An AI-based adaptive phase change liquid cooling fault control method and system

Abstract

The present invention relates to the field of adaptive control technology, specifically an AI-based adaptive phase change liquid cooling fault control method and system, comprising the following steps: obtaining coolant temperature, flow rate, and phase change state data; real-time monitoring of the liquid to gel phase change time under different flow rate and temperature conditions; calculating the phase change lag time increments of adjacent cycles; and obtaining phase change lag time change trend analysis results. In the present invention, by real-time monitoring of the phase change time of the coolant under different flow rate and temperature conditions, the heat exchange behavior of the coolant can be more accurately predicted, so that the downward trend of the cooling capacity has the ability to be quantitatively analyzed. Based on the construction and change rate analysis of the heat absorption rate time series, combined with the degree of abnormal fluctuation of the heat absorption rate, the coolant pump speed, flow distribution, and flow path are dynamically adjusted to adapt to the heat dissipation requirements under different working conditions, so that the cooling scheme can be adaptively optimized according to the operating trend.

Description

Technical Field

[0001] The present invention relates to the field of adaptive control technology, and in particular to an AI-based adaptive phase-change liquid cooling fault control method and system. Background Art

[0002] The field of adaptive control technology encompasses control methods and strategies that adjust in real time to changes in a system's dynamic characteristics. The core content involves using algorithms to adjust control parameters to maintain system performance stability in complex environments or under uncertain conditions. Adaptive control is commonly used in a variety of fields, including power systems, robotic control, autonomous driving, and thermal management systems. Its methods include model-referenced adaptive control, artificial intelligence-based adaptive control, and parameter adaptive adjustment. In recent years, with the improvement of computing power and the development of artificial intelligence, adaptive control has gradually been combined with technologies such as deep learning and reinforcement learning, expanding its control capabilities in nonlinear systems and complex environments.

[0003] Among them, the AI-based adaptive phase change liquid cooling fault control method refers to the use of artificial intelligence technology to detect and analyze the faults of the phase change liquid cooling system, and adaptively adjust the cooling strategy. This method mainly identifies and models the heat transfer characteristics, phase change process, and key parameter changes that may lead to cooling failure in the phase change liquid cooling system. By using sensors to collect data such as temperature, flow, and pressure of the cooling system in real time, and combining machine learning models to predict the phase change state of the cooling medium. Based on the prediction results, the operation mode of the liquid cooling system is adaptively adjusted, including adjusting the circulation rate of the coolant, optimizing the working parameters of the heat exchange structure, and dynamically allocating cooling resources to ensure the stable operation of the system under different working conditions.

[0004] Existing fault control systems for phase-change liquid cooling systems lack the ability to dynamically monitor the coolant's phase change process. Relying solely on thresholds for single parameters such as temperature and flow, they struggle to accurately identify potential hysteresis effects during the phase change process, limiting real-time adjustments to the phase change's heat absorption capacity. Due to a lack of tracking of phase change lag time increments, existing solutions struggle to predict cooling capacity decline, resulting in passive adjustments only after cooling capacity declines, impacting heat dissipation stability. Existing technologies primarily rely on static models to monitor heat absorption rates, making it difficult to capture abnormal fluctuations in the coolant's heat absorption capacity. This results in suboptimal heat dissipation efficiency under certain extreme operating conditions. Coolant pump speed, flow distribution, and flow path adjustment methods are relatively fixed, lacking adaptive optimization capabilities for the completeness of the coolant's phase change, resulting in reduced cooling resource utilization under varying operating conditions. Existing technologies struggle to comprehensively analyze the rate of cooling capacity decline, hindering adequate prediction of cooling system failures before they occur, increasing maintenance costs and impacting ongoing equipment operation. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose an AI-based adaptive phase change liquid cooling fault control method.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an AI-based adaptive phase change liquid cooling fault control method, comprising the following steps:

[0007] S1: Acquire coolant temperature, flow rate, and phase change state data, monitor the phase change time from liquid to gel under different flow rate and temperature conditions in real time, calculate the phase change lag time increments of adjacent cycles, and obtain the phase change lag time change trend analysis results;

[0008] S2: Based on the phase change lag time change trend analysis results, calculate the phase change lag time growth rate, heat load change rate, and flow rate dynamic change rate, determine the cooling capacity decrease rate, and obtain a cooling capacity decrease rate evaluation result;

[0009] S3: Based on the cooling capacity decrease rate evaluation result, calculate the coolant phase change heat absorption rate and the heat absorption rate change rate per unit time, determine the degree of heat absorption rate deviation, and obtain the heat absorption rate abnormal fluctuation degree analysis result;

[0010] S4: Based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, and the coolant pump speed, flow distribution and flow path switching are adjusted to obtain a coolant flow rhythm adjustment record;

[0011] S5: Based on the coolant flow rhythm adjustment record, calculate the heat load balance, heat dissipation efficiency change and local temperature rise reduction before and after the adjustment, analyze and judge the stability of the cooling process, and generate an adaptive cooling control plan.

[0012] As a further solution of the present invention, the phase change lag time change trend analysis results include phase change lag time series, phase change lag time increment data, and phase change lag time growth trend; the cooling capacity decrease rate evaluation results include phase change lag time growth rate, heat load change rate, flow rate dynamic change rate, and cooling capacity decrease rate; the heat absorption rate abnormal fluctuation degree analysis results include heat absorption rate time series, heat absorption rate change rate, and heat absorption rate deviation degree analysis results; the coolant flow rhythm adjustment record includes coolant pump speed adjustment record, flow distribution adjustment record, and dynamic flow path switching record; the adaptive cooling control scheme includes heat load balance, heat dissipation efficiency change, local temperature rise reduction amplitude data, and cooling process stability judgment results.

[0013] As a further embodiment of the present invention, the specific steps of obtaining coolant temperature, flow rate, and phase change state data, monitoring the phase change time from liquid to gel under different flow rate and temperature conditions in real time, calculating the phase change lag time increments of adjacent cycles, and obtaining the phase change lag time change trend analysis results are as follows:

[0014] S111: acquiring coolant temperature, flow rate, and phase change state data in the phase change liquid cooling system, monitoring the phase change time from liquid to gel under different flow rate and temperature conditions, recording the phase change lag time series of the coolant under different conditions, and obtaining phase change lag time series data;

[0015] S112: Based on the phase change lag time series data, the formula:

[0016]

[0017] Calculate the phase change lag time increment ΔT φ , output phase change lag time increment data, where T φ (t) represents the phase change lag time of the current cycle, T φ (t-1) represents the phase change lag time of the previous cycle, V i represents the flow rate of the coolant in group i, T i represents the coolant temperature corresponding to group i, n represents the total number of flow rate groups, T φ (j) represents the phase change lag time of group j, represents the mean of all measured phase transition lag times, and m represents the total number of data sets;

[0018] S113: Based on the phase change lag time increment data, analyzing the change pattern of the phase change lag time, determining the change rate and trend characteristics of the phase change lag time, and obtaining a phase change lag time change trend analysis result.

