operating condition database. The retrieval algorithm searches for historical data records with similar charging demands and environmental conditions, extracting corresponding historical temperature change features. These historical features represent the system's past behavior patterns under similar conditions. All these parameters—including real-time collected peak charging demand, ambient temperature and humidity readings, calculated current temperature increments, and the current actual set flow rate of the auxiliary circulation loop—are fed into a pre-trained deep learning model.

[0073] This deep learning model employs a deep neural network architecture, containing multiple hidden layers and nonlinear activation functions, enabling it to capture the complex nonlinear relationship between input parameters and temperature changes. During training, the model utilizes a large amount of historical data, iteratively adjusting its internal weight parameters to learn how to predict future temperature change trends based on input conditions. When current parameters are input, the model performs multi-layered transformations and feature extraction on this data during forward propagation, ultimately outputting a predicted temperature change feature. This feature value represents the expected rate and trend of system temperature change under current conditions. The calculation of the predicted temperature change feature can be represented using the following relationship:

[0074]

[0075] in, Represents the predicted temperature change characteristics. Represents the current peak charging demand. Represents ambient temperature and humidity. Represents the current temperature increment. This represents the initial set flow rate of the auxiliary circulation control loop. (Function) This represents the nonlinear mapping relationship established by the deep learning model. The entire prediction process is dynamic and continuous; as new real-time data is continuously input, the prediction results are updated accordingly, ensuring that the system always makes judgments based on the latest information.

[0076] Example 2: See Figure 3 In the operation of the dual-cycle adaptive liquid cooling system of the charging pile, the primary control module undertakes the crucial task of initial flow regulation based on prediction results. The control determination unit of the primary control module first processes the predicted temperature change characteristic data from the feature prediction module. This predicted characteristic value represents the system's estimate of the temperature change rate over a future period. By multiplying it with the first time reference parameter extracted from the operating condition database, the predicted temperature offset is calculated. This offset quantifies the total temperature change amplitude that may occur throughout the entire temperature change cycle. The system adds the first temperature reference threshold to this predicted temperature offset to obtain the control determination parameter, which actually represents the final temperature level that the system may reach under the current predicted conditions. The control determination unit compares this calculated control determination parameter with a pre-set safe temperature range, which is determined based on the material properties and operating requirements of the charging pile power module. If the determination parameter exceeds the upper or lower limit of the safe range, the system will trigger a flow regulation command. This means that the current heat dissipation conditions may not be able to control the temperature within the allowable range, and it is necessary to intervene in the temperature change process by adjusting the auxiliary circulation flow rate.

[0077] After determining that flow regulation is necessary, the system further calculates the demand adjustment amount. This process involves subtracting the lower limit of the safe temperature range from the first temperature reference threshold; the difference represents the temperature drop that needs to be avoided. This demand adjustment amount reflects the amount of temperature change the system needs to compensate for. The system divides this demand adjustment amount by the first time reference parameter to obtain the target regulation characteristic. This target regulation characteristic actually defines a new, more moderate target rate of temperature change, according to which the system can eventually stabilize within the safe temperature range. The entire calculation process fully considers the system's thermal inertia and response characteristics, ensuring that the regulation target can achieve temperature control without causing excessive oscillations in the system.

[0078] The flow rate calculation unit of the primary control module then begins operation, its task being to transform the calculated target regulation features into specific flow rate adjustment commands. This unit again invokes the deep learning model, but this time in a different manner than in the feature prediction stage. The flow rate calculation unit passes the currently monitored peak charging demand, ambient temperature and humidity readings, and the calculated current temperature increment as input parameters to the model. Then, using the target regulation features as output constraints, it solves for the required auxiliary circulation target flow rate value through the model's inversion calculation process. This inversion calculation process essentially seeks a set of input parameters that enable the model to output a result that matches the target regulation features. Mathematically, this is equivalent to solving an optimization problem: finding the optimal combination of input parameters given the output target.

[0079] During the inversion calculation process, the deep learning model adjusts its internal parameters and weights, iteratively approaching the target output value. Each iteration evaluates the difference between the current output and the target regulation feature, adjusting the input parameters accordingly. This process continues until a combination of input parameters is found that allows the model output to closely approximate the target regulation feature, including the target flow velocity value of the auxiliary circulation. This calculation process fully considers the complex interrelationships between various system parameters, ensuring that the obtained flow velocity value not only achieves the temperature control target but also coordinates with other system parameters.

