Server power temperature control method, system, and storage medium

Abstract

The application relates to the technical field of server temperature control, and provides a server power supply temperature control method, a system and a storage medium. The server power supply temperature control method comprises the following steps: acquiring a current computing power load parameter of a main server and a current temperature parameter inside a main shell; judging whether the main server satisfies preset vacuumization starting conditions based on the current computing power load parameter and the current temperature parameter, and obtaining a judgment result; based on the judgment result satisfying the preset vacuumization starting conditions, controlling a power supply module to supply power to a vacuumization device; based on the reduced current absolute pressure value, determining an actual boiling temperature of cooling liquid contained in the main shell under the reduced current absolute pressure value; and based on the actual boiling temperature being lower than a preset frequency reduction threshold temperature, driving the cooling liquid to absorb heat generated by the main server and to evaporate by phase change before the temperature of the main server rises to the preset frequency reduction threshold temperature, so as to take away the heat of the main server.

Description

Technical Field

[0001] This application relates to the field of server temperature control technology, and in particular to server power supply temperature control methods, systems and storage media. Background Technology

[0002] As a core power supply component of server systems, the power density of server power supplies continuously increases with the growing computing power demands of servers. During the process of converting AC to DC to power various server components, server power supplies generate a significant amount of Joule heat. With the surge in server computing load, the output power of the server power supply increases accordingly, causing a sharp rise in the heat generated by its internal power devices (such as MOSFETs, transformers, and inductors). If the heat generated by the server power supply cannot be dissipated in time, its temperature will continue to rise, leading to decreased power conversion efficiency, poor output stability, and in severe cases, even triggering the power supply's overheat protection mechanism and forcibly shutting down the system, causing server system crashes.

[0003] Currently, a common approach to cooling server power supplies is to immerse both the power supply and the main server in a coolant-filled casing, utilizing direct contact between the coolant and the power supply / main server for heat transfer. In this method, a single-phase immersion coolant (such as mineral oil or synthetic ester oil) is typically used. This coolant utilizes its own sensible heat (i.e., the heat absorbed by the increased temperature) to dissipate the heat generated by the power supply and main server. Specifically, the main casing is filled with single-phase coolant, and the power supply and main server are completely immersed in it. The heat generated by the power supply and main server is directly transferred to the coolant in contact with it, causing the coolant temperature to rise. The heated coolant is then driven by an external circulation pump through an external heat exchanger, releasing the heat to the external environment before returning to the main casing, thus achieving continuous cooling of the power supply and main server.

[0004] However, the existing single-phase immersion cooling method has a major drawback: the heat exchange efficiency of the coolant is highly dependent on its specific heat capacity and circulation flow rate. Single-phase coolants rely solely on sensible heat for heat exchange, meaning the amount of heat removed per unit volume is extremely limited. When the server power supply outputs high power instantaneously due to a surge in computing power demand, a large amount of heat is released rapidly within a very short time, causing the coolant temperature to rise quickly. Because single-phase coolants cannot undergo a phase change (i.e., they cannot utilize latent heat of vaporization), their heat exchange is strictly limited to the sensible heat range corresponding to the coolant temperature rise. This causes the coolant temperature inside the main casing to easily and rapidly rise to the preset throttling threshold temperature of the server power supply or main server. Once this preset throttling threshold temperature is reached, the server power supply or main server will be forced to reduce its operating frequency to decrease heat generation, directly leading to a decrease in the overall computing power output of the server system, making it unable to meet the continuous computing power demands of high-load application scenarios.

[0005] Therefore, how to provide a method that can overcome the limitation of single-phase coolant sensible heat transfer efficiency and effectively delay the server power supply or main server from reaching the preset throttling threshold temperature when the server power supply generates instantaneous high heat due to a surge in computing power demand is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] This application aims to address at least one of the technical problems existing in the related art.

[0007] According to an embodiment of the first aspect of this application, a server power supply temperature control method is applied to a server system. The server system includes a power module, a main housing, a main server, a coolant, and a vacuum pump. The main server and the coolant are disposed inside the main housing, and the vacuum pump is disposed on the outer wall of the main housing and communicates with the interior of the main housing. The control method includes: S100. Based on the current operating status of the main server, obtain the current computing power load parameters of the main server and the current temperature parameters inside the main casing; S200. Based on the current computing power load parameters and the current temperature parameters, determine whether the main server meets the preset vacuum start-up conditions, and obtain the determination result; S300. Based on the judgment result, if the preset vacuuming start condition is met, control the power module to supply power to the vacuuming device so that the vacuuming device performs a vacuuming action on the main housing, reduces the absolute pressure inside the main housing, and obtains the reduced current absolute pressure value. S400. Based on the reduced current absolute pressure value, determine the actual boiling temperature of the coolant contained inside the main housing at the reduced current absolute pressure value, wherein the actual boiling temperature is lower than the preset frequency reduction threshold temperature corresponding to the main server. S500: Based on the fact that the actual boiling temperature is lower than the preset frequency reduction threshold temperature, the coolant is driven to absorb the heat generated by the operation of the main server and undergo phase change evaporation before the temperature of the main server rises to the preset frequency reduction threshold temperature, so as to remove the heat of the main server, and at the same time, the real-time pressure change state inside the main casing caused by the phase change evaporation and the real-time temperature of the main server are obtained. S600. Based on the real-time pressure change status and the real-time temperature of the main server, control the operation of the vacuum pump to maintain the absolute pressure inside the main housing within a preset target pressure range.

[0008] The control system according to a second aspect of this application includes a control component, which is used to perform the control method described above.

[0009] An electronic device according to a third aspect of this application includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the control method described above.

[0010] According to a fourth aspect of this application, a non-transitory computer-readable storage medium includes a computer program that, when executed by the processor, implements the control method described above.

[0011] According to a fifth aspect embodiment of the present application, the computer program product includes a computer program that, when executed by the processor, implements the control method described above.

[0012] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1This is a schematic flowchart of the server power supply temperature control method provided in the embodiments of this application; Figure 2 This is a schematic flowchart of the automatic temperature control method for power supplies provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0015] The embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but should not be used to limit the scope of this application.

[0016] In the description of the embodiments of this application, it should be noted that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the embodiments of this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this application. In addition, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0017] In the description of the embodiments of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application based on the specific circumstances.

[0018] In the embodiments of this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0019] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the embodiments of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0020] According to the embodiments of the first aspect of this application, such as Figure 1 As shown, this application provides a server power supply temperature control method. The control method is applied to a server system, which includes a power supply, a main casing, a main server, a coolant, and a vacuum pump. The main server and the coolant are disposed inside the main casing, and the vacuum pump is disposed on the outer wall of the main casing and communicates with the interior of the main casing.

[0021] The control method includes the following steps: S100. Based on the current operating status of the main server, obtain the current computing load parameters of the main server and the current temperature parameters inside the main casing. Specifically, the current computing load parameters of the main server can characterize the power consumption level or task processing intensity of the main server at the current moment. This computing load parameter can be obtained by reading the real-time output power data of the main server's management controller (such as BMC, i.e., baseboard management controller) or power management unit (PMU). The current temperature parameters inside the main casing can be collected in real time by a temperature sensor located inside the main casing (e.g., near the main server or below the coolant level). This temperature parameter can reflect the heat generation of the main server in the current operating state and the current temperature level of the coolant inside the main casing.

[0022] S200. Based on the current computing load parameters and the current temperature parameters, determine whether the main server meets the preset vacuuming startup conditions, and obtain the determination result. Specifically, the preset vacuuming startup conditions are: when the current computing load parameters are greater than a preset computing load threshold (e.g., 80% of the rated power) and the current temperature parameters are greater than a preset temperature threshold (e.g., 5°C below the throttling threshold), the startup conditions are determined to be met. This combination of conditions can identify in advance that the main server is about to enter a high-heat operating condition and the temperature is close to the throttling threshold, requiring early intervention to start phase change cooling. When either condition is not met, such as a low computing load or the current temperature being within a safe range, the startup conditions are determined not to be met, and the system maintains its current state of operation. This judgment logic ensures that the vacuuming action is only triggered when truly necessary, avoiding unnecessary power consumption and mechanical wear.

[0023] S300: Based on the judgment result, if the preset vacuuming start-up condition is met, the control power supply supplies power to the vacuuming device, so that the vacuuming device performs a vacuuming action on the main housing, reducing the absolute pressure inside the main housing, and obtaining the reduced current absolute pressure value. Specifically, when step S200 determines that the preset vacuuming start-up condition is met, the controller (e.g., the management controller of the server system or an independent temperature control module) sends a power supply command to the power supply, and the power supply outputs operating voltage to the vacuuming device. The vacuuming device (e.g., a vacuum pump) begins to evacuate the inside of the main housing, extracting the gas (including air and any water vapor that may be generated) inside the main housing, causing the absolute pressure inside the main housing to gradually decrease from atmospheric pressure (approximately 101 kPa). During the vacuuming process, the pressure sensor installed inside the main housing monitors and feeds back the current absolute pressure value in real time. When the absolute pressure drops to the preset start-up pressure value (e.g., 40 kPa), the control component records the reduced current absolute pressure value and can pause the vacuuming as needed to observe the cooling effect or continue evacuating to a lower pressure.

[0024] S400. Based on the reduced current absolute pressure value, determine the actual boiling temperature of the coolant contained inside the main casing at the reduced current absolute pressure value. This actual boiling temperature is lower than the preset throttling threshold temperature corresponding to the main server. Specifically, according to the phase change principle, the boiling temperature of a liquid decreases as the ambient pressure decreases. The boiling temperatures of different coolants (e.g., mineral oil, synthetic ester oil, fluorinated liquid, etc.) at different absolute pressures can be determined by their saturated vapor pressure curves. The control unit has a pre-stored saturated vapor pressure-temperature lookup table or fitting formula for the currently used coolant. After obtaining the reduced current absolute pressure value, the control unit obtains the actual boiling temperature of the coolant at that pressure by looking up the table or calculating. This actual boiling temperature is designed to be lower than the preset throttling threshold temperature of the main server (e.g., if the preset throttling threshold temperature is 85°C, the actual boiling temperature can be reduced to 50°C to 60°C), thereby ensuring that the coolant begins to boil and absorb heat before the server temperature reaches the throttling threshold.

[0025] S500: Based on the fact that the actual boiling temperature is lower than the preset frequency reduction threshold temperature, the coolant is driven to absorb the heat generated by the main server's operation and undergo phase change evaporation before the main server's temperature rises to the preset frequency reduction threshold temperature, thereby removing the heat from the main server. Simultaneously, the real-time pressure change state inside the main casing due to the phase change evaporation and the real-time temperature of the main server are acquired. Specifically, when the absolute pressure inside the main casing is reduced to a level that makes the actual boiling temperature of the coolant lower than the preset frequency reduction threshold temperature, the heat generated by the main server's operation is transferred to the coolant in contact with it. Since the temperature of the main server (i.e., the heat source temperature) rises rapidly during power surges, when the local temperature of the main server or the coolant temperature reaches the actual boiling temperature, the coolant begins to boil on the surface of the main server or inside the main casing, undergoing a liquid-to-gas phase transition. This phase change process absorbs a large amount of latent heat, and the heat removed by a unit volume of coolant is far greater than the heat removed by a single-phase coolant relying on sensible heat at the same temperature rise. Therefore, before the main server's temperature rises to the preset frequency reduction threshold temperature, the boiling phase change effectively removes a large amount of heat generated by the main server. During the phase change evaporation process, the gas generated by the evaporation of coolant will cause the pressure inside the main casing to change. The control component obtains the pressure change status in real time through a pressure sensor, and at the same time obtains the real-time temperature of the main server through a temperature sensor, providing feedback data for the closed-loop control of step S600.

[0026] S600. Based on the real-time pressure change and the real-time temperature of the main server, the operation of the vacuum pump is controlled to maintain the absolute pressure inside the main casing within the preset target pressure range. Specifically, after the coolant undergoes phase change evaporation, the pressure inside the main casing gradually increases due to gas generation, causing the actual boiling temperature of the coolant to rise accordingly. The control unit dynamically adjusts the operating state of the vacuum pump (e.g., starts vacuuming, stops vacuuming, or adjusts the pumping rate) according to the real-time pressure change and the real-time temperature of the main server, so that the absolute pressure inside the main casing is always maintained within the preset target pressure range. This preset target pressure range is a pressure range whose corresponding coolant boiling temperature range is generally lower than the preset frequency reduction threshold temperature, while avoiding excessive pressure leading to excessive boiling of the coolant or excessive negative pressure on the main casing structure. When the real-time pressure rises to the upper limit of the target pressure range, the control unit starts or accelerates the vacuum pump to reduce the pressure back to within the range; when the real-time pressure drops to the lower limit of the target pressure range, the control unit stops the vacuum pump, allowing the pressure to rise naturally. This closed-loop control ensures that the coolant can continuously absorb heat through boiling phase change during high-power operation of the main server, effectively delaying or preventing the main server temperature from reaching the preset throttling threshold temperature.

