Server circulating heat dissipation temperature control system and method thereof

Abstract

This invention discloses a server circulating heat dissipation temperature control system and method, comprising a cold plate assembly, circulating pipes, a circulating pump, a heat exchanger, a liquid storage tank, and a control unit. It adopts a multi-branch parallel liquid cooling architecture, with each cooling branch correspondingly equipped with a cold plate assembly, a flow control valve, and an outlet temperature sensor. The control unit calculates the temperature rise rate based on the branch outlet temperature, derives the predicted temperature and predicted temperature difference, and generates valve control commands by combining absolute temperature closed-loop quantities and balance correction quantities. Simultaneously, it periodically integrates the global highest predicted temperature difference, the valve opening degree in high-opening intervals, and the valve percentage to regulate the circulating pump speed. This invention introduces a global branch balance correction mechanism, calculating the relative thermal deviation based on the weighted average of the predicted temperature differences of all effective branches, and generating a correction quantity by combining the thermal response coefficient and adjustment gain. This drives the coolant to preferentially flow towards hot spot branches, solving the problems of ineffective current occupation in cold branches and insufficient heat dissipation in hot spots, and reducing the risk of overheating of core components.

Description

Technical Field

[0001] This invention relates to the field of heat dissipation and temperature control technology, and in particular to a server circulating heat dissipation and temperature control system and method. Background Technology

[0002] With the rapid development of cloud computing, artificial intelligence, and high-performance computing industries, the computing density and device integration of servers continue to increase. The heat flux density of core components such as CPUs, GPUs, memory modules, and power modules has increased significantly, making traditional air cooling insufficient to meet the heat dissipation requirements of high-density servers. Liquid cooling, due to its high heat exchange efficiency and strong heat dissipation capacity, has become the core heat dissipation solution for high heat density servers. To balance the reliability and energy efficiency of heat dissipation under all operating conditions, a combined circulating heat dissipation architecture of "liquid cooling as the main method and air cooling as a supplement" is widely used in servers and server racks with multiple heat-generating components.

[0003] In current multi-branch parallel liquid cooling systems, each heat-generating component has an independent cold plate branch, and each branch is equipped with a flow control valve to regulate the flow rate. The system provides circulation power through a variable frequency circulating pump and is supplemented by an air-cooled module to enhance heat dissipation. Existing temperature control technologies are mainly divided into two categories: independent branch temperature control and temperature-pressure coordinated temperature control. The former equips each liquid cooling branch with an independent temperature closed-loop controller, independently adjusting the corresponding valve opening based on the outlet temperature of a single branch, while uniformly regulating the circulating pump speed based on the total return water temperature or the average temperature of the branch. The latter collects multiple parameters such as the system supply and return water pressure difference and branch temperature, and synchronously adjusts the valve opening and pump output power to achieve a coordinated balance between flow rate and pressure.

[0004] However, existing temperature control solutions have the following problems in practical engineering applications:

[0005] Firstly, the cooling capacity of branch circuits lacks a coordinated allocation mechanism, resulting in poor suppression of local hot spots. Under the existing independent temperature control mode, each branch circuit valve adjusts independently based on its own temperature, without establishing a dynamic scheduling logic for cooling capacity among the branches. Branches with lower temperatures continuously occupy excessive coolant flow, while hot spots with high heat loads and rapid temperature rise cannot obtain sufficient cooling resources first. If the circulating pump adjusts only based on the average temperature, it is even more likely that local hot spots will be masked by the average temperature, leading to the risk of device overheating.

[0006] Secondly, multi-actuator coupled regulation is prone to system oscillation, resulting in insufficient control stability and energy efficiency. When the flow control valve and the circulating pump perform closed-loop integral regulation synchronously, their control actions interfere with each other, easily causing an oscillation phenomenon of "chasing" between the pump and valve. This increases the system's ineffective energy consumption, leads to a decrease in temperature control accuracy, and causes the system to remain in a fluctuating state for a long time. At the same time, the existing solution does not set a hierarchical control sequence for the valve and the circulating pump, making it impossible to maintain the valve's opening margin during regulation. This can easily lead to the problem of losing regulation capability after the valve is fully open or fully closed.

[0007] Thirdly, the auxiliary air-cooling start-stop logic is flawed, making it difficult to balance operational reliability and energy efficiency. Some solutions have excessively low air-cooling start-up thresholds, leading to prolonged low-load operation of the air-cooling module and a decrease in system energy efficiency. Other solutions have delayed start-up conditions, failing to replenish heat dissipation capacity promptly after liquid cooling capacity is saturated, exacerbating the risk of overheating. Furthermore, most solutions lack reasonable start-stop hysteresis intervals and minimum operating / stop time limits, causing frequent start-stop cycles of the air-cooling module due to small temperature fluctuations, severely shortening the lifespan of components such as fans.

[0008] Fourth, the system exhibits lag in response under dynamic loads, resulting in insufficient fault tolerance and fault diagnosis capabilities. Existing control schemes often rely solely on feedback adjustments based on the current measured temperature, failing to consider the rate of temperature rise to predict temperature trends. When faced with rapid temperature increases due to sudden changes in server computing load, the control response suffers from inherent lag, making short-term overheating a common occurrence. Furthermore, most fault diagnosis schemes rely on branch flow meters for flow monitoring, leading to high hardware costs. In scenarios without branch flow meters, there is a lack of effective means to identify abnormal branch cooling effects, and there is a lack of comprehensive safety degradation control logic when sensors fail, making it difficult to guarantee system reliability.

[0009] In summary, there is an urgent need for a server circulating heat dissipation temperature control method that is compatible with multi-branch parallel liquid cooling architecture. This method should be able to achieve priority scheduling of cooling capacity among branches, hierarchical coordinated adjustment of pumps and valves, and on-demand delayed start and stop of auxiliary air cooling, so as to comprehensively improve the dynamic response performance, operational stability and energy efficiency of high-density server heat dissipation systems. Summary of the Invention

[0010] The purpose of this invention is to address the shortcomings of the prior art by providing a server circulating heat dissipation temperature control system and method.