[0019] As a further solution of the present invention, based on the phase change lag time change trend analysis results, the phase change lag time growth rate, the heat load change rate, and the flow rate dynamic change rate are calculated to determine the cooling capacity decrease rate. The specific steps for obtaining the cooling capacity decrease rate evaluation result are as follows:

[0020] S211: Calculating the phase change lag time growth rate based on the phase change lag time change trend analysis result, monitoring the phase change lag time data at different time points, calculating the change amount at adjacent time points, and calculating the phase change lag time growth rate based on the normalized time interval to obtain the phase change lag time growth rate;

[0021] S212: Obtain server thermal power consumption data, calculate the thermal power consumption change value per unit time, and use the formula:

[0022]

[0023] Calculate the heat load change rate R q , where Q i Represents the thermal power consumption at the i-th time point, Q i-1 Represents the thermal power consumption at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, Q j Represents the thermal power consumption data of group j, Represents the average value of all measured thermal power consumption, Q max represents the maximum thermal power consumption value, and M represents the total number of data groups;

[0024] S213: Calculate the dynamic change rate of the flow rate in combination with the coolant flow rate data, compare the phase change lag time growth rate and the heat load change rate, analyze the decrease in cooling capacity, and obtain a cooling capacity decrease rate evaluation result.

[0025] As a further embodiment of the present invention, based on the cooling capacity decrease rate evaluation result, the specific steps of calculating the coolant phase change heat absorption rate and the heat absorption rate change rate per unit time, determining the degree of heat absorption rate deviation, and obtaining the heat absorption rate abnormal fluctuation degree analysis result are as follows:

[0026] S311: Based on the cooling capacity decrease rate evaluation result, obtaining the phase change heat absorption, specific heat capacity, and flow thermal resistance data of the coolant, calculating the coolant phase change heat absorption rate per unit time, and obtaining the coolant phase change heat absorption rate data;

[0027] S312: Based on the coolant phase change heat absorption rate data, construct time series data using the formula:

[0028]

[0029] Calculate the rate of change of heat absorption rate R qφ , where q φ,i represents the coolant heat absorption rate at the i-th time point, q φ,i-1 Represents the coolant heat absorption rate at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, q φ,max represents the maximum heat absorption rate value, represents the mean of all measured heat absorption rates;

[0030] S313: Based on the heat absorption rate change rate, determine the degree to which the heat absorption rate deviates from the optimal heat absorption range, analyze the deviation amplitude and change pattern, and obtain an analysis result of the abnormal fluctuation degree of the heat absorption rate.

[0031] As a further solution of the present invention, based on the analysis results of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, and the coolant pump speed, flow distribution and flow path switching are adjusted to obtain the coolant flow rhythm adjustment record. The specific steps are:

[0032] S411: Based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, the phase change heat of the coolant under different flow rate conditions is monitored, and the phase change heat absorption ratio per unit volume is calculated to obtain coolant phase change completion degree data;

[0033] S412: Based on the coolant phase change completion data, determine whether the current flow rate is too fast or too slow, using the formula:

[0034]

[0035] Calculate the flow velocity deviation D v , where v c Represents the current coolant flow rate, v opt represents the optimal coolant flow rate, q φ,i represents the coolant heat absorption rate at the i-th time point, represents the mean of all measured heat absorption rates, q φ,max represents the maximum heat absorption rate value, and M represents the total number of data sets;

[0036] S413: Based on the flow velocity deviation, adjust the coolant pump speed and flow distribution, perform dynamic flow path switching, record all adjusted flow parameters, and obtain a coolant flow rhythm adjustment record.

[0037] As a further solution of the present invention, based on the coolant flow rhythm adjustment record, the heat load balance, the change in heat dissipation efficiency, and the reduction in local temperature rise before and after the adjustment are calculated, the stability of the cooling process is analyzed and determined, and the specific steps for generating an adaptive cooling control scheme are as follows:

[0038] S511: Based on the coolant flow rhythm adjustment record, calculating the heat load balance degree during the cooling process before and after the adjustment, obtaining heat load distribution data during the cooling process, calculating the heat load difference value of each local area, and performing normalization processing to obtain heat load balance degree data;

[0039] S512: Based on the heat load balance data, calculate the change in heat dissipation efficiency, obtain heat dissipation power data during the cooling process, compare the change in heat dissipation per unit time under different flow rates, calculate the adjustment range of heat dissipation efficiency, and analyze the dynamic change trend of heat dissipation capacity to obtain heat dissipation efficiency change data;

[0040] S513: Based on the heat dissipation efficiency change data, calculate the local temperature rise reduction range, analyze the overall operation trend of the cooling process, and determine the stability of the cooling process. Comprehensively adjust the cooling parameters to obtain an adaptive cooling control solution.

[0041] An AI-based adaptive phase-change liquid cooling fault control system, comprising:

[0042] The cooling parameter monitoring module obtains coolant temperature, flow rate, and phase change state data, calls the embedded sensor node to monitor the phase change time from liquid to gel, records the phase change lag time series, calculates the phase change lag time increments of adjacent cycles, and analyzes the phase change lag time trend analysis results;

[0043] The phase change lag calculation module calculates the phase change lag time growth rate based on the phase change lag time change trend analysis results, calculates the heat load change rate in combination with the server heat power consumption data, and compares the coolant flow rate dynamic change rate to obtain the cooling capacity degradation rate assessment result;

[0044] The cooling capacity evaluation module, based on the cooling capacity decrease rate evaluation result, calls the coolant phase change heat absorption, specific heat capacity, and flow thermal resistance data, calculates the heat absorption rate time series, analyzes the heat absorption rate change rate, and obtains the heat absorption rate abnormal fluctuation degree analysis result;

[0045] The heat absorption rate analysis module calculates the coolant phase change completion degree at the current flow rate based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, determines whether the flow rate is abnormal, adjusts the pump speed, flow rate and path, and obtains the coolant flow rhythm adjustment record;

[0046] The flow rhythm adjustment module calculates the heat load balance, the change in heat dissipation efficiency, and the reduction in local temperature rise before and after the adjustment based on the coolant flow rhythm adjustment record, determines the stability of the cooling process, and generates an adaptive cooling control plan.

