An energy management method and system for a liquid cooling system based on multi-modal fusion

Abstract

This invention discloses an energy management method and system for liquid cooling systems based on multimodal fusion, belonging to the field of intelligent thermal management and energy-saving control technology for liquid cooling systems. The invention collects multi-point temperature data of the area to be cooled using a distributed temperature sensor array, constructs a temperature prediction model, and enables early prediction of temperature trends. It accurately locates high-temperature areas based on multi-level temperature thresholds and load conditions. The system maintains continuous spraying operation throughout the process and dynamically adjusts the coolant flow rate according to real-time temperature and load conditions. After spraying, the coolant is temperature-detected and filtered, then uniformly returned to the liquid cooling distribution unit. It is then distributed according to the temperature requirements of each area and sent to a dry cooler for heat exchange and cooling before re-entering the circulation pipeline for reuse. This invention solves the problems of low coolant utilization and high energy consumption in existing liquid cooling systems, achieving closed-loop recovery and efficient recycling of coolant. It is suitable for liquid cooling heat dissipation scenarios such as data centers, new energy storage, and high-power electrical equipment.

Description

Technical Field

[0001] This invention belongs to the field of intelligent thermal management and energy-saving control technology for liquid cooling systems, and particularly relates to an energy management method and system for liquid cooling systems based on multimodal fusion. Background Technology

[0002] With the rapid development of new energy storage, data centers, and high-power electrical equipment, the operating power of equipment is constantly increasing, and the heat generated is increasing dramatically. This places higher demands on the heat dissipation efficiency, temperature control accuracy, and energy utilization efficiency of cooling systems. Liquid cooling systems, due to their advantages such as high heat dissipation efficiency, uniform temperature control, and adaptability to high-power loads, have been widely used in thermal management scenarios for various high-heat-generating equipment. Their energy consumption and temperature control performance directly affect the operational stability, service life, and overall energy consumption level of the equipment.

[0003] Currently, the energy management methods of existing liquid cooling systems still have many shortcomings, making it difficult to meet the refined and energy-saving thermal management needs of high-power equipment.

[0004] Temperature control is lagging and emergency cooling methods are crude. Existing liquid cooling systems mostly adopt passive circulation heat dissipation or single threshold-triggered cooling methods, which can only start the cooling operation after the temperature exceeds the safety threshold. They lack the ability to predict temperature change trends, resulting in a delayed cooling response and easy occurrence of local overheating. Moreover, emergency cooling often uses fixed flow rate spraying or overall cooling mode, which cannot accurately locate high temperature areas. This not only results in low cooling efficiency but also causes a large waste of coolant and energy.

[0005] The coolant flow rate regulation lacks adaptive capability; the coolant flow rate of existing liquid cooling systems is mostly fixed or manually adjusted, and cannot be dynamically adjusted according to the real-time temperature of the area to be cooled and the load operation status. When the load is low and the temperature is low, a high flow rate is still maintained, resulting in redundant energy consumption; when the load suddenly increases and the temperature rises sharply, the flow rate cannot be increased in time, making it difficult to cool down quickly and affecting the safe operation of the equipment.

[0006] The energy utilization efficiency is low, and the coolant recovery rate is insufficient. The coolant in the existing liquid cooling system is mostly discharged directly after single use, or after simple circulation, it is not subjected to targeted temperature screening and optimized distribution. As a result, the waste heat of the coolant is not fully utilized, which not only increases the cost of coolant replenishment, but also increases the overall energy consumption of the system. At the same time, the system lacks a closed-loop optimization mechanism for its own energy consumption, and cannot dynamically adjust the operating parameters according to the heat dissipation effect and energy consumption data, making it difficult to achieve the global optimization of heat dissipation effect and energy consumption.

[0007] The system has a low level of intelligence and insufficient overall control capabilities. The temperature acquisition, spray control, and flow rate regulation modules of the existing liquid cooling system mostly operate independently, lacking a unified central control unit, which makes it impossible for the modules to work together. Furthermore, it lacks remote monitoring, fault warning, and parameter adaptive optimization functions, requiring frequent manual intervention for adjustments. This not only increases labor costs but also makes the system prone to malfunctions due to human error, affecting heat dissipation and energy management efficiency.