[0011] The objective of this invention is achieved through the following technical solution: a server circulating heat dissipation temperature control system, comprising a cold plate assembly, a circulating pipeline, a circulating pump, a heat exchanger, a liquid storage tank, and a control unit, wherein the cold plate assembly is in thermal contact with the target heat dissipation device; each cold plate assembly is provided with an outlet temperature sensor on its outlet side for collecting the outlet temperature; The circulation pipeline includes a supply main pipe, a return main pipe, and several parallel cooling branches. Each cold plate assembly is respectively set in a cooling branch. The coolant flows through the supply main pipe, the cooling branch and the corresponding cold plate assembly, the return main pipe, the circulation pump, the heat exchanger, and the storage tank in sequence before returning to the supply main pipe, forming a cooling circulation loop. Each cooling branch is equipped with a flow control valve to independently regulate the coolant flow rate of the corresponding cooling branch; the control unit is communicatively connected to the temperature sensor group, the circulating pump, and the flow control valve, and the control unit is configured to execute the following temperature control logic: Calculate the temperature rise rate of each cooling branch based on the outlet side temperature of each cooling branch, and obtain the predicted temperature and predicted temperature difference based on the temperature rise rate; calculate the absolute temperature closed-loop quantity and balance correction quantity, and calculate the control command of the flow control valve based on the absolute temperature closed-loop quantity and balance correction quantity. The circulating pump speed is periodically adjusted by comprehensively calculating the circulating pump speed command based on the global highest predicted temperature difference, the opening value of the flow control valve in the preset high opening range, and the valve ratio in the preset high opening range, and then controlling the operation of the circulating pump.

[0012] Preferably, an auxiliary air-cooling module is also included; the auxiliary air-cooling module includes a fan and a guide air duct, the inlet of the guide air duct is connected to the hot air area inside the server chassis, the heat exchanger is located in the flow section of the guide air duct, and the fan is used to drive the hot air inside the server chassis to flow through the air-cooled heat exchanger and then be discharged outward. The control unit analyzes the liquid cooling capacity saturation status in real time and determines whether the system meets the cooling mode switching conditions. If the cooling mode switching conditions are met, the auxiliary air cooling module is started, and the fan of the auxiliary air cooling module is subjected to closed-loop speed regulation.

[0013] Preferably, the cold plate assembly is provided with a microchannel, and the two ends of the microchannel are respectively connected to the liquid inlet distribution chamber and the liquid outlet collection chamber.

[0014] A server circulating heat dissipation temperature control method includes the following specific steps: S1: Collect outlet temperature data for each cooling branch, calculate the temperature rise rate of each cooling branch, and calculate the predicted temperature and predicted temperature difference for each cooling branch in the future based on the temperature rise rate. S2: Calculate the absolute temperature closed-loop control quantity of the flow control valve corresponding to each cooling branch, and calculate the balance correction quantity based on the predicted temperature difference of each cooling branch. Calculate the original valve control command based on the absolute temperature closed-loop control quantity and the balance correction quantity. The original valve control command is output to each flow control valve after being limited by amplitude and rate of change. S3: Adjust the circulating pump based on the preset control cycle, extract the opening value of the flow control valve in the preset high opening range, calculate the valve ratio in the preset high opening range, and calculate the circulating pump speed command based on the global maximum predicted temperature difference, the high opening value of the flow control valve in the preset high opening range, and the valve ratio in the preset high opening range, and control the operation of the circulating pump. S4: Analyze the liquid cooling capacity saturation status in real time and determine whether the system meets the cooling mode switching conditions; if the cooling mode switching conditions are met, start the auxiliary air cooling module and perform closed-loop speed regulation on the fan of the auxiliary air cooling module; shut down the auxiliary air cooling module after the system meets the exit conditions.

[0015] Preferably, in step S1, the original outlet temperature signals of each cooling branch collected by the outlet temperature sensor are subjected to validity verification, median filtering and first-order filtering to obtain outlet side temperature data. No. The temperature rise rate of each cooling branch at the k-th sampling time The calculation is based on temperature data obtained from m consecutive sampling times, and the formula is as follows: ; In the formula, For temperature sampling period, Let be the filtered outlet temperature of the i-th cooling branch at the k-th sampling time. For the first The filtered outlet temperature at each sampling time; Predicted temperature at the k-th sampling time for: ; In the formula, For the amplitude limiting function, This is the upper limit threshold for the rate of temperature rise. For prediction time constant; The predicted temperature difference is the difference between the predicted temperature of the current cooling branch and the target outlet temperature.

[0016] Preferably, in step S2, the inlet side temperature of the cold plate assembly is collected, and the heat load of the current cooling branch is calculated based on the inlet side temperature, the outlet side temperature and the pre-calibrated pump speed-valve opening-flow mapping relationship. The feedforward flow demand is calculated based on the heat load, and the feedforward flow demand is converted into the bias of the flow control valve and superimposed on the absolute temperature closed-loop control quantity.

[0017] Preferably, the calculation process for the balance correction in step S2 is as follows: Calculate the weighted average of the predicted temperature differences for all effective cooling branches to obtain the average predicted temperature difference; The difference between the predicted temperature difference of each cooling branch and the average predicted temperature difference is used as the relative thermal deviation. The balance correction amount for each cooling branch is calculated based on the relative thermal deviation. The calculation formula is as follows: ; In the formula, To effectively increase the number of cooling branches, For the first The relative thermal response coefficient of each cooling branch; To balance the gain of the branch, This represents the relative thermal deviation of the i-th cooling branch. The subscript is used for traversal, representing each valid cooling branch involved in the calculation; Represents an effective cooling branch; This represents the relative thermal deviation of the j-th effective cooling branch. Let represent the relative thermal response coefficient of the j-th effective cooling branch; the algebraic sum of the balance corrections for all effective cooling branches is zero. When the command of the flow control valve reaches the upper or lower limit of the opening, bounded projection is performed. The specific method is as follows: First, the flow control valves that have reached the boundary are limited and a residual correction amount is generated; the residual correction amount generated by the limiting is redistributed according to the residual opening margin of the flow control valves that have not reached the opening boundary, until the residual correction amount is less than the set threshold or all flow control valves reach the opening boundary; the opening of the flow control valves of all cooling branches is not lower than the preset minimum safe opening.