[0080] After obtaining the target flow rate value, the system updates the initial set flow rate of the auxiliary circulation regulating loop to this newly calculated target value. This update process sends commands to the flow regulating valve via the control interface, adjusting its opening and thus changing the flow rate of the electrolyte in the auxiliary circulation loop. The flow regulating valve uses a high-precision actuator, capable of precisely adjusting its opening according to the control signal to achieve smooth flow rate changes. Changes in flow rate directly affect the heat exchange efficiency in the heat exchanger, thereby altering the heat dissipation rate of the coolant in the main circulation loop. The entire adjustment process employs a gradual adjustment strategy to avoid sudden, large changes that could cause system oscillations.

[0081] While adjusting the flow rate, the system continuously monitors changes in the main circulation outlet temperature, comparing the actual temperature changes with the predicted values. This monitoring process provides feedback for potential further adjustments and also accumulates new operational data for the operating condition database. All parameter changes, command execution, and temperature response data during the adjustment process are meticulously recorded. This data will be used to optimize the accuracy of the deep learning model and the precision of the inversion calculations. Through this continuous learning and optimization mechanism, the system can continuously improve the accuracy and response speed of its flow control, gradually adapting to various operating conditions and environmental changes.

[0082] The entire primary control process embodies a combination of predictive control and feedback regulation. The system not only reacts based on the current state but, more importantly, intervenes in advance using predictive information. This method effectively overcomes the system's thermal inertia delay, implementing regulatory measures before the temperature significantly deviates from the safe range, thereby maintaining temperature stability within the ideal range. Through inversion calculations using a deep learning model, the system can find the optimal flow rate value to achieve the temperature control target, avoiding the oscillation and overshoot problems that may arise from traditional trial-and-error regulation. This intelligent regulation method ensures that the charging pile power module receives appropriate thermal management under various operating conditions, preventing performance degradation or damage caused by overheating and avoiding energy waste caused by excessive cooling.

[0083] Example 3: See Figure 4 During the operation of the dual-cycle adaptive liquid cooling system of the charging pile, the secondary control module operates on the basis of the primary control. By continuously monitoring the actual temperature change and comparing it with the predicted value, it achieves dynamic fine-tuning of the auxiliary circulation flow rate to cope with the deviation between the actual operating conditions and the prediction model.

[0084] The error analysis unit of the secondary control module collects main circulation outlet temperature data at fixed time intervals, starting from the initial temperature monitoring point. This acquisition process is performed at a high frequency to ensure the capture of detailed temperature change characteristics. The system calculates the temperature change between two adjacent time points and divides it by the time interval to obtain the actual temperature change characteristic value. This calculation process is continuous, forming a series of data points reflecting the instantaneous rate of temperature change. Simultaneously, the system compares these actually measured temperature change characteristic values ​​with the predicted temperature change characteristics output by the previous feature prediction module point by point, calculating the absolute difference between the two. This difference quantifies the deviation between the actual system behavior and the expected behavior and is defined as the dynamic adjustment judgment quantity. When the dynamic adjustment judgment quantity continuously exceeds a preset error threshold, it indicates a significant difference between the actual temperature change trend and the predicted value, and the system therefore triggers a secondary dynamic flow rate adjustment command. This error threshold is derived from statistical analysis of historical system operating data and reflects the acceptable deviation range between the predicted and actual values ​​under normal operating conditions.

[0085] Upon receiving an adjustment command, the dynamic adjustment unit of the secondary control module executes corresponding adjustment operations based on the comparison between the actual temperature change characteristics and the predicted values. When the actual temperature change characteristics are lower than the predicted temperature change characteristics, it indicates that the system's heat dissipation effect is better than expected. In this case, the current flow rate of the auxiliary circulation control loop can be appropriately reduced to decrease energy consumption and avoid overcooling. Conversely, when the actual temperature change characteristics are higher than the predicted temperature change characteristics, it indicates that the system's temperature rise rate exceeds expectations, and the auxiliary circulation flow rate needs to be increased to enhance heat dissipation capacity. The adjustment range of the flow rate is determined by a preset ratio, which is calculated using the following formula:

[0086]

[0087] in: This represents the proportionality coefficient for adjusting the flow rate. Represents the preset weighting coefficient. This is the dynamic adjustment decision quantity. The preset weighting coefficient is a parameter optimized and determined based on system characteristics and historical adjustment effects; it determines the system's sensitivity to deviations and the strength of its response. The dynamic adjustment decision quantity reflects the magnitude of the deviation between the current actual value and the predicted value. The product of the two ensures that the adjustment amplitude is proportional to the degree of deviation, achieving a smooth adjustment that is neither too aggressive nor too conservative.