[0008] To address the problems of lagging temperature control, inflexible flow rate adjustment, low energy utilization, and insufficient intelligence in the existing technologies, there is an urgent need for an energy management method and system for liquid cooling systems that can achieve precise spray cooling in high-temperature areas, adaptive adjustment of coolant flow rate, coolant recycling and reuse, and closed-loop optimization of energy consumption. This would improve the temperature control accuracy, heat dissipation efficiency, and energy utilization efficiency of liquid cooling systems, reduce energy consumption and operating costs, and ensure the stable and reliable operation of high-power equipment. Summary of the Invention

[0009] The purpose of this invention is to provide an energy management method and system for liquid cooling systems based on multimodal fusion, which solves the problems of temperature control lag, crude emergency cooling methods, lack of adaptive ability in cold liquid flow rate regulation, low energy utilization efficiency, insufficient cold liquid recovery and utilization rate, low system intelligence, and insufficient overall control capability in the energy management of existing liquid cooling systems.

[0010] To achieve the above objectives, the present invention provides an energy management method for liquid cooling systems based on multimodal fusion, comprising the following steps: S1. Collect real-time temperature data of multiple points in the liquid cooling system's cooling area through a distributed temperature sensor array, and simultaneously obtain the operating parameters of the load to be cooled. Based on a genetic algorithm, construct a temperature change prediction model to predict the temperature change trend within a preset time period and output the predicted temperature value. S2. Preset safe temperature threshold, warning temperature threshold and emergency cooling threshold. Compare the real-time temperature value and predicted temperature value collected by S1 with each threshold to prepare for or trigger the spraying. S3. Based on the temperature difference between the real-time temperature of the target area and the safe temperature threshold, the predicted temperature change trend and the operating parameters of the load to be cooled, a coolant flow rate adjustment function is constructed, and the coolant flow rate is adjusted in real time according to the temperature and load conditions. S4. During the spray cooling process, continuously collect temperature data of the target area and energy consumption data of the liquid cooling system, compare the temperature feedback value with the safe temperature threshold, and iteratively optimize the coolant flow rate adjustment function based on the energy consumption data using a genetic algorithm. S5. The coolant after spraying is returned to the liquid cooling distribution unit. After being temporarily stored and distributed by the liquid cooling distribution unit, it is sent to the dry cooler for heat exchange and cooling. The cooled coolant re-enters the spraying circulation pipeline and is distributed to the corresponding spraying unit according to the temperature requirements of the area.

[0011] Preferably, the deployment density of the distributed temperature sensor array in S1 is positively correlated with the heat load density of the area to be cooled, and the acquisition frequency of each sensor is automatically adjusted according to the load operating status; the temperature change prediction model is generated by training through historical temperature data, load operating parameters and ambient temperature data.

[0012] Preferably, in S2, the real-time temperature value and the predicted temperature value collected in S1 are compared with each threshold, and the heat dissipation requirement is determined in combination with the load operating status. Based on the temperature sensor location information, the high-temperature target area is accurately located, and the corresponding area spraying is started. The whole-area spraying unit keeps running continuously.

[0013] Preferably, the coolant flow rate regulation function in S3 uses temperature difference as the core variable and is modified by combining load power, ambient temperature, and initial coolant temperature. The expression is as follows: ; In the formula, For the target spray flow rate, This is the flow rate adjustment coefficient. The difference between the real-time temperature and the safe temperature threshold. This represents the real-time power of the load to be cooled. The initial temperature of the coolant. The ambient temperature.

[0014] Preferably, the coolant flow rate regulation function automatically calculates the target spray flow rate, and dynamically adjusts the spray coolant flow rate through the coordinated control of the flow rate regulation valve and the variable frequency pump, so as to linearly match the temperature difference and the flow rate.

[0015] Preferably, the closed-loop feedback regulation in S4 adopts a hierarchical control strategy: If the temperature feedback value is lower than the safe temperature threshold, the spray flow rate will be automatically reduced until the spraying stops and a coolant recovery command will be triggered. At the same time, based on energy consumption data, the coolant flow rate adjustment function will be iteratively optimized through a genetic algorithm to achieve the global optimal balance between heat dissipation and energy consumption.