[0047] Compared with the prior art, the advantages and positive effects of the present invention are:

[0048] In the present invention, by real-time monitoring of the phase change time of the coolant under different flow rates and temperature conditions and calculating the phase change lag time increment, the microscopic changes in the phase change process can be dynamically identified, and the heat exchange behavior of the coolant can be predicted more accurately. The calculation of the phase change lag time growth rate, the heat load change rate, and the dynamic flow rate change rate is combined to enable quantitative analysis of the downward trend of the cooling capacity, which helps to take active adjustment strategies before the cooling performance declines and reduce the risk of failure. Based on the construction and change rate analysis of the heat absorption rate time series, the nonlinear fluctuations of the cooling process can be identified, thereby ensuring that the heat absorption capacity of the coolant is always in the high-efficiency range. According to the degree of abnormal fluctuation of the heat absorption rate, the coolant pump speed, flow distribution and flow path are dynamically adjusted to adapt to the heat dissipation requirements under different working conditions. Based on the heat load balance, heat dissipation efficiency change and local temperature rise reduction before and after the cooling process adjustment, a cooling stability analysis system for dynamic environments is constructed, so that the cooling scheme can be adaptively optimized according to the operating trend, thereby improving the stability of long-term heat dissipation performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flow chart of the main steps of the present invention;

[0050] Figure 2 This is a flow chart of step S1 of the present invention;

[0051] Figure 3 This is a flow chart of step S2 of the present invention;

[0052] Figure 4 This is a flow chart of step S3 of the present invention;

[0053] Figure 5 This is a flow chart of step S4 of the present invention;

[0054] Figure 6 This is a flow chart of step S5 of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0056] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0057] See also Figure 1 , an AI-based adaptive phase change liquid cooling fault control method, comprising the following steps:

[0058] S1: Acquire the coolant temperature, flow rate, and phase change state data in the phase change liquid cooling system. Use embedded sensor nodes to monitor the phase change time from liquid to gel under different flow rate and temperature conditions in real time. Record the phase change lag time series, calculate the phase change lag time increments of adjacent cycles, and obtain the phase change lag time change trend analysis results.

[0059] S2: Based on the phase change lag time trend analysis results, calculate the phase change lag time growth rate. Combined with the server thermal power consumption data, calculate the heat load change rate. Synchronously combine the coolant flow rate data to calculate the dynamic flow rate change rate. Compare the phase change lag time growth rate, heat load change rate, and flow rate dynamic change rate, and comprehensively analyze the cooling capacity reduction rate to obtain the cooling capacity reduction rate assessment result.

[0060] S3: Based on the cooling capacity degradation rate assessment results, obtain the coolant phase change heat absorption, specific heat capacity, and flow thermal resistance data, calculate the coolant phase change heat absorption rate per unit time, construct a heat absorption rate time series, calculate the heat absorption rate change rate, determine the degree to which the heat absorption rate deviates from the optimal heat absorption range, and obtain the analysis results of the abnormal fluctuation degree of the heat absorption rate;

[0061] S4: Based on the analysis results of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, and it is determined whether the current flow rate is too fast or too slow. The coolant pump speed and flow distribution are adjusted, and dynamic flow path switching is performed to obtain a coolant flow rhythm adjustment record;

[0062] S5: Based on the coolant flow rhythm adjustment record, calculate the heat load balance, heat dissipation efficiency change, and local temperature rise reduction in the cooling process before and after the adjustment, analyze the overall operating trend of the cooling process, determine the stability of the cooling process, and generate an adaptive cooling control plan.

[0063] The analysis results of the phase change lag time change trend include the phase change lag time series, the phase change lag time increment data, and the phase change lag time growth trend. The cooling capacity decrease rate assessment results include the phase change lag time growth rate, the heat load change rate, the flow velocity dynamic change rate, and the cooling capacity decrease rate. The analysis results of the abnormal fluctuation degree of the heat absorption rate include the heat absorption rate time series, the heat absorption rate change rate, and the heat absorption rate deviation degree analysis results. The coolant flow rhythm adjustment records include the coolant pump speed adjustment records, the flow distribution adjustment records, and the dynamic flow path switching records. The adaptive cooling control scheme includes the heat load balance, the heat dissipation efficiency change, the local temperature rise reduction amplitude data, and the cooling process stability judgment results.

[0064] See also Figure 2 , S1 step is:

[0065] S111: acquiring coolant temperature, flow rate, and phase change state data in the phase change liquid cooling system, monitoring the phase change time from liquid to gel under different flow rate and temperature conditions, recording the phase change lag time series of the coolant under different conditions, and obtaining phase change lag time series data;

[0066] The coolant temperature, flow rate, and phase change state data in the phase change liquid cooling system are obtained, and the phase change behavior of the coolant under different flow rate and temperature conditions is monitored in real time using embedded sensor nodes. The temperature range of the monitoring point is set to -10°C to 80°C, the flow rate range is 0.1m / s to 5m / s, and the measurement data is recorded at a frequency of 10 times per second. In actual applications, the initial temperature of the coolant in a liquid cooling system is set to 25°C and the flow rate is 2m / s. During the cooling process, when the temperature drops to a certain threshold (such as 10°C), the coolant may Entering the supercooling state, when the temperature is further reduced to 5°C, gel structure begins to form in some areas. In order to determine the phase change lag time, the time series from liquid to gel state is recorded. Each phase change moment is determined by the sensor detecting a sudden change in coolant viscosity or a change in optical sensing characteristics. During the experiment, the phase change moments under different flow rates and temperature conditions were recorded. For example, under the conditions of a flow rate of 2m / s and an initial temperature of 25°C, the coolant undergoes a phase change after the temperature drops to 3°C, with a lag time of 15 seconds. All experimental data constitute the phase change lag time series data.

[0067] S112: Based on the phase transition lag time series data, the formula is:

[0068]

[0069] Calculate the phase change lag time increment ΔT φ , output phase change lag time increment data, where T φ (t) represents the phase change lag time of the current cycle, T φ(t-1) represents the phase change lag time of the previous cycle, V i represents the flow rate of the coolant in group i, T i represents the coolant temperature corresponding to group i, n represents the total number of flow rate groups, T φ (j) represents the phase change lag time of group j, represents the mean of all measured phase transition lag times, and m represents the total number of data sets;

[0070] Based on the phase change lag time series data, the phase change lag time increments of adjacent cycles are calculated. The time cycle interval is set to 10 seconds, the phase change moment of each cycle is obtained, and the increment calculation formula is used for calculation.