[0088] The entire dynamic adjustment process employs a gradual strategy to avoid system oscillations caused by large abrupt changes. After each adjustment, the system continues to monitor temperature changes, evaluate the adjustment effect, and implement further fine-tuning as needed. This continuous feedback adjustment mechanism enables the system to adapt to various unexpected situations, such as sudden changes in ambient temperature, fluctuations in charging load, or other unforeseen disturbances. All adjustment operations are recorded in detail, including adjustment time, adjustment magnitude, and temperature changes before and after adjustment. These records provide valuable field data for system optimization and model correction.

[0089] Through the implementation of the two-level control module, the system can not only make forward-looking adjustments based on predictions, but also make real-time corrections according to actual operating conditions, forming a complete closed-loop control. This dual control mechanism greatly enhances the system's robustness and adaptability, ensuring that the temperature remains stable within the ideal range under various complex operating conditions. The system's self-correcting capability allows it to gradually reduce its reliance on predictive models and rely more on real-time feedback information to make adjustment decisions, thereby continuously improving the accuracy and reliability of control.

[0090] In actual operation, the secondary control module works in collaboration with the primary control module to form a hierarchical control structure. The primary control module provides the basic speed setting, while the secondary control module is responsible for fine-grained real-time adjustments. This division of labor and cooperation ensures both a grasp of the system's macroscopic state and a rapid response to microscopic fluctuations. The entire control process demonstrates the advantages of intelligent control systems in handling complex dynamic systems. By combining various control strategies, including prediction and feedback, macroscopic and microscopic, and look-ahead and real-time approaches, it achieves efficient and precise management of the charging pile's heat dissipation system.

[0091] Example 4: See Figure 5 At the regional collaborative management level of the dual-cycle adaptive liquid cooling system for charging piles, the collaborative decision-making module integrates the operating status of multiple charging piles within the region to achieve load balancing and optimized temperature control. The system collects real-time operating parameters of each charging pile through a status synchronization unit, including current peak charging demand, ambient temperature and humidity readings, main circulation outlet temperature monitoring values, and real-time flow rate data of the auxiliary circulation loop. This data is uploaded from the local controller of each charging pile to the collaborative decision-making center at specific time intervals, forming a panoramic view of the regional operating status.

[0092] When the main circulation outlet temperature of a charging pile consistently exceeds the first preset temperature difference threshold, the system marks the node as being in a high-load state, indicating that the charging pile is operating under significant heat dissipation pressure. Conversely, when the main circulation outlet temperature of a charging pile is detected to be significantly lower than the second preset temperature difference threshold, the system node is marked as being in a low-load state, meaning that its heat dissipation capacity has a certain margin. The identification of these two states provides a basic basis for subsequent coordinated control, as detailed in Table 1.

[0093] Table 1: Monitoring Data of Regional Charging Pile Cluster Operation Status

[0094]

[0095] The strategy generation unit analyzes the load correlation characteristics between charging piles based on historical operating data stored in the operating condition database. By analyzing the frequency and duration of simultaneous high or low load states of different charging piles in historical data, the system generates a dynamic weight vector to quantify the degree of mutual influence between nodes. This weight vector reflects the thermal coupling relationship of the charging pile cluster in the area; nodes that frequently experience simultaneous high or low temperatures are considered to have higher correlation weights.