[0018] Preferably, when the temperature of each cooling branch is in a stable state, a self-calibration process for the thermal response coefficient is performed, as follows: Keeping the circulating pump and fan in constant condition, the opening of the flow control valve corresponding to each individual cooling branch is adjusted sequentially. After the branch thermal delay, the original thermal response coefficient of the current cooling branch is calculated. The original thermal response coefficient is the ratio of the steady-state change of the outlet temperature of the cooling branch to the change of the opening of the flow control valve. The original thermal response coefficients of all cooling branches are calculated sequentially using the above method. The median of all original thermal response coefficients is then calculated, and the ratio of the original thermal response coefficient of the current cooling branch to the median of all original thermal response coefficients is taken as the relative thermal response coefficient of the current cooling branch.

[0019] Preferably, in step S4, the cooling mode switching conditions include: 1. The continuous time during which the highest predicted global temperature difference high pressure reaches the air-cooled start-up temperature difference threshold and the liquid cooling capacity is in a saturated state reaches the preset start-up confirmation time. 2. Any predicted temperature reaches the preset high-temperature protection threshold; 3. The temperature rise rate of any cooling branch exceeds the rapid temperature rise threshold, and the speed of the circulating pump reaches the preset high speed and the opening degree of the flow control valve of the cooling branch reaches the preset high opening degree. If any of the above conditions are met, the auxiliary air-cooling module will be activated; If the global highest predicted temperature difference is lower than the air-cooling exit temperature difference threshold and the liquid cooling capacity remains unsaturated for a period of time that reaches the preset exit confirmation time, then the exit condition is met and the auxiliary air-cooling module is shut down; the air-cooling exit temperature difference threshold is less than the air-cooling start-up temperature difference threshold.

[0020] Preferably, in step S4, the closed-loop speed control method of the fan is as follows: the difference between the global maximum predicted temperature difference and the preset air-cooled reference temperature difference is used as a variable input to the dead zone function to obtain the fan adjustment error; based on the fan proportional control gain and the fan adjustment error as the proportional adjustment amount, the original speed command is generated by superimposing the fan minimum starting speed, the proportional adjustment amount and the integral term; the original speed command is output to the fan of the auxiliary air-cooled module after integral anti-saturation processing, speed limiting and rate of change limiting; the fan running time each time is not less than the minimum fan running time; the fan shutdown time is not less than the preset minimum shutdown time.

[0021] The beneficial effects of this invention are: 1. This invention breaks through the technical limitations of traditional independent closed-loop control of each branch. On the basis of single-branch absolute temperature closed-loop control, a global branch balance correction mechanism is added to realize dynamic scheduling of cooling capacity among branches. Based on the weighted average of the predicted temperature difference of all effective branches, the relative thermal deviation of each branch is calculated. Combined with the relative thermal response coefficient of the branch and the balance adjustment gain, a balance correction amount is generated and superimposed on the valve control command. This drives the coolant flow to preferentially tilt towards hot spots with high heat load and rapid temperature rise, while low-temperature and low-load branches actively release redundant flow. The coolant resource allocation is optimized from a global perspective, which solves the problems of ineffective flow occupation of cold branches and insufficient cooling of hot spots under traditional independent temperature control. It avoids the risk of average temperature masking local hot spots and significantly reduces the probability of overheating of core components such as CPU and GPU.

[0022] 2. This invention adopts a hierarchical time-sequential control architecture. The flow control valve performs real-time continuous closed-loop temperature regulation, and the circulating pump performs periodic speed regulation according to a preset control cycle. Through peak-shifting regulation in the time dimension, control interference caused by synchronous integral regulation of both is avoided, significantly improving the stability and accuracy of system temperature control and reducing ineffective fluctuation energy consumption. The circulating pump speed adopts multi-parameter comprehensive control logic, using the global maximum predicted temperature difference, the opening value of valves in the high-opening range, and the proportion of high-opening valves as joint inputs, rather than coarsely adjusting based on average temperature: when most valves are in the low-opening range, the pump speed is actively reduced to reduce redundant head; when the proportion of high-opening valves increases and the global temperature difference increases, the pump speed is gradually increased to supplement the system flow pressure, always maintaining sufficient opening adjustment margin for valves, avoiding loss of branch regulation capability after valves are fully open / closed, realizing dynamic matching of system pressure, branch flow and heat dissipation demand, effectively reducing the operating energy consumption of the circulating pump and improving the overall energy efficiency ratio.

[0023] 3. This invention constructs a complete auxiliary air-cooling start-stop and speed regulation logic. The fan speed is adjusted in a closed loop proportional-integral manner based on the global highest predicted temperature difference, so as to realize the dynamic matching between air-cooling heat dissipation capacity and system heat load. When the liquid cooling capacity is insufficient, the heat dissipation power is accurately supplemented, taking into account both heat dissipation reliability and system energy efficiency.

[0024] 4. This invention designs an online self-calibration mechanism for branch thermal response coefficients. Under steady-state system temperature conditions, the opening degree of valves in a single branch is adjusted sequentially. The original thermal response coefficient is calculated based on the ratio of the steady-state change in outlet temperature to the change in valve opening. Then, the relative thermal response coefficient of each branch is obtained by normalizing the median of the original thermal response coefficients of all branches. This allows for accurate quantification of branch heat dissipation characteristics without the need to configure flow meters in each branch, significantly reducing system hardware costs and pipeline wiring complexity. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall structure of the server circulating heat dissipation temperature control system of the present invention.

[0026] Figure 2 This is a schematic diagram of the microchannel structure of the cold plate assembly.

[0027] Figure 3 This is a flowchart of the circulating heat dissipation temperature control method of the present invention.

[0028] In the diagram, 101 is the cold plate assembly, 103 is the microchannel, 104 is the liquid inlet distribution chamber, 105 is the liquid outlet collection chamber, 201 is the liquid supply main pipe, 202 is the liquid return main pipe, 203 is the cooling branch, 204 is the quick-connect sealing joint, 205 is the heat insulation layer, 300 is the circulating pump, 400 is the heat exchanger, 410 is the auxiliary air cooling module, 411 is the fan, 412 is the air duct, 500 is the liquid storage tank, 501 is the exhaust valve, 502 is the drain port, 600 is the flow control valve, 601 is the outlet temperature sensor, 700 is the control unit, 800 is the server chassis, and 801 is the coolant inlet. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0030] Those skilled in the art should understand that, in the disclosure of this invention, the terms "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 this invention 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, the above terms should not be construed as limiting this invention.