[0071] If the lag time of a cycle is 15 seconds and the lag time of the previous cycle is 18 seconds, then the time increment is |15-18|=3 seconds. For the calculation of flow rate and temperature, assuming that the average temperature under the five flow rate conditions (0.5, 1.0, 2.0, 3.0, and 4.0 m / s) is (8, 6, 4, 5, and 7°C), the calculation items are:

[0072]

[0073] If the mean phase transition lag time is 10 seconds and the sum of the squared differences of the five sets of data is 80, then the standard deviation term is:

[0074]

[0075] The final calculated phase change lag time increment is 3+12.2-8.94=6.26 seconds.

[0076] S113: Analyzing the change pattern of the phase change lag time based on the phase change lag time increment data, determining the change rate and trend characteristics of the phase change lag time, and obtaining the phase change lag time change trend analysis results;

[0077] Based on the changing trend of phase change lag time, the changing pattern of phase change lag time is analyzed, the changing rate and trend characteristics of phase change lag time are obtained, and the judgment standard is set: if the absolute value of the lag time increment is greater than 5 seconds, it is a drastic change stage, and if it is less than 2 seconds, it is a stable stage. The standard is set based on the time fluctuation range of the phase change process under different flow rates and temperature conditions. According to the lag time data under multiple experimental environments, the phase change lag time increments in different stages are statistically analyzed, and the boundary value of the extreme change of the phase change rate is set to ensure that the setting is applicable to the working conditions of flow rate 0.1m / s to 5m / s and temperature range -10℃ to 80℃. The phase change lag time changes at maximum and minimum flow rates are analyzed, and the experimental data show It shows that when the flow rate is 0.1m / s and the temperature is -5℃, the phase change lag time increment can be as high as 7.2 seconds, while when the flow rate is 5m / s and the temperature is 80℃, the lag time increment can be as low as 0.8 seconds. Therefore, 5 seconds are set as the drastic change limit and 2 seconds as the stable limit, so that this standard covers the phase change trend under all flow rate and temperature conditions. In this experiment, the calculated value of the phase change lag time increment is 6.26 seconds, which meets the judgment standard of the drastic change stage. The lag time increment values ​​of all cycles are plotted into a trend graph and the trend curve is observed. If the lag time gradually decreases, it indicates that the phase change is accelerated. If the lag time increases, it indicates that the phase change of the coolant is greatly affected by the outside world. In this way, the phase change lag time change trend analysis results are obtained.

[0078] See also Figure 3 , step S2 is:

[0079] S211: Calculating the phase change lag time growth rate based on the phase change lag time change trend analysis result, monitoring the phase change lag time data at different time points, calculating the change amount at adjacent time points, and calculating the phase change lag time growth rate based on the normalized time interval to obtain the phase change lag time growth rate;

[0080] Based on the results of the phase change lag time change trend analysis, the phase change lag time growth rate is calculated. First, the phase change lag time data at multiple time points are monitored, and the phase change time data in different time periods are recorded. For example, under different server load conditions, the phase change state change time of the coolant in the pipeline is 12.5s, 13.8s, 15.2s, and 16.5s, respectively, and the time interval is set to 5s. On this basis, the phase change lag time changes at adjacent time points are compared, that is, the phase change time change in each time period is calculated. For example, the phase change lag time changes at adjacent time points are 1.3s, 1.4s, and 1.3s, respectively, and the time interval normalization is used to calculate the phase change lag time growth rate. The normalization calculation uses the ratio of the time change to the time interval for normalization, such as: Finally, a set of phase change lag time growth rates at multiple time points is obtained, and trend calculation is performed based on the set to finally obtain the phase change lag time growth rate.

[0081] S212: Obtain server thermal power consumption data, calculate the thermal power consumption change value per unit time, and use the formula:

[0082]

[0083] Calculate the heat load change rate R q , where Q i Represents the thermal power consumption at the i-th time point, Q i-1 Represents the thermal power consumption at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, Q j Represents the thermal power consumption data of group j, Represents the average value of all measured thermal power consumption, Q max represents the maximum thermal power consumption value, and M represents the total number of data groups;

[0084] Obtain the server thermal power consumption data and calculate the thermal power consumption change rate per unit time based on the thermal power consumption at adjacent time points. As shown in Table 2.1, set the thermal power consumption data at multiple time points and calculate its change value.

[0085] Table 2.1 Server thermal power consumption data table

[0086]

[0087]

[0088] According to the data in Table 2.1, the data is used to calculate the change in heat load at adjacent moments. For example, |125-120|=5W, |130-125|=5W, |138-130|=8W, the total change is 5+5+8=18W, and the total time interval is 5+5+5=15s. The result of the first calculation is:

[0089]

[0090] Calculate the mean Maximum thermal power consumption Q max =138W, calculate the second term:

[0091]

[0092] Calculated:

[0093]

[0094] Finally, the rate of change of heat load is:

[0095] R q =1.2+6.65=7.85;

[0096] The calculated heat load change rate can be used to calculate the subsequent cooling capacity reduction rate.

[0097] S213: Calculating the dynamic change rate of the flow rate based on the coolant flow rate data, comparing the phase change hysteresis time growth rate and the heat load change rate, analyzing the cooling capacity degradation, and obtaining a cooling capacity degradation rate assessment result;

[0098] The dynamic change rate of the flow rate is calculated synchronously with the coolant flow rate data, and the growth rate of the phase change lag time, the heat load change rate, and the dynamic change rate of the flow rate are compared to obtain the cooling capacity decrease rate. As shown in Table 2.2, the coolant flow rate at multiple time points is set and its change rate is calculated.

[0099] Table 2.2 Coolant flow rate change data table

[0100]

[0101] Calculate the dynamic rate of change of flow velocity, such as:

[0102]

[0103] Substitute into numerical calculations:

[0104]

[0105] Calculate the dynamic change rate of flow velocity R v =0.08, then, the phase change lag time growth rate, heat load change rate, and flow rate dynamic change rate are comprehensively compared, such as R φ =0.26, R q =7.85, R v =0.08, and the normalized weight is used to calculate the comprehensive cooling capacity decrease rate:

[0106]

[0107] Finally, the cooling capacity decrease rate evaluation results are obtained.