[0096] Based on the calculated dynamic weight vector, the system performs a weighted average of the temperature increment data from all high-load nodes to obtain the global temperature compensation. The calculation of the global temperature compensation can be represented by the following formula:

[0097]

[0098] in, This represents the global temperature compensation amount. Representing the Dynamic weights of high-load nodes Representing the Temperature increment of a high-load node This represents the total number of high-load nodes within the region. This compensation amount represents the concentration of overall heat dissipation demand in the region; a larger value indicates a more concentrated heat dissipation pressure. Based on the calculated global temperature compensation amount and the current operating parameters of each low-load node, the strategy generation unit calculates the flow rate compensation coefficient for each low-load node. This coefficient determines the proportion of auxiliary circulation flow rate that each low-load node needs to increase. The calculation process considers both global heat dissipation demands and the actual operating status and capacity margin of each node.

[0099] The calculated flow rate compensation coefficients are distributed to the primary control modules of the corresponding charging piles. These local controllers adjust the flow rate setpoints of the auxiliary circulation loop based on the received coefficients. By appropriately improving the heat dissipation capacity of low-load nodes, the system achieves heat redistribution within the area, effectively alleviating the heat dissipation pressure on high-load nodes. This adjustment process employs a gradual adjustment strategy to avoid system oscillations caused by sudden flow rate changes, ensuring the stability of temperature regulation. Throughout the entire coordinated adjustment process, the system continuously monitors the temperature changes of each charging pile and evaluates the adjustment effect in real time. If a significant deviation is found in the temperature response of a node compared to expectations, the system dynamically adjusts the weighting coefficient of that node to optimize the accuracy of subsequent coordinated decisions. All adjustment operations and effect evaluation data are recorded in the operating condition database, providing a more accurate historical reference for subsequent coordinated decisions.

[0100] This regional collaborative temperature management strategy fully leverages the overall regulation capabilities of the charging pile cluster, improving system operating efficiency through optimized resource allocation. High-load nodes benefit from the auxiliary cooling provided by low-load nodes, avoiding the risk of individual node overheating, while low-load nodes improve equipment utilization efficiency by participating in collaborative regulation. By implementing regional collaborative control, the system can achieve dynamic balance of thermal load among different charging piles, preventing localized overheating or resource idleness.

[0101] Example 5: The optimization execution unit of the collaborative decision-making module continuously monitors the changes in the main circulation outlet temperature of each charging pile within the area after it applies the flow rate compensation coefficient distributed by the strategy generation unit. The optimization execution unit collects continuous readings from temperature sensors and calculates the rate of change by real-time monitoring of the main circulation outlet temperature changes of each charging pile within the area after applying the flow rate compensation coefficient. This monitoring data is acquired using high-frequency sampling to ensure that subtle trends and dynamic characteristics of temperature changes are captured. The system compares the actual observed overall temperature change rate of the area with the expected convergence rate. This expected value is a reference standard established based on historical best operating data, reflecting the temperature regulation speed that the system should achieve under ideal conditions.

[0102] When monitoring data shows that the overall temperature change rate of the area fails to reach the expected convergence rate, the optimization execution unit initiates a recalculation process. This process first analyzes the current temperature response data of each charging pile to identify nodes with slow response or unsatisfactory performance. The system reassesses the thermal coupling relationship between nodes and updates the dynamic weight vector based on the latest operating data. The new weight vector more accurately reflects the degree of mutual influence between charging piles under the current operating conditions. Nodes with temperature responses that deviate significantly from expectations are assigned adjusted weight values ​​to ensure that subsequent calculations better reflect the actual system behavior.

[0103] Based on the updated dynamic weight vector, the system iteratively generates a new global temperature compensation value. This recalculation process considers the actual effects of previous adjustment measures as well as the current real-time state of the system. The new global temperature compensation value not only reflects the overall heat dissipation demand of the region but also incorporates empirical data accumulated during previous adjustments, making the compensation calculation more accurate and in line with actual needs. This iterative calculation process may be performed multiple times, adjusting the calculation parameters each time based on the latest monitoring data until the optimal compensation value is found.

[0104] The iteratively generated global temperature compensation is input to the secondary control modules of each charging pile. These local controllers dynamically fine-tune the flow rate based on the received compensation. The secondary control modules utilize their fine-tuning capabilities to adjust the flow rate of the auxiliary circulation loop according to the magnitude and direction of the compensation. This adjustment process employs a small-step, gradual approach to avoid excessive disturbance to the system. Each charging pile's secondary control module, based on its specific operating status and capacity limitations, transforms the global compensation into adjustment commands suitable for local conditions, achieving an organic combination of global goals and local conditions. Throughout the optimization process, the system establishes a complete feedback adjustment loop. The optimization execution unit continuously monitors the adjustment effect, compares the actual temperature change data with the expected target, and determines whether a new round of optimization calculation needs to be initiated based on the deviation. This cyclical process enables the system to continuously adapt to changing operating conditions, gradually approaching the optimal operating state. All decision data, adjustment commands, and effect records during the optimization process are saved in the operating condition database, providing a reference for future optimization decisions.