[0031] It is understood that the term "a" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0032] like Figures 1 to 2 As shown, a server circulating heat dissipation temperature control system includes a cold plate assembly 101, a circulation pipeline, a circulation pump 300, a heat exchanger 400, a liquid storage tank 500, and a control unit 700. The cold plate assembly 101 is in thermal contact with the target heat dissipation device. Each cold plate assembly 101 is provided with an outlet temperature sensor 601 on its outlet side for collecting the outlet temperature. The circulation pipeline includes a liquid supply main pipe 201, a liquid return main pipe 202, and several parallel cooling branch pipes 203. Each cold plate assembly 101 is respectively installed in a cooling branch pipe 203. The coolant flows sequentially through the liquid supply main pipe 201, the cooling branch pipes 203 and the corresponding cold plate assembly 101, the liquid return main pipe 202, the circulation pump 300, the heat exchanger 400, and the liquid storage tank 500 before returning to the liquid supply main pipe 201, forming a cooling circulation loop. Each cooling branch 203 is equipped with a corresponding flow control valve 600 for independently adjusting the coolant flow rate of the corresponding cooling branch 203; the control unit 700 is communicatively connected to the temperature sensor group, the circulating pump 300, and the flow control valve 600, and the control unit 700 is configured to execute the following temperature control logic: The temperature rise rate of each cooling branch 203 is calculated based on the outlet side temperature of each cooling branch 203. The predicted temperature and predicted temperature difference are obtained based on the temperature rise rate. The absolute temperature closed-loop quantity and balance correction quantity are calculated. The control command of the flow control valve 600 is calculated based on the absolute temperature closed-loop quantity and balance correction quantity. The speed of the circulating pump 300 is periodically adjusted. The speed command of the circulating pump 300 is calculated based on the global highest predicted temperature difference, the opening value of the flow control valve 600 located in the preset high opening range, and the valve ratio in the preset high opening range, and the operation of the circulating pump 300 is controlled.

[0033] The target heat dissipation devices are the main heat-generating chips such as CPU and GPU, as well as power devices such as memory, power modules, and power supply modules.

[0034] The circulating pump 300 can be a pump that supports frequency conversion, PWM, or stepped speed regulation. Each cooling branch pipe is equipped with a flow control valve 600, which can be a proportional electric regulating valve. The flow control valve 600 can be located on the outlet or inlet side of the cold plate assembly 101.

[0035] In some embodiments of this application, an inlet temperature sensor is also provided on the inlet side of the cold plate assembly 101 to detect the inlet temperature of the cold plate assembly 101.

[0036] The supply main pipe 201, return main pipe 202, and cooling branch pipes are all connected by quick-connect sealing joints 204, and a heat insulation layer 205 is installed around the outer perimeter of the pipes. The side or back panel of the server chassis 800 is equipped with a coolant inlet 801 and a coolant outlet, the positions of which correspond to those of the supply main pipe 201 and return main pipe 202. The storage tank 500 is equipped with an vent valve 501 and a drain port 502.

[0037] The cold plate assembly 101 is provided with a microchannel 103, and the two ends of the microchannel 103 are respectively connected to the liquid inlet distribution chamber 104 and the liquid outlet collection chamber 105.

[0038] In this application, the heat exchanger 400 is an air-cooled heat exchanger 400.

[0039] This circulating heat dissipation and temperature control system also includes an auxiliary air-cooling module 410; the auxiliary air-cooling module 410 includes a fan 411 and a guide air duct 412. The inlet of the guide air duct 412 connects to the hot air area inside the server chassis 800. The heat exchanger 400 is located within the flow section of the guide air duct 412. The fan 411 drives the hot air inside the server chassis 800 to flow through the air-cooled heat exchanger 400 and then exhaust it outwards. The control unit 700 analyzes the liquid cooling capacity saturation state in real time and determines whether the system meets the cooling mode switching conditions. If the cooling mode switching conditions are met, the auxiliary air-cooling module 410 is activated, and closed-loop speed regulation is performed on the fan 411 of the auxiliary air-cooling module 410. A windproof or sealed structure is provided between the heat exchanger 400 and the guide air duct 412 to reduce air bypass.

[0040] like Figure 3 As shown, the present invention also provides a server circulating heat dissipation temperature control method, including the following specific steps: S1: Collect outlet temperature data for each cooling branch, calculate the temperature rise rate of each cooling branch, and calculate the predicted temperature and predicted temperature difference for each cooling branch in the future based on the temperature rise rate.

[0041] In this application, the server circulating heat dissipation temperature control system is configured with a total of There are 1 cooling branch, and the cooling branch number is... .

[0042] Before system startup, preset the target outlet temperature for each cooling branch. High temperature protection threshold Emergency temperature threshold These preset parameters include the upper limit of the flow control valve opening, the lower limit of the flow control valve opening, the upper limit of the circulating pump speed, and the lower limit of the circulating pump speed; among them, .

[0043] Simultaneously define the dead-time function. : ; dead zone width, For input variables. It is a symbolic function.

[0044] Upon system startup, the control unit reads the target outlet temperature of each cooling branch, the high-temperature protection threshold, the emergency temperature threshold, the upper and lower limits of the flow control valve opening, the upper and lower limits of the circulating pump speed, and checks the parameter logic relationships. Each flow control valve is then set to its preset startup opening. The circulating pump is set to the preset starting speed. The fans of the auxiliary air-cooling module should be kept off.

[0045] Upon system startup, a pre-circulation process is executed first. During this phase, the control unit circulates the coolant within each cooling branch and expels residual gases through the reservoir and vent valve. Integral regulation of the flow control valve is paused during the pre-circulation period. Once the pre-circulation is complete and the number of effective outlet temperature sensors meets the requirements, the system enters the liquid-cooled main control state. In this state, the fans of the auxiliary air-cooling module remain off, and cooling is achieved solely through the circulation of coolant in the cooling branches.

[0046] In this step, the raw outlet temperature signals of each cooling branch collected by the outlet temperature sensor are subjected to validity verification, median filtering, and first-order filtering to obtain the outlet side temperature data. The specific methods for median filtering and first-order filtering are as follows: The median of 3 to 7 consecutive raw outlet temperature signals is taken. : ; For the first The cooling branch is in the first The original signal of the outlet temperature at each sampling time. For the first The cooling branch is in the first -1 sampling time of the original signal of the outlet temperature.