[0108] See also Figure 4 , S3 steps are:

[0109] S311: Based on the cooling capacity degradation rate evaluation result, obtain the phase change heat absorption, specific heat capacity, and flow thermal resistance data of the coolant, calculate the coolant phase change heat absorption rate per unit time, and obtain the coolant phase change heat absorption rate data;

[0110] Based on the evaluation results of the cooling capacity decrease rate, the phase change heat absorption of the coolant is first obtained. Usually, experimental measurement methods can be used to obtain heat absorption data in different temperature ranges. For example, the phase change heat absorption measured at three typical temperature points of 25℃, 40℃ and 55℃ are 220kJ / kg, 185kJ / kg and 160kJ / kg respectively. The specific heat capacity parameters are obtained. Taking water-based coolant as an example, its specific heat capacity is about 4.2kJ / kg·K. For some special coolants, such as coolants containing nanoparticles, its specific heat capacity may be 3.8kJ / kg·K. According to the flow state of the coolant in the cooling system, its flow thermal resistance is determined. The flow thermal resistance can be calculated by measuring the pressure difference and flow velocity at both ends of the pipe. For example, the flow thermal resistance of a coolant at a speed of 2m / s is measured to be 0.15K·m 2 / W, then based on the above parameters, calculate the coolant phase change heat absorption rate per unit time, using the formula Q = lC p ΔT calculation, where l represents the coolant mass flow rate per unit time, C p represents specific heat capacity, ΔT represents temperature, and taking 0.5 kg / s as the measured value, the heat absorption rate at 40°C is calculated as:

[0111] Q=0.5×4.2×(60-40)=42kJ / s;

[0112] Applying this calculation to different temperature conditions, for example at 25°C:

[0113] Q=0.5×4.2×(60-25)=73.5kJ / s;

[0114] Finally, the coolant phase change heat absorption rate data is obtained.

[0115] S312: Based on the coolant phase change heat absorption rate data, construct time series data using the formula:

[0116]

[0117] Calculate the rate of change of heat absorption rate R qφ , where q φ,i represents the coolant heat absorption rate at the i-th time point, q φ,i-1 Represents the coolant heat absorption rate at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, q φ,max represents the maximum heat absorption rate value, represents the mean of all measured heat absorption rates;

[0118] Based on the coolant phase change heat absorption rate, time series data is constructed to record the heat absorption rate in different time periods. Data is collected at intervals of 10 seconds. For example, the heat absorption rates recorded at 10 seconds, 20 seconds, and 30 seconds are 42 kJ / s, 39 kJ / s, and 35 kJ / s, respectively. They are organized into a time series, and the change in heat absorption rate at adjacent time points is calculated.

[0119] N = 3 (number of data points), q φ,max =42kJ / s (maximum heat absorption rate),

[0120] Calculate each item:

[0121]

[0122] Finally, the rate of change of heat absorption rate is obtained.

[0123] Table 3.1 lists the data on the change of coolant heat absorption rate:

[0124] Table 31 Heat absorption rate time series data

[0125]

[0126]

[0127] As shown in Table 3.1, the heat absorption rate varies greatly in different time periods, especially the decline between 20s and 30s is large. Therefore, it is necessary to further judge the extent to which the heat absorption rate deviates from the optimal heat absorption range.

[0128] S313: Based on the heat absorption rate change rate, determine the degree to which the heat absorption rate deviates from the optimal heat absorption range, analyze the deviation amplitude and change pattern, and obtain the analysis result of the abnormal fluctuation degree of the heat absorption rate;

[0129] Based on the rate of change of heat absorption rate, the degree of heat absorption rate deviation from the optimal heat absorption range is judged. First, the optimal heat absorption range is defined. The setting of this range is based on the heat absorption capacity test of the coolant under different temperature, flow rate and pipe diameter conditions to ensure that the heat absorption rate maintains efficient heat exchange under stable working conditions. The specific test process is as follows: Under the conditions of coolant flow rate of 2m / s, ambient temperature of 25℃ and pipe diameter of 0.02m, the heat absorption rate of the coolant is measured by a heat flow meter, and the heat absorption stability range is calculated based on multiple experimental data. The mean and standard deviation of the heat absorption rate are obtained from 100 sets of measurement data. The collected heat absorption rate data range is 38-52kJ / s, and the mean heat absorption rate is calculated. Calculate the standard deviation of the heat absorption rate σ q =2.6kJ / s.

[0130] Based on the mean and standard deviation, set the optimal range of heat absorption rate:

[0131]

[0132] Therefore, the optimal heat absorption rate range is set to 40-50kJ / s. When the heat absorption rate is lower than 40kJ / s, the cooling efficiency decreases. Calculate the deviation of the current heat absorption rate from the optimal range:

[0133] Δq=q opt,min -q φ,min ;

[0134] in:

[0135] q opt,min =40kJ / s (minimum value in the optimal range);

[0136] q φ,min =35kJ / s (actual minimum heat absorption rate);

[0137] Δq=40-35=5kJ / s;

[0138] This indicates that the heat absorption rate has dropped by 5 kJ / s, which is 12.5% ​​lower than the optimal range. To set the deviation threshold, it is necessary to determine it based on the operating stability of the cooling system and the heat exchange performance of the coolant. The specific method is as follows: Set the average heat absorption rate under normal operating conditions of the system to And calculate its relative standard deviation, that is Calculated:

[0139]

[0140] In order to ensure that the fluctuation of heat absorption rate remains within the normal range and to avoid short-term disturbances affecting the determination of heat absorption capacity, 1.5 times the relative standard deviation is usually used as a reasonable threshold of heat absorption rate deviation, that is,

[0141] δ q =1.5×CV=1.5×5.75%=8.625%;

[0142] Considering the operational safety of the cooling system, the deviation threshold was rounded to 10%. This value indicates that when the heat absorption rate falls below the optimal range by more than 10%, the system enters an inefficient heat exchange state. A value of 12.5% ​​greater than 10% indicates that the system has entered an inefficient heat exchange state, ultimately yielding the analysis results for the degree of abnormal fluctuation in the heat absorption rate.

[0143] See also Figure 5 , step S4 is:

[0144] S411: Based on the analysis results of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, the phase change heat of the coolant under different flow rate conditions is monitored, and the phase change heat absorption ratio per unit volume is calculated to obtain the coolant phase change completion degree data;

[0145] Based on the analysis results of the abnormal fluctuation degree of heat absorption rate, the coolant phase change completion degree under the current flow rate conditions is obtained, the phase change heat absorption ratio of the coolant under different flow rate conditions is calculated, and the time series of phase change heat is established to calculate the phase change heat ratio of unit volume of coolant. In the specific calculation process, the coolant flow rate v c Specific heat capacity C p and the phase change heat absorption q φ , assuming that the specific heat capacity of a test coolant is C p =4.18kJ / (kg·K), flow rate is v c =2.5m / s, the heat absorbed by the phase change is q φ =225kJ / kg, then the mass m of coolant flowing through per unit time can be calculated by the flow velocity and the cross-sectional area of ​​the pipe. Assuming that the cross-sectional area of ​​the pipe is A = 0.002m 2 , then the mass passing through per unit time is:

[0146] m c =ρ·v c ·A·t;

[0147] Assume that the coolant density is ρ = 1000 kg / m 3 , the flow time is t = 1s, then the calculation is:

[0148] m c =1000×2.5×0.002×1=5kg;

[0149] The phase change heat absorption per unit time is:

[0150] Q φ =m c ·q φ =5×225=1125kJ;

[0151] This value is used to determine the current coolant phase change heat absorption ratio and is used in subsequent cooling efficiency calculations to obtain the coolant phase change completion data.