[0027] The beneficial effects of this implementation method are significant. First, by pre-vacuuming before a power surge in the main server to reduce the absolute pressure inside the main casing, the actual boiling temperature of the coolant is lowered to below the preset throttling threshold temperature, thus initiating phase change cooling while the server temperature is still within a safe range. The coolant absorbs a large amount of latent heat of vaporization during boiling and vaporization, and its heat transfer per unit volume far exceeds the heat that traditional single-phase coolants can remove by relying solely on sensible heat. This effectively overcomes the heat dissipation bottleneck of single-phase immersion cooling, promptly removing the instantaneous high heat flux density generated by the server power supply or the main server due to a surge in computing power demand, delaying or even preventing it from reaching the throttling threshold temperature, thereby ensuring the server system's continuous computing power output capability under high load scenarios. Second, by monitoring the pressure changes inside the main casing and the real-time temperature of the main server in real time, the operation of the vacuum equipment is controlled in a closed loop to maintain the absolute pressure inside the main casing within the preset target pressure range, ensuring that the actual boiling temperature of the coolant is always below the throttling threshold temperature, achieving continuous and stable operation of phase change cooling. Furthermore, this control method activates the vacuum equipment only when preset vacuuming start-up conditions are met, avoiding ineffective operation of the vacuum equipment under low-load conditions, reducing system energy consumption and equipment wear, and extending the service life of the vacuum equipment. In addition, this method requires no hardware modification to the main casing and coolant; it can be implemented simply by adding the vacuum equipment and corresponding control logic, offering advantages such as low modification cost and wide applicability. This control method can be widely applied to high-power-density server systems, server systems, and high-performance computing devices, effectively improving the system's thermal management capabilities and operational stability under high-load conditions.

[0028] In some examples, the preset vacuuming start condition may also include the current computing load change rate of the main server, that is, when the computing load rises sharply in a short period of time, the vacuuming is started in advance to compensate for system response delay. The coolant may be a single-phase immersion coolant with a high boiling point and stable chemical properties, the vacuuming equipment may be a rotary vane vacuum pump or a Roots vacuum pump, and the preset frequency reduction threshold temperature may be set according to the specific model of the main server and the working environment.

[0029] In one embodiment of this application, step S100 specifically includes the following steps: S101. Based on the current operating status of the main server, obtain the current computing load parameters of the main server. The current computing load parameters represent the amount of computing tasks undertaken by the main server at the current moment. Specifically, the controller (e.g., the server's baseboard management controller (BMC) or a separate temperature management module) obtains parameters that reflect the amount of computing tasks, such as the power consumption, CPU / GPU utilization, or task queue depth of the main server at the current moment, by reading the real-time output data of the main server's power management unit (PMU) or system management interface, as the current computing load parameters.

[0030] S102. Based on a temperature sensor installed inside the main housing, the current temperature parameter inside the main housing is collected. The current temperature parameter represents the instantaneous temperature value inside the main housing at the current moment. Specifically, the temperature sensor is installed inside the main housing and immersed in the coolant, for example, near the heat-generating area of ​​the main server or on the coolant return path. The controller reads the output signal of the temperature sensor at a preset sampling frequency to obtain the current instantaneous temperature value inside the main housing.

[0031] S103. Based on the current computing load parameters, determine the preset frequency reduction threshold temperature corresponding to the main server. The preset frequency reduction threshold temperature is the critical temperature value at which the main server triggers frequency reduction operation due to overheating. Specifically, the critical temperature for triggering frequency reduction may differ for main servers of different models or configurations under different computing loads. The controller stores a table of correspondence between each computing load level and the frequency reduction threshold temperature. The controller determines the preset frequency reduction threshold temperature corresponding to the main server under the current load condition by looking up the table or interpolating the calculation based on the current computing load parameters.

[0032] S104. Based on the current computing load parameters, determine the preset vacuuming start condition corresponding to the main server. The preset vacuuming start condition is a criterion used to determine whether to start the vacuuming equipment to perform vacuuming. Specifically, the preset vacuuming start condition is a criterion dynamically adjusted according to the current computing load parameters, including a combination of computing load thresholds and temperature thresholds required to trigger vacuuming. For example, under low computing load conditions, the trigger condition can be set to a higher computing load threshold and a higher temperature threshold; under high computing load conditions, the trigger condition can be set to a lower computing load threshold and a lower temperature threshold to start vacuuming in advance for phase change cooling preparation. The controller dynamically determines this criterion based on the current computing load parameters.

[0033] S105. Based on the current computing power load parameters and the preset vacuum start-up conditions, determine the computing power load level corresponding to the current computing power load parameters. Specifically, the controller maps the current computing power load parameters to the corresponding computing power load level, such as "low load," "medium load," "high load," or "peak load," according to a preset computing power load level classification standard (e.g., based on the percentage of current power consumption to rated power). This computing power load level is used to select the corresponding pressure configuration in subsequent steps.

[0034] S106. Based on the computing power load level, select a preset target pressure range corresponding to the computing power load level from the preset pressure configuration table. The preset target pressure range is the range of absolute pressure values ​​to be achieved inside the main casing after the vacuuming operation is completed. Specifically, the preset pressure configuration table stores the correspondence between each computing power load level and the target pressure range—under different load levels, the heat generation and heat flux density of the main server are different, and the actual boiling temperature of the coolant required to initiate the boiling phase change is also different, thus corresponding to different target absolute pressure ranges for the main casing. The controller retrieves and selects the preset target pressure range corresponding to the current computing power load level from the preset pressure configuration table as the target value for subsequent vacuuming control.

[0035] The beneficial effects of this implementation method are significant. First, by simultaneously determining the preset throttling threshold temperature, preset vacuuming start conditions, computing load level, and corresponding preset target pressure range for the main server during the current computing load parameter acquisition step, deep coupling between the vacuuming control logic and the real-time computing load of the main server is achieved. The heating characteristics of the main server differ under different load levels, requiring different pressure conditions to initiate phase change cooling. Selecting the target pressure range based on the differentiated computing load level allows for more precise matching of current heat dissipation needs. This avoids excessive vacuuming and energy waste under low loads and establishes a lower pressure environment in advance under high loads to reserve a larger margin for phase change heat absorption. Second, the correlation between the preset throttling threshold temperature and the current computing load parameters allows the temperature control strategy to adapt to the actual throttling characteristics of different main server models under different load conditions, improving the compatibility and universality of the control strategy. Third, the preset vacuuming start conditions are dynamically adjusted according to the current computing load parameters, making the triggering timing of the vacuuming action more reasonable—intervening earlier under high load conditions and delaying intervention under low load conditions, balancing heat dissipation needs and system energy consumption. This step, together with subsequent steps such as vacuuming, pressure maintenance, and phase change cooling, constitutes an intelligent phase change cooling temperature control method that can be finely adjusted according to computing load.

[0036] In one embodiment of this application, step S200, which determines whether the main server meets the preset vacuum start-up conditions based on the current computing power load parameters and the current temperature parameters, specifically includes the following sub-steps: S201. Based on the current computing power load parameters, compare the current computing power load parameters with a preset load threshold to obtain a load comparison result. The preset load threshold is a pre-set critical computing power value used to distinguish between high load and low load states. Specifically, the controller reads the preset load threshold (e.g., 80% of the main server's rated power or 75% of the CPU utilization) from memory or registers, compares the current computing power load parameters obtained in step S101 with the preset load threshold, determines whether the current computing power load has entered the high load range, and obtains the load comparison result.

[0037] S202. Based on the current temperature parameter, compare the current temperature parameter with a preset temperature threshold to obtain a temperature comparison result. The preset temperature threshold is a pre-set critical temperature value used to determine whether the internal temperature of the main housing is close to the preset frequency reduction threshold temperature. Specifically, the controller reads the preset temperature threshold (e.g., the preset frequency reduction threshold temperature minus 5°C) from memory or a register, compares the current temperature parameter collected in step S102 with the preset temperature threshold, and determines whether the current temperature inside the main housing is close to the frequency reduction threshold temperature to obtain a temperature comparison result.

[0038] S203. Based on the load comparison result and the temperature comparison result, determine whether the load comparison result indicates that the current computing power load parameter is greater than the preset load threshold, and obtain a first determination sub-result. Specifically, the controller performs logical parsing on the load comparison result. If the current computing power load parameter is greater than the preset load threshold, the first determination sub-result is "yes" (or logical value "1"), indicating that the main server is currently in a high-load operation state; otherwise, it is "no", indicating that the current computing power load has not yet reached the level that requires phase change cooling intervention.

[0039] S204. Based on the load comparison result and the temperature comparison result, determine whether the temperature comparison result indicates that the current temperature parameter is greater than the preset temperature threshold, and obtain a second determination sub-result. Specifically, the controller performs logical parsing on the temperature comparison result. If the current temperature parameter is greater than the preset temperature threshold, the second determination sub-result is "yes," indicating that the temperature inside the main casing has approached or reached the warning level that requires phase change cooling intervention; otherwise, it is "no," indicating that the current temperature is still within the safe range.

[0040] S205. Based on the first determination sub-result and the second determination sub-result, when the first determination sub-result indicates that the current computing power load parameter is greater than the preset load threshold and the second determination sub-result indicates that the current temperature parameter is greater than the preset temperature threshold, the determination result is determined to meet the preset vacuum start-up condition. Specifically, the controller performs a logical AND operation on the first determination sub-result and the second determination sub-result—only when both determination sub-results are "yes" will the controller finally determine that the main server meets the preset vacuum start-up condition and output the determination result to step S300 to trigger vacuum start-up; if either determination sub-result is "no", it is determined that the preset vacuum start-up condition is not met, the system maintains the current operating state and continues to monitor.

[0041] The beneficial effects of this implementation are significant. First, by setting two independent judgment conditions—a preset load threshold and a preset temperature threshold—and using a logical AND operation, a two-factor joint judgment of the vacuum pumping start condition is achieved. Vacuum pumping is only triggered when the main server is simultaneously under high computing load (i.e., with high heat potential) and the internal temperature of the main casing has risen to a warning level close to the throttling threshold temperature. This dual-condition design effectively avoids the accidental activation of the vacuum pumping device when the main server temperature rises due to a brief fluctuation but the computing load is not high, reducing system energy consumption and unnecessary equipment operation. Second, the high load condition judgment ensures that the vacuum pumping action is only initiated when the main server actually has a continuous high heat demand, avoiding excessive intervention in low-load idle states. Third, the near-throttling threshold temperature judgment ensures that the vacuum pumping action intervenes in advance before the main server temperature reaches the throttling threshold temperature. By utilizing the time window for vacuum pumping to lower the boiling point of the coolant, phase change cooling can be established in time before the temperature reaches the throttling threshold, providing an effective temperature buffer margin for the main server. This dual-condition judgment logic works in conjunction with the subsequent steps of vacuuming in S300, boiling temperature determination in S400, and phase change cooling in S500 to form an intelligent phase change cooling start-up strategy based on "load-temperature" dual-state perception, ensuring that the vacuuming action is initiated at the appropriate time and plays its role under the appropriate conditions.

[0042] In one embodiment of this application, step S300, which involves controlling the vacuuming equipment to perform a vacuuming action and obtaining the reduced current absolute pressure value based on the determination result meeting the preset vacuuming start condition, specifically includes the following steps: S301. Based on the judgment result indicating that the preset vacuuming start condition is met, a vacuuming start command is generated. This vacuuming start command is used to trigger the power supply to supply power to the vacuuming equipment. Specifically, after the controller determines in step S200 that the main server meets the preset vacuuming start condition, the controller's internal logic module generates a vacuuming start command signal. This command signal is sent to the power supply's control terminal to request the power supply to provide operating power to the vacuuming equipment.

[0043] S302. Based on the vacuum start command, the power supply is controlled to supply electrical energy to the vacuum equipment, so that the vacuum equipment starts and performs vacuuming operations at a preset initial power. The preset initial power is the pre-set starting operating power of the vacuum equipment when it begins vacuuming. Specifically, after receiving the vacuum start command, the power supply supplies power to the drive motor of the vacuum equipment (e.g., a vacuum pump) according to preset voltage and current output parameters. The vacuum equipment starts operating at the preset initial power (e.g., 60% of the rated power) and begins evacuating the inside of the main housing. Setting the preset initial power avoids the inrush current and mechanical shock that may occur when the vacuum equipment starts at full power, which helps to extend the equipment's lifespan.

[0044] S303. Based on the vacuuming device performing a vacuuming action at the preset initial power, the current absolute pressure value inside the main housing is acquired in real time, thus obtaining the real-time absolute pressure value during the vacuuming process. Specifically, during the continuous vacuuming process, a pressure sensor installed inside the main housing continuously collects the absolute pressure value inside the main housing at a preset sampling frequency, and sends the collected real-time data to the controller, enabling the controller to dynamically grasp the pressure drop trend and current pressure state inside the main housing.