[0047] Then, a first-order filter is performed to obtain the... The cooling branch is in the first outlet temperature after median filtering at each sampling time : ; In the formula, These are the filter coefficients. For the first The cooling branch is in the first The outlet temperature after median filtering at each sampling time.

[0048] A larger filtering coefficient is used when the temperature rises and a smaller filtering coefficient is used when the temperature falls, thereby achieving a rapid response during the heating phase and a slow removal of cooling capacity during the cooling phase.

[0049] No. The temperature rise rate of each cooling branch at the k-th sampling time The calculation is based on temperature data obtained from m consecutive sampling periods, and the formula is as follows: ; In the formula, For temperature sampling period, The outlet temperature of the i-th cooling branch after filtering at the k-th sampling time is... The outlet temperature of the i-th cooling branch after filtering at the (k-1)-th sampling time; Predicted temperature at time k for: ; In the formula, This is a limiting function; The upper limit threshold for the rate of temperature rise is set in advance; Let be the predicted time constant for the i-th cooling branch; Predicted temperature difference This is the difference between the predicted temperature and the target outlet temperature of the current cooling branch, i.e.: .

[0050] Predicted temperature difference from all effective cooling branches Find the highest predicted temperature difference globally : .

[0051] S2: Calculate the absolute temperature closed-loop control quantity of the flow control valve corresponding to each cooling branch. Simultaneously, the balance correction amount is calculated based on the predicted temperature difference of each cooling branch. The original control command for the valve is calculated based on the absolute temperature closed-loop control quantity and the balance correction quantity. The original control command for the valve is then output to each flow control valve after being limited by amplitude and rate of change.

[0052] In this step, the absolute temperature closed-loop control quantity : ; The proportional gain coefficient for the i-th cooling branch is... The deviation function with dead zone dead zone width, This is the gain coefficient for the temperature rise rate. Let be the predicted time constant for the i-th cooling branch. Let be the temperature rise rate of the i-th cooling branch. This represents the actual temperature difference. .

[0053] Points Pre-Update Value for: ; The final integral term of the previous moment, The integral gain coefficient of the i-th cooling branch, The temperature sampling period; The calculation process for the balance correction is as follows: Calculate the weighted average of the predicted temperature differences for all effective cooling branches to obtain the average predicted temperature difference. ; ; in, This represents the priority weight of the i-th cooling branch, or the confidence level of the outlet temperature sensor for the corresponding cooling branch. Represents an effective cooling branch; The difference between the predicted temperature difference of each cooling branch and the average predicted temperature difference is used as the relative thermal deviation. : ; The balance correction amount for each cooling branch is calculated based on the relative thermal deviation. The calculation formula is as follows: ; In the formula, To effectively increase the number of cooling branches, For the first The relative thermal response coefficient of each cooling branch, when not calibrated, is taken as 1; To balance the gain of the branch, This represents the relative thermal deviation of the i-th cooling branch. The subscript is used for traversal, representing each valid cooling branch involved in the calculation; Represents an effective cooling branch; This represents the relative thermal deviation of the j-th effective cooling branch. This represents the relative thermal response coefficient of the j-th effective cooling branch; The algebraic sum of the balance corrections for all effective cooling branches is zero.

[0054] When the opening of a cooling branch with a higher temperature or a faster temperature rises is increased, the opening of a cooling branch with a lower temperature is reduced accordingly. This allows the existing cooling capacity to be redistributed among the branches first, rather than immediately opening all flow control valves and increasing the pump speed simultaneously.

[0055] The original valve control command for the flow control valve corresponding to the i-th cooling branch. for: ; in, The preset steady-state base opening for the i-th cooling branch; The pump speed feedforward gain for the i-th cooling branch; This represents the change in the speed of the circulating pump. ; Let K be the normalized rotational speed of the circulating pump at the Kth sampling time. The normalized rotational speed of the circulating pump at the (K-1)th sampling time.

[0056] Receive valve original control command Then, the valve's final control command is obtained after limiting the amplitude and rate of change of the original control command. , ; The rate of change constraint function, For the amplitude limiting function, To preset the maximum safe opening, This is the preset minimum safe opening degree.

[0057] Final valve command The output is sent to each flow control valve.

[0058] When the flow control valve command reaches the upper or lower opening limit, bounded projection is executed. The specific method is as follows: First, the flow control valves that have reached the boundary are limited and a residual correction amount is generated; the residual correction amount generated by the limiting is redistributed according to the residual opening margin of the flow control valves that have not reached the opening boundary, until the residual correction amount is less than the set threshold or all flow control valves have reached the opening boundary; the opening of the flow control valves in all cooling branches is not lower than the preset minimum safe opening. .

[0059] Furthermore, the inlet temperature of the cold plate assembly is collected by an inlet temperature sensor installed on the inlet side of the cold plate assembly. Based on the inlet temperature, outlet temperature, and a pre-calibrated pump speed-valve opening-flow mapping relationship, the heat load of the current cooling branch is calculated. ; Among them, in calculating the heat load of the current cooling branch First, calculate the inlet and outlet temperature difference of the current cooling branch. Inlet and outlet temperature difference This is the difference between the inlet-side temperature, the outlet-side temperature, and the inlet-side temperature. The pump speed-valve opening-flow mapping relationship characterizes the correspondence between the circulating pump speed, the flow control valve opening, and the flow rate in the cooling branch. Here, the current circulating pump speed and the flow control valve opening are known quantities; by combining these with the pump speed-valve opening-flow mapping relationship, the current flow rate in the cooling branch can be obtained. The heat load of the current cooling branch is then determined. The calculation formula is as follows: ; In the formula, and These are the density and specific heat capacity at constant pressure of the coolant in the cooling branch, respectively. This represents the current flow rate of the cooling branch.

[0060] Feedforward flow demand calculated based on heat load The calculation formula is as follows: ; The allowable temperature rise value is the preset value.

[0061] The feedforward flow demand is converted into the bias of the flow control valve and added to the absolute temperature closed-loop control quantity.

[0062] S3: Adjust the circulating pump based on the preset control cycle, extract the opening value of the flow control valve located in the preset high opening range, calculate the valve ratio in the preset high opening range, and comprehensively calculate the circulating pump speed command based on the global maximum predicted temperature difference, the opening value of the flow control valve located in the preset high opening range, and the valve ratio in the preset high opening range, and control the operation of the circulating pump.

