A smart heat dissipation control method and system for charging piles

Abstract

This application relates to the field of charging piles, and more particularly to an intelligent heat dissipation control method and system for charging piles. The method includes the following steps: constructing a liquid-cooled pipe network; the liquid-cooled pipe network includes several flow channels and valves, and refrigerant flows through the flow channels to cool the charging piles; constructing a graph neural network about the charging piles and the liquid-cooled pipe network, and outputting the anomaly probability of the charging piles based on the graph neural network; calculating the cooling priority score of each charging pile based on the anomaly probability; setting a combination of opening or closing schemes for each valve, calculating the refrigerant flow rate of each charging pile under any combination of schemes, and summing the product of the priority score of each charging pile and the corresponding flow rate as the heat dissipation score of the scheme; selecting the combination of schemes with the highest heat dissipation score and controlling the corresponding valve to open. This application dynamically adjusts the refrigerant flow distribution strategy according to the actual operating status of the charging pile, improves the efficiency and accuracy of cooling, and reduces the safety hazard of overheating of the charging pile.

Description

Technical Field

[0001] This application relates to the field of charging piles, and in particular to an intelligent heat dissipation control method and system for charging piles. Background Technology

[0002] With the rapid development of new energy vehicles, charging piles, as an important supporting facility, have been widely used. However, in actual use, the electronic components inside the charging pile generate a lot of heat during high-power charging, causing the equipment temperature to rise. Poor heat dissipation will directly affect the working efficiency, stability, and lifespan of the charging pile, and may even cause safety hazards.

[0003] Existing charging stations typically employ several heat dissipation methods to manage the heat generated during operation: natural cooling, air cooling, and liquid cooling. Currently, most ordinary AC charging stations use natural cooling or forced air cooling, while high-power DC fast charging stations rely more heavily on liquid cooling. The core principle of liquid cooling is to utilize the high specific heat capacity and good thermal conductivity of liquids to absorb and dissipate heat. Coolant flows from a circulation pump through pipes (channels) on a cold plate, absorbing heat from the heat-generating elements, increasing its temperature. The heated coolant then enters the radiator, releasing heat through heat exchange with the outside air.

[0004] Traditional liquid cooling systems typically employ a fixed refrigerant flow distribution strategy. In high-power charging scenarios, some high-load charging stations may not receive sufficient cooling resources, leading to excessively high temperatures. Furthermore, the inability to promptly identify charging stations with abnormal temperature rises can trigger overheating risks, potentially causing equipment malfunctions or damage. Conversely, for low-load or lower-temperature charging stations, excessive refrigerant supply results in energy waste. Therefore, determining how to tailor a cooling strategy based on the specific conditions of each charging station is a crucial issue that needs to be addressed. Summary of the Invention

[0005] To address the aforementioned technical problem of how to set cooling strategies based on the actual conditions of charging piles, this application provides an intelligent heat dissipation control method and system for charging piles.

[0006] Firstly, this application provides an intelligent heat dissipation control method for charging piles, employing the following technical solution:

[0007] A smart heat dissipation control method for charging piles includes the following steps: constructing a liquid-cooled pipe network; the liquid-cooled pipe network includes several flow channels and valves, with refrigerant flowing through the flow channels to cool the charging piles, and valves controlling the on / off state of the cooling circuit in their respective flow channels; constructing a graph neural network relating the charging piles and the liquid-cooled pipe network, and outputting the anomaly probability of the charging piles based on the graph neural network; calculating the cooling priority score of each charging pile based on the anomaly probability; setting combinations of valve opening or closing schemes, calculating the refrigerant flow rate of each charging pile under any scheme combination, and summing the product of the priority score of each charging pile and the corresponding flow rate as the heat dissipation score of that scheme, and selecting the appropriate heat dissipation method. The scheme combination with the highest heat score controls the opening of the corresponding valve; the priority score is calculated as follows: construct the unit time temperature rise rate sequence of each charging pile within a set time window, calculate the mean of the temperature rise rate sequence of all charging piles, and obtain the degree of temperature anomaly by positively normalizing the ratio of the temperature rise sequence of a single charging pile to the mean; calculate the temperature difference between the inside and outside of the charging pile, and set the environmental factor to 1 in response to the temperature difference being less than or equal to 0, otherwise, obtain the environmental factor by negatively normalizing the temperature difference between the inside and outside of the charging pile; the product of the anomaly probability, the degree of anomaly, and the environmental factor is used as the priority score of the charging pile.