[0152] S412: Based on the coolant phase change completion data, determine whether the current flow rate is too fast or too slow, using the formula:

[0153]

[0154] Calculate the flow velocity deviation D v , where v cRepresents the current coolant flow rate, v opt represents the optimal coolant flow rate, q φ,i represents the coolant heat absorption rate at the i-th time point, represents the mean of all measured heat absorption rates, q φ,max represents the maximum heat absorption rate value, and M represents the total number of data sets;

[0155] Based on the coolant phase change completion data, determine whether the current flow rate is in the too fast or too slow range, calculate the degree of flow rate deviation, and the flow rate judgment needs to be based on the optimal coolant flow rate v opt Perform deviation calculations.

[0156] v opt Depends on the heat load and the heat exchange capacity of the coolant. In the server cooling system, the optimal flow rate is usually v opt =3.0m / s, assuming that the current coolant flow rate in a certain environment is v c =2.5m / s, calculate its velocity deviation:

[0157]

[0158] If the measured heat absorption rate data is q φ =[220, 225, 230, 240] kJ / kg, calculate its mean:

[0159]

[0160] Calculate the standardized deviation:

[0161]

[0162]

[0163] Substituting into the formula:

[0164]

[0165] This value is used to determine the degree of deviation of the coolant flow rate and to assist in adjusting the coolant pump speed and flow distribution to obtain the flow rate deviation.

[0166] S413: Based on the flow velocity deviation, the coolant pump speed and flow distribution are adjusted, dynamic flow path switching is performed, and all adjusted flow parameters are recorded to obtain a coolant flow rhythm adjustment record;

[0167] Based on the flow rate deviation, the coolant pump speed and flow distribution are adjusted, and dynamic flow path switching is performed. During the specific adjustment process, different cooling strategies need to be selected according to different flow rate ranges. If the flow rate deviation D v If it is lower than 0.1, the current coolant flow rate is maintained. If 0.1≤Dv ≤0.3, then the pump speed needs to be adjusted by 5%-10%. v >0.3, dynamic flow path switching is required. The setting basis of the two thresholds of 0.1 and 0.3 is the nonlinear change law of fluid heat transfer efficiency. When the flow rate deviates from the optimal range by less than 10%, its impact on the overall heat transfer efficiency is less than 3%. When it deviates by more than 30%, the heat transfer efficiency drops by more than 15%. Therefore, 0.1 and 0.3 are reasonable as the setting ranges to reduce the impact of uneven heat transfer in the system. For example, in a cooling system, the initial pump speed is set to 1200rpm, and D is calculated. v =0.203, the pump speed needs to be adjusted:

[0168] B′=B×(1+ΔP);

[0169] Where ΔP is the pump speed adjustment ratio, which is 5%, then:

[0170] B′=1200×(1+0.05)=1260rpm;

[0171] Adjust the flow distribution simultaneously. In a multi-path cooling system, the flow adjustment needs to be based on the heat load of the coolant flow path. Assuming the initial flow distribution is [L1, L2, L3] = [0.3, 0.4, 0.3], due to the large flow rate deviation, the flow needs to be redistributed so that the path with a larger heat load receives more flow. For example, after adjustment:

[0172] [L′1, L′2, L′3]=[0.35, 0.45, 0.2];

[0173] Finally, all adjusted flow parameters are recorded to obtain the coolant flow rhythm adjustment record.

[0174] Table 4.1 Coolant parameter values

[0175] parameter symbol Value unit Coolant density ρ 1000 <![CDATA[kg / m 3 ]]> Coolant specific heat capacity <![CDATA[C p ]]> 4.18 kJ / (kg·K) Optimal flow rate <![CDATA[v opt ]]> 3.0 m / s Current flow rate <![CDATA[v c ]]> 2.5 m / s Phase change heat absorption <![CDATA[q φ ]]> 225 kJ / kg Initial pump speed B 1200 rpm Pump speed adjustment ratio ΔP 5％ -

[0176] As shown in Table 4.1, the coolant related parameters used in the embodiment are listed. Combined with the data in the table, further flow rate adjustment calculations can be performed.

[0177] See also Figure 6 , step S5 is:

[0178] S511: Based on the coolant flow rhythm adjustment record, calculating the heat load balance degree during the cooling process before and after the adjustment, obtaining heat load distribution data during the cooling process, calculating the heat load difference value of each local area, and performing normalization processing to obtain heat load balance degree data;

[0179] Based on the coolant flow rhythm adjustment record, the cooling system parameters before and after the adjustment are obtained. First, the heat load distribution of the coolant at different flow rates is monitored, the temperature of each monitoring point is measured, and the heat load data of each area is calculated. For example, if the initial temperature of a certain area is 65°C, the temperature drops to 45°C after cooling, and the flow rate is 0.5L / s, then the heat load of the area can be calculated through the heat capacity. At the same time, the temperature distribution of each area is compared, the local heat load deviation value is calculated, and then normalized to obtain the overall heat load balance. Assuming that the average heat load of a system is 500W, and the heat load deviation of a certain area is ±50W, the calculated balance can indicate whether the area is in a heat load optimization state. Combined with the monitoring point data, the heat load distribution of the entire system is calculated to obtain the heat load balance data.

[0180] S512: Based on the heat load balance data, calculate the change in heat dissipation efficiency, obtain heat dissipation power data during the cooling process, compare the change in heat dissipation per unit time under different flow rates, calculate the adjustment range of heat dissipation efficiency, and analyze the dynamic change trend of heat dissipation capacity to obtain heat dissipation efficiency change data;

[0181] Based on the degree of heat load balance, the heat dissipation power before and after adjustment is obtained, the change in the cooling system's heat dissipation capacity per unit time is calculated, and the heat dissipation at each monitoring point is measured. If the initial heat dissipation of a heat dissipation area is 1200W and is 1350W after adjustment, the change in heat dissipation efficiency can be calculated using the heat dissipation power ratio. Furthermore, by comparing the heat dissipation per unit time under different flow rate conditions, for example, when the flow rate is 1.2m / s, the heat dissipation power increases by 10%, while when the flow rate is 0.8m / s, the heat dissipation power decreases by 5%. The trend of heat dissipation efficiency change is calculated, and the adjustment range of the cooling system's heat dissipation capacity under different working conditions is comprehensively evaluated to obtain the heat dissipation efficiency change data.