[0063] In this step, the value located within the preset high opening range is first calculated based on the preset high opening range. High opening value of flow control valve : ; In the formula, N is the total number of flow control valves located within the preset high opening range.

[0064] In this application, a high opening range is preset. The value ranges from 0.75 to 0.90.

[0065] Then calculate the valve proportion within the preset high opening range. The valve ratio The calculation method is: the total number of flow control valves located within the preset high opening range divided by the total number of effective flow control valves.

[0066] Then, the global maximum predicted temperature difference and the high opening value of the flow control valve are analyzed. and the proportion of valves in the preset high opening range Normalization is performed to obtain the normalized value of the highest predicted temperature difference globally. Normalized value of high opening degree of flow control valve and valve proportional normalization value The details are as follows:

[0067]

[0068]

[0069] In the formula, The desired high valve opening degree is the preset value. This is the normalized reference value for temperature difference. The normalized reference value for opening degree, As the weight of the temperature term, As the weight of the opening margin term, Weight of the high opening ratio item.

[0070] Original speed command of the circulating pump for: ; In the formula, The preset base steady-state speed, For the proportional gain of the circulating pump, This is the effective integral term for the circulating pump.

[0071] Among them, the pre-update value of the integral term of the circulating pump for:

[0072] In the formula, For the valid integral term of the previous period, For the integral gain of the circulating pump, This refers to the slow-layer control cycle of the circulating pump.

[0073] To avoid the valve statistics having an inappropriate negative effect on the circulating pump speed under hot conditions, this invention sets a temperature-forced pump speed lower limit. When the circulating pump is running, its speed is not lower than the temperature-forced pump speed lower limit.

[0074] When the adjustment amount of the circulating pump exceeds the preset value in a single operation, the integral term in the control command of the flow control valve is frozen within 2 to 20 seconds to reduce the oscillation caused by the simultaneous operation of the two integral components of the circulating pump and the flow control valve.

[0075] S4: Analyze the liquid cooling capacity saturation status in real time and determine whether the system meets the cooling mode switching conditions; if the cooling mode switching conditions are met, start the auxiliary air cooling module and perform closed-loop speed regulation on the fan of the auxiliary air cooling module; shut down the auxiliary air cooling module after the system meets the exit conditions.

[0076] In this step, The conditions for determining the saturation state of liquid cooling capacity are as follows:

[0077] in, For air-cooled start-up permit conditions, This is the current speed of the circulating pump. This refers to the threshold speed for air-cooled pump startup. This represents the selection criteria for branches where the predicted temperature difference is positive.

[0078] The conditions for air-cooled start-up are: the maximum value of the flow control valve opening in each branch where the current speed of the circulating pump is greater than or equal to the air-cooled start-up pump speed threshold and the predicted temperature difference is positive is greater than or equal to the air-cooled start-up opening threshold.

[0079] The cooling mode switching conditions include: 1. The continuous time during which the highest predicted global temperature difference high pressure reaches the air-cooled start-up temperature difference threshold and the liquid cooling capacity is in a saturated state reaches the preset start-up confirmation time. 2. Any predicted temperature reaches the preset high-temperature protection threshold; 3. The temperature rise rate of any cooling branch exceeds the rapid temperature rise threshold, and the speed of the circulating pump reaches the preset high speed and the opening degree of the flow control valve of the cooling branch reaches the preset high opening degree.

[0080] If any of the above conditions are met, the auxiliary air-cooling module will be activated; If the global highest predicted temperature difference is lower than the air-cooling exit temperature difference threshold and the liquid cooling capacity remains unsaturated for a period of time that reaches the preset exit confirmation time, then the exit condition is met and the auxiliary air-cooling module is shut down; the air-cooling exit temperature difference threshold is less than the air-cooling start-up temperature difference threshold.

[0081] The closed-loop speed control method for the fan is as follows: the difference between the global maximum predicted temperature difference and the preset air-cooled reference temperature difference is used as a variable input to the dead-zone function to obtain the fan adjustment error. ; ; In the formula, This is the reference temperature difference for air cooling.

[0082] The original speed command is generated by superimposing the fan proportional control gain and fan adjustment error as proportional adjustment quantities, and the minimum starting speed of the fan, the proportional adjustment quantity, and the integral term. ; ; In the formula, For the effective integral term of the fan, Minimum fan speed for startup. This is the fan proportional gain coefficient.

[0083] The original speed command is output to the fan of the auxiliary air-cooling module after integral anti-saturation processing, speed limiting, and rate of change limitation; the fan's running time each time is not less than the minimum fan running time; the fan's shutdown time is not less than the preset minimum shutdown time.

[0084] When the temperature of each cooling branch is stable, a self-calibration process for the thermal response coefficient is performed, as follows: Keeping the circulating pump and fan in constant condition, the opening of the flow control valve corresponding to each individual cooling branch is adjusted sequentially. After the branch thermal delay, the original thermal response coefficient of the current cooling branch is calculated. The original thermal response coefficient is the ratio of the steady-state change of the outlet temperature of the cooling branch to the change of the opening of the flow control valve. The original thermal response coefficients of all cooling branches are calculated sequentially using the above method. The median of all original thermal response coefficients is then calculated, and the ratio of the original thermal response coefficient of the current cooling branch to the median of all original thermal response coefficients is taken as the relative thermal response coefficient of the current cooling branch.

[0085] During operation, the control integral term of the flow control valve, circulating pump and fan is subjected to inverse calculation anti-saturation treatment; according to the sensor validity and the cooling response status of the cooling branch, when the sensor fails, the branch cooling is abnormal or the temperature is too high, it enters the degraded operation or emergency protection state.

[0086] Specifically, when the outlet temperature sensor of each cooling branch fails, the flow control valve of the corresponding cooling branch is set to a preset safe opening. The signal of that cooling branch is excluded from the global statistics, the circulating pump is maintained at a speed no lower than the minimum degraded speed, and the remaining effective cooling branches continue to be controlled in a closed loop.

[0087] When any cooling branch reaches the emergency temperature threshold, multiple critical sensors in the cooling branches fail simultaneously, or the system continues to heat up rapidly, an emergency state is entered: the circulating pump and fan are increased to the maximum allowable value, the valves of the hot spot branch are fully opened, the remaining cooling branches are maintained at an opening not lower than the safe opening, and an over-temperature alarm and load reduction request are output; under the condition that the circulating pump and air cooling are stable, when the opening increment of the flow control valve of a certain cooling branch exceeds the test threshold and the temperature rise rate does not decrease after passing through the response window, and this is repeated a set number of times, it is determined that the cooling effect of that cooling branch is abnormal, an alarm is output and the system enters a degraded state.

