Method and system for intelligent temperature control regulation of single-phase immersion liquid cooling with latent heat release

Abstract

The application discloses a phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control regulation method and system, and relates to the technical field of electronic equipment thermal management. The phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control regulation method and system comprises the following steps: S1, collecting phase change liquid cooling temperature control data, and performing pretreatment on the phase change liquid cooling temperature control data; S2, evaluating the running state of a server, assigning a thermal response level identifier based on the evaluation result, and constructing a thermal response state data set; S3, evaluating the cooling slow-release capability under a temperature control early warning state and a thermal shock state, generating and executing a self-adaptive cooling regulation strategy in stages; and S4, after the execution of the cooling regulation strategy, evaluating the response efficiency of the cooling strategy, generating a cooling strategy optimization instruction, and realizing strategy closed-loop regulation based on the change trend of the evaluation result in a monitoring continuous regulation period. The application solves the problