Anti-condensation control method and device for liquid cooling system

Abstract

The application discloses a condensation control method and device of a liquid cooling system, and relates to the technical field of condensation prevention. The condensation point position is calibrated and the condensation control device is installed through the whole link of the liquid cooling system. A thermal inertia response model is constructed for each condensation point position. According to the load scheduling prediction signal, the surface temperature and dew point margin of the condensation point position are outputted. The condensation risk of the condensation point position is judged. According to the number of condensation point positions in the early warning area and the high-risk area, the flow of the branch where the condensation point position is located is adjusted. When the flow adjustment reaches the safety lower limit, the dehumidification unit executes the dehumidification operation according to the dew point margin of the condensation point position. The application realizes the function of accurately sensing the condensation risk by designing the calibrated condensation point position and calculating the dew point margin. A multi-point distributed sensing system is constructed. The data of the parts prone to condensation are collected in real time. The dew point margin is used as the only judgment standard. The accuracy of the condensation detection is improved.

Description

Technical Field

[0001] This invention relates to the field of anti-condensation technology, specifically to an anti-condensation control method and device for a liquid cooling system. Background Technology

[0002] With the rapid iteration and upgrading of high power density computing devices, energy storage converters and power electronic devices, the heat density of devices continues to rise. Liquid cooling systems, with their high heat dissipation efficiency, excellent temperature control accuracy and low noise, are gradually replacing traditional air cooling solutions and becoming the mainstream heat dissipation solution. During the operation of liquid cooling systems, when the temperature of components such as the surface of the cold plate, pipe joints and the inner wall of the electrical cavity is lower than the dew point temperature of the surrounding air, water vapor in the air will condense on the surface of the components to form condensation, which can easily cause faults such as electrical short circuits, metal corrosion and reduced insulation performance.

[0003] Existing anti-condensation control methods for liquid cooling systems rely on global temperature and humidity parameters at the computer room and cabinet level to assess condensation risk. This cannot cover the micro-environmental differences at localized points prone to condensation, such as the edges of cold plates, quick-connect fittings in pipes, and the inner walls of electrical cavities. Furthermore, adjustments are only initiated when temperature and humidity parameters reach the condensation threshold. However, liquid cooling plates and circulating water have significant thermal inertia, and condensation often occurs on component surfaces before adjustments are completed, resulting in insufficient reliability of anti-condensation measures.

[0004] Patent CN114253325B discloses an anti-condensation and temperature safety management system and an anti-condensation method. The patent realizes the monitoring of temperature and humidity in the cabinet equipment through sensors, and further controls the humidity control device through the control circuit board to absorb water molecules in the cabinet equipment, which can effectively prevent water molecules in the cabinet equipment from condensing and forming condensation, and greatly improve the anti-condensation effect of the cabinet equipment.

[0005] The aforementioned patent uses a thermal infrared temperature matrix sensor to collect data on the internal ambient temperature and inner surface temperature of the cabinet equipment, and a humidity sensor to collect humidity data from the polymer humidity-controlling material sheet. The data interface transmits the collected data on the internal ambient temperature, inner surface temperature, and humidity from the polymer humidity-controlling material sheet to the control circuit board. After analysis, the control circuit board controls the alarm components and fan to work in real time. This allows the polymer humidity-controlling material sheet to absorb water molecules inside the cabinet equipment when humidity is high, improving the anti-condensation effect of the cabinet equipment. It also enables real-time monitoring of the temperature value inside the cabinet equipment, effectively improving the working stability and safety of the cabinet equipment. However, there is room for improvement in the accuracy of condensation risk detection.

[0006] Therefore, this application proposes a method and device for preventing condensation in a liquid cooling system that accurately detects the risk of condensation. Summary of the Invention

[0007] The purpose of this invention is to provide a method and apparatus for preventing condensation in a liquid cooling system, so as to solve the technical problems mentioned in the background art, which rely on global temperature and humidity parameters and have blind spots in condensation detection.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a method for preventing condensation in a liquid cooling system, comprising the following steps:

[0009] S1: Combine thermal simulation with historical fault review to sort out the condensation risk of the entire liquid cooling system, mark the condensation points, deploy miniature integrated sensing units at the condensation points, divide the liquid cooling system into branches according to the internal physical partitions and configure execution units, deploy dehumidification units in high-risk cavities, and establish the mapping relationship between condensation points and their respective branches and between condensation points and their corresponding dehumidification units.

[0010] S2: The miniature integrated sensing unit collects data from the corresponding dew point, calculates the local dew point temperature and dew point margin for each dew point, classifies the condensation risk of the corresponding dew point based on the dew point margin, and performs differentiated control on the dew point.

[0011] S3: Construct an independent thermal inertial response model for each condensation point. Based on the load scheduling prediction signal, the thermal inertial response model outputs the surface temperature and dew point margin of the corresponding condensation point, and the execution unit adjusts the branch flow.

[0012] S4: Based on the condensation risk at the condensation point, calculate the flow adjustment weight for each branch, and the execution unit adjusts the flow of the branch.

[0013] S5: When the adjustment of the execution unit reaches the safety lower limit, the dehumidification unit performs dehumidification operation according to the dew point margin of the condensation point.

[0014] Preferably, S1 specifically comprises:

[0015] S11: By combining thermal simulation and historical fault review, the risk of condensation in the entire liquid cooling system is ranked, and the condensation points of cold plates, pipes and cavities are calibrated.

[0016] S12: A miniature integrated sensor unit is deployed at each condensation point. The data collected by the miniature integrated sensor unit is wirelessly transmitted to the anti-condensation controller.

[0017] S13: Based on the load-temperature test data from the calibration phase, initialize the thermal inertial response model parameters for each condensation point;

[0018] S14: Divide the liquid cooling system into branches according to the internal physical partitions, and install an electric proportional regulating valve on the liquid supply pipe of each branch.

[0019] S15: Dehumidification units are deployed in high-risk enclosed cavities within the liquid cooling system, with each dehumidification unit corresponding to 2-4 adjacent condensation points.

[0020] Preferably, S2 specifically includes:

[0021] S21: The miniature integrated sensing unit collects the ambient temperature, relative humidity, and surface temperature at the corresponding dew point, and calculates the local dew point temperature at the corresponding dew point using the Magnus dew point calculation formula. ;

[0022] S22: According to the formula Calculate the dew point margin independently for each condensation point. ,in The measured surface temperature at this condensation point;

[0023] S23: Determine the risk level of a condensation point based on its dew point margin. When the temperature is above 5℃, the condensation point is in a safe zone with no risk of condensation; when the temperature is below 2℃... When the temperature is ≤5℃, this condensation point is in the warning zone, indicating a potential risk of condensation. When the temperature is below 2℃, this condensation point is in a high-risk area and there is a risk of condensation.

[0024] S24: When the dew point is in the safe zone, the operating parameters of the liquid cooling system remain unchanged. When the dew point is in the warning zone, the dew point margin of the dew point is obtained for 5 sampling cycles. The execution unit adjusts the branch flow rate based on the rate of change of the dew point margin. When the dew point is in the high-risk zone, the execution unit adjusts the branch flow rate with the goal of raising the dew point margin of the dew point to 3°C.

[0025] Preferably, S3 specifically includes:

[0026] S31: Construct an independent thermal inertial response model for each condensation point, inputting the branch flow rate, load power, and initial surface temperature corresponding to the condensation point, and then applying the formula... Calculate the surface temperature at the condensation point at time t. ,in, For time variables, This refers to the equilibrium temperature of the condensation point under stable load operation. The condensation point potential heat time constant, It is a natural constant;

[0027] S32: Obtain the computing power scheduling instructions of the data center in the liquid cooling system, the BMS power instructions of the energy storage system, and the start / stop signals of the converter, and obtain the load power change trend in the next 1-5 minutes;

[0028] S33: Based on the load power change trend, the load input thermal inertia response model is predicted to predict the surface temperature and dew point margin of the condensation point within the next 30s-5min. The execution unit adjusts the branch flow rate according to the dew point margin.

[0029] S34: During normal operation, the surface temperature of the condensation point collected by the micro-integrated sensing unit is used as a reference to correct the thermal inertial response model parameters corresponding to the condensation point.

[0030] Preferably, S4 specifically comprises:

[0031] S41: Divide the liquid cooling secondary side supply pipeline into independent branches according to the internal physical zoning of the liquid cooling system, and install an electric proportional regulating valve in each branch.

[0032] S42: Calculate the flow regulation weight of the corresponding branch based on the number of condensation points in the warning zone and high-risk zone within each branch. For each condensation point in the warning zone within the branch, the flow regulation weight increases by 0.3, and for each condensation point in the high-risk zone, the flow regulation weight increases by 1.0.

[0033] S43: Based on the flow adjustment weight of each branch, for every increase of 0.3 in the flow adjustment weight, the execution unit reduces the corresponding branch flow by 5%.

[0034] Preferably, S5 specifically includes:

[0035] S51: When the flow rate of the branch corresponding to the condensation point is adjusted to the lower safety limit or the temperature of the core component reaches the upper limit, the dehumidification unit corresponding to the condensation point will be started.

[0036] S52: Dehumidifiers located at condensation points in the warning zone operate at 30% of their rated power, while dehumidifiers located at condensation points in the high-risk zone operate at 100% of their rated power.

[0037] S53: After the dew point margin at this condensation point rises above 5°C, the dehumidification unit will be shut off after a 2-minute delay.

[0038] Preferably, the high-risk cavity includes a power module cavity for installing liquid-cooled cold plates, a server chassis cavity for laying liquid-cooled branch pipes and quick-connect connectors, a high-voltage power distribution cavity for liquid-cooled pipes passing through the wall, and an electrical cavity for energy storage converters.

[0039] Preferably, the lower limit of the flow regulation safety of each branch is 60% of the rated flow. The core components include a heating chip, a power module, and high-voltage electrical terminals. The upper limit of the temperature of the heating chip is 80-90℃, the upper limit of the temperature of the power module is 75-85℃, and the upper limit of the temperature of the high-voltage electrical terminals is 80-90℃.

[0040] Preferably, the condensation points for cold plates are the surface of the cold plate adjacent to the printed circuit board, the inlet and outlet interfaces of the cold plate, and the low-temperature area at the edge of the cold plate; the condensation points for pipelines are the quick-connect fittings of branch pipelines, the wall-penetrating pipe sections of cavities, and the bends in pipelines; and the condensation points for cavities are the cold areas on the inner wall of electrical cavities, the periphery of liquid-cooled conductive terminals, and the bottom of the inner wall of sealed cavities.

[0041] Preferably, the anti-condensation control device includes: an anti-condensation controller, a miniature integrated sensing unit, an execution unit, and a dehumidification unit;

[0042] The miniature integrated sensing unit consists of a MEMS micro-environment temperature and humidity sensor and a patch surface temperature sensor. The MEMS micro-environment temperature and humidity sensor is used to collect the air temperature and humidity within 10mm around the condensation point, and the patch surface temperature sensor is used to collect the actual wall temperature of the surface at the condensation point.

[0043] The actuator consists of a branch electric proportional regulating valve, a variable frequency pump, a total flow regulating valve, and a feedback module, and is used for continuous regulation and closed-loop feedback of branch flow.

[0044] The dehumidification unit consists of a miniature semiconductor condensation dehumidification module, a miniature silent centrifugal fan, a condensate discharge structure, and a PWM power regulation drive module, and is used for dehumidifying the microenvironment around the condensation point.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] 1. This invention achieves accurate perception of condensation risk by designing calibrated condensation points and calculating dew point margin for risk assessment. It solves the problems of blind spots in condensation detection, distorted condensation judgment logic, and lag in condensation adjustment. It can construct a multi-point distributed sensing system, collect data from condensation-prone areas in real time, and use dew point margin as the sole criterion for judgment, thereby improving the accuracy and real-time performance of condensation detection.

[0047] 2. This invention achieves directional flow regulation by designing a dynamic flow regulation system that maps condensation points to their respective branches. This solves the problems of poor global flow or water temperature regulation accuracy, inability to handle local condensation risks, and impact on the heat dissipation capacity of the liquid cooling system. By regulating the flow of branches corresponding to condensation points with condensation risks and adjusting the flow regulation range according to the number of condensation points with condensation risks in the branch, the device response speed and anti-condensation reliability are improved.

[0048] 3. This invention, by designing a thermal inertial response model, predicts and adjusts branch flow rates in advance based on changes in equipment load, thereby achieving the function of predicting condensation risk. It solves the problems of thermal inertia causing condensation risk to occur faster than adjustment, blind spots in the detection of local cold spots, and insufficient long-term accuracy of static models. It independently constructs a thermal inertial response model for each condensation point, predicts changes in condensation risk based on changes in equipment load, eliminates condensation risk in advance, and optimizes the corresponding thermal inertial response model with actual measured data of condensation points, thereby improving the reliability of condensation prevention.

[0049] 4. This invention, by designing a dehumidification unit that activates after the branch flow rate regulation reaches the lower limit, achieves the function of fixed-point dehumidification compensation. It solves the problems that liquid cooling regulation cannot completely eliminate the risk of condensation, the high energy consumption of global dehumidification, and the inability of liquid cooling and dehumidification to coordinate. Dehumidification is activated only when the branch flow rate regulation reaches the safe lower limit. The dehumidification unit only covers the microenvironment of the corresponding point, eliminating the protection blind spot of liquid cooling regulation. Flow rate regulation and dehumidification are seamlessly connected, improving the protection effect against the risk of condensation. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the anti-condensation control method of the present invention;

[0051] Figure 2 This is a schematic diagram of the condensation point calibration and device layout of the present invention;

[0052] Figure 3 This is a schematic diagram illustrating the risk assessment of condensation points according to the present invention.

[0053] Figure 4 This is a schematic diagram of the thermal inertial response model prediction of the present invention;

[0054] Figure 5 This is a schematic diagram of the branch flow regulation of the present invention;

[0055] Figure 6 This is a schematic diagram of the operation of the dehumidification unit of the present invention. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Example 1: Please refer to Figure 1 , Figure 2 and Figure 3 A method for preventing condensation in a liquid cooling system, S1 specifically includes:

[0058] S11: By combining thermal simulation and historical fault review, the risk of condensation in the entire liquid cooling system is ranked, and the condensation points of cold plates, pipes and cavities are calibrated.

[0059] S12: A miniature integrated sensor unit is deployed at each condensation point. The data collected by the miniature integrated sensor unit is wirelessly transmitted to the anti-condensation controller.

[0060] S13: Based on the load-temperature test data from the calibration phase, initialize the thermal inertial response model parameters for each condensation point;

[0061] S14: Divide the liquid cooling system into branches according to the internal physical partitions, and install an electric proportional regulating valve on the liquid supply pipe of each branch.

[0062] S15: Dehumidification units are deployed in high-risk enclosed cavities within the liquid cooling system, with each dehumidification unit corresponding to 2-4 adjacent condensation points;

[0063] S2 specifically refers to:

[0064] S21: The miniature integrated sensing unit collects the ambient temperature, relative humidity, and surface temperature at the corresponding dew point, and calculates the local dew point temperature at the corresponding dew point using the Magnus dew point calculation formula. ;

[0065] S22: According to the formula Calculate the dew point margin independently for each condensation point. ,in The measured surface temperature at this condensation point;

[0066] S23: Determine the risk level of a condensation point based on its dew point margin. When the temperature is above 5℃, the condensation point is in a safe zone with no risk of condensation; when the temperature is below 2℃... When the temperature is ≤5℃, this condensation point is in the warning zone, indicating a potential risk of condensation. When the temperature is below 2℃, this condensation point is in a high-risk area and there is a risk of condensation.

[0067] S24: When the dew point is in the safe zone, the operating parameters of the liquid cooling system remain unchanged. When the dew point is in the warning zone, the dew point margin of the dew point is obtained for 5 sampling cycles. The execution unit adjusts the branch flow rate based on the rate of change of the dew point margin. When the dew point is in the high-risk zone, the execution unit adjusts the branch flow rate with the goal of raising the dew point margin of the dew point to 3°C.

[0068] For cold plates, condensation points are the surface adjacent to the cold plate and the printed circuit board, the inlet and outlet interfaces of the cold plate, and the low-temperature area at the edge of the cold plate. For pipes, condensation points are the quick-connect fittings of branch pipes, the pipe sections that pass through the wall in the cavity, and the bends in the pipes. For cavities, condensation points are the cold area on the inner wall of the electrical cavity, the area around the liquid-cooled conductive terminals, and the bottom of the inner wall of the sealed cavity.

[0069] Furthermore, through thermal simulation and historical fault review, the condensation risk of the entire liquid cooling system—including the cold plate, piping, and cavity—was prioritized. The following areas were calibrated: the surface adjacent to the printed circuit board, the cold plate inlet / outlet interfaces, the low-temperature zone at the edge of the cold plate, quick-connect fittings in branch piping, wall-penetrating sections of the cavity, pipe bends, cold zones on the inner wall of the electrical cavity, the periphery of the liquid cooling conductive terminals, and the bottom of the inner wall of the sealed cavity. These high-risk condensation points were designated as condensation points. A miniature integrated sensor unit was deployed at each condensation point, and all miniature integrated sensor units were connected to an anti-condensation controller. Branching was then performed according to the internal physical partitions of the liquid cooling system. An execution unit is configured for each branch to establish a mapping relationship between condensation points and their respective branches. When the liquid cooling system is in normal safe operation, it operates under rated load, resulting in high surface temperatures on the cold plates and pipes. The anti-condensation controller collects surface temperature and local temperature and humidity data for all condensation points at a frequency of 1Hz. The controller calculates the local dew point temperature and dew point margin for each condensation point in real time. At this point, the dew point margin for all condensation points is greater than 5℃, placing it within the safe zone. The anti-condensation controller maintains the rated flow rate and rated supply temperature of the liquid cooling system without triggering adjustments, ensuring maximum heat dissipation efficiency of the liquid cooling system. When the load on the liquid cooling system drops suddenly or the ambient humidity rises, the surface temperature of the cold plate or pipes gradually decreases. If the dew point margin at a certain condensation point drops to the range (2, 5), that condensation point is in the warning zone. The anti-condensation controller extracts the dew point margin for five consecutive sampling cycles at that point, calculates the rate of decrease, and judges the risk development trend. If the rate of decrease exceeds 0.3℃ / s, the anti-condensation controller issues an adjustment command to the execution unit of the corresponding branch at that condensation point, slightly reducing the flow rate of that branch. If, after the flow rate adjustment of the corresponding branch is completed, the dew point margin at that condensation point continues to drop to... Below 2℃, the condensation point is in a high-risk zone with an extremely high risk of condensation. The anti-condensation controller aims to raise the dew point margin of the condensation point to 3℃. The control execution unit adjusts the flow rate of the branch where the condensation point is located and uses a PID algorithm to dynamically correct the flow rate adjustment until the dew point margin of the condensation point rises above 3℃, maintaining the adjustment parameters stable and preventing condensation. When the liquid cooling system load increases, the cold plate temperature rises, or the ambient humidity decreases, the dew point margin of the condensation point rises above 5℃, and the condensation point is in a safe zone. The anti-condensation controller then gradually restores the flow rate of each branch to the rated value.

[0070] Example 2: Please refer to Figure 1 and Figure 5A method for preventing condensation in a liquid cooling system, S4 specifically includes:

[0071] S41: Divide the liquid cooling secondary side supply pipeline into independent branches according to the internal physical zoning of the liquid cooling system, and install an electric proportional regulating valve in each branch.

[0072] S42: Calculate the flow regulation weight of the corresponding branch based on the number of condensation points in the warning zone and high-risk zone within each branch. For each condensation point in the warning zone within the branch, the flow regulation weight increases by 0.3, and for each condensation point in the high-risk zone, the flow regulation weight increases by 1.0.

[0073] S43: Based on the flow adjustment weight of each branch, for every increase of 0.3 in the flow adjustment weight, the execution unit reduces the flow of the corresponding branch by 5%;

[0074] The lower limit of flow regulation for each branch is 60% of the rated flow. The core components include a heating chip, a power module, and high-voltage electrical terminals. The upper limit of the temperature of the heating chip is 80-90℃, the upper limit of the temperature of the power module is 75-85℃, and the upper limit of the temperature of the high-voltage electrical terminals is 80-90℃.

[0075] Furthermore, the liquid cooling secondary side supply pipeline is divided into multiple independent branches according to the internal physical partitions of the liquid cooling system. Each branch corresponds to one server or a group of cold plates. An electric proportional regulating valve is installed on the supply pipeline of each branch to establish a mapping relationship between condensation points and their respective branches, clearly defining the condensation points corresponding to each branch. When the liquid cooling system is in a normal and safe operation phase, the flow regulation weight of all branches is 0, the electric proportional regulating valve maintains 100% of its rated opening, and the branches operate at rated flow to ensure maximum heat dissipation efficiency. When the computing power of a server decreases and the load drops sharply, the temperature of the cold plate in the corresponding branch decreases. Within this branch, there is one condensation point in the warning zone and one condensation point in the high-risk zone. The anti-condensation controller, based on the mapping relationship between condensation points and their respective branches, locates the branch to which the risky condensation point belongs and eliminates the risk of condensation without wind. Interference from the high-risk branch is addressed by calculating its flow regulation weight, which is set to 1.3. Based on this weight, the anti-condensation controller reduces the branch's opening by 20%, resulting in a 20% decrease in flow. When the branch's flow is reduced to 60% of its rated flow, even if condensation points still exist in the warning or high-risk zones, the anti-condensation controller stops further flow regulation to prevent insufficient cooling capacity of the liquid cooling system. If the GPU chip temperature corresponding to this branch rises to 85°C during regulation, flow regulation is immediately stopped to prevent insufficient cooling of core components. Once the liquid cooling system returns to normal operation, the anti-condensation controller adjusts the branch's flow at a rate of 5% increase every 5 seconds to prevent sudden changes in flow from causing pressure fluctuations in the liquid cooling system.

[0076] Example 3: Please refer to Figure 1 and Figure 4 A method for preventing condensation in a liquid cooling system, comprising the following steps:

[0077] S1: Combine thermal simulation with historical fault review to sort out the condensation risk of the entire liquid cooling system, mark the condensation points, deploy miniature integrated sensing units at the condensation points, divide the liquid cooling system into branches according to the internal physical partitions and configure execution units, deploy dehumidification units in high-risk cavities, and establish the mapping relationship between condensation points and their respective branches and between condensation points and their corresponding dehumidification units.

[0078] S2: The miniature integrated sensing unit collects data from the corresponding dew point, calculates the local dew point temperature and dew point margin for each dew point, classifies the condensation risk of the corresponding dew point based on the dew point margin, and performs differentiated control on the dew point.

[0079] S3: Construct an independent thermal inertial response model for each condensation point. Based on the load scheduling prediction signal, the thermal inertial response model outputs the surface temperature and dew point margin of the corresponding condensation point, and the execution unit adjusts the branch flow.

[0080] S4: Based on the condensation risk at the condensation point, calculate the flow adjustment weight for each branch, and the execution unit adjusts the flow of the branch.

[0081] S5: When the adjustment of the execution unit reaches the safety lower limit, the dehumidification unit performs dehumidification operation according to the dew point margin of the condensation point.

[0082] S3 specifically refers to:

[0083] S31: Construct an independent thermal inertial response model for each condensation point, inputting the branch flow rate, load power, and initial surface temperature corresponding to the condensation point, and then applying the formula... Calculate the surface temperature at the condensation point at time t. ,in, For time variables, This refers to the equilibrium temperature of the condensation point under stable load operation. The condensation point potential heat time constant, It is a natural constant;

[0084] S32: Obtain the computing power scheduling instructions of the data center in the liquid cooling system, the BMS power instructions of the energy storage system, and the start / stop signals of the converter, and obtain the load power change trend in the next 1-5 minutes;

[0085] S33: Based on the load power change trend, the load input thermal inertia response model is predicted to predict the surface temperature and dew point margin of the condensation point within the next 30s-5min. The execution unit adjusts the branch flow rate according to the dew point margin.

[0086] S34: During normal operation, the surface temperature of the condensation point collected by the micro-integrated sensing unit is used as a reference to correct the thermal inertial response model parameters corresponding to the condensation point.

[0087] Furthermore, after calibrating the condensation points of the liquid cooling system, load-temperature calibration tests were conducted at each point. Under different load power (20%, 50%, 80%, and 100%) and different branch flow rates, complete surface temperature change curves at the condensation points were collected. The thermal inertial response model corresponding to each condensation point was fitted to determine the initial parameters of the equilibrium temperature and thermal time constant. During stable operation of the liquid cooling system, the miniature integrated sensing unit continuously collected real-time surface temperature data at each condensation point at a frequency of 1Hz. The anti-condensation controller would then... The measured temperature is compared with the theoretical temperature output by the thermal inertial response model. Model parameters such as the thermal time constant are corrected using the recursive least squares method to eliminate drift caused by component aging, pipe scaling, and changes in ambient temperature. The anti-condensation controller connects in real-time to the upstream BMS power dispatch system. Upon receiving a predicted command that the converter will reduce its load from 100% full load to 20% after 30 seconds, it immediately extracts core parameters such as the load change amplitude and the time of change. The predicted load power, current branch flow rate, and initial surface temperature of the condensation point are then substituted into the corresponding thermal inertial response model for that condensation point. Based on the model, the surface temperature change curve of the condensation point within the next 30 seconds to 5 minutes is calculated. Combined with the current local microenvironment temperature and humidity, the dew point margin change trend of the condensation point in the future period is calculated simultaneously. If it is predicted that the dew point margin of a certain condensation point will drop below 2°C in the next 40 seconds, the anti-condensation controller issues an adjustment command. Through the mapping relationship between the condensation point and its corresponding branch, the branch corresponding to the condensation point is located, and the execution unit of that branch reduces the branch flow by 15% in advance. This pre-intervention is completed before the load decreases and the cold plate temperature drops, thus offsetting the interaction between the cold plate and the water. Due to thermal inertia delay, after the load on the liquid cooling system decreases, the miniature integrated sensing unit collects the actual temperature of the condensation point in real time to calculate the dew point margin. The predicted value is compared with the measured value, and the deviation data is input into the model correction module to optimize the prediction accuracy under the same operating conditions. When a signal of load recovery is received, the thermal inertia response model predicts that the temperature of the condensation point will continue to rise and the dew point margin will steadily rise to above 5°C. The anti-condensation controller controls the execution unit to gradually increase the branch flow rate to the rated value at a rate of 5% increase every 5 seconds, restoring the normal heat dissipation mode of the liquid cooling system.

[0088] Example 4: Please refer to Figure 1 and Figure 6A method for preventing condensation in a liquid cooling system, comprising the following steps:

[0089] S1: Combine thermal simulation with historical fault review to sort out the condensation risk of the entire liquid cooling system, mark the condensation points, deploy miniature integrated sensing units at the condensation points, divide the liquid cooling system into branches according to the internal physical partitions and configure execution units, deploy dehumidification units in high-risk cavities, and establish the mapping relationship between condensation points and their respective branches and between condensation points and their corresponding dehumidification units.

[0090] S2: The miniature integrated sensing unit collects data from the corresponding dew point, calculates the local dew point temperature and dew point margin for each dew point, classifies the condensation risk of the corresponding dew point based on the dew point margin, and performs differentiated control on the dew point.

[0091] S3: Construct an independent thermal inertial response model for each condensation point. Based on the load scheduling prediction signal, the thermal inertial response model outputs the surface temperature and dew point margin of the corresponding condensation point, and the execution unit adjusts the branch flow.

[0092] S4: Based on the condensation risk at the condensation point, calculate the flow adjustment weight for each branch, and the execution unit adjusts the flow of the branch.

[0093] S5: When the adjustment of the execution unit reaches the safety lower limit, the dehumidification unit performs dehumidification operation according to the dew point margin of the condensation point.

[0094] S5 specifically refers to:

[0095] S51: When the flow rate of the branch corresponding to the condensation point is adjusted to the lower safety limit or the temperature of the core component reaches the upper limit, the dehumidification unit corresponding to the condensation point will be started.

[0096] S52: Dehumidifiers located at condensation points in the warning zone operate at 30% of their rated power, while dehumidifiers located at condensation points in the high-risk zone operate at 100% of their rated power.

[0097] S53: After the dew point margin at this condensation point rises to above 5°C, the dehumidification unit will be shut off after a 2-minute delay.

[0098] High-risk cavities include power module cavities where liquid-cooled cold plates are installed, server chassis cavities where liquid-cooled branch pipes and quick-connect fittings are laid, high-voltage power distribution cavities where liquid-cooled pipes pass through walls, and electrical cavities for energy storage converters.

[0099] The lower limit of flow regulation for each branch is 60% of the rated flow. The core components include a heating chip, a power module, and high-voltage electrical terminals. The upper limit of the temperature of the heating chip is 80-90℃, the upper limit of the temperature of the power module is 75-85℃, and the upper limit of the temperature of the high-voltage electrical terminals is 80-90℃.

[0100] Furthermore, after condensation point calibration, high-risk enclosed cavities such as server chassis and power module cavities were identified. Dehumidification units were deployed in each high-risk enclosed cavity, with each unit covering 2-4 adjacent condensation points within the cavity. A mapping relationship was established between condensation points and corresponding dehumidification units. During normal operation of the liquid cooling system, if the flow rate adjustment on the liquid cooling side does not reach the lower safety limit and the core component temperature does not reach the upper limit, the dehumidification unit remains in standby mode. The anti-condensation controller adjusts the branch flow rate based on the risk assessment of the condensation points. When the ambient temperature continues to rise and the server is operating under high load, the branch flow rate corresponding to condensation points in the warning and high-risk zones is continuously adjusted. If the corresponding branch flow rate is reduced to 60% of the rated flow rate and the condensation point is still in the warning or high-risk zone, it is determined that there is no further adjustment margin on the liquid cooling side, and condensation occurs. When the dehumidification unit corresponding to the condensation point is activated, and the GPU chip temperature corresponding to the branch rises to 85℃, the flow rate can no longer be reduced, the anti-condensation controller controls the dehumidification unit corresponding to the condensation point to start. After the dehumidification unit starts, if the condensation point is in the warning zone, the dehumidification unit operates at a low load of 30% of the rated power to slowly reduce the local humidity of the cavity. If the condensation point is in the high-risk zone, the dehumidification unit operates at full load of 100% of the rated power to quickly reduce the relative humidity of the microenvironment around the condensation point, increase the dew point margin, and quickly eliminate the risk of condensation. When the condensation point rises from the high-risk zone back to the warning zone, the anti-condensation controller reduces the dehumidification unit from full power to 30% of the rated power. When the dew point margin of the condensation point rises back to above 5℃, the anti-condensation controller controls the dehumidification unit to continue running for 2 minutes and then shuts it off to avoid the risk of repeated humidity rebound in the cavity.

[0101] Example 5: Please refer to Figure 1 A method for preventing condensation in a liquid cooling system, comprising the following steps:

[0102] S1: Combine thermal simulation with historical fault review to sort out the condensation risk of the entire liquid cooling system, mark the condensation points, deploy miniature integrated sensing units at the condensation points, divide the liquid cooling system into branches according to the internal physical partitions and configure execution units, deploy dehumidification units in high-risk cavities, and establish the mapping relationship between condensation points and their respective branches and between condensation points and their corresponding dehumidification units.

[0103] S2: The miniature integrated sensing unit collects data from the corresponding dew point, calculates the local dew point temperature and dew point margin for each dew point, classifies the condensation risk of the corresponding dew point based on the dew point margin, and performs differentiated control on the dew point.

[0104] S3: Construct an independent thermal inertial response model for each condensation point. Based on the load scheduling prediction signal, the thermal inertial response model outputs the surface temperature and dew point margin of the corresponding condensation point, and the execution unit adjusts the branch flow.

[0105] S4: Based on the condensation risk at the condensation point, calculate the flow adjustment weight for each branch, and the execution unit adjusts the flow of the branch.

[0106] S5: When the adjustment of the execution unit reaches the safety lower limit, the dehumidification unit performs dehumidification operation according to the dew point margin of the condensation point.