[0008] Optionally, the nodes of the graph neural network represent the real-time temperature, estimated charging time, and valve opening status of each charging pile; the edges represent the connection relationship of the charging piles in the flow channel; the input of the graph neural network is the real-time temperature, estimated charging time, and valve opening status of the charging pile, and the output is the probability of temperature anomalies of the charging pile.

[0009] Optionally, the method for obtaining the environmental factor after negatively normalizing the temperature difference between the inside and outside of the charging pile is as follows: calculate the difference between the ambient temperature outside the charging pile and the temperature inside the charging pile; obtain the maximum internal and external temperature difference among all charging piles; add 1 to the internal and external temperature difference and take the logarithm for smoothing; compare the smoothed internal and external temperature difference with the smoothed maximum internal and external temperature difference to obtain the ratio, and subtract the ratio from 1 to perform negatively normalization to obtain the environmental factor.

[0010] Optionally, a sequence of temperature rise rates per unit time for each charging pile can be constructed within a set time window, where the window can be a fixed window or a dynamic window.

[0011] Optionally, the method for setting up a dynamic window is as follows: set an initial window length, calculate the current temperature rise rate, use the negative number of the temperature rise rate as the power and the natural logarithm as the base to obtain the window adjustment coefficient, and use the product of the window adjustment coefficient and the initial window length as the current window length.

[0012] Optionally, the method for calculating the refrigerant flow rate of each charging pile under any combination of schemes is as follows: In any combination of schemes, for any charging pile, take the average of multiple sets of historical flow data passing through the charging pile in the combination of schemes as the refrigerant flow rate of the charging pile.

[0013] Optional: Calculate the mean of the temperature rise rate sequence of all charging piles, and then perform positive correlation normalization on the ratio of the temperature rise sequence of a single charging pile to the mean to obtain the degree of temperature rise anomaly. The positive correlation normalization method is maximum-minimum value normalization or standardized normalization.

[0014] Secondly, this application provides an intelligent heat dissipation control system for a charging pile, which adopts the following technical solution:

[0015] A smart heat dissipation control system for a charging pile includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the smart heat dissipation control method for the charging pile described above is implemented.

[0016] The beneficial effects are: the intelligent heat dissipation control method of the above-mentioned charging pile is generated into a computer program and stored in the memory so that it can be loaded and executed by the processor, thereby making the system easy to use by creating a system based on the memory and processor.

[0017] This application has the following technical effects:

[0018] 1. By dynamically adjusting the refrigerant flow distribution strategy, cooling efficiency and accuracy are effectively improved. Based on the actual operating status of the charging piles, the system intelligently assesses the heat dissipation needs of each charging pile and prioritizes providing sufficient cooling resources to charging piles with high loads and abnormal temperature rises, thereby significantly reducing the safety hazards of overheating of charging piles.

[0019] 2. Energy efficiency has been optimized, avoiding energy waste caused by excessive supply of coolant to charging piles under low load or low temperature. By comprehensively considering the rate of temperature rise, environmental factors, and probability of anomalies, the stability and reliability of the entire system are ensured.

[0020] 3. By using graph neural networks to analyze the connection relationships and operating status between charging piles, the system's ability to identify potential anomalies and its response speed are further improved, providing a more reliable heat dissipation solution for high-power charging scenarios. Attached Figure Description

[0021] Figure 1 This is a flowchart of an intelligent heat dissipation control method for a charging pile according to an embodiment of this application.