[0016] An energy management system for liquid cooling systems based on multimodal fusion, employing an energy management method for liquid cooling systems based on multimodal fusion as described above, includes: a temperature acquisition and prediction module, a spray control module, a flow rate regulation module, a closed-loop optimization module, a coolant reflux heat exchange module, and a central control module. The temperature acquisition and prediction module includes a distributed temperature sensor array and a prediction processing unit. The distributed temperature sensor array is used to acquire real-time temperature data of multiple points in the area to be cooled, and the prediction processing unit is used to obtain the operating parameters of the load to be cooled, construct a temperature change prediction model based on a genetic algorithm, and output real-time temperature values ​​and predicted temperature values. The spray control module includes a spray unit, a positioning unit, and a triggering unit. The spray unit includes several independently controlled spray heads. The positioning unit is used to locate the target area with excessively high temperature based on the position information of the temperature sensor. The triggering unit is used to trigger a spray preparation command or a spray cooling command based on the temperature comparison result, and control the corresponding spray head to start. The flow rate regulation module includes a variable frequency liquid pump, a flow rate regulating valve, and a flow rate calculation unit. The flow rate calculation unit is used to construct a coolant flow rate regulation function and calculate the target spray flow rate based on temperature difference, predicted temperature trend, and load operating parameters. The variable frequency liquid pump and the flow rate regulating valve work together to dynamically regulate the flow rate of the sprayed coolant. The closed-loop optimization module includes a feedback acquisition unit and an optimization processing unit. The feedback acquisition unit is used to continuously collect temperature data and system energy consumption data of the target area. The optimization processing unit is used to compare the temperature feedback value with the safe temperature threshold, adjust the spray flow rate, and iteratively optimize the coolant flow rate adjustment function through a genetic algorithm. The coolant return heat exchange module includes a liquid cooling distribution unit, a dry cooler, a temperature detection unit, and a circulation pipeline. The temperature detection unit is used to detect the temperature of the coolant after spraying, the filtration unit is used to filter the coolant, and the circulation pipeline is used to collect the coolant after spraying through the return pipeline into the liquid cooling distribution unit, and then send it to the dry cooler for heat exchange and cooling, and then supply it to the spraying circulation. The central control module is electrically connected to the temperature acquisition and prediction module, the spray control module, the flow rate regulation module, the closed-loop optimization module, and the coolant reflux heat exchange module, respectively. It is used to receive signals from each module, control the coordinated operation of each module, and realize the automated and refined management of the energy of the liquid cooling system.

[0017] Preferably, the spray head of the spray unit adopts a rotatable structure, and each spray head corresponds to an independent flow rate control branch, which independently adjusts the spray angle and flow rate according to the size and temperature distribution of the target area; the distributed temperature sensor array adopts a waterproof high-precision sensor and has a self-calibration function, which automatically completes a calibration every 24 hours.

[0018] Preferably, the flow rate regulation module also includes a flow rate monitoring unit, which is used to monitor the actual value of the spray flow rate in real time, compare the actual flow rate value with the target flow rate value, and automatically adjust the opening of the flow rate regulation valve and the speed of the variable frequency liquid pump when the deviation exceeds the set threshold to ensure the accuracy of flow rate regulation; the cold liquid circulation and recovery module also includes a heat insulation layer, which is used to insulate the circulation pipeline and reduce the temperature loss of the cold liquid during the circulation process.

[0019] Preferably, the central control module also includes a data storage unit and a remote interaction unit. The data storage unit is used to store temperature data, energy consumption data, flow rate adjustment parameters and optimization results; the remote interaction unit is used to realize remote monitoring, parameter setting and fault alarm functions, and provide real-time feedback on the system operating status. When the system malfunctions, it immediately sends an alarm signal to the management personnel terminal.

[0020] Therefore, the present invention employs the above-mentioned energy management method and system for liquid cooling systems based on multimodal fusion, which has the following beneficial effects: (1) This invention collects multi-point temperature data of the area to be cooled in real time through a distributed temperature sensor array, and constructs a temperature change prediction model by combining a genetic algorithm. It can predict the temperature change trend in the future within a preset time period, realize the early intervention of cooling operation, and solve the drawback of the slow response of traditional passive cooling. At the same time, it can accurately locate the high temperature target area according to the location information of the temperature sensor, control the corresponding spray unit to start spraying, replace the existing fixed flow rate spraying or overall cooling mode, quickly realize the directional cooling of the high temperature area, avoid the ineffective consumption of coolant and energy, ensure that the equipment to be cooled will not have operational failures due to local high temperature, and improve the operational stability of the equipment. (2) Based on the temperature difference between the real-time temperature of the target area and the safe temperature threshold, the temperature prediction trend, and the operating parameters of the load to be cooled, the present invention constructs a dedicated coolant flow rate adjustment function, automatically calculates the optimal spray flow rate, and achieves linear matching between the spray flow rate and the temperature requirement through the coordinated control of the variable frequency liquid pump and the flow rate adjustment valve. This solves the problems of energy waste and untimely cooling caused by fixed coolant flow rate or manual adjustment in the prior art. When the load is low and the temperature is low, the flow rate is automatically reduced; when the load increases suddenly and the temperature rises sharply, the flow rate is quickly increased. Under the premise of ensuring that the heat dissipation requirements are met, the energy consumption of the liquid cooling system is reduced to the maximum extent. (3) The present invention adds a coolant reflux heat exchange module to perform temperature detection and filtration on the coolant after spraying. The coolant whose temperature does not exceed the preset recovery threshold is uniformly refluxed to the liquid cooling distribution unit (CDU). After being reasonably distributed according to the temperature requirements of the area to be cooled, it is sent to the dry cooler to complete the heat exchange and cooling, and then re-enters the spray circulation pipeline for continuous use. This effectively solves the problems of low coolant utilization, high replenishment cost and unreasonable circulation path of the existing liquid cooling system. The temperature, load and energy consumption data are collected in real time through the closed-loop feedback mechanism. Combined with the genetic algorithm to iteratively optimize the coolant flow rate adjustment function, the flow rate is dynamically matched with the load and temperature changes to accurately adjust the flow rate, so as to achieve the global optimal heat dissipation stability and energy consumption, and further reduce the overall energy consumption and operating cost of the system. (4) This invention uses a central control module to manage all functional modules such as temperature acquisition and prediction, spray control, flow rate adjustment, and cold liquid circulation and recovery, so as to realize the fully automated operation of the process without the need for frequent manual intervention to adjust parameters. At the same time, it is equipped with remote interaction and fault alarm functions, which can provide real-time feedback on the system operation status. When the system is abnormal, an alarm signal is immediately sent to the management personnel terminal, which solves the problems of insufficient overall management and control capabilities, high dependence on manual labor, and easy operation errors in the existing system. This reduces labor costs and improves the stability and reliability of the system operation.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] Figure 1 This is a flowchart of an energy management method for liquid cooling systems based on multimodal fusion according to the present invention. Figure 2 This is a schematic diagram of the overall framework of an energy management system for liquid cooling systems based on multimodal fusion according to the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages disclosed in the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not intended to limit the embodiments of the present invention. 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 this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0024] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0025] The following is combined Figures 1-2 The embodiments of the present invention will be described in detail below.

[0026] Example 1 like Figure 1 The diagram shown is a flowchart of an energy management method for liquid cooling systems based on multimodal fusion, according to the present invention. It is applied to a liquid cooling system for data center racks, where the area to be cooled is the inside of the server rack, and the load to be cooled consists of multiple server hosts. The steps are as follows: S1. Temperature Acquisition and Prediction: A distributed temperature sensor array is deployed within the server rack according to heat load density, with a spacing of 3cm in high-heat CPU areas and 10cm in low-heat areas. The sensor acquisition frequency automatically adjusts with the load: 10Hz under high load and 2Hz under low load. Real-time temperatures are collected from multiple points, and operating parameters such as server power, fan speed, and ambient temperature are read. A genetic algorithm temperature prediction model is trained using a historical temperature-power-ambient temperature dataset to predict the temperature trend for the next 30 seconds, outputting both real-time and predicted temperature values.

[0027] S2. Spray triggering and precise positioning preset thresholds: safe temperature 45℃, warning temperature 55℃, emergency cooling temperature 65℃.

[0028] The system maintains continuous spraying operation. When the real-time temperature / predicted temperature reaches the warning temperature or high-temperature control temperature, the system accurately locates the third unit in the middle of the cabinet based on the high-temperature sensor position, strengthens the spraying supply in that area, and maintains basic spraying status in other areas.

[0029] S3. Adaptive flow rate regulation uses a coolant flow rate regulation function: In the formula: Take an empirical coefficient of 0.85; =Real-time temperature -45℃; Real-time power consumption of the server; The initial temperature of the coolant is 20℃; The ambient temperature is 25℃. The flow rate calculation unit outputs the target flow rate, which is adjusted in conjunction with the variable frequency liquid pump and the flow control valve. The greater the temperature difference and the higher the load, the faster the coolant flow rate, achieving a precise match between heat dissipation requirements and energy consumption.

[0030] S4. Closed-loop feedback and dynamic optimization: Continuously collect target area temperature and system energy consumption data during the spraying process. If the temperature feedback value is below 45℃: reduce the flow rate in stages, first to 50%, and if it is still below the threshold after 10 seconds, stop spraying; Energy consumption data is collected synchronously. With the goal of "lowest energy consumption per unit of heat dissipation", the flow rate coefficient k is iteratively optimized through a genetic algorithm. The parameters are updated every 30 seconds to achieve global optimization of heat dissipation and energy consumption.

[0031] S5. Coolant Recirculation, CDU Distribution and Dry Cooler Heat Exchange: The temperature of the sprayed coolant is detected and filtered. Coolant whose temperature does not exceed the preset recovery threshold is uniformly recirculated into the liquid cooling distribution unit (CDU). The CDU distributes the coolant according to the real-time temperature requirements of each area, and then delivers the distributed coolant to the dry cooler for heat exchange and cooling. The cooled coolant re-enters the spray circulation pipeline and is continuously supplied to each spray unit, realizing the closed-loop recycling of coolant.

[0032] Example 2 This embodiment is a practical hardware system, such as... Figure 2 As shown, the method described in Example 1 includes six main modules and a central control module: The temperature acquisition and prediction module uses a waterproof, high-precision temperature sensor with an accuracy of ±0.1℃ and is automatically calibrated every 24 hours. The prediction processing unit receives sensor data and load parameters, runs a genetic algorithm model, and outputs the real-time temperature and the predicted temperature.

[0033] Each spray head in the spray control module corresponds to an independent flow rate branch; the positioning unit locks the high-temperature area based on the sensor coordinates; the triggering unit controls the continuous operation of the spray throughout the process and adjusts the area spray intensity according to the temperature and load status.

[0034] The flow rate regulation module includes a variable frequency pump, a flow rate regulation valve, a flow rate calculation unit, and a flow rate monitoring unit. The velocity calculation unit calculates the target velocity according to a function; The flow rate monitoring unit compares the actual flow rate with the target flow rate in real time. If the deviation is greater than ±5%, it automatically corrects the pump speed and valve opening to ensure adjustment accuracy.

[0035] The closed-loop optimization module's feedback acquisition unit collects temperature, power, and energy consumption in real time; the optimization processing unit executes a graded temperature control strategy and iteratively optimizes the flow rate adjustment parameters to form a closed loop of temperature control, flow rate, and energy consumption.

[0036] The coolant circulation and recovery module includes a temperature detection unit, a filtration unit, a return pipeline, a liquid cooling distribution unit (CDU), a dry cooler, and a pipeline insulation layer. After spraying, the coolant flows through the return pipeline into the CDU for unified distribution, and then is sent to the dry cooler for full heat exchange and cooling. Qualified low-temperature coolant re-enters the circulation. The insulation layer reduces the temperature loss of the coolant during transportation and heat exchange.