[0045] S304. Based on the real-time absolute pressure value during the vacuuming process, the real-time absolute pressure value is compared with a preset vacuuming termination pressure threshold to obtain a termination comparison result. The preset vacuuming termination pressure threshold is a pre-set absolute pressure critical value used to determine the cessation of the vacuuming action. Specifically, the controller compares each collected real-time absolute pressure value with the preset vacuuming termination pressure threshold (e.g., 30 kPa) stored in the controller. This threshold is set comprehensively based on the saturated vapor pressure characteristics of the coolant, the frequency reduction threshold temperature of the main server, and a safety margin. When the real-time absolute pressure value is greater than the threshold, the comparison result is "termination condition not met," and the vacuuming equipment continues to operate; when the real-time absolute pressure value is not greater than the threshold, the comparison result is "termination condition met."

[0046] S305. Based on the termination comparison result indicating that the real-time absolute pressure value is not greater than the preset vacuum termination pressure threshold, the controller stops supplying power to the vacuuming equipment, causing the vacuuming equipment to stop performing vacuuming operations, and records the real-time absolute pressure value as the reduced current absolute pressure value. Specifically, when the controller determines that the real-time absolute pressure value has reached or fallen below the preset vacuum termination pressure threshold, the controller sends a stop power supply command to the power supply, cutting off the power supply to the vacuuming equipment, and the vacuuming equipment stops operating. Simultaneously, the controller stores the current real-time absolute pressure value in a register or memory as the reduced current absolute pressure value, for use in the subsequent step S400 to calculate the actual boiling temperature of the coolant.

[0047] The beneficial effects of this implementation method are significant. First, by starting the vacuum equipment with a preset initial power and gradually reducing the internal pressure of the main casing, the current and mechanical shocks during full-power startup are avoided. This ensures a smooth and controllable pressure reduction process, providing a gradually adaptable pressure change environment for the coolant and main server inside the main casing, preventing violent boiling and splashing of the coolant or stress impact on the main casing structure caused by a sudden pressure drop. Second, by real-time monitoring of the absolute pressure inside the main casing and comparing it with a preset vacuum termination pressure threshold in a closed-loop manner, precise termination control of the vacuuming action is achieved. This ensures that the internal pressure of the main casing is precisely reduced to near the target value, avoiding excessive vacuuming that could cause excessive negative pressure stress on the main casing or excessive evaporation of the coolant. It also avoids insufficient vacuuming that could result in the actual boiling temperature of the coolant remaining above the frequency reduction threshold temperature, thus preventing effective phase change cooling. Third, this termination judgment logic precisely controls the operating time of the vacuum equipment within the required range, preventing the vacuum equipment from continuing ineffective operation after reaching the target pressure, reducing system energy consumption and equipment wear, and extending the service life of the vacuum equipment.

[0048] In one embodiment of this application, a boiling temperature verification step is further included between step S400 and step S500, specifically including the following sub-steps: S401. Based on the reduced current absolute pressure value, the actual boiling temperature of the coolant at the reduced current absolute pressure value is obtained by consulting a property parameter lookup table. The property parameter lookup table is a pre-stored set of mapping relationships characterizing the boiling temperatures of the coolant under different absolute pressures. Specifically, the controller internally stores a saturated vapor pressure-temperature lookup table for the currently used coolant (e.g., mineral oil, synthetic ester oil, or fluorinated liquid), which is obtained through experimental calibration or from property data provided by the coolant supplier. The controller uses the reduced current absolute pressure value recorded in step S305 as a lookup index to retrieve the property parameter lookup table and calculates the actual boiling temperature of the coolant corresponding to that pressure value through lookup or interpolation.

[0049] S402. Based on the actual boiling temperature and the preset frequency reduction threshold temperature, the actual boiling temperature is numerically compared with the preset frequency reduction threshold temperature to obtain a boiling temperature comparison result. The boiling temperature comparison result is used to determine whether the actual boiling temperature is lower than the preset frequency reduction threshold temperature. Specifically, the controller compares the actual boiling temperature obtained from the physical property parameter lookup table with the preset frequency reduction threshold temperature corresponding to the main server determined in step S103 to determine whether the actual boiling temperature is lower than the frequency reduction threshold temperature.

[0050] S403. Based on the boiling temperature comparison result indicating that the actual boiling temperature is lower than the preset frequency reduction threshold temperature, the coolant is allowed to continuously absorb the heat generated by the main server during the entire heating process before the main server temperature rises to the preset frequency reduction threshold temperature and undergo phase change evaporation. Specifically, after verifying that the actual boiling temperature is indeed lower than the preset frequency reduction threshold temperature, the controller confirms that the current reduced absolute pressure value meets the start-up conditions for phase change cooling. At this time, the coolant can continuously absorb heat and undergo phase change evaporation throughout the entire process of the main server temperature rising from the current value to the frequency reduction threshold temperature, as long as the main server temperature reaches or exceeds the actual boiling temperature, using the latent heat to remove the heat generated by the main server.

[0051] Following step S500, a pressure maintenance and temperature hysteresis control step is also included, specifically comprising the following sub-steps: S501. Based on the real-time pressure change state inside the main housing caused by the phase change evaporation, when the real-time pressure change state indicates that the absolute pressure rise rate inside the main housing is greater than a preset pressure rise rate threshold, the power supply is controlled to re-supply the vacuum pump, so that the vacuum pump starts and performs a supplementary vacuum pumping action. The preset pressure rise rate threshold is a pre-set critical rate value used to determine that the pressure inside the main housing rises sharply due to a large amount of phase change evaporation. Specifically, during the phase change evaporation of the coolant, the gas generated by the coolant evaporation will cause the absolute pressure inside the main housing to gradually increase. The controller monitors the absolute pressure change rate (i.e., the pressure increment per unit time) inside the main housing in real time through a pressure sensor. When this rise rate exceeds the preset pressure rise rate threshold, it indicates that the evaporation rate of the coolant is too fast, and the pressure inside the main housing is rising rapidly. If not intervened in time, the actual boiling temperature of the coolant will rise above the frequency reduction threshold temperature, thereby losing the phase change cooling capacity. Based on the determination result, the controller sends a power supply command to the power source again, causing the vacuum pumping equipment to restart and perform supplementary vacuuming to reduce the internal pressure of the main housing back to the target range.

[0052] S502. Based on the real-time temperature of the main server, when the real-time temperature of the main server drops to a preset restart temperature threshold, the power supply is controlled to stop supplying power to the vacuum pump, so that the vacuum pump stops working. The preset restart temperature threshold is a pre-set temperature lower than the preset frequency reduction threshold and is used to determine that the thermal state of the main server has recovered to a safe level. Specifically, when the computing load of the main server decreases or the phase change cooling plays a sufficient role, the real-time temperature of the main server begins to drop. The controller continuously monitors the real-time temperature of the main server. When the temperature drops to the preset restart temperature threshold, it indicates that the main server has left the high-heat state and no longer needs the continuous intervention of phase change cooling. The controller then sends a stop power supply command to the power supply, cutting off the power supply to the vacuum pump. The vacuum pump stops working, and the internal pressure of the main casing gradually rises to normal pressure or is maintained at a low pressure level under natural conditions. The system returns to standby or low-load operation mode. Setting a preset restart temperature threshold (e.g., the frequency reduction threshold temperature minus 10°C to 15°C) prevents the vacuum equipment from frequently starting and stopping when the main server temperature fluctuates briefly, thus improving the system's operational stability and equipment lifespan.

[0053] The beneficial effects of this implementation are significant. First, by adding a boiling temperature verification step between steps S400 and S500, the reduced current absolute pressure value is converted into the actual boiling temperature of the coolant using a physical property parameter comparison table, and compared with the preset frequency reduction threshold temperature. This ensures that the actual boiling temperature is indeed lower than the frequency reduction threshold temperature before phase change evaporation is allowed, avoiding the risk of phase change cooling failure due to insufficient pressure reduction. Second, by monitoring the rate of pressure rise inside the main casing in real time, supplementary vacuuming is initiated promptly when the pressure rises sharply due to a large amount of phase change evaporation. This maintains the absolute pressure inside the main casing at a level that keeps the coolant boiling temperature below the frequency reduction threshold temperature, ensuring the continuous effectiveness of phase change cooling and preventing the coolant boiling temperature from rising above the frequency reduction threshold temperature due to pressure rebound, thus losing its heat dissipation advantage. Third, by using temperature hysteresis control with a preset restart temperature threshold, the vacuuming equipment is stopped promptly after the main server temperature drops to a safe level. This avoids frequent start-stop of the vacuuming equipment during short-term fluctuations in the main server temperature, reducing equipment wear and energy consumption, while ensuring the system's operational stability during temperature fluctuations. The pressure maintenance and temperature hysteresis control steps, together with the aforementioned vacuuming start-up, boiling temperature determination, and phase change cooling steps, constitute a complete vacuuming-phase change cooling closed-loop control system, ensuring that the main server temperature is continuously controlled below the frequency reduction threshold temperature under high load conditions.

[0054] In one embodiment of this application, step S600, based on the real-time pressure change and the real-time temperature of the main server, controls the operation of the vacuum pump to maintain the absolute pressure inside the main housing within a preset target pressure range. This specifically includes the following sub-steps: S601. Based on the real-time pressure change state inside the main casing caused by the phase change evaporation, obtain the current real-time absolute pressure value inside the main casing. Specifically, after the coolant undergoes phase change evaporation, the absolute pressure inside the main casing continuously changes due to the generation of steam. The controller collects the current absolute pressure value in real time at a preset sampling frequency using a pressure sensor installed inside the main casing, as the basis data for subsequent interval judgment.

[0055] S602. Based on the real-time temperature of the main server, the real-time temperature of the main server is compared with the preset frequency reduction threshold temperature to obtain a real-time temperature comparison result. The real-time temperature comparison result is used to determine whether the real-time temperature of the main server is still lower than the preset frequency reduction threshold temperature. Specifically, the controller obtains the current temperature of the main server in real time through a temperature sensor located near the main server or inside the main casing, and compares this temperature with the preset frequency reduction threshold temperature. This comparison result is used to determine whether the main server is still in a heat-generating state requiring phase change cooling protection.

[0056] S603. Based on the current real-time absolute pressure value and the preset target pressure range, the current real-time absolute pressure value is compared with the upper and lower limits of the preset target pressure range to obtain a pressure range comparison result. Specifically, the controller reads the preset target pressure range (with upper and lower limits) selected in step S106 from memory, and compares the current real-time absolute pressure value with the upper and lower limits respectively to determine whether the current pressure is within the range, higher than the upper limit, or lower than the lower limit.

[0057] S604. Based on the pressure range comparison result and the real-time temperature comparison result, when the pressure range comparison result indicates that the current real-time absolute pressure value is lower than the lower limit of the preset target pressure range, the power supply is controlled to stop supplying power to the vacuuming equipment, so that the vacuuming equipment stops the vacuuming operation, and the absolute pressure inside the main casing naturally recovers due to the continuous phase change evaporation of the coolant. Specifically, when the controller determines that the current real-time absolute pressure value has dropped below the lower limit of the target pressure range, it indicates that the pressure inside the main casing is too low, the coolant is over-boiling, or the vacuuming is excessive. At this time, the controller sends a stop power supply command to the power supply, and the vacuuming equipment stops operating. As the coolant continues to undergo phase change evaporation to generate steam under the continuous heating condition of the main server, the absolute pressure inside the main casing will naturally recover until it re-enters the target pressure range.

[0058] S605. Based on the pressure range comparison result and the real-time temperature comparison result, when the pressure range comparison result indicates that the current real-time absolute pressure value is higher than the upper limit of the preset target pressure range and the real-time temperature comparison result indicates that the real-time temperature of the main server is lower than the preset frequency reduction threshold temperature, the power supply is controlled to supply power to the vacuum pumping device, so that the vacuum pumping device starts at a preset maintenance power and performs a vacuuming action, thereby reducing the absolute pressure inside the main casing to within the preset target pressure range. Specifically, when the controller simultaneously determines that the current real-time absolute pressure value has risen above the upper limit of the target pressure range, and the real-time temperature of the main server is still lower than the preset frequency reduction threshold temperature, it indicates that the pressure inside the main casing has increased due to the generation of a large amount of steam. If intervention is not timely, the actual boiling temperature of the coolant will rise above the frequency reduction threshold temperature, resulting in the failure of phase change cooling. The controller sends a power supply command to the power supply, causing the vacuum pumping equipment to restart at a preset maintenance power (usually a lower operating power) and perform a supplementary vacuum pumping action. This extracts excess steam from inside the main housing, reducing the absolute pressure back to the target pressure range, thereby maintaining the actual boiling temperature of the coolant below the frequency reduction threshold temperature.