[0022] Figure 2This is a flowchart of a method for calculating priority scores in an intelligent heat dissipation control method for a charging pile according to an embodiment of this application. Detailed Implementation

[0023] This application discloses an intelligent heat dissipation control method for charging piles, referring to... Figure 1 The steps include:

[0024] S1: Construct a liquid cooling pipeline network; the liquid cooling pipeline network includes several flow channels and valves. Refrigerant flows through the flow channels to cool the charging piles, and the valves control the on / off of the cooling circuit in their respective flow channels.

[0025] In one embodiment, at least one charging pile and a valve are provided on a flow channel. When the valve is open, the flow channel is open, allowing refrigerant to flow through the channel to the corresponding charging pile for cooling. When the valve is closed, the flow channel is closed, preventing refrigerant from flowing through the channel to the corresponding charging pile. Multiple flow channels are arranged in parallel and are interconnected. Refrigerant is distributed to multiple flow channels through a total inflow channel, and refrigerant passing through the charging pile flows out through a total outflow channel.

[0026] S2: Construct a graph neural network about charging piles and liquid cooling pipelines, and output the abnormal probability of charging piles based on the graph neural network.

[0027] In this application, to more accurately assess the heat dissipation requirements and potential anomalies of charging piles, a graph neural network (GNN) specifically designed for charging piles and liquid cooling pipe networks needs to be constructed. This network analyzes the connections between charging piles and their operating status to output the probability of temperature anomalies for each charging pile.

[0028] The nodes in the graph neural network represent the key operating parameters of each charging pile, including: real-time temperature, reflecting the current internal temperature level of the charging pile; estimated charging time, predicting the working duration of the charging pile in the future based on the current load and charging plan; and valve opening status, recording the on / off state of the corresponding valves on the flow channel to understand whether the refrigerant is flowing to the charging pile.

[0029] Edges represent the physical connections between charging stations formed by flow channels in the liquid-cooled piping network. The existence of edges reflects the flow path of the refrigerant between different charging stations, providing topology information for global heat dissipation optimization.

[0030] The graph neural network (Graph Neural Network) takes the charging pile's real-time temperature, estimated charging time, and valve opening status as input, and outputs the probability of temperature anomalies for each charging pile. The Graph Neural Network learns and analyzes the input data to output the probability of temperature anomalies for each charging pile. The construction and training of the Graph Neural Network are existing technologies and will not be elaborated upon here.

[0031] For example, suppose there are 3 charging stations (denoted as N1, N2, and N3) in a liquid cooling network. Each node has the following characteristics: real-time temperature Current internal temperature of the charging station (unit: °C); estimated charging time : Expected working time in the near future (unit: minutes); Valve opening status The states correspond to the valves in the flow path (0 indicates closed, 1 indicates open); the edge relationships are as follows: there is one edge between N1 and N2; there is one edge between N2 and N3; there is no direct edge between N1 and N3. This can be represented using an adjacency matrix: ,in Represents a node and nodes There is a border between them. It indicates no boundary.

[0032] Output the probability of temperature anomalies for each node. , This indicates that the probability of temperature anomaly at node N1 is 0.2. The probability of temperature anomaly at node N2 is 0.8; The probability of temperature anomaly at node N3 is 0.1.

[0033] S3: Calculate the priority score for cooling of each charging pile based on the anomaly probability.

[0034] Reference Figure 2 The calculation method for priority scores includes steps S30-S32, as detailed below:

[0035] S30: Construct a unit time temperature rise rate sequence for each charging pile within a set time window, calculate the mean of the temperature rise rate sequences of all charging piles, and obtain the degree of temperature rise anomaly by positively normalizing the ratio of the temperature rise sequence of a single charging pile to the mean.

[0036] In constructing the temperature rise rate sequence per unit time for each charging pile within a set time window, the window can be a fixed window or a dynamic window.

[0037] In one embodiment, a fixed window length (such as 5 minutes or 10 minutes) is set, during which the real-time temperature change of the charging pile is recorded, and the rate of temperature rise per unit time is calculated.

[0038] In one embodiment, the dynamic window is set as follows: An initial window length (e.g., 5 minutes or 10 minutes) is set; the current temperature rise rate is calculated; the negative of the temperature rise rate is raised to the power of the natural logarithm to obtain a window adjustment coefficient; and the product of the window adjustment coefficient and the initial window length is used as the current window length. When the temperature rise rate is fast, the response efficiency can be improved by shortening the window length.