[0088] Compared with existing technologies, the present invention has the following advantages: 1. This invention breaks through the technical limitations of traditional independent closed-loop control of each branch. On the basis of single-branch absolute temperature closed-loop control, a global branch balance correction mechanism is added to realize dynamic scheduling of cooling capacity among branches. Based on the weighted average of the predicted temperature difference of all effective branches, the relative thermal deviation of each branch is calculated. Combined with the relative thermal response coefficient of the branch and the balance adjustment gain, a balance correction amount is generated and superimposed on the valve control command. This drives the coolant flow to preferentially tilt towards hot spots with high heat load and rapid temperature rise, while low-temperature and low-load branches actively release redundant flow. The coolant resource allocation is optimized from a global perspective, which solves the problems of ineffective flow occupation of cold branches and insufficient cooling of hot spots under traditional independent temperature control. It avoids the risk of average temperature masking local hot spots and significantly reduces the probability of overheating of core components such as CPU and GPU.

[0089] 2. This invention adopts a hierarchical time-sequential control architecture. The flow control valve performs real-time continuous closed-loop temperature regulation, and the circulating pump performs periodic speed regulation according to a preset control cycle. Through peak-shifting regulation in the time dimension, control interference caused by synchronous integral regulation of both is avoided, significantly improving the stability and accuracy of system temperature control and reducing ineffective fluctuation energy consumption. The circulating pump speed adopts multi-parameter comprehensive control logic, using the global maximum predicted temperature difference, the opening value of valves in the high-opening range, and the proportion of high-opening valves as joint inputs, rather than coarsely adjusting based on average temperature: when most valves are in the low-opening range, the pump speed is actively reduced to reduce redundant head; when the proportion of high-opening valves increases and the global temperature difference increases, the pump speed is gradually increased to supplement the system flow pressure, always maintaining sufficient opening adjustment margin for valves, avoiding loss of branch regulation capability after valves are fully open / closed, realizing dynamic matching of system pressure, branch flow and heat dissipation demand, effectively reducing the operating energy consumption of the circulating pump and improving the overall energy efficiency ratio.

[0090] 3. This invention constructs a complete auxiliary air-cooling start-stop and speed regulation logic. The fan speed is adjusted in a closed loop proportional-integral manner based on the global highest predicted temperature difference, so as to realize the dynamic matching between air-cooling heat dissipation capacity and system heat load. When the liquid cooling capacity is insufficient, the heat dissipation power is accurately supplemented, taking into account both heat dissipation reliability and system energy efficiency.

[0091] 4. This invention designs an online self-calibration mechanism for branch thermal response coefficients. Under steady-state system temperature conditions, the opening degree of valves in a single branch is adjusted sequentially. The original thermal response coefficient is calculated based on the ratio of the steady-state change in outlet temperature to the change in valve opening. Then, the relative thermal response coefficient of each branch is obtained by normalizing the median of the original thermal response coefficients of all branches. This allows for accurate quantification of branch heat dissipation characteristics without the need to configure flow meters in each branch, significantly reducing system hardware costs and pipeline wiring complexity.

[0092] This invention is not limited to the preferred embodiments described above. Anyone can derive other products in various forms under the guidance of this invention. However, regardless of any changes in shape or structure, any technical solution that is the same as or similar to this application falls within the protection scope of this invention.

Claims

1. A server circulating heat dissipation temperature control system, characterized in that, It includes a cold plate assembly, circulation pipeline, circulation pump, heat exchanger, liquid storage tank, and control unit. The cold plate assembly is in thermal contact with the target heat dissipation device. Each cold plate assembly is equipped with an outlet temperature sensor on its outlet side for collecting the outlet temperature. The circulation pipeline includes a supply main pipe, a return main pipe, and several parallel cooling branches. Each cold plate assembly is respectively set in a cooling branch. The coolant flows through the supply main pipe, the cooling branch and the corresponding cold plate assembly, the return main pipe, the circulation pump, the heat exchanger, and the storage tank in sequence before returning to the supply main pipe, forming a cooling circulation loop. Each cooling branch is equipped with a flow control valve to independently regulate the coolant flow rate of the corresponding cooling branch; the control unit is communicatively connected to the temperature sensor group, the circulating pump, and the flow control valve, and the control unit is configured to execute the following temperature control logic: Calculate the temperature rise rate of each cooling branch based on the outlet side temperature of each cooling branch, and obtain the predicted temperature and predicted temperature difference based on the temperature rise rate; calculate the absolute temperature closed-loop quantity and balance correction quantity, and calculate the control command of the flow control valve based on the absolute temperature closed-loop quantity and balance correction quantity. The circulating pump speed is periodically adjusted by comprehensively calculating the circulating pump speed command based on the global maximum predicted temperature difference, the opening value of the flow control valve in the preset high opening range, and the valve ratio in the preset high opening range, and then controlling the operation of the circulating pump. It also includes an auxiliary air-cooling module; the auxiliary air-cooling module includes a fan and a guide air duct, the inlet of the guide air duct is connected to the hot air area inside the server chassis, the heat exchanger is located in the flow section of the guide air duct, and the fan is used to drive the hot air inside the server chassis to flow through the air-cooled heat exchanger and then be discharged outward. The control unit analyzes the liquid cooling capacity saturation status in real time and determines whether the system meets the cooling mode switching conditions. If the cooling mode switching conditions are met, the auxiliary air cooling module is started, and the fan of the auxiliary air cooling module is subjected to closed-loop speed regulation.

2. The server circulating heat dissipation temperature control system according to claim 1, characterized in that, The cold plate assembly is provided with microchannels, and the two ends of the microchannels are respectively connected to the liquid inlet distribution chamber and the liquid outlet collection chamber.