[0105] This iterative optimization mechanism enables the system to learn and adapt. By analyzing historical optimization data, the system can identify the optimal adjustment strategy under different operating conditions, gradually reducing its reliance on the initial model and making decisions based more on actual operating data. The optimization execution unit also identifies nodes that consistently perform poorly, marks these nodes specially, and gives them special attention in subsequent optimizations, even suggesting equipment inspection or maintenance when necessary. By implementing this continuous optimization mechanism, the system can maintain efficient temperature control performance in complex operating environments. Even under challenging conditions of drastic fluctuations in charging load and frequent changes in environmental conditions, the system can maintain stable temperature control through rapid iterative adjustments. This dynamic optimization capability not only improves the system's operational reliability but also maximizes energy utilization efficiency, avoiding energy waste caused by over-adjustment. The entire optimization process demonstrates the advantages of intelligent control systems in handling multivariable and nonlinear problems. By combining real-time monitoring, dynamic adjustment, and continuous learning, it achieves a high level of management of the charging pile's heat dissipation system.

[0106] The implementation of optimized execution units transforms regional collaborative regulation from static rule enforcement to a dynamic optimization process. The system no longer simply executes preset strategies but continuously adjusts and improves control measures based on real-time feedback. This evolutionary management approach significantly enhances the system's adaptability and robustness, ensuring that the charging pile cluster maintains optimal thermal management under various operating conditions, providing strong support for the reliable operation of charging infrastructure.

[0107] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0108] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A charging pile double-circulation adaptive liquid cooling system, characterized in that, The application relates to a temperature control system for charging piles, comprising: a main circulation cooling loop for basic heat dissipation of charging pile power modules, containing a cooling liquid circulating pump and a radiator; an auxiliary circulation regulation loop for dynamically regulating electrolyte heat exchange rate, containing an electrolyte storage tank, a flow regulating valve and a heat exchanger; a working condition database storing historical charging process records, containing charging demand peak value, environmental temperature and humidity, main circulation inlet temperature, main circulation outlet temperature and auxiliary circulation flow rate; a feature prediction module for collecting charging demand features and environmental temperature and humidity of the current charging pile, combining the working condition database and a deep learning model to output predicted temperature features of the current charging process; a primary regulation module for judging whether to start primary flow regulation according to the predicted temperature features and calculating initial flow rate adjustment value of the auxiliary circulation regulation loop; a secondary regulation module for calculating actual temperature change features based on actual outlet temperature change of the main circulation cooling loop and comparing the actual temperature change features with the predicted temperature features to judge whether to start secondary dynamic flow rate regulation; a collaborative decision module for receiving state data of multiple charging piles in a region, making collaborative decisions based on the working condition database, generating optimal temperature balancing strategies and distributing the strategies to the primary regulation modules and the secondary regulation modules of the charging piles; a strategy generation unit of the collaborative decision module performs the following operations: generating a dynamic weight vector according to historical load correlation of each charging pile in the working condition database; weighting and averaging temperature increments of all high-load nodes in the region based on the dynamic weight vector to obtain a global temperature compensation amount; calculating flow rate compensation coefficients of each low-load node according to the global temperature compensation amount; distributing the flow rate compensation coefficients to the primary regulation modules of the corresponding charging piles.

2. The dual-cycle adaptive liquid cooling system of claim 1, wherein, a data collection unit of the working condition database performs the following operations: collecting complete records of each charging process, including charging start time, charging demand peak value, environmental temperature and humidity, main circulation inlet temperature time sequence, main circulation outlet temperature time sequence and auxiliary circulation set flow rate; screening records in which the maximum and minimum values of the main circulation outlet temperature time sequence are within a safe temperature range to form a standard record set; calculating time intervals from the charging start time to the temperature stable time of each record in the standard record set as a first time reference parameter; extracting the average increment of the main circulation outlet temperature from the charging start time to the temperature stable time in the standard record set as a first temperature reference threshold value; calculating the average rate of change of the main circulation outlet temperature per unit time interval in each record as a historical temperature change feature.