[0037] The central control module, as the master control unit, is electrically connected to all the above modules, and uniformly dispatches commands and uploads data; it has a built-in data storage unit to store data such as temperature, energy consumption, and flow rate parameters for a period of one year; it is equipped with a remote interaction unit to support remote monitoring, parameter modification, and over-temperature / fault alarm push notifications via PC / mobile phone.

[0038] The system operates fully automatically after power-on: temperature acquisition, prediction, positioning, spraying, flow rate adjustment, closed-loop optimization, and coolant recovery. No manual intervention is required throughout the process, achieving continuous, stable, refined, and energy-efficient energy management of the spray-type liquid cooling system.

[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An energy management method for liquid cooling systems based on multimodal fusion, characterized in that, Includes the following steps: S1. Collect real-time temperature data of multiple points in the liquid cooling system's cooling area through a distributed temperature sensor array, and simultaneously obtain the operating parameters of the load to be cooled. Based on a genetic algorithm, construct a temperature change prediction model to predict the temperature change trend within a preset time period and output the predicted temperature value. S2. Preset safe temperature threshold, warning temperature threshold and emergency cooling threshold. Compare the real-time temperature value and predicted temperature value collected by S1 with each threshold to prepare for or trigger the spraying. S3. Based on the temperature difference between the real-time temperature of the target area and the safe temperature threshold, the predicted temperature change trend and the operating parameters of the load to be cooled, a coolant flow rate adjustment function is constructed, and the coolant flow rate is adjusted in real time according to the temperature and load conditions. S4. During the spray cooling process, continuously collect temperature data of the target area and energy consumption data of the liquid cooling system, compare the temperature feedback value with the safe temperature threshold, and iteratively optimize the coolant flow rate adjustment function based on the energy consumption data using a genetic algorithm. S5. The coolant after spraying is returned to the liquid cooling distribution unit. After being temporarily stored and distributed by the liquid cooling distribution unit, it is sent to the dry cooler for heat exchange and cooling. The cooled coolant re-enters the spraying circulation pipeline and is distributed to the corresponding spraying unit according to the temperature requirements of the area.

2. The energy management method for liquid cooling systems based on multimodal fusion according to claim 1, characterized in that: The deployment density of the distributed temperature sensor array in S1 is positively correlated with the heat load density of the area to be cooled, and the acquisition frequency of each sensor is automatically adjusted according to the load operating status; the temperature change prediction model is generated by training through historical temperature data, load operating parameters and ambient temperature data.

3. The energy management method for liquid cooling systems based on multimodal fusion according to claim 2, characterized in that: In S2, the real-time temperature value and predicted temperature value collected by S1 are compared with each threshold. Combined with the load operation status, the heat dissipation demand is judged. Based on the temperature sensor location information, the high temperature target area is accurately located, and the corresponding area spraying is started. The whole area spraying unit keeps running continuously.

4. The energy management method for liquid cooling systems based on multimodal fusion according to claim 3, characterized in that: In S3, the coolant flow rate regulation function uses temperature difference as the core variable and is modified by combining load power, ambient temperature, and initial coolant temperature. The expression is as follows: ； In the formula, For the target spray flow rate, This is the flow rate adjustment coefficient. The difference between the real-time temperature and the safe temperature threshold. This represents the real-time power of the load to be cooled. The initial temperature of the coolant. The ambient temperature.

5. The energy management method for liquid cooling systems based on multimodal fusion according to claim 4, characterized in that: The coolant flow rate regulation function automatically calculates the target spray flow rate. Through the coordinated control of the flow rate regulation valve and the variable frequency liquid pump, the flow rate of the spray coolant is dynamically adjusted, and the temperature difference and flow rate are linearly matched.

6. The energy management method for liquid cooling systems based on multimodal fusion according to claim 5, characterized in that, S4 employs a hierarchical control strategy for closed-loop feedback regulation. If the temperature feedback value is lower than the safe temperature threshold, the spray flow rate will be automatically reduced until the spraying stops and a coolant recovery command will be triggered. At the same time, based on energy consumption data, the coolant flow rate adjustment function will be iteratively optimized through a genetic algorithm to achieve the global optimal balance between heat dissipation and energy consumption.