[0059] This implementation method offers significant advantages. First, by comparing the current real-time absolute pressure value with the upper and lower limits of the preset target pressure range, hysteresis control of the internal pressure of the main casing—stopping at the lower limit and starting at the upper limit—is achieved. When the pressure falls below the lower limit, vacuuming stops, allowing the pressure to rise naturally, avoiding excessive vacuuming that leads to energy waste and excessive coolant evaporation. When the pressure exceeds the upper limit and the main server is still at a temperature requiring phase change cooling protection, supplementary vacuuming is automatically initiated to pull the pressure back within the range, ensuring that the coolant boiling temperature is always maintained below the frequency reduction threshold temperature, continuously leveraging the phase change heat absorption effect. Second, by introducing the real-time temperature of the main server as an auxiliary criterion for vacuuming initiation, ineffective actions such as initiating vacuuming due to pressure rise when the main server temperature has already dropped to a safe level are avoided, further optimizing system energy consumption. Third, the preset maintenance power setting allows the vacuuming equipment to operate at a lower power level, reducing equipment wear and energy consumption while maintaining the target pressure. This hysteresis pressure control strategy, together with the aforementioned vacuum start-up, boiling temperature determination, and phase change cooling steps, constitutes a complete and efficient closed-loop control system for server power supply phase change cooling.

[0060] In one embodiment of this application, a pre-barreling startup step based on historical temperature data is included before step S300. Specifically, the controller first acquires multiple historical temperature sampling values ​​of the main server over a continuous time series based on the current operating state of the main server. These multiple historical temperature sampling values ​​are temperature value sequences collected by the internal temperature sensor of the main housing at multiple consecutive historical moments. The controller continuously collects the main server temperature at a set sampling period (e.g., once per second) and stores the temperature data from the most recent few moments in an internal buffer to form a temperature value sequence.

[0061] Subsequently, based on the multiple historical temperature sampling values, the temperature change of the main server per unit time is calculated to obtain the real-time temperature rise rate value. This real-time temperature rise rate value characterizes how quickly the temperature of the main server rises at the current moment. The controller performs a difference operation on the temperature value sequence in the buffer, for example, by dividing the difference between the current temperature value and the previous temperature value by the sampling time interval, or by fitting a temperature curve using the least squares method and calculating its slope to obtain the real-time temperature rise rate value at the current moment.

[0062] Then, based on the real-time temperature rise rate value, it is compared with a preset temperature rise rate threshold to obtain a rate comparison result. The preset temperature rise rate threshold is a pre-set critical rate value used to determine that the main server is about to experience a rapid temperature rise. This preset temperature rise rate threshold is pre-calibrated based on the thermal time constant of the main server and the response delay of the cooling system. When the temperature rise rate exceeds this threshold, it indicates that the main server is experiencing a rapid power surge, and conventional heat exchange methods may not be able to respond in time.

[0063] Based on the rate comparison result indicating that the real-time temperature rise rate is greater than the preset temperature rise rate threshold, a pre-voltage reduction start signal is generated. This pre-voltage reduction start signal is used to trigger the power supply to the vacuum pump before the current temperature of the main server reaches the preset temperature threshold. That is, before the main server temperature rises to the preset temperature threshold set in step S202, the controller has already identified the impending high-heat state by monitoring the temperature rise rate.

[0064] Subsequently, based on the pre-pressure reduction start signal, when the current temperature parameter has not reached the preset temperature threshold and the current computing power load parameter has not exceeded the preset load threshold, the power supply is controlled to supply electrical energy to the vacuum pumping device. This enables the vacuum pumping device to perform a pre-vacuuming action at a preset reserve power, and to pre-reduce the absolute pressure inside the main casing to a reserve pressure value. The reserve pressure value is greater than the preset vacuuming termination pressure threshold, and the corresponding coolant pre-boiling temperature is higher than the current temperature parameter but lower than the preset frequency reduction threshold temperature. Specifically, the controller starts the vacuum pumping device when the temperature has not reached the warning value and the computing power load has not exceeded the load threshold, but operates at a lower power to reduce the pressure inside the main casing to an intermediate value. The coolant boiling temperature corresponding to this reserve pressure value is higher than the current temperature but lower than the frequency reduction threshold temperature, meaning that the coolant has not yet boiled at the current temperature, but the pressure is already in a preset reserve state. If the temperature continues to rise, further vacuuming is sufficient to quickly reduce the boiling temperature below the current temperature and initiate phase change cooling.

[0065] Finally, based on the fact that the absolute pressure inside the main casing has been reduced to the pre-set pressure value, the judgment step in S200 continues to be executed. When the judgment result meets the preset vacuuming start condition, the power supply is controlled to supply the vacuuming device with a preset full power greater than the preset pre-set power, so that the vacuuming device switches from the preset pre-set power to the preset full power greater than the preset pre-set power to perform subsequent vacuuming actions. That is, after the vacuuming device completes the pre-pressure reduction at the pre-set power, the condition judgment in step S200 continues to be executed. When the current temperature and computing load of the main server both reach the trigger condition, the controller increases the operating power of the vacuuming device from the pre-set power to the full power, quickly completes the remaining vacuuming process, further reduces the pressure inside the main casing from the pre-set pressure value to the preset vacuuming termination pressure threshold, and lowers the actual boiling temperature of the coolant below the current temperature, thus initiating phase change evaporative cooling.

[0066] The beneficial effects of this implementation method are significant. First, by calculating the real-time temperature rise rate based on historical temperature sampling data, it is possible to identify an impending rapid temperature rise trend before the main server temperature reaches the preset temperature threshold, generate a pre-depressurization start signal, and activate the vacuum pump to perform pre-vacuuming, reducing the internal pressure of the main casing to the preset pressure value in advance. This pre-depressurization process utilizes the time window from temperature rise to triggering frequency reduction, advancing the vacuuming action that should have started after the temperature reached the threshold to the temperature rise process, effectively shortening the response time from triggering vacuuming to the actual boiling of the coolant, ensuring that phase change cooling has been established and begins to function before the main server temperature reaches the frequency reduction threshold. Secondly, by employing a preset reserve power instead of full power for pre-vacuuming, the initial pressure reduction is completed with lower energy consumption and mechanical load during the pre-pressure reduction process. This avoids the energy waste and equipment wear caused by prolonged full-power operation of the vacuum equipment. Simultaneously, the design ensures that the coolant's pre-boiling temperature, corresponding to the reserve pressure value, is higher than the current temperature but lower than the frequency reduction threshold temperature. This prevents the coolant from prematurely boiling at the current temperature, thus avoiding unnecessary phase change evaporation, and keeps the system in a "standby state," ready to quickly initiate phase change cooling should the temperature rise further. Thirdly, after pre-pressure reduction, the condition judgment in step S200 is performed again. When the condition is met, the vacuum equipment is switched to full-power operation, enabling the pre-pressured system to quickly complete deep vacuuming. This forms a two-stage response strategy from pre-pressure reduction to full-speed vacuuming, ensuring both response speed and optimizing system energy consumption and equipment lifespan. This pre-pressure reduction start-up step, together with the aforementioned vacuum start-up condition judgment, boiling temperature determination and phase change cooling steps, constitutes a complete feedforward phase change cooling control link from temperature rise trend prediction, pre-pressure reduction preparation to deep vacuum start-up, further improving the server system's temperature response capability and thermal management efficiency under high power surge scenarios.

[0067] In one embodiment of this application, the server system further includes a secondary housing, a secondary server, a steam pipe, and a solenoid valve. The secondary server is disposed inside the secondary housing, the steam pipe connects the interior of the primary housing and the interior of the secondary housing, and the solenoid valve is disposed on the steam pipe for controlling the on / off state of the steam pipe. After step S500, the method further includes the following steam discharge and condensate reflux steps: Based on the steam generated inside the main casing due to the phase change evaporation of the coolant, the current steam pressure value inside the main casing at the current moment is obtained. This current steam pressure value represents the absolute pressure increase inside the main casing due to steam generation. Specifically, after the coolant undergoes phase change evaporation to generate steam, the steam accumulates inside the main casing, causing the absolute pressure inside the main casing to gradually increase. The controller obtains the current steam pressure value in real time through a pressure sensor installed inside the main casing to determine whether steam needs to be released externally.

[0068] Based on the current steam pressure value, the current steam pressure value is compared with a preset steam discharge pressure threshold to obtain a steam pressure comparison result. The preset steam discharge pressure threshold is a pre-set critical pressure value used to determine that the steam pressure inside the main shell is too high and steam needs to be discharged to the auxiliary shell. This preset steam discharge pressure threshold can be set to be higher than the upper limit of the preset target pressure range selected in step S106 but lower than the safe pressure bearing limit of the main shell. When the pressure inside the main shell reaches this threshold, it indicates that the pumping rate of the vacuum equipment is insufficient to offset the steam generation rate, and auxiliary pressure relief is required with the help of the auxiliary shell.

[0069] Based on the steam pressure comparison result indicating that the current steam pressure value is greater than the preset steam discharge pressure threshold, the solenoid valve is controlled to open, thereby opening the steam pipeline and allowing steam inside the main casing to enter the secondary casing through the steam pipeline. Specifically, when the controller determines that the pressure inside the main casing has reached the critical value requiring discharge, the controller sends an opening command to the solenoid valve, and the valve core of the solenoid valve actuates to open the steam pipeline. High-temperature steam inside the main casing flows into the secondary casing along the steam pipeline under the action of the pressure difference.

[0070] The steam enters the secondary housing through the steam pipe. Utilizing the lower ambient temperature inside the secondary housing compared to the steam temperature, the steam condenses on the inner wall of the secondary housing and the outer surface of the secondary server. This transfers the heat carried by the steam to the interior of the secondary housing, simultaneously producing a condensed liquid coolant. Specifically, the secondary housing typically houses a secondary server (which can be another server operating under lower load or in standby mode). The temperature inside the secondary housing is significantly lower than the temperature of the incoming steam. When the high-temperature steam contacts the inner wall of the secondary housing or the outer surface of the secondary server, the steam releases its latent heat of vaporization and condenses into a liquid coolant. The heat released during condensation is absorbed by the environment of the secondary server inside the secondary housing, achieving heat transfer.

[0071] Based on the condensed liquid coolant, the condensed liquid coolant is returned from the sub-shell to the main shell through the steam pipe to replenish the coolant lost in the main shell due to phase change evaporation. Specifically, the liquid coolant condensed inside the sub-shell accumulates at the bottom of the sub-shell under gravity and is returned to the main shell through the steam pipe (or dedicated return pipe), thereby compensating for the loss of coolant volume in the main shell due to continuous coolant evaporation, maintaining a relatively stable coolant level and total volume in the main shell, and ensuring that the main server is always adequately submerged in coolant.

[0072] Based on the current steam pressure inside the main housing dropping below the preset steam discharge pressure threshold, the solenoid valve is controlled to close, thereby cutting off the steam pipeline and stopping the steam flow between the main housing and the auxiliary housing. Specifically, when the pressure inside the main housing drops below the preset steam discharge pressure threshold due to steam discharge and subsequent reduced boiling of the coolant, the controller sends a closing command to the solenoid valve. The solenoid valve cuts off the steam pipeline, the steam flow between the main housing and the auxiliary housing stops, and the system returns to the normal operating mode where pressure is maintained by a vacuum device.

[0073] The beneficial effects of this implementation are significant. First, by using the secondary housing as a temporary steam condensation and pressure relief device when the vacuum equipment's pressure maintenance capacity is insufficient, an additional steam discharge and condensation channel is provided to the primary housing. This effectively prevents the problem of a rapid increase in pressure due to excessive steam accumulation inside the primary housing, which could cause the coolant boiling temperature to rise above the throttling threshold. This provides a redundant protection mechanism for pressure control within the primary housing. Second, the steam condensation process inside the secondary housing transfers a large amount of heat generated by the primary server to the secondary server environment through latent heat, achieving cross-regional heat transfer and preventing heat accumulation inside the primary housing, further enhancing the thermal management capabilities of the entire server system. Third, the condensed liquid coolant flows back to the primary housing through steam pipes or return pipes, achieving coolant recycling. This avoids the need for external replenishment due to continuous evaporation reducing the coolant level inside the primary housing, thus reducing system operating costs. Furthermore, the controlled opening and closing of the solenoid valves ensures that steam discharge and condensation return are automatically activated only when the pressure inside the primary housing is too high and automatically closed after the pressure returns to normal. This achieves intelligent control of discharge on demand, reducing unnecessary heat loss and interference of steam circulation with the operation of the secondary server. The steam discharge and condensation reflux steps, together with the aforementioned vacuum start-up, phase change cooling and pressure maintenance steps, constitute a complete main shell phase change cooling pressure management system.

[0074] In one embodiment of this application, the server system further includes an intake valve, which is disposed on the outer wall of the main housing and communicates with the interior of the main housing, for controlling the entry of external gas into the interior of the main housing. After step S600, the method further includes the following dynamic pressure regulation step based on a load-pressure mapping table: Based on the fact that the current absolute pressure value inside the main casing is maintained within the preset target pressure range, the controller continuously acquires the current real-time computing load parameters of the main server and the current real-time absolute pressure value inside the main casing. Specifically, after the pressure inside the main casing is stably maintained within the preset target pressure range in step S600, the controller does not stop monitoring, but continues to acquire the real-time computing load of the main server and the real-time absolute pressure inside the main casing at a preset sampling frequency, providing real-time data for subsequent dynamic adjustments.