3. The dual-cycle adaptive liquid cooling system of claim 2, wherein, a feature extraction unit of the feature prediction module performs the following operations: real-time monitoring of the charging start time and the main circulation outlet temperature time sequence of the current charging process; when the main circulation outlet temperature exceeds the first temperature reference threshold value, marking the time as a temperature monitoring start point; calculating the difference between the main circulation outlet temperature corresponding to the temperature monitoring start point and the charging start point temperature as the current temperature increment; extracting matching historical temperature change features from the working condition database according to the current charging demand peak value and the environmental temperature and humidity; The current charging demand peak, environmental temperature and humidity, current temperature increment, and initial set flow rate of the auxiliary circulation regulation loop are input into the deep learning model, and a predicted temperature change feature is output.

4. The dual-cycle adaptive liquid cooling system of claim 3, wherein, The regulation determination unit of the primary regulation module performs the following operations: The product of the predicted temperature change feature and the first time reference parameter is calculated as a predicted temperature offset; The first temperature reference threshold is added to the predicted temperature offset to obtain a regulation determination parameter; When the regulation determination parameter exceeds the safe temperature range, a primary flow regulation instruction is triggered; The difference between the lower limit of the safe temperature range and the first temperature reference threshold is calculated as a demand adjustment amount; The ratio of the demand adjustment amount to the first time reference parameter is calculated as a target adjustment feature.

5. The dual-cycle adaptive liquid cooling system of claim 4, wherein, The flow rate calculation unit of the primary regulation module performs the following operations: The current charging demand peak, environmental temperature and humidity, and current temperature increment are input into the deep learning model; The target adjustment feature is used as an output constraint to obtain the target flow rate value of the auxiliary circulation regulation loop through inverse calculation of the deep learning model; The initial set flow rate of the auxiliary circulation regulation loop is updated to the target flow rate value.

6. The dual-cycle adaptive liquid cooling system of claim 5, wherein, The error analysis unit of the secondary regulation module performs the following operations: The main circulation outlet temperature is collected at fixed time intervals from the temperature monitoring starting point; The main circulation outlet temperature change rate of adjacent time intervals is calculated as an actual temperature change feature; The absolute difference between the actual temperature change feature and the predicted temperature change feature is calculated as a dynamic adjustment determination amount; When the dynamic adjustment determination amount exceeds the preset error threshold, a secondary dynamic flow rate adjustment instruction is triggered.

7. The dual-cycle adaptive liquid cooling system of claim 6, wherein, The dynamic adjustment unit of the secondary regulation module performs the following operations: When the actual temperature change feature is lower than the predicted temperature change feature, the current flow rate of the auxiliary circulation regulation loop is reduced by a preset proportion; When the actual temperature change feature is higher than the predicted temperature change feature, the current flow rate of the auxiliary circulation regulation loop is increased by a preset proportion; The preset proportion is equal to the product of a preset weight coefficient and the dynamic adjustment determination amount.

8. The dual-cycle adaptive liquid cooling system of claim 1, wherein, The state synchronization unit of the collaborative decision module performs the following operations: The current charging demand peak, environmental temperature and humidity, main circulation outlet temperature, and auxiliary circulation real-time flow rate of each charging pile in the region are received; When the main circulation outlet temperature of any charging pile exceeds the first preset temperature difference threshold, the charging pile is marked as a high-load node; When the main circulation outlet temperature of any charging pile is lower than the second preset temperature difference threshold, the charging pile is marked as a low-load node.

9. The dual-cycle adaptive liquid cooling system of claim 1, wherein, The optimization execution unit of the collaborative decision module performs the following operations: The main circulation outlet temperature change rate of each charging pile after executing the flow rate compensation coefficient is monitored in real time; When the overall temperature change rate of the region does not reach the expected convergence rate, the dynamic weight vector is recalculated; A new global temperature compensation amount is iteratively generated based on the updated dynamic weight vector; The iteratively generated global temperature compensation amount is input into the secondary regulation module for dynamic flow rate fine-tuning.

Images