[0182] Table 5.1 Heat dissipation efficiency change data table

[0183]

[0184] As shown in Table 5.1, the changes in heat dissipation efficiency at different monitoring points vary with the adjustment of the coolant flow rhythm.

[0185] S513: Based on the heat dissipation efficiency change data, calculate the local temperature rise reduction range, analyze the overall operation trend of the cooling process, determine the stability of the cooling process, comprehensively adjust the cooling parameters, and obtain an adaptive cooling control solution;

[0186] Based on the change in heat dissipation efficiency, the reduction in local temperature rise is calculated, and the temperature changes in each key area before and after adjustment are obtained. Assuming that the temperature in a key area of ​​a certain device is 85°C before adjustment and drops to 72°C after adjustment, the reduction in local temperature rise can be calculated to be 13°C. Further analysis of the temperature change trend in each area determines the overall stability of the cooling process, and comprehensive adjustments are made to the flow rate distribution, coolant flow rate, and circulation strategy to ensure that the system temperature changes are within a stable range. Ultimately, an adaptive cooling control solution is obtained.

[0187] Table 5.2 Temperature Rise Reduction Range Data Table

[0188]

[0189] As shown in Table 5.2, the temperature rise and reduction at each monitoring point are different, indicating that the cooling system needs to adjust the flow rate and flow rate according to the heat dissipation requirements of specific areas to optimize temperature stability.

[0190] An AI-based adaptive phase-change liquid cooling fault control system, comprising:

[0191] The cooling parameter monitoring module obtains coolant temperature, flow rate, and phase change state data, calls the embedded sensor node to monitor the phase change time from liquid to gel, records the phase change lag time series, calculates the phase change lag time increments of adjacent cycles, and analyzes the phase change lag time trend analysis results;

[0192] The phase change lag calculation module calculates the phase change lag time growth rate based on the phase change lag time change trend analysis results. It also calculates the heat load change rate based on the server thermal power consumption data and compares the dynamic change rate of the coolant flow rate to obtain the cooling capacity degradation rate assessment result.

[0193] Based on the cooling capacity degradation rate assessment results, the cooling capacity assessment module calls the coolant phase change heat absorption, specific heat capacity, and flow thermal resistance data to calculate the heat absorption rate time series, analyze the heat absorption rate change rate, and obtain the analysis results of the abnormal fluctuation degree of the heat absorption rate;

[0194] The heat absorption rate analysis module calculates the coolant phase change completion degree at the current flow rate based on the analysis results of the abnormal fluctuation degree of the heat absorption rate, determines whether the flow rate is abnormal, adjusts the pump speed, flow rate and path, and obtains the coolant flow rhythm adjustment record;

[0195] The flow rhythm adjustment module calculates the heat load balance, the change in heat dissipation efficiency, and the reduction in local temperature rise before and after the adjustment based on the coolant flow rhythm adjustment record, determines the stability of the cooling process, and generates an adaptive cooling control plan.

[0196] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An AI-based adaptive phase change liquid cooling fault control method, characterized in that: The following steps are involved: S1: Acquire coolant temperature, flow rate, and phase change state data, monitor the phase change time from liquid to gel under different flow rate and temperature conditions in real time, calculate the phase change lag time increments of adjacent cycles, and obtain the phase change lag time change trend analysis results; S2: Based on the phase change lag time change trend analysis results, calculate the phase change lag time growth rate, heat load change rate, and flow rate dynamic change rate, determine the cooling capacity decrease rate, and obtain a cooling capacity decrease rate evaluation result; S3: Based on the cooling capacity decrease rate evaluation result, calculate the coolant phase change heat absorption rate and the heat absorption rate change rate per unit time, determine the degree of heat absorption rate deviation, and obtain the heat absorption rate abnormal fluctuation degree analysis result; S4: Based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, and the coolant pump speed, flow distribution and flow path switching are adjusted to obtain a coolant flow rhythm adjustment record; S5: Based on the coolant flow rhythm adjustment record, calculate the heat load balance, heat dissipation efficiency change and local temperature rise reduction before and after the adjustment, analyze and judge the stability of the cooling process, and generate an adaptive cooling control plan.

2. The AI-based adaptive phase change liquid cooling fault control method according to claim 1 is characterized in that: The phase change lag time change trend analysis results include phase change lag time series, phase change lag time increment data, and phase change lag time growth trend; the cooling capacity decrease rate assessment results include phase change lag time growth rate, heat load change rate, flow rate dynamic change rate, and cooling capacity decrease rate; the heat absorption rate abnormal fluctuation degree analysis results include heat absorption rate time series, heat absorption rate change rate, and heat absorption rate deviation degree analysis results; the coolant flow rhythm adjustment record includes coolant pump speed adjustment record, flow distribution adjustment record, and dynamic flow path switching record; the adaptive cooling control scheme includes heat load balance, heat dissipation efficiency change, local temperature rise reduction amplitude data, and cooling process stability judgment results.

3. The AI-based adaptive phase change liquid cooling fault control method according to claim 1 is characterized in that: The specific steps for obtaining coolant temperature, flow rate, and phase change state data, monitoring the phase change time from liquid to gel under different flow rate and temperature conditions in real time, calculating the phase change lag time increments of adjacent cycles, and obtaining the phase change lag time change trend analysis results are as follows: S111: acquiring coolant temperature, flow rate, and phase change state data, monitoring the phase change time from liquid to gel under different flow rate and temperature conditions, recording the phase change lag time series of the coolant under different conditions, and obtaining phase change lag time series data; S112: Based on the phase change lag time series data, the formula: Calculate the phase change lag time increment ΔT φ , output phase change lag time increment data, where T φ (t) represents the phase change lag time of the current cycle, T φ (t-1) represents the phase change lag time of the previous cycle, V i represents the flow rate of the coolant in group i, T i represents the coolant temperature corresponding to group i, n represents the total number of flow rate groups, T φ (j) represents the phase change lag time of group j, represents the mean of all measured phase transition lag times, and m represents the total number of data sets; S113: Based on the phase change lag time increment data, analyzing the change pattern of the phase change lag time, determining the change rate and trend characteristics of the phase change lag time, and obtaining a phase change lag time change trend analysis result.