[0039] The mean of the temperature rise rate sequence of all charging piles is calculated. The degree of temperature anomaly is obtained by positively normalizing the ratio of the temperature rise rate sequence of a single charging pile to the mean. The positive correlation normalization method can be either maximum / minimum value normalization or standardized normalization, which are existing techniques and will not be elaborated further.

[0040] S31: Calculate the temperature difference between the inside and outside of the charging pile. If the temperature difference is less than or equal to 0, the environmental factor is set to 1. Otherwise, the environmental factor is obtained by negatively normalizing the temperature difference between the inside and outside of the charging pile.

[0041] In one embodiment, the difference between the ambient temperature outside the charging pile and the temperature inside the charging pile is calculated. , , This indicates the internal temperature of the charging station. Indicates the external ambient temperature; obtains the maximum internal and external temperature difference among all charging stations.

[0042] When the internal and external temperature difference is not greater than 0, the influence of environmental factors on priority calculation is not considered. When the internal and external temperature difference is greater than 0, the difference is increased by 1 and then the logarithm is taken for smoothing. , The smoothed internal and external temperature difference value The maximum internal and external temperature difference after smoothing The ratios are obtained by comparison, and the environmental factors are obtained by subtracting the ratios from 1 for negative correlation normalization. , External temperature difference refers to the average temperature of the entire environment. The hyperparameter is used to prevent the denominator from being 0. The hyperparameter can be adjusted according to the actual application scenario.

[0043] S32: The product of the anomaly probability, the anomaly degree, and the environmental factors is used as the priority score for the charging pile.

[0044] Specifically, the temperature anomaly probability output by the graph neural network has a value range of [0, 1]. When the temperature anomaly probability increases, the priority score is higher, indicating that the charging pile needs a higher cooling priority.

[0045] The degree of abnormality of the charging pile's temperature rise rate ranges from [0, 1] and is positively correlated with the priority score. If the temperature rise rate of a charging pile is significantly higher than the average level of other charging piles, its degree of abnormality will be higher, and its priority score will also increase accordingly.

[0046] The environmental factor is the result of negative correlation normalization, and its value range is [0, 1]. Taking into account both active heat dissipation (liquid cooling system) and passive heat dissipation, if the passive heat dissipation conditions are good (i.e., large temperature difference), the charging pile can effectively reduce the temperature through natural heat dissipation. Therefore, its cooling priority can be appropriately reduced. At this time, the environmental factor decreases and the priority score is low.

[0047] S4: Set the combination of opening or closing schemes for each valve, calculate the refrigerant flow rate of each charging pile under any combination of schemes, and sum the product of the priority score of each charging pile and the corresponding flow rate as the heat dissipation score of the scheme. Select the combination of schemes with the highest heat dissipation score and control the corresponding valve to open.

[0048] Specifically, assuming the system has There are 10 charging stations, and each charging station corresponds to a valve on a flow channel. The state of each valve can be represented in binary: When the valve is opened, the refrigerant can flow through this channel to the corresponding charging station. When the valve is closed, the refrigerant cannot flow through the channel to the corresponding charging station. Therefore, the state of all valves can be represented as a length of... binary vector: ,according to The various possible states of a valve can generate multiple combinations of schemes, each corresponding to a specific valve opening / closing configuration.

[0049] In any combination of schemes, for any charging pile, the average of multiple sets of historical flow data passing through the charging pile in the combination of schemes is taken as the refrigerant flow rate of the charging pile.

[0050] For each combination of schemes, the sum of the product of the priority scores of all charging piles and the corresponding refrigerant flow rates is used as the heat dissipation score for that scheme. The heat dissipation scores of all scheme combinations are compared, and the opening or closing state of the corresponding valve is controlled according to the selected optimal scheme combination. For example: if the optimal scheme combination is... If so, then valves N1 and N3 will be opened, and valve N2 will be closed.