7. An energy management system for liquid cooling systems based on multimodal fusion, characterized in that, An energy management method for a liquid cooling system based on multimodal fusion as described in any one of claims 1-6 is provided, comprising: a temperature acquisition and prediction module, a spray control module, a flow rate regulation module, a closed-loop optimization module, a coolant reflux heat exchange module, and a central control module; The temperature acquisition and prediction module includes a distributed temperature sensor array and a prediction processing unit. The distributed temperature sensor array is used to acquire real-time temperature data of multiple points in the area to be cooled, and the prediction processing unit is used to obtain the operating parameters of the load to be cooled, construct a temperature change prediction model based on a genetic algorithm, and output real-time temperature values ​​and predicted temperature values. The spray control module includes a spray unit, a positioning unit, and a triggering unit. The spray unit includes several independently controlled spray heads. The positioning unit is used to locate the target area with excessively high temperature based on the position information of the temperature sensor. The triggering unit is used to trigger a spray preparation command or a spray cooling command based on the temperature comparison result, and control the corresponding spray head to start. The flow rate regulation module includes a variable frequency liquid pump, a flow rate regulating valve, and a flow rate calculation unit. The flow rate calculation unit is used to construct a coolant flow rate regulation function and calculate the target spray flow rate based on temperature difference, predicted temperature trend, and load operating parameters. The variable frequency liquid pump and the flow rate regulating valve work together to dynamically regulate the flow rate of the sprayed coolant. The closed-loop optimization module includes a feedback acquisition unit and an optimization processing unit. The feedback acquisition unit is used to continuously collect temperature data and system energy consumption data of the target area. The optimization processing unit is used to compare the temperature feedback value with the safe temperature threshold, adjust the spray flow rate, and iteratively optimize the coolant flow rate adjustment function through a genetic algorithm. The coolant return heat exchange module includes a liquid cooling distribution unit, a dry cooler, a temperature detection unit, and a circulation pipeline. The temperature detection unit is used to detect the temperature of the coolant after spraying, the filtration unit is used to filter the coolant, and the circulation pipeline is used to collect the coolant after spraying through the return pipeline into the liquid cooling distribution unit, and then send it to the dry cooler for heat exchange and cooling, and then supply it to the spraying circulation. The central control module is electrically connected to the temperature acquisition and prediction module, the spray control module, the flow rate regulation module, the closed-loop optimization module, and the coolant reflux heat exchange module, respectively. It is used to receive signals from each module, control the coordinated operation of each module, and realize the automated and refined management of the energy of the liquid cooling system.

8. The energy management system for liquid cooling systems based on multimodal fusion according to claim 7, characterized in that: The spray head of the spray unit adopts a rotatable structure, and each spray head corresponds to an independent flow rate control branch. The spray angle and flow rate are independently adjusted according to the size and temperature distribution of the target area. The distributed temperature sensor array uses waterproof high-precision sensors and has a self-calibration function, which automatically completes a calibration every 24 hours.

9. An energy management system for liquid cooling systems based on multimodal fusion according to claim 8, characterized in that: The flow rate regulation module also includes a flow rate monitoring unit, which monitors the actual value of the spray flow rate in real time, compares the actual flow rate value with the target flow rate value, and automatically adjusts the opening of the flow rate regulation valve and the speed of the variable frequency liquid pump when the deviation exceeds the set threshold to ensure the accuracy of flow rate regulation; the cold liquid circulation and recovery module also includes a heat insulation layer, which is used to insulate the circulation pipeline and reduce the temperature loss of the cold liquid during the circulation process.

10. An energy management system for a liquid cooling system based on multimodal fusion according to claim 9, characterized in that: The central control module also includes a data storage unit and a remote interaction unit. The data storage unit is used to store temperature data, energy consumption data, flow rate regulation parameters and optimization results; the remote interaction unit is used to realize remote monitoring, parameter setting and fault alarm functions, and provide real-time feedback on the system operating status. When the system malfunctions, it immediately sends an alarm signal to the management personnel terminal.

Images