3. A server circulating heat dissipation temperature control method, based on the server circulating heat dissipation temperature control system of claim 1, characterized in that, The specific steps include the following: S1: Collect outlet temperature data for each cooling branch, calculate the temperature rise rate of each cooling branch, and calculate the predicted temperature and predicted temperature difference for each cooling branch in the future based on the temperature rise rate. S2: Calculate the absolute temperature closed-loop control quantity of the flow control valve corresponding to each cooling branch, and calculate the balance correction quantity based on the predicted temperature difference of each cooling branch. Calculate the original valve control command based on the absolute temperature closed-loop control quantity and the balance correction quantity. The original valve control command is output to each flow control valve after being limited by amplitude and rate of change. S3: Adjust the circulating pump based on the preset control cycle, extract the opening value of the flow control valve in the preset high opening range, calculate the valve ratio in the preset high opening range, and calculate the circulating pump speed command based on the global maximum predicted temperature difference, the high opening value of the flow control valve in the preset high opening range, and the valve ratio in the preset high opening range, and control the operation of the circulating pump. S4: Analyze the liquid cooling capacity saturation status in real time and determine whether the system meets the cooling mode switching conditions; if the cooling mode switching conditions are met, start the auxiliary air cooling module and perform closed-loop speed regulation on the fan of the auxiliary air cooling module. The auxiliary air-cooling module will be shut down once the system meets the exit conditions.

4. The server circulating heat dissipation temperature control method according to claim 3, characterized in that, In step S1, the original outlet temperature signals of each cooling branch collected by the outlet temperature sensor are subjected to validity verification, median filtering and first-order filtering to obtain outlet side temperature data. No. The temperature rise rate of each cooling branch at the k-th sampling time The calculation is based on temperature data obtained from m consecutive sampling times, and the formula is as follows: ； In the formula, For temperature sampling period, Let be the filtered outlet temperature of the i-th cooling branch at the k-th sampling time. For the first The filtered outlet temperature at each sampling time; Predicted temperature at the k-th sampling time for: ； In the formula, For the amplitude limiting function, This is the upper limit threshold for the rate of temperature rise. For prediction time constant; The predicted temperature difference is the difference between the predicted temperature of the current cooling branch and the target outlet temperature.

5. The server circulating heat dissipation temperature control method according to claim 3, characterized in that, In step S2, the inlet side temperature of the cold plate assembly is collected. Based on the inlet side temperature, outlet side temperature and the pre-calibrated pump speed-valve opening-flow mapping relationship, the heat load of the current cooling branch is calculated. Based on the heat load, the feedforward flow demand is calculated, the feedforward flow demand is converted into the bias of the flow control valve, and then superimposed on the absolute temperature closed-loop control quantity.

6. The server circulating heat dissipation temperature control method according to claim 3, characterized in that, In step S2, the calculation process for the balance correction is as follows: Calculate the weighted average of the predicted temperature differences for all effective cooling branches to obtain the average predicted temperature difference; The difference between the predicted temperature difference of each cooling branch and the average predicted temperature difference is used as the relative thermal deviation. The balance correction amount for each cooling branch is calculated based on the relative thermal deviation. The calculation formula is as follows: ； In the formula, To effectively increase the number of cooling branches, For the first The relative thermal response coefficient of each cooling branch; To balance the gain of the branch, This represents the relative thermal deviation of the i-th cooling branch. The subscript is used for traversal, representing each valid cooling branch involved in the calculation; Represents an effective cooling branch; This represents the relative thermal deviation of the j-th effective cooling branch. Let represent the relative thermal response coefficient of the j-th effective cooling branch; the algebraic sum of the balance corrections for all effective cooling branches is zero. When the command of the flow control valve reaches the upper or lower limit of the opening, bounded projection is performed. The specific method is as follows: First, the flow control valves that have reached the boundary are limited and a residual correction amount is generated; the residual correction amount generated by the limiting is redistributed according to the residual opening margin of the flow control valves that have not reached the opening boundary, until the residual correction amount is less than the set threshold or all flow control valves reach the opening boundary; the opening of the flow control valves of all cooling branches is not lower than the preset minimum safe opening.

7. The server circulating heat dissipation temperature control method according to claim 6, characterized in that, When the temperature of each cooling branch is stable, a self-calibration process for the thermal response coefficient is performed, as follows: Keeping the circulating pump and fan in constant condition, the opening of the flow control valve corresponding to each individual cooling branch is adjusted sequentially. After the branch thermal delay, the original thermal response coefficient of the current cooling branch is calculated. The original thermal response coefficient is the ratio of the steady-state change of the outlet temperature of the cooling branch to the change of the opening of the flow control valve. The original thermal response coefficients of all cooling branches are calculated sequentially using the above method. The median of all original thermal response coefficients is then calculated, and the ratio of the original thermal response coefficient of the current cooling branch to the median of all original thermal response coefficients is taken as the relative thermal response coefficient of the current cooling branch.

8. The server circulating heat dissipation temperature control method according to claim 3, characterized in that, In step S4, the cooling mode switching conditions include:

1. The continuous time during which the highest predicted global temperature difference high pressure reaches the air-cooled start-up temperature difference threshold and the liquid cooling capacity is in a saturated state reaches the preset start-up confirmation time.

2. Any predicted temperature reaches the preset high-temperature protection threshold; 3. The temperature rise rate of any cooling branch exceeds the rapid temperature rise threshold, and the speed of the circulating pump reaches the preset high speed and the opening degree of the flow control valve of the cooling branch reaches the preset high opening degree. If any of the above conditions are met, the auxiliary air-cooling module will be activated; If the global highest predicted temperature difference is lower than the air-cooling exit temperature difference threshold and the liquid cooling capacity remains unsaturated for a period of time that reaches the preset exit confirmation time, then the exit condition is met and the auxiliary air-cooling module is shut down; the air-cooling exit temperature difference threshold is less than the air-cooling start-up temperature difference threshold.

9. The server circulating heat dissipation temperature control method according to claim 3, characterized in that, In step S4, the closed-loop speed control method of the fan is as follows: the difference between the global maximum predicted temperature difference and the preset air-cooled reference temperature difference is used as a variable input to the dead zone function to obtain the fan adjustment error; based on the fan proportional control gain and the fan adjustment error as the proportional adjustment amount, the original speed command is generated by superimposing the minimum fan starting speed, the proportional adjustment amount and the integral term; the original speed command is output to the fan of the auxiliary air-cooled module after integral anti-saturation processing, speed limiting and rate of change limiting; the fan running time each time is not less than the minimum fan running time; the fan shutdown time is not less than the preset minimum shutdown time.

Images