[0051] This application also discloses an intelligent heat dissipation control system for a charging pile, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent heat dissipation control method for the charging pile according to this application is implemented.

[0052] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0053] In this application, the aforementioned memory can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), etc., or any other medium that can be used to store required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0054] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A smart heat dissipation control method for a charging pile, characterized in that, Including the following steps: Construct a liquid cooling pipeline network; the liquid cooling pipeline network includes several flow channels and valves. Refrigerant flows through the flow channels to cool the charging piles, and the valves control the on / off of the cooling circuit in their respective flow channels. Construct a graph neural network about charging piles and liquid cooling pipelines, and output the anomaly probability of charging piles based on the graph neural network; Calculate the priority score for cooling of each charging pile based on the probability of anomalies; Set up combinations of opening or closing schemes for each valve, calculate the refrigerant flow rate of each charging pile under any combination of schemes, and sum the product of the priority score of each charging pile and the corresponding flow rate as the heat dissipation score of the scheme. Select the combination of schemes with the highest heat dissipation score and control the corresponding valve to open. The priority score is calculated as follows: a unit time temperature rise rate sequence of each charging pile is constructed within a set time window, the mean of the temperature rise rate sequence of all charging piles is calculated, and the degree of temperature rise anomaly is obtained by positively normalizing the ratio of the temperature rise sequence of a single charging pile to the mean. The temperature difference between the inside and outside of the charging pile is calculated. When the temperature difference is less than or equal to 0, the environmental factor is set to 1. Otherwise, the environmental factor is obtained by negatively normalizing the temperature difference between the inside and outside of the charging pile. The product of the anomaly probability, the anomaly degree, and the environmental factor is used as the priority score of the charging pile.

2. The intelligent heat dissipation control method for charging piles according to claim 1, characterized in that, The nodes of the graph neural network represent the real-time temperature, estimated charging time, and valve opening status of each charging pile; the edges represent the connection relationship of the charging piles in the flow channel; the input of the graph neural network is the real-time temperature, estimated charging time, and valve opening status of the charging pile, and the output is the probability of temperature anomalies of the charging pile.

3. The intelligent heat dissipation control method for charging piles according to claim 1, characterized in that, The method for obtaining environmental factors by negatively normalizing the temperature difference between the inside and outside of the charging pile is as follows: Calculate the temperature difference between the external environment and the internal temperature of the charging pile; obtain the maximum internal and external temperature difference among all charging piles. Add 1 to the temperature difference between the inside and outside and then take the logarithm for smoothing. The smoothed internal and external temperature difference value is compared with the smoothed maximum internal and external temperature difference value to obtain the ratio. The ratio is then subtracted from 1 to perform negative correlation normalization to obtain the environmental factor.

4. The intelligent heat dissipation control method for charging piles according to claim 1, characterized in that, In constructing the temperature rise rate sequence per unit time for each charging pile within a set time window, the window can be a fixed window or a dynamic window.

5. The intelligent heat dissipation control method for charging piles according to claim 4, characterized in that, The method for setting up a dynamic window is as follows: set an initial window length, calculate the current temperature rise rate, use the negative number of the temperature rise rate as the power and the natural logarithm as the base to obtain the window adjustment coefficient, and use the product of the window adjustment coefficient and the initial window length as the current window length.

6. The intelligent heat dissipation control method for charging piles according to claim 1, characterized in that, The method for calculating the refrigerant flow rate of each charging pile under any combination of schemes is as follows: In any combination of schemes, for any charging pile, take the average of multiple sets of historical flow data passing through the charging pile in the combination of schemes as the refrigerant flow rate of the charging pile.

7. The intelligent heat dissipation control method for charging piles according to claim 1, characterized in that, The mean of the temperature rise rate sequence of all charging piles is calculated. The ratio of the temperature rise rate sequence of a single charging pile to the mean is positively correlated and normalized to obtain the degree of temperature rise anomaly. The positive correlation normalization method is maximum-minimum value normalization or standardized normalization.

8. An intelligent heat dissipation control system for a charging pile, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the intelligent heat dissipation control method for a charging pile according to any one of claims 1-7.

Images