4. The AI-based adaptive phase change liquid cooling fault control method according to claim 1 is characterized in that: Based on the phase change lag time change trend analysis results, the phase change lag time growth rate, the heat load change rate, and the flow rate dynamic change rate are calculated to determine the cooling capacity decrease rate. The specific steps for obtaining the cooling capacity decrease rate evaluation result are as follows: S211: Calculating the phase change lag time growth rate based on the phase change lag time change trend analysis result, monitoring the phase change lag time data at different time points, calculating the change amount at adjacent time points, and calculating the phase change lag time growth rate based on the normalized time interval to obtain the phase change lag time growth rate; S212: Obtain server thermal power consumption data, calculate the thermal power consumption change value per unit time, and use the formula: Calculate the heat load change rate R q , where Q i Represents the thermal power consumption at the i-th time point, Q i-1 Represents the thermal power consumption at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, Q j Represents the thermal power consumption data of group j, Represents the average value of all measured thermal power consumption, Q max represents the maximum thermal power consumption value, and M represents the total number of data groups; S213: Calculate the dynamic change rate of the flow rate in combination with the coolant flow rate data, compare the phase change lag time growth rate and the heat load change rate, analyze the decrease in cooling capacity, and obtain a cooling capacity decrease rate evaluation result.

5. The AI-based adaptive phase change liquid cooling fault control method according to claim 1 is characterized in that: Based on the cooling capacity decrease rate evaluation result, the specific steps of calculating the coolant phase change heat absorption rate and the heat absorption rate change rate per unit time, judging the degree of heat absorption rate deviation, and obtaining the heat absorption rate abnormal fluctuation degree analysis result are as follows: S311: Based on the cooling capacity decrease rate evaluation result, obtaining the phase change heat absorption, specific heat capacity, and flow thermal resistance data of the coolant, calculating the coolant phase change heat absorption rate per unit time, and obtaining the coolant phase change heat absorption rate data; S312: Based on the coolant phase change heat absorption rate data, construct time series data using the formula: Calculate the rate of change of heat absorption rate R qφ , where q φ,i represents the coolant heat absorption rate at the i-th time point, q φ,i-1 Represents the coolant heat absorption rate at the previous time point, Δt i represents the i-th time interval, N represents the total number of time intervals, q φ,max represents the maximum heat absorption rate value, represents the mean of all measured heat absorption rates; S313: Based on the heat absorption rate change rate, determine the degree to which the heat absorption rate deviates from the optimal heat absorption range, analyze the deviation amplitude and change pattern, and obtain an analysis result of the abnormal fluctuation degree of the heat absorption rate.

6. The AI-based adaptive phase change liquid cooling fault control method according to claim 1, characterized in that: Based on the analysis results of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, and the coolant pump speed, flow distribution and flow path switching are adjusted. The specific steps for obtaining the coolant flow rhythm adjustment record are as follows: S411: Based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, the coolant phase change completion degree at the current flow rate is calculated, the phase change heat of the coolant under different flow rate conditions is monitored, and the phase change heat absorption ratio per unit volume is calculated to obtain coolant phase change completion degree data; S412: Based on the coolant phase change completion data, determine whether the current flow rate is too fast or too slow, using the formula: Calculate the flow velocity deviation D v , where v c Represents the current coolant flow rate, v opt represents the optimal coolant flow rate, q φ,i represents the coolant heat absorption rate at the i-th time point, represents the mean of all measured heat absorption rates, q φ,max represents the maximum heat absorption rate value, and M represents the total number of data sets; S413: Based on the flow velocity deviation, adjust the coolant pump speed and flow distribution, perform dynamic flow path switching, record all adjusted flow parameters, and obtain a coolant flow rhythm adjustment record.

7. The AI-based adaptive phase change liquid cooling fault control method according to claim 1 is characterized in that: Based on the coolant flow rhythm adjustment record, the heat load balance, the change in heat dissipation efficiency, and the reduction in local temperature rise before and after the adjustment are calculated, the stability of the cooling process is analyzed and judged, and the specific steps for generating an adaptive cooling control scheme are as follows: S511: Based on the coolant flow rhythm adjustment record, calculating the heat load balance degree during the cooling process before and after the adjustment, obtaining heat load distribution data during the cooling process, calculating the heat load difference value of each local area, and performing normalization processing to obtain heat load balance degree data; S512: Based on the heat load balance data, calculate the change in heat dissipation efficiency, obtain heat dissipation power data during the cooling process, compare the change in heat dissipation per unit time under different flow rates, calculate the adjustment range of heat dissipation efficiency, and analyze the dynamic change trend of heat dissipation capacity to obtain heat dissipation efficiency change data; S513: Based on the heat dissipation efficiency change data, calculate the local temperature rise reduction range, analyze the overall operation trend of the cooling process, and determine the stability of the cooling process. Comprehensively adjust the cooling parameters to obtain an adaptive cooling control solution.

8. An AI-based adaptive phase change liquid cooling fault control system, characterized in that: The system is used to perform the method according to any one of claims 1 to 7, comprising: The cooling parameter monitoring module obtains coolant temperature, flow rate, and phase change state data, calls the embedded sensor node to monitor the phase change time from liquid to gel, records the phase change lag time series, calculates the phase change lag time increments of adjacent cycles, and analyzes the phase change lag time trend analysis results; The phase change lag calculation module calculates the phase change lag time growth rate based on the phase change lag time change trend analysis results, calculates the heat load change rate in combination with the server heat power consumption data, and compares the coolant flow rate dynamic change rate to obtain the cooling capacity degradation rate assessment result; The cooling capacity evaluation module, based on the cooling capacity decrease rate evaluation result, calls the coolant phase change heat absorption, specific heat capacity, and flow thermal resistance data, calculates the heat absorption rate time series, analyzes the heat absorption rate change rate, and obtains the heat absorption rate abnormal fluctuation degree analysis result; The heat absorption rate analysis module calculates the coolant phase change completion degree at the current flow rate based on the analysis result of the abnormal fluctuation degree of the heat absorption rate, determines whether the flow rate is abnormal, adjusts the pump speed, flow rate and path, and obtains the coolant flow rhythm adjustment record; The flow rhythm adjustment module calculates the heat load balance, the change in heat dissipation efficiency, and the reduction in local temperature rise before and after the adjustment based on the coolant flow rhythm adjustment record, determines the stability of the cooling process, and generates an adaptive cooling control plan.

Images