[0075] Based on the current real-time computing load parameters of the main server, the target absolute pressure value corresponding to the current real-time computing load parameters is queried from a preset load-pressure mapping table. The preset load-pressure mapping table is a pre-stored set of mapping relationships characterizing the correspondence between different computing load levels and optimal heat dissipation absolute pressure values. Specifically, the controller stores a load-pressure mapping table obtained through experimental calibration or theoretical calculation. This mapping table divides the computing load of the main server into several levels, each level corresponding to an optimal absolute pressure value. This optimal pressure value is the pressure value that maximizes the phase change cooling efficiency while keeping the actual boiling temperature of the coolant just below the frequency reduction threshold temperature under that load condition. When the computing load increases, the corresponding target absolute pressure value decreases, causing the coolant boiling temperature to drop to cope with the higher heat flux density; when the computing load decreases, the corresponding target absolute pressure value increases to reduce unnecessary vacuuming energy consumption.

[0076] Based on the current real-time absolute pressure value and the target absolute pressure value, a numerical comparison is performed between the two values ​​to obtain a pressure deviation comparison result. This result indicates whether the current real-time absolute pressure value is higher, lower, or equal to the target absolute pressure value. Specifically, the controller calculates the difference between the current real-time absolute pressure value and the target absolute pressure value, and determines the current pressure state based on the sign and magnitude of this difference to decide whether to perform a pressure increase or decrease operation.

[0077] Based on the pressure deviation comparison result indicating that the current real-time absolute pressure value is higher than the target absolute pressure value, the controller controls the power supply to supply power to the vacuum pumping device, so that the vacuum pumping device performs a compensating vacuum pumping action at a preset compensation power, thereby reducing the absolute pressure inside the main housing to the target absolute pressure value. Specifically, when the current pressure is higher than the target value, the controller performs a compensating vacuum pumping through the vacuum pumping device to extract excess steam and gas from inside the main housing, reducing the pressure to the target value.

[0078] Based on the pressure deviation comparison result indicating that the current real-time absolute pressure value is lower than the target absolute pressure value, the controller controls the intake valve to open to a preset degree to introduce external gas into the main housing, thereby raising the absolute pressure inside the main housing to the target absolute pressure value. Specifically, when the current pressure is lower than the target value, the controller sends an opening command to the intake valve, allowing external air (or other gas) to enter the main housing through the intake valve, causing the pressure to naturally rise to the target value. The preset opening degree of the intake valve is dynamically adjusted according to the difference between the current pressure and the target pressure to avoid pressure overshoot.

[0079] Based on the adjustment of the absolute pressure inside the main housing to the target absolute pressure value, the power supply to the vacuum pump is stopped to halt the compensating vacuum pumping operation, and the air inlet valve is closed to maintain the absolute pressure inside the main housing at the target absolute pressure value. Specifically, once the pressure is adjusted to the target value, the controller simultaneously stops the vacuum pump and closes the air inlet valve, keeping the absolute pressure inside the main housing stable near the target value until the next load change triggers a new round of adjustment.

[0080] The beneficial effects of this implementation method are significant. First, by using a preset load-pressure mapping table, the target absolute pressure inside the main casing can be dynamically adjusted according to changes in the real-time computing load of the main server, achieving precise matching between the phase change cooling capacity (i.e., the coolant boiling temperature) and the current heat generation of the main server. When the main server's computing load increases and heat generation increases, the target pressure is automatically reduced to lower the coolant boiling temperature and increase the phase change heat absorption capacity; when the load decreases and heat generation decreases, the target pressure is automatically increased to raise the coolant boiling temperature, avoiding energy waste caused by excessive vacuuming and achieving refined energy-saving control of "cooling on demand". Second, by setting an intake valve and cooperating with a vacuuming device, bidirectional precise adjustment of the absolute pressure inside the main casing is achieved. Not only can the pressure be reduced by vacuuming, but the pressure can also be increased by introducing external gas through the intake valve, so that pressure control is no longer limited to unidirectional pressure reduction, ensuring that the pressure can be quickly restored to the target value when the load decreases, avoiding unnecessary losses due to excessive boiling of the coolant caused by excessive pressure. Furthermore, when load changes cause the pressure to deviate from the target value, the system can automatically perform compensatory vacuuming or pressurization operations, and automatically stop and close the valves after adjustment to maintain stable internal pressure in the main casing, achieving closed-loop adaptive pressure regulation. This dynamic pressure regulation step, together with the aforementioned vacuuming start-up, phase change cooling, and pressure maintenance steps, constitutes an adaptive phase change cooling pressure management system based on real-time computing power load feedback, further improving the accuracy of server power supply temperature control and energy utilization efficiency.

[0081] In one embodiment of this application, the server system further includes a bubble sensor disposed inside the main housing and near the main server, for monitoring the bubble state parameters generated by the coolant during phase change evaporation. After step S500, the method further includes the following film boiling monitoring and recovery steps: Based on the phase change evaporation of the coolant inside the main casing, the bubble sensor acquires the bubble density and average bubble size parameters of the coolant at the current moment. The bubble density parameter represents the number of bubbles contained in a unit volume of coolant, and the average bubble size parameter represents the average value of the bubble diameter in the coolant. Specifically, the bubble sensor can be an optical bubble sensor (e.g., a high-speed camera with image recognition) or an ultrasonic bubble sensor, placed in the coolant near the heating surface of the main server. When the coolant boils on the surface of the main server, the sensor detects the density (number of bubbles per unit volume) and average size (average value of the diameter of each bubble) of the generated bubbles in real time and transmits the detection data to the controller. In the nucleus boiling state, the bubble density is relatively low and the bubble size is small, allowing the bubbles to detach from the heating surface in time; in the film boiling state, the bubble density increases sharply, the bubble size increases significantly, and a continuous vapor film forms on the surface of the main server.

[0082] Based on the bubble density parameter and the average bubble size parameter, the bubble density parameter is compared with a preset bubble density threshold to obtain a density comparison result, and the average bubble size parameter is compared with a preset bubble size threshold to obtain a size comparison result. The preset bubble density threshold is a pre-set critical bubble density value used to determine whether the coolant has entered the film boiling state, and the preset bubble size threshold is a pre-set critical bubble size value used to determine whether the coolant has entered the film boiling state. Specifically, the controller internally stores preset bubble density thresholds and preset bubble size thresholds obtained through experimental calibration. When both the bubble density and average size exceed their respective thresholds, it indicates that the boiling state of the coolant has changed from efficient nucleus boiling to inefficient film boiling.

[0083] Based on the density comparison result indicating that the bubble density parameter is greater than the preset bubble density threshold and the size comparison result indicating that the average bubble size parameter is greater than the preset bubble size threshold, the coolant is determined to be in a film boiling state, resulting in a boiling state determination result. The film boiling state is defined as the boiling state in which the coolant forms a continuous vapor film on the surface of the main server, leading to a sharp decrease in heat exchange efficiency. Specifically, the controller performs a logical AND operation on the density comparison result and the size comparison result. When both conditions are simultaneously met, the coolant is determined to have entered a film boiling state. This determination logic ensures that a film boiling determination is made only when both bubble density and size reach critical levels, avoiding misjudgment based on a single indicator.

[0084] Based on the boiling state determination result indicating that the coolant is in the film boiling state, the controller opens the air intake valve to introduce external gas into the main housing, thereby increasing the absolute pressure inside the main housing to a high-pressure recovery value higher than the preset target pressure range. The coolant's recovery boiling temperature corresponding to this high-pressure recovery value is higher than the current real-time temperature of the main server. Specifically, when film boiling is detected, the controller sends an opening command to the air intake valve, allowing external gas to enter the main housing and rapidly increasing the absolute pressure inside. The design goal of this high-pressure recovery value is to ensure that the coolant's boiling temperature at this pressure is higher than the current real-time temperature of the main server, thereby stopping the coolant's phase change evaporation and creating conditions for the elimination of the vapor film.

[0085] Based on the increase in absolute pressure inside the main casing to the high-pressure recovery value, the coolant stops phase change evaporation to eliminate the continuous vapor film on the surface of the main server. This causes the coolant to switch from film boiling to nucleation boiling, where the coolant forms discrete bubbles on the main server surface and has a higher heat transfer efficiency than film boiling. Specifically, when the actual boiling temperature of the coolant exceeds the current temperature of the main server due to the increased pressure, the coolant stops boiling, and the continuous vapor film on the main server surface gradually ruptures and dissipates in the absence of a continuous vapor supply. Subsequently, as the main server continues to generate heat, its surface temperature rises again to the boiling temperature of the coolant at the current pressure, and the coolant begins to boil again. Since there is no pre-existing vapor film on the main server surface at this time, boiling occurs again in the form of discrete bubble nucleation, and the heat transfer efficiency returns to normal levels.

[0086] Based on the coolant switching to the nucleus boiling state, the bubble sensor re-acquires updated bubble density and average bubble size parameters of the coolant. When the updated bubble density parameter is less than the preset bubble density threshold and the updated average bubble size parameter is less than the preset bubble size threshold, the air inlet valve is closed and the power supply is restarted to the vacuum equipment, causing the vacuum equipment to restart and restore the absolute pressure inside the main casing to the preset target pressure range. Specifically, after the coolant resumes nucleus boiling, the controller continuously collects updated bubble density and size data through the bubble sensor. When both fall below the threshold, it indicates that a stable nucleus boiling has been re-established on the surface of the main server, and the vapor film has been effectively eliminated. At this time, the controller closes the air inlet valve to stop the introduction of external gas and restarts the vacuum equipment, gradually reducing the absolute pressure inside the main casing from the high pressure recovery value back to the preset target pressure range, allowing the coolant to maintain efficient nucleus boiling heat dissipation at a lower boiling temperature.

[0087] The benefits of this implementation are significant. First, by monitoring bubble density and size parameters in real time using bubble sensors, it is possible to accurately distinguish between nucleation boiling and film boiling, two boiling states with drastically different heat transfer efficiencies, providing a reliable detection method for timely intervention in film boiling. When film boiling is detected, external gas is introduced into the main casing through the air inlet valve, increasing the absolute pressure inside the main casing and temporarily stopping the coolant from boiling, thereby eliminating the continuous vapor film on the surface of the main server. This "pressure-locking and film-eliminating" operation fundamentally breaks the continuous conditions for film boiling, allowing the coolant to recover from film boiling to the more efficient nucleation boiling state, effectively restoring the heat dissipation capacity of the phase change cooling system. Second, the timely elimination of film boiling avoids the risk of a sharp deterioration in heat transfer on the main server surface and a rapid temperature rise to the throttling threshold temperature caused by the vapor film, ensuring the server's continuous computing power output under high-power operation. Furthermore, after returning to nucleus boiling, the system automatically closes the intake valve and restarts the vacuum equipment, restoring the internal pressure of the main casing to the target pressure range. This lowers the coolant boiling temperature back to a level suitable for the current load, and the system returns to normal phase change cooling operation mode, balancing system safety and operational economy. This film boiling monitoring and recovery step, in conjunction with the aforementioned vacuum startup, phase change cooling, and pressure maintenance steps, constitutes a closed-loop management system for adaptive identification, proactive intervention, and automatic recovery of film boiling. This provides reliable technical assurance for the phase change cooling safety of server power supplies under high heat flux density operating conditions.

[0088] Secondly, this application also provides an automatic power supply temperature control method, applied to a target power supply system. The target power supply system includes multiple power modules, each distributed across different racks or areas of the server system, for providing power support to server equipment within the server system. For example... Figure 2 As shown, the automatic temperature control method for the power supply includes: A100 acquires real-time operating status data and environmental status data for each power module. Specifically, the real-time operating status data includes parameters such as output power, input voltage, input current, internal temperature, fan speed, and conversion efficiency of each power module, which can be acquired in real time through the sensors and power management chips built into each power module. The environmental status data includes parameters such as the inlet air temperature, outlet air temperature, air velocity, and internal temperature and humidity of each power module's location, which can be acquired through environmental monitoring sensors deployed within the server system. The controller continuously reads the above data at a preset sampling period (e.g., every 5 seconds or every 10 seconds) to provide an accurate data basis for subsequent temperature prediction and load adjustment.

[0089] A200. Based on the real-time operating status data, a preset thermal characteristic model is used to predict the temperature change trend of each power module within a preset future time window. Specifically, the thermal characteristic model is a pre-established mathematical model used to characterize the dynamic relationship between the temperature change of the power module and its operating parameters and environmental parameters. It can be established based on the thermodynamic characteristics of the power module (including physical parameters such as heat capacity, thermal resistance, and heat dissipation efficiency), or it can be implemented through a machine learning model trained on historical operating data. The controller inputs the real-time operating status data and environmental status data of each power module acquired at the current moment into the thermal characteristic model. Based on the current heat generation and heat dissipation conditions of each module, the model calculates and outputs the temperature change curve of each module within a preset future time window (e.g., the next 30 to 120 seconds), including trend information such as the rate of temperature rise, peak temperature, and the time when the temperature reaches each key node. This prediction process provides a forward-looking data foundation for active temperature control, enabling the controller to identify power modules that are about to experience temperature anomalies in advance.