[0107] The lower limit of flow regulation for each branch is 60% of the rated flow. The core components include a heating chip, a power module, and high-voltage electrical terminals. The upper limit of the temperature of the heating chip is 80-90℃, the upper limit of the temperature of the power module is 75-85℃, and the upper limit of the temperature of the high-voltage electrical terminals is 80-90℃.

[0108] Furthermore, when the dew point margin at all condensation points is higher than 5°C, the anti-condensation controller maintains the rated flow rate of all branches, the dehumidification unit is in standby mode, and the liquid cooling system operates at the minimum energy consumption for basic heat dissipation. When a condensation point is in the warning zone, the anti-condensation controller only processes the branch containing that condensation point based on the mapping relationship between the condensation point and its corresponding branch, calculates the flow rate adjustment weight of that branch, slightly adjusts the flow rate of that branch, and sets a lower limit for branch flow rate adjustment to avoid the liquid cooling system increasing its cooling capacity and generating secondary energy consumption after the chip overheats. When a condensation point is in the high-risk zone, if the flow rate adjustment of the branch corresponding to that condensation point has not reached the lower limit and the core component temperature has not reached the upper limit, the anti-condensation controller still adjusts the flow rate of that branch based on the flow rate adjustment weight, keeps the dehumidification unit in standby mode, and does not increase energy consumption. The dehumidification unit is only activated when the branch flow rate is adjusted to the safe lower limit or the core component temperature reaches the upper temperature limit to reduce energy consumption.

[0109] Working principle: The miniature integrated sensing unit collects surface temperature and local temperature and humidity data of the condensation point. The anti-condensation controller calculates the local dew point temperature and dew point margin of each condensation point in real time. If the dew point margin of the condensation point is greater than 5℃, the condensation point is in the safe zone and no flow regulation is triggered. If the dew point margin of the condensation point is in the range of (2, 5], the condensation point is in the warning zone. The dew point margin decrease rate of the condensation point is calculated. If the dew point margin decrease rate exceeds 0.3℃ / s, the execution unit slightly reduces the flow rate of the corresponding branch of the condensation point. If the dew point margin of the condensation point is less than 2℃, the condensation point is in the high-risk zone. With the goal of raising the dew point margin of the condensation point to 3℃, the execution unit adjusts the flow rate of the corresponding branch.

[0110] The anti-condensation controller receives the equipment load scheduling prediction signal, inputs the load change into the thermal inertial response model, calculates the surface temperature and dew point margin of the condensation point, and adjusts the flow rate of the corresponding branch based on the dew point margin risk judgment of the condensation point.

[0111] The flow regulation weight is calculated based on the number of condensation points in the warning zone and high-risk zone within the branch. For each condensation point in the warning zone, the flow regulation weight increases by 0.3; for each condensation point in the high-risk zone, the flow regulation weight increases by 1.0. For every 0.3 increase in the flow regulation weight, the execution unit reduces the flow of that branch by 5%, with a safe lower limit of 60% for the flow reduction.

[0112] If the flow rate is adjusted to the lower limit or the temperature of the core component reaches the upper limit, the anti-condensation controller starts the dehumidification unit. It dehumidifies the condensation points in the warning zone at 30% of the rated power and the condensation points in the high-risk zone at 100% of the rated power. After the dew point margin of the condensation point rises to above 5°C, the dehumidification unit continues to process for 2 minutes and then shuts off.

[0113] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. A method for preventing condensation control in a liquid cooling system, characterized in that: Includes the following steps: S1: Combine thermal simulation with historical fault review to sort out the condensation risk of the entire liquid cooling system, mark the condensation points, deploy miniature integrated sensing units at the condensation points, divide the liquid cooling system into branches according to the internal physical partitions and configure execution units, deploy dehumidification units in high-risk cavities, and establish the mapping relationship between condensation points and their respective branches and between condensation points and their corresponding dehumidification units. S2: The miniature integrated sensing unit collects data from the corresponding dew point, calculates the local dew point temperature and dew point margin for each dew point, classifies the condensation risk of the corresponding dew point based on the dew point margin, and performs differentiated control on the dew point. S3: Construct an independent thermal inertial response model for each condensation point. Based on the load scheduling prediction signal, the thermal inertial response model outputs the surface temperature and dew point margin of the corresponding condensation point, and the execution unit adjusts the branch flow. S4: Based on the condensation risk at the condensation point, calculate the flow adjustment weight for each branch, and the execution unit adjusts the flow of the branch. S5: When the adjustment of the execution unit reaches the safety lower limit, the dehumidification unit performs dehumidification operation according to the dew point margin of the condensation point.

2. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: Specifically, S1 is: S11: By combining thermal simulation and historical fault review, the risk of condensation in the entire liquid cooling system is ranked, and the condensation points of cold plates, pipes and cavities are calibrated. S12: A miniature integrated sensor unit is deployed at each condensation point. The data collected by the miniature integrated sensor unit is wirelessly transmitted to the anti-condensation controller. S13: Based on the load-temperature test data from the calibration phase, initialize the thermal inertial response model parameters for each condensation point; S14: Divide the liquid cooling system into branches according to the internal physical partitions, and install an electric proportional regulating valve on the liquid supply pipe of each branch. S15: Dehumidification units are deployed in high-risk enclosed cavities within the liquid cooling system, with each dehumidification unit corresponding to 2-4 adjacent condensation points.

3. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: Specifically, S2 is: S21: The miniature integrated sensing unit collects the ambient temperature, relative humidity, and surface temperature at the corresponding dew point, and calculates the local dew point temperature at the corresponding dew point using the Magnus dew point calculation formula. ; S22: According to the formula Calculate the dew point margin independently for each condensation point. ,in The measured surface temperature at this condensation point; S23: Determine the risk level of a condensation point based on its dew point margin. When the temperature is above 5℃, the condensation point is in a safe zone with no risk of condensation; when the temperature is below 2℃... When the temperature is ≤5℃, this condensation point is in the warning zone, indicating a potential risk of condensation. When the temperature is below 2℃, this condensation point is in a high-risk area and there is a risk of condensation. S24: When the dew point is in the safe zone, the operating parameters of the liquid cooling system remain unchanged. When the dew point is in the warning zone, the dew point margin of the dew point is obtained for 5 sampling cycles. The execution unit adjusts the branch flow rate based on the rate of change of the dew point margin. When the dew point is in the high-risk zone, the execution unit adjusts the branch flow rate with the goal of raising the dew point margin of the dew point to 3°C.

4. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: Specifically, S3 is: S31: Construct an independent thermal inertial response model for each condensation point, inputting the branch flow rate, load power, and initial surface temperature corresponding to the condensation point, and then applying the formula... Calculate the surface temperature at the condensation point at time t. ,in, For time variables, This refers to the equilibrium temperature of the condensation point under stable load operation. The condensation point potential heat time constant, It is a natural constant; S32: Obtain the computing power scheduling instructions of the data center in the liquid cooling system, the BMS power instructions of the energy storage system, and the start / stop signals of the converter, and obtain the load power change trend in the next 1-5 minutes; S33: Based on the load power change trend, the load input thermal inertia response model is predicted to predict the surface temperature and dew point margin of the condensation point within the next 30s-5min. The execution unit adjusts the branch flow rate according to the dew point margin. S34: During normal operation, the surface temperature of the condensation point collected by the micro-integrated sensing unit is used as a reference to correct the thermal inertial response model parameters corresponding to the condensation point.

5. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: Specifically, S4 is: S41: Divide the liquid cooling secondary side supply pipeline into independent branches according to the internal physical zoning of the liquid cooling system, and install an electric proportional regulating valve in each branch. S42: Calculate the flow regulation weight of the corresponding branch based on the number of condensation points in the warning zone and high-risk zone within each branch. For each condensation point in the warning zone within the branch, the flow regulation weight increases by 0.3, and for each condensation point in the high-risk zone, the flow regulation weight increases by 1.

0. S43: Based on the flow adjustment weight of each branch, for every increase of 0.3 in the flow adjustment weight, the execution unit reduces the corresponding branch flow by 5%.

6. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: Specifically, S5 is: S51: When the flow rate of the branch corresponding to the condensation point is adjusted to the lower safety limit or the temperature of the core component reaches the upper limit, the dehumidification unit corresponding to the condensation point will be started. S52: Dehumidifiers located at condensation points in the warning zone operate at 30% of their rated power, while dehumidifiers located at condensation points in the high-risk zone operate at 100% of their rated power. S53: After the dew point margin at this condensation point rises above 5°C, the dehumidification unit will be shut off after a 2-minute delay.

7. The anti-condensation control method for a liquid cooling system according to claim 1, characterized in that: The high-risk cavities include power module cavities with liquid-cooled cold plates, server chassis cavities with liquid-cooled branch pipes and quick-connect fittings, high-voltage power distribution cavities with liquid-cooled pipes passing through walls, and electrical cavities for energy storage converters.

8. The anti-condensation control method for a liquid cooling system according to claim 6, characterized in that: The lower limit of the flow regulation safety for each branch is 60% of the rated flow. The core components include a heating chip, a power module, and high-voltage electrical terminals. The upper limit of the temperature of the heating chip is 80-90℃, the upper limit of the temperature of the power module is 75-85℃, and the upper limit of the temperature of the high-voltage electrical terminals is 80-90℃.

9. The anti-condensation control method for a liquid cooling system according to claim 2, characterized in that: The condensation points for cold plates are the surfaces adjacent to the printed circuit board, the inlet and outlet interfaces of the cold plates, and the low-temperature areas at the edges of the cold plates. The condensation points for pipelines are the quick-connect fittings of branch pipelines, the wall-penetrating sections of the cavities, and the bends in the pipelines. The condensation points for cavities are the cold areas on the inner walls of electrical cavities, the periphery of liquid-cooled conductive terminals, and the bottom of the inner walls of sealed cavities.

10. A condensation control device for a liquid cooling system, applicable to the condensation control method for a liquid cooling system as described in any one of claims 1-9, characterized in that: The anti-condensation control device includes: an anti-condensation controller, a miniature integrated sensing unit, an execution unit, and a dehumidification unit; The miniature integrated sensing unit consists of a MEMS micro-environment temperature and humidity sensor and a patch surface temperature sensor. The MEMS micro-environment temperature and humidity sensor is used to collect the air temperature and humidity within 10mm around the condensation point, and the patch surface temperature sensor is used to collect the actual wall temperature of the surface at the condensation point. The actuator consists of a branch electric proportional regulating valve, a variable frequency pump, a total flow regulating valve, and a feedback module, and is used for continuous regulation and closed-loop feedback of branch flow. The dehumidification unit consists of a miniature semiconductor condensation dehumidification module, a miniature silent centrifugal fan, a condensate discharge structure, and a PWM power regulation drive module, and is used for dehumidifying the microenvironment around the condensation point.

Images