[0090] A300. Based on the temperature change trend and the preset temperature safety range, determine the real-time thermal status of each power module. Specifically, the preset temperature safety range is a lower and upper temperature limit set in advance according to the specifications, rated operating temperature range, and server system operating requirements of each power module. When the temperature of a power module is within this range, it indicates normal operation. The controller compares the predicted temperature change trend of each power module in the future time window obtained in step A200 with the preset temperature safety range to determine whether each module is within the normal operating temperature range at the current moment and in the future, whether it is approaching the upper or lower temperature limit, and whether the rate of temperature change is too fast, thereby classifying the real-time thermal status of each power module into different levels such as "normal," "warning," or "critical." For example, if it is predicted that the temperature of a power module will continue to rise and approach the upper temperature limit within the next 30 seconds, the thermal status of the module is determined to be "warning"; if it is predicted that its temperature will reach or exceed the upper temperature limit within the next 30 seconds, its thermal status is determined to be "critical."

[0091] A400. Based on the real-time operating status data and the real-time thermal status, determine the load adjustment strategy for each power module. Specifically, the controller comprehensively formulates a load adjustment scheme based on the real-time operating status data such as the current output power and conversion efficiency of each power module obtained in step A100, and the real-time thermal status of each module determined in step A300. The load adjustment strategy includes: for power modules with a real-time thermal status of "warning" or "critical", their load allocation ratio should be reduced to reduce heat generation; for power modules with a real-time thermal status of "normal" and a low current load rate, their load allocation ratio can be appropriately increased to take over the load transferred from high-temperature modules; for modules with a thermal status of "warning" but a high current conversion efficiency, their heat dissipation equipment (such as fan speed) can be controlled first to cool them down, rather than immediately reducing the load. The controller's decision is based on multi-objective optimization, that is, under the premise of ensuring that the temperature of all power modules is within a safe range, the load distribution of each power module is as balanced as possible or the modules operate in the highest conversion efficiency range as possible.

[0092] A500. Based on the load adjustment strategy, the load distribution ratio of each power module in the target power system is dynamically adjusted, and corresponding heat dissipation control commands are generated to control the temperature of each power module. Specifically, the controller converts the load adjustment strategy determined in step A400 into specific execution commands: for power modules that need to reduce load, the controller sends a load reduction command to the power management unit of that module to reduce its output power; for modules that need to increase load, the controller sends a load increase command to them to handle more load; for modules that need enhanced heat dissipation, the controller sends a heat dissipation control command to the fan drive circuit or liquid cooling valve controller of that module to increase the fan speed or increase the coolant flow rate to enhance heat dissipation capacity. The above load adjustment and heat dissipation control are executed in tandem, reducing heat generation at the source (by adjusting load distribution) and improving heat dissipation efficiency (by enhancing heat dissipation), achieving proactive and precise control of the power module temperature through a two-pronged approach. This adjustment process can continue continuously, and as the thermal state of each power module changes, the load distribution ratio and heat dissipation control parameters are also dynamically updated accordingly, forming a closed-loop control.

[0093] The beneficial effects of this implementation method are significant. First, by acquiring real-time operating status data and environmental status data of each power module, and using a preset thermal characteristic model to predict the temperature change trend of each module within a preset time window, temperature control is transformed from the traditional passive response type (taking measures after detecting excessive temperature) to an active predictive type (predicting the trend and intervening in advance before the temperature exceeds the limit). This effectively solves the problem of delayed response in passive temperature regulation in related technologies, significantly improving the timeliness and effectiveness of temperature control. Second, based on the comparison between the predicted temperature change trend and the preset safe temperature range, the real-time thermal state of each power module is accurately determined. This allows the controller to accurately identify modules that are about to experience temperature anomalies and to differentiate the treatment of modules with different thermal states, avoiding a "one-size-fits-all" uniform control strategy and improving the accuracy and efficiency of control. Third, by comprehensively determining the load adjustment strategy by integrating real-time operating status data and real-time thermal state, the load allocation decision not only considers the temperature state but also takes into account the current conversion efficiency, load rate, and other operating parameters of each module. Under the premise of ensuring temperature safety, the power system operates in the high-efficiency range as much as possible, improving the overall energy utilization efficiency of the server system. Finally, the coordinated execution of load adjustment and thermal control simultaneously regulates the power module temperature from two dimensions: the heat source (load) and the heat dissipation path (heat dissipation equipment). Compared to single-method cooling, this approach offers more significant cooling effects and a faster response, ensuring that each power module remains within a safe and efficient operating temperature range under complex and variable load conditions. This guarantees the stability and reliability of the server system's power supply system. This control method can be widely applied to large-scale server systems, edge computing nodes, and various scenarios requiring high-reliability power management.

[0094] In one embodiment of this application, the real-time operating status data and environmental status data of each power module in step A100 specifically include the following: First, the input electrical parameters, output electrical parameters, and temperature sampling values ​​of key internal components of each power module are collected. Specifically, the input electrical parameters include the input voltage, input current, and input power of each power module, which can be obtained in real time through a voltage divider resistor network and a current sampling resistor (or Hall current sensor) located at the input of the power module in conjunction with an analog-to-digital converter. The output electrical parameters include the output voltage, output current, and output power of each power module, which can be obtained in real time through a voltage detection circuit and a current detection circuit located at the output of the power module. The temperature sampling values ​​of key internal components include the surface temperature of key heat-generating components such as power switching devices (e.g., MOSFETs), magnetic components (e.g., transformers, inductors), rectifier diodes, and electrolytic capacitors within each power module, which can be obtained in real time through NTC thermistors, thermocouples, or digital temperature sensors that are pre-attached or embedded near the aforementioned key components. The aforementioned electrical parameters and temperature sampling values ​​are collected by the power management unit or independent monitoring unit of each power module at a preset sampling frequency (e.g., every 100 milliseconds or every 500 milliseconds) and transmitted to the central controller via a communication bus (e.g., PMBus, I²C or CAN bus).

[0095] Secondly, the intake air temperature, ambient humidity, and ambient air pressure of the target power system's environment are collected. Specifically, the intake air temperature can be collected by temperature sensors installed at the air inlets of each power module. This temperature characterizes the cooling air inlet temperature of the cabinet or server room where the power module is located and is one of the key environmental parameters affecting the power module's heat dissipation efficiency. The ambient humidity can be collected by humidity sensors installed in the server system's server room. This humidity reflects the water vapor content in the air and has a certain impact on the insulation performance and heat dissipation effect of the power module. The ambient air pressure can be collected by air pressure sensors installed in the server system's server room. This air pressure characterizes the absolute atmospheric pressure in the server room. This parameter directly affects the air density and cooling capacity of the forced air cooling system, thus affecting the power module's heat dissipation efficiency. The above environmental status data are continuously collected by environmental monitoring sensors deployed in various areas of the server system at a preset sampling frequency and aggregated to the central controller through the data acquisition unit.

[0096] The benefits of this implementation method are significant. First, by collecting the input and output electrical parameters of each power module, the current load power and conversion efficiency of each module can be accurately calculated, providing an accurate electrical status data foundation for subsequent load adjustment strategies. Second, by collecting temperature sampling values ​​of key components inside the modules (rather than relying solely on the module casing temperature or outlet temperature), the controller can grasp the actual temperature status of the hottest spots (i.e., power devices and magnetic components) inside each module, providing more sensitive and representative temperature data for temperature change trend prediction, which is beneficial for identifying potential temperature anomalies earlier. Third, by collecting multi-dimensional environmental status data such as inlet air temperature, ambient humidity, and ambient air pressure, the thermal characteristic model can comprehensively consider the impact of environmental conditions on the heat dissipation capacity of the power module, improving the accuracy and environmental adaptability of temperature change trend prediction and avoiding prediction deviations caused by ignoring environmental changes. This data acquisition method provides comprehensive, accurate, and timely data support for subsequent thermal characteristic model prediction, thermal status determination, and load adjustment strategies, and is a fundamental guarantee for realizing active and refined temperature control of server system power supplies.

[0097] In one embodiment of this application, step A200, based on the real-time operating status data, uses a preset thermal characteristic model to predict the temperature change trend of each power module within a future preset time window, specifically including the following steps: First, historical heat loss data and heat dissipation coefficients of each power module are obtained to construct a first-order thermal resistance-capacitance network model. Specifically, the historical heat loss data refers to the historical heat loss data of the power modules under the same or similar operating conditions (e.g., the same output power and the same input voltage), which can be retrieved from the historical operating database stored internally by the controller. The heat dissipation coefficient characterizes the heat conduction and convection capacity of the power module to transfer heat from the module's interior to the surrounding environment, and can be determined based on the power module's structural design, material properties, and heat sink specifications. Based on the above data, the controller establishes a first-order thermal resistance-capacitance network model. This model equates the heat-generating element (e.g., power switching device) of the power module to a thermal capacitance element, the path of heat transfer from the heat-generating element to the external environment to a thermal resistance element, and heat loss to a heat flow source injected into the thermal capacitance element. The mathematical expression of this first-order thermal resistance-capacitance network model is a first-order linear differential equation with constant coefficients, describing the dynamic relationship between the power module's temperature change over time and its internal heat generation and external heat dissipation conditions.

[0098] Then, the real-time power loss in the real-time operating status data is used as the heat flux input, and the boundary conditions of the first-order thermal resistive-capacitive network model are corrected in conjunction with the environmental state data. Specifically, the controller extracts the current power loss (i.e., the difference between input power and output power) of each module from the real-time operating status data of each power module obtained in step A100, and applies this loss value as the heat flux input to the heat flux source end of the first-order thermal resistive-capacitive network model. At the same time, the controller corrects the boundary conditions of the model based on the environmental state data (including inlet air temperature, ambient humidity, and ambient air pressure) obtained in step A100. For example, it adjusts the ambient reference temperature based on the inlet air temperature and corrects the thermal conductivity and convective heat transfer coefficient of the air based on the ambient air pressure and humidity, so that the model can reflect the actual heat dissipation capacity of the power module under the current environmental conditions.

[0099] Finally, the differential equation corresponding to the modified thermal resistive-capacitive network model is solved to obtain the predicted temperature change curve of each power module within a future preset time window. Specifically, the controller uses the temperature sampling value of each power module at the current moment as the initial condition of the differential equation, and the modified environmental parameters and current power loss as the input of the equation. The differential equation is solved by analytical methods (e.g., directly solving a first-order linear differential equation) or numerical methods (e.g., the Euler method or the Runge-Kutta method). The solution results in a functional relationship between the temperature of each power module and time, i.e., the temperature change curve within a future preset time window (e.g., from the next 30 seconds to 120 seconds). This curve contains the predicted temperature value and its rate of change at each future time point, which is used in subsequent step A300 to compare with the preset temperature safety range to determine the real-time thermal state of each power module.

[0100] The beneficial effects of this implementation method are significant. First, by constructing a first-order thermal resistance-capacitance network model to characterize the thermal dynamic characteristics of each power module, this model is built based on thermodynamic principles rather than being purely data-driven. Its physical meaning is clear, and it can accurately reflect the temperature response characteristics of the power modules when power loss changes and environmental conditions change, resulting in high accuracy and interpretability of temperature prediction. Second, by using real-time power loss as heat flow input, the model can instantly respond to changes in heat generation caused by changes in the current load of the power modules, achieving dynamic synchronization between temperature prediction and real-time operating status. Third, by incorporating environmental condition data (inlet air temperature, ambient humidity, and ambient air pressure) to correct the model's boundary conditions, the model can adapt to differences in heat dissipation capacity under different environmental conditions—the predicted temperature rise rate will increase when the inlet air temperature rises, and decrease when the inlet air temperature falls, thereby improving the environmental adaptability and accuracy of temperature change trend prediction. Finally, solving the differential equation yields a continuous temperature change curve rather than discrete point predictions. This allows the controller to obtain accurate temperature predictions at various time points, as well as trend information in more dimensions, such as the rate of temperature change and the time to peak temperature. This provides richer and more accurate decision-making basis for subsequent precise thermal state determination and load adjustment strategies. This thermal characteristic model prediction method, together with the aforementioned data acquisition and subsequent thermal state determination steps, constitutes a technical link for proactive temperature prediction and dynamic control, realizing a leap from passive response to proactive prediction.

[0101] In one embodiment of this application, step A400, based on the real-time operating status data and the real-time thermal state, determines the load adjustment strategy for each power module, specifically including the following steps: First, the conversion efficiency and remaining heat dissipation capacity of each power module under the current real-time operating state are calculated. Specifically, the conversion efficiency is the ratio of the output power to the input power of each power module, reflecting the efficiency with which the power module converts input electrical energy into output electrical energy. The higher the conversion efficiency, the less power loss and heat generation the power module generates when outputting the same power. The controller calculates the current conversion efficiency of each module in real time based on the input voltage, input current, output voltage, and output current data collected in step A100. The remaining heat dissipation capacity is the heat increment that each power module can withstand before rising from the current temperature to the upper limit of the temperature safety range (i.e., the upper limit of the preset temperature safety range, for example, 85°C) under the current temperature conditions. This parameter can be calculated based on the prediction results of the first-order thermal resistance-capacitance network model in step A200, combined with the difference between the current temperature and the upper limit of the safety range temperature, the current temperature rise rate, and the heat dissipation conditions. The greater the remaining heat dissipation capacity, the farther the module is from the temperature upper limit and the greater its potential to handle more load.

[0102] Then, based on the conversion efficiency and the remaining heat dissipation capacity, the power modules are ranked according to their energy efficiency. Specifically, the controller ranks the power modules according to a comprehensive index that combines conversion efficiency from high to low with remaining heat dissipation capacity from large to small. The ranking priority can be: first compare conversion efficiency, with modules having higher conversion efficiency ranked first; when conversion efficiencies are similar, compare remaining heat dissipation capacity, with modules having larger remaining heat dissipation capacity ranked first. This ranking result represents the priority order of each power module in accepting additional load in the current state—the modules ranked higher are the most suitable modules in the system to receive load transfer, possessing both high conversion efficiency (small impact on overall energy efficiency after load transfer) and large heat dissipation margin (able to maintain a safe temperature even after bearing more load).

[0103] Finally, if a power module's predicted temperature exceeds the first warning threshold, a load reduction strategy is generated, and part of the load on that power module is transferred to other power modules ranked higher in energy efficiency. Specifically, the first warning threshold is a warning temperature value below the upper limit of the safe temperature range (e.g., the upper limit is 85°C, and the first warning threshold is 75°C), used to identify power modules approaching the safe temperature boundary in advance. When the controller determines, based on the predicted temperature change curve in step A200, that the predicted temperature of a power module will exceed the first warning threshold within a preset time window, the controller generates a load reduction strategy for that module—specifically, reducing the module's output power and transferring the load reduction required by that module to other power modules ranked higher in energy efficiency. This load transfer prevents the high-temperature module from continuing to heat up and allows the receiving module to operate at higher efficiency, thereby reducing additional heat loss caused by the load transfer.

[0104] The load transfer amount can be calculated comprehensively based on the predicted temperature rise rate and remaining heat dissipation capacity of the high-temperature module, the remaining heat dissipation capacity and conversion efficiency of the receiving module, ensuring that the predicted temperature of the receiving module remains within a safe range after accepting the transferred load. The above load adjustment strategy can be continuously executed for each power module. When the predicted temperature recovers to below the first warning threshold, the controller can maintain the current load allocation or gradually restore the original allocation ratio.

[0105] The benefits of this implementation method are significant. First, by calculating the conversion efficiency and remaining heat dissipation capacity of each power module under current operating conditions, load adjustment decisions are based on quantitative indicators. This ensures that the load transfer strategy considers not only temperature conditions but also the energy efficiency level and heat dissipation margin of each module, achieving synergy between temperature control and energy efficiency optimization. Second, by ranking power modules according to their conversion efficiency and remaining heat dissipation capacity and prioritizing load transfer to modules ranked higher, the system ensures that during load adjustment, it always prioritizes modules with high conversion efficiency and sufficient heat dissipation margin. This reduces additional heat loss and overall energy consumption caused by load transfer, improving the power supply efficiency of the server system. Third, by setting a first warning threshold (below the upper limit of the safe temperature range), load adjustment actions are executed before the temperature reaches the critical value. Compared to the traditional approach of taking cooling measures only after the temperature exceeds the limit, this method utilizes predictive information to achieve pre-load scheduling, effectively preventing power modules from triggering passive load reduction or overheat protection due to short-term power surges. This further improves the operational stability and reliability of the server system's power system under high dynamic load conditions. This load adjustment strategy, together with the aforementioned temperature prediction and thermal state determination steps, constitutes a feedforward load and temperature collaborative control system based on prediction information.

[0106] In one embodiment of this application, the method further includes an energy-saving sleep control step, as detailed below: When the predicted temperature of all power modules is detected to be below the second energy-saving threshold, and the current total load power is below the preset sleep trigger power, the minimum number of working modules required to meet system redundancy requirements is calculated. The second energy-saving threshold is a temperature value (e.g., 55°C) below the first warning threshold, used to determine whether the current thermal state of each power module is sufficiently safe to allow some modules to enter a low-power sleep state. The preset sleep trigger power is a power threshold pre-set based on the total server load currently running on the server system. When the total load power is below this threshold, it indicates that the current system is in a low-load operating state, and there is room for some power modules to sleep without affecting power supply reliability. After the controller meets both of the above conditions (temperature is below the second energy-saving threshold and total load power is below the sleep trigger power), it calculates the minimum number of working modules required to meet the redundancy requirements under the current load conditions based on the system's preset redundancy requirements (e.g., N+1 redundancy or N+N redundancy, i.e., at least N power modules need to work simultaneously to meet the power supply demand). For example, if the current total load is 40% of the rated total capacity, and the rated power of a single power module is 20% of the total capacity, then theoretically at least 2 modules are needed to meet the load demand. Combined with N+1 redundancy, the minimum number of working modules is 3.

[0107] Then, the remaining power modules exceeding the minimum number of working modules are controlled to enter a low-power sleep state. For example, if there are 8 power modules in the system, and calculations show that only 3 modules are needed to meet the redundant power supply requirements under the current load conditions, the controller will send instructions to the power management units of the remaining 5 power modules to perform soft shutdown or standby operations, entering a low-power sleep state. The sleep state can reduce the standby power consumption of these power modules to a lower level of rated power consumption (e.g., less than 5% of full power).

[0108] Finally, the load distribution ratio of the remaining working power modules is increased to ensure they operate within a preset maximum efficiency range. The controller redistributes the load tasks originally allocated to the dormant modules to the remaining working power modules, raising the load rate of each working module to the preset maximum efficiency range (e.g., 50%~80% of rated power, where power modules typically have the highest conversion efficiency). By concentrating the total load on a smaller number of power modules, each working module operates within its high-efficiency range, reducing the problem of low overall conversion efficiency caused by all modules operating at low load rates (e.g., 20%~30% load rate).

[0109] The beneficial effects of this implementation method are significant. First, by setting a second energy-saving threshold and a dual conditional judgment based on the sleep trigger power, the energy-saving sleep operation is only initiated when the temperature of all power modules is at a safe low level and the total system load is indeed low. This avoids the risk of the remaining modules overheating due to the sleep module causing the load to increase, thus achieving energy saving while ensuring temperature safety. Second, by calculating the minimum number of working modules that meet redundancy requirements, the number of working power modules is minimized while ensuring the reliability of the server system's power supply. This allows as many modules as possible to enter a low-power sleep state, reducing the overall standby power consumption of the system. Third, by centrally distributing the load to the remaining working modules, they operate in the highest efficiency range, improving the conversion efficiency of the working modules and further reducing energy loss caused by low load operation. This step works in conjunction with the aforementioned load adjustment strategy: dynamic load distribution is used for temperature control when the load is high, and module sleep is used for energy saving optimization when the load is low. Together, they constitute a refined power management scheme covering the entire load range, significantly reducing the power system energy consumption of the server system. This energy-saving sleep control method can be widely applied to power supply systems that require high reliability and fluctuating loads, such as large server systems and edge computing nodes.

[0110] In some embodiments, generating the corresponding heat dissipation control command in step A500 specifically includes the following steps: First, the heat dissipation duct impedance characteristics of the power module are obtained, and the minimum heat dissipation airflow required to maintain the safe temperature range is calculated. Specifically, the heat dissipation duct impedance characteristics are the resistance characteristics generated by the internal heat dissipation ducts of the power module (including the air inlet, air outlet, radiator fin gaps, and the airflow path formed by the arrangement of internal components) during airflow. These characteristics can be obtained through wind tunnel testing or computational fluid dynamics simulation and are stored in the controller in the form of a wind pressure-airflow characteristic curve. The minimum heat dissipation airflow is the minimum airflow required to maintain the power module temperature below the upper limit of the safe temperature range under the current environmental conditions and the power module's heating state. Based on the temperature change trend of each power module within a preset time window predicted in step A200 and the environmental state data such as the inlet air temperature obtained in step A100, the controller, combined with the heat generation of the power module and the heat dissipation duct impedance characteristics, calculates the minimum heat dissipation airflow required to maintain the safe temperature range by back-calculating using the heat balance equation.

[0112] Then, based on the minimum cooling airflow and the impedance characteristics of the cooling duct, the base fan speed is determined. Specifically, the controller substitutes the calculated minimum cooling airflow into the air pressure-airflow characteristic curve of the cooling duct, retrieves the corresponding air pressure value that the fan needs to provide at that airflow, and then combines it with the fan's own PQ characteristic curve (the relationship between airflow and air pressure that the fan can provide at different speeds) to determine the corresponding fan speed that meets the airflow and air pressure requirements, i.e., the base fan speed. This base fan speed is the basic speed value required to maintain the temperature safety of the power module under steady-state conditions.

[0113] Then, the slope of the predicted temperature change curve is extracted as a feedforward control term. Based on this feedforward control term, the base fan speed is compensated and corrected to generate the final fan speed control command. Specifically, the controller extracts the current temperature change slope (i.e., the derivative of temperature with respect to time, reflecting the rate of temperature rise or fall) from the predicted temperature change curves of each power module obtained in step A200. The mathematical meaning of this temperature change slope as a feedforward control term is as follows: a positive and large slope indicates that the temperature is rising rapidly, requiring an increase in fan speed in advance to enhance heat dissipation and prevent temperature overshoot; a negative slope indicates that the temperature is falling, allowing for an appropriate reduction in fan speed to reduce unnecessary energy consumption and noise. The controller superimposes the base fan speed with this feedforward compensation term (for example, multiplying it by the temperature change slope using a certain proportional coefficient k and then superimposing it onto the base speed), so that the final output fan speed control command includes both the basic speed component required to maintain a steady-state temperature and a predictive compensation component to address temperature change trends. This allows the fan to adjust its speed in advance before the actual temperature change, achieving feedforward-feedback composite control.

[0114] The beneficial effects of this implementation method are significant. First, by calculating the minimum cooling airflow required to maintain a safe temperature range and determining the base fan speed based on the impedance characteristics of the cooling duct, the fan speed setting is based on accurate heat balance calculations rather than empirical estimations. This avoids the problems of insufficient cooling due to setting the fan speed too low or wasted energy due to setting it too high, achieving a precise match between cooling airflow supply and cooling demand, and improving the scientific and economical nature of fan control. Second, by extracting the slope of the predicted temperature change curve as a feedforward control term and compensating for and correcting the base fan speed, the fan speed control has an anticipatory adjustment capability. When the slope of the temperature change is positive and large (indicating that the temperature is about to rise rapidly), the fan speed increases in advance before the temperature reaches the level requiring increased airflow. This effectively avoids the response lag and temperature overshoot problems common in feedback control, resulting in smaller temperature fluctuations and a smoother temperature curve for the power module. Furthermore, the combination of feedforward control and base speed control ensures that fan speed adjustment balances the accuracy of steady-state heat dissipation requirements (base speed) with the timeliness of response to dynamic temperature changes (feedforward compensation), achieving precise and rapid control of the power module temperature. This heat dissipation control command generation method, together with the aforementioned temperature prediction steps and load adjustment strategies, forms a complete closed loop of "load-heat dissipation" coordinated regulation of the server system power system, significantly improving the temperature stability and energy efficiency of the power module under high dynamic load conditions.

[0115] In some embodiments, the automatic power supply temperature control method further includes a system-level linkage optimization step. Specifically: First, the real-time operating status parameters of the server system's cooling system are acquired. The cooling system includes equipment such as chillers, cooling towers, chilled water pumps, cooling water pumps, and air conditioning terminals (e.g., precision air conditioners or liquid-cooled distribution units). Its real-time operating status parameters include: the chiller's outlet water temperature, return water temperature, and load rate; the speed and operating frequency of the chilled water pumps and cooling water pumps; the cooling tower's heat dissipation capacity and outlet water temperature; the precision air conditioner's supply air temperature, return air temperature, and fan speed; and the total instantaneous power consumption of the cooling system. These parameters are collected in real-time by the server system infrastructure management system (DCIM) or the communication interfaces of each device (e.g., Modbus, BACnet) and aggregated to the central controller.

[0116] Then, a coupled model is established between the power supply system's heat dissipation energy consumption and the server system's cooling system energy consumption. This coupled model characterizes the relationship between the fan speed and power consumption of each power module in the power supply system, as well as the impact of the heat dissipated by the power module fans on the server system rack intake air temperature and the precision air conditioner return air temperature. This establishes a quantitative relationship between changes in power module fan speed and changes in the cooling system's cooling load. Specifically, the power supply module's heat dissipation energy consumption is a function of the fan speed of each power module (determined by the fan's PQ and power consumption characteristics), while the server system's cooling system energy consumption is a function of the air temperature entering the rack and the total heat that the cooling system needs to remove. Since the hot air discharged by the power module fans raises the ambient temperature inside the rack, thereby increasing the precision air conditioner's return air temperature and cooling load, there is a coupled relationship between the power module's heat dissipation behavior and the cooling system's energy consumption. This coupled model can be established using thermodynamic equations and server system thermal simulation.

[0117] Finally, the load allocation ratio and the heat dissipation control command are globally optimized by minimizing the sum of the total energy consumption of the target power system and the energy consumption of the server system's cooling system as the objective function. Specifically, the objective function is: Objective function = Total energy consumption of the power system (including conversion losses of all power modules + fan power consumption + standby power consumption of hibernation modules) + Total energy consumption of the cooling system (including power consumption of chillers, water pumps, cooling towers, and precision air conditioners). This objective function is subject to the following constraints: the temperature of each power module does not exceed the upper limit of the safe temperature range, the load rate of each power module does not exceed its rated capacity, system redundancy requirements are met, and all equipment in the cooling system operates within a safe operating range. The controller uses a global optimization algorithm (such as particle swarm optimization, genetic algorithm, or gradient-based sequential quadratic programming algorithm) to iteratively optimize the objective function, solving for the load allocation ratio of each power module and the fan speed control command (i.e., heat dissipation control command) that minimizes the total energy consumption under the above constraints. This achieves coordinated energy-saving optimization of the power system and cooling system while ensuring the power supply safety and temperature safety of the server system.

[0118] The beneficial effects of this implementation method are significant. First, by establishing a coupled model of the power system's heat dissipation energy consumption and the server system's cooling system energy consumption, the heat dissipation behavior of the power module and the operating status of the server system's cooling system are incorporated into a unified analysis framework. This overcomes the limitations of traditional methods where the power system's heat dissipation and cooling systems operate independently and are uncoordinated, achieving cross-system collaborative optimization. Second, the global optimization solution with minimizing total energy consumption as the objective function ensures that the determination of load distribution ratios and heat dissipation control commands is no longer limited to the power system itself (e.g., simply reducing fan speed). Instead, it comprehensively weighs the mutual influence between the power system's fan power consumption and the cooling system's energy consumption from a global perspective. For example, appropriately increasing the power module fan speed may increase the power system's heat dissipation energy consumption, but it may also lower the return air temperature entering the precision air conditioner, reduce the load and energy consumption of the cooling system, and potentially reduce overall energy consumption. Through global optimization, the optimal balance point between the power supply fan speed and the cooling capacity of the cooling system can be found, achieving the lowest overall energy consumption. Furthermore, this integrated optimization method fully utilizes predicted temperature change trend data and real-time operating status data, enabling the optimization decision-making to be forward-looking and dynamically adaptable. It can update the optimization results in real time based on changes in server system load and environmental conditions, continuously maintaining the system's high energy efficiency. This step, together with the aforementioned temperature prediction, load adjustment, and heat dissipation control steps, constitutes a full-link energy efficiency optimization system covering the internal power module and the overall server system environment, significantly reducing the server system's power utilization efficiency (PUE) and demonstrating significant economic and environmental benefits.

[0119] According to an embodiment of the third aspect of this application, the control system includes a control component for performing the control method described above.

[0120] According to the embodiments of the third aspect of this application, such as Figure 3 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute the aforementioned control method.

[0121] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] On the other hand, this application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer is able to perform the control methods provided by the above methods.

[0123] According to an embodiment of the fifth aspect of this application, the application further includes a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the control methods provided above.

[0124] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0125] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate this application and are not intended to limit this application. Although this application has been described in detail with reference to the embodiments, those skilled in the art should understand that various combinations, modifications, or equivalent substitutions of the technical solutions of this application do not depart from the spirit and scope of the technical solutions of this application and should be covered within the scope of the claims of this application.

Claims

1. A method for controlling the temperature of a server power supply, characterized in that, The system is applied to a server system, which includes a power module, a main housing, a main server, a coolant, and a vacuum pump. The main server and the coolant are disposed inside the main housing, and the vacuum pump is disposed on the outer wall of the main housing and communicates with the interior of the main housing. The control method includes: S100. Based on the current operating status of the main server, obtain the current computing power load parameters of the main server and the current temperature parameters inside the main casing; S200. Based on the current computing power load parameters and the current temperature parameters, determine whether the main server meets the preset vacuum start-up conditions, and obtain the determination result; S300. Based on the judgment result, if the preset vacuuming start condition is met, control the power module to supply power to the vacuuming device so that the vacuuming device performs a vacuuming action on the main housing, reduces the absolute pressure inside the main housing, and obtains the reduced current absolute pressure value. S400. Based on the reduced current absolute pressure value, determine the actual boiling temperature of the coolant contained inside the main housing at the reduced current absolute pressure value, wherein the actual boiling temperature is lower than the preset frequency reduction threshold temperature corresponding to the main server. S500: Based on the fact that the actual boiling temperature is lower than the preset frequency reduction threshold temperature, the coolant is driven to absorb the heat generated by the operation of the main server and undergo phase change evaporation before the temperature of the main server rises to the preset frequency reduction threshold temperature, so as to remove the heat of the main server, and at the same time, the real-time pressure change state inside the main casing caused by the phase change evaporation and the real-time temperature of the main server are obtained. S600. Based on the real-time pressure change status and the real-time temperature of the main server, control the operation of the vacuum pump to maintain the absolute pressure inside the main housing within a preset target pressure range.

2. The server power module temperature control method according to claim 1, characterized in that, Step S100 specifically includes: S101. Based on the current operating status of the main server, obtain the current computing power load parameter of the main server, wherein the current computing power load parameter represents the amount of computing tasks undertaken by the main server at the current moment; S102. Based on the temperature sensor installed inside the main housing, the current temperature parameter inside the main housing is collected, and the current temperature parameter represents the instantaneous temperature value inside the main housing at the current moment. S103. Based on the current computing power load parameters, determine the preset frequency reduction threshold temperature corresponding to the main server. The preset frequency reduction threshold temperature is the critical temperature value at which the main server triggers a frequency reduction operation due to overheating. S104. Based on the current computing power load parameters, determine the preset vacuuming start condition corresponding to the main server. The preset vacuuming start condition is a criterion for determining whether to start the vacuuming equipment to perform vacuuming action. S105. Based on the current computing power load parameters and the preset vacuum start conditions, determine the computing power load level corresponding to the current computing power load parameters; S106. Based on the computing power load level, select a preset target pressure range corresponding to the computing power load level from the preset pressure configuration table. The preset target pressure range is the range of absolute pressure values ​​that the main housing should reach after the vacuuming action is completed.

3. The server power module temperature control method according to claim 1, characterized in that, Step S200 specifically includes: S201. Based on the current computing power load parameters, compare the current computing power load parameters with a preset load threshold to obtain a load comparison result. The preset load threshold is a pre-set critical computing power value used to distinguish between high load and low load states. S202. Based on the current temperature parameter, compare the current temperature parameter with a preset temperature threshold to obtain a temperature comparison result. The preset temperature threshold is a pre-set critical temperature value used to determine whether the internal temperature of the main housing is close to the preset frequency reduction threshold temperature. S203. Based on the load comparison result and the temperature comparison result, determine whether the load comparison result indicates that the current computing power load parameter is greater than the preset load threshold, and obtain the first determination sub-result; S204. Based on the load comparison result and the temperature comparison result, determine whether the temperature comparison result indicates that the current temperature parameter is greater than the preset temperature threshold, and obtain a second determination sub-result; S205. Based on the first determination sub-result and the second determination sub-result, when the first determination sub-result indicates that the current computing power load parameter is greater than the preset load threshold and the second determination sub-result indicates that the current temperature parameter is greater than the preset temperature threshold, the determination result is determined to meet the preset vacuuming start condition.

4. The server power module temperature control method according to claim 1, characterized in that, Step S300 specifically includes: S301. Based on the judgment result, if the preset vacuuming start condition is met, a vacuuming start command is generated. The vacuuming start command is used to trigger the vacuuming equipment to start working. S302. Based on the vacuuming start command, control the vacuuming equipment to start and perform vacuuming action with a preset initial power, wherein the preset initial power is the preset starting operating power of the vacuuming equipment when it starts vacuuming. S303. Based on the vacuuming device, perform vacuuming action with the preset initial power, and obtain the current absolute pressure value inside the main shell in real time to obtain the real-time absolute pressure value during the vacuuming process. S304. Based on the real-time absolute pressure value during the vacuuming process, the real-time absolute pressure value is compared with a preset vacuuming termination pressure threshold to obtain a termination comparison result. The preset vacuuming termination pressure threshold is a pre-set absolute pressure critical value used to determine the cessation of the vacuuming action. S305. Based on the termination comparison result indicating that the real-time absolute pressure value is not greater than the preset vacuum termination pressure threshold, control the vacuuming equipment to stop performing the vacuuming action, and record the real-time absolute pressure value as the reduced current absolute pressure value.

5. The server power module temperature control method according to claim 1, characterized in that, Between step S400 and step S500, the following steps are also included: S401. Based on the reduced current absolute pressure value, by querying the physical property parameter lookup table of the coolant, the actual boiling temperature of the coolant corresponding to the reduced current absolute pressure value is obtained. The physical property parameter lookup table is a pre-stored set of mapping relationships used to characterize the boiling temperature of the coolant under different absolute pressures. S402. Based on the actual boiling temperature and the preset frequency reduction threshold temperature, the actual boiling temperature is compared with the preset frequency reduction threshold temperature to obtain a boiling temperature comparison result. The boiling temperature comparison result is used to determine whether the actual boiling temperature is lower than the preset frequency reduction threshold temperature. S403. Based on the boiling temperature comparison result, the actual boiling temperature is lower than the preset frequency reduction threshold temperature, allowing the coolant to continuously absorb the heat generated by the operation of the main server and undergo phase change evaporation during the entire heating process before the temperature of the main server rises to the preset frequency reduction threshold temperature. The following steps are included after step S500: S501. Based on the real-time pressure change state inside the main housing caused by the phase change evaporation, when the real-time pressure change state indicates that the absolute pressure rise rate inside the main housing is greater than a preset pressure rise rate threshold, control the power module to re-supply the vacuum pumping device so that the vacuum pumping device can start and perform a supplementary vacuum pumping action. The preset pressure rise rate threshold is a pre-set critical rate value used to determine the rapid increase in pressure inside the main housing due to a large amount of phase change evaporation. S502. Based on the real-time temperature of the main server, when the real-time temperature of the main server drops to a preset restart temperature threshold, control the power module to stop supplying power to the vacuum pumping device so that the vacuum pumping device stops working. The preset restart temperature threshold is a temperature hysteresis value that is preset below the preset frequency reduction threshold and is used to determine that the thermal state of the main server has recovered to a safe level.

6. The server power module temperature control method according to claim 1, characterized in that, Step S600 specifically includes: S601. Based on the real-time pressure change state inside the main shell caused by the phase change evaporation, obtain the current real-time absolute pressure value inside the main shell; S602. Based on the real-time temperature of the main server, the real-time temperature of the main server is compared with the preset frequency reduction threshold temperature to obtain a real-time temperature comparison result. The real-time temperature comparison result is used to determine whether the real-time temperature of the main server is still lower than the preset frequency reduction threshold temperature. S603. Based on the current real-time absolute pressure value and the preset target pressure range, compare the current real-time absolute pressure value with the upper limit and lower limit of the preset target pressure range respectively to obtain the pressure range comparison result; S604. Based on the pressure range comparison result and the real-time temperature comparison result, when the pressure range comparison result indicates that the current real-time absolute pressure value is lower than the lower limit of the preset target pressure range, control the power module to stop supplying power to the vacuuming device, so that the vacuuming device stops the vacuuming action, thereby allowing the absolute pressure inside the main housing to naturally rise due to the continuous phase change evaporation of the coolant. S605. Based on the pressure range comparison result and the real-time temperature comparison result, when the pressure range comparison result indicates that the current real-time absolute pressure value is higher than the upper limit of the preset target pressure range and the real-time temperature comparison result indicates that the real-time temperature of the main server is lower than the preset frequency reduction threshold temperature, the power module is controlled to supply power to the vacuum pumping device so that the vacuum pumping device starts at a preset maintenance power and performs a vacuuming action, thereby reducing the absolute pressure inside the main casing to within the preset target pressure range.

7. A control system, characterized in that, It includes a control component for performing the control method as described in any one of claims 1 to 6.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the control method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the control method according to any one of claims 1 to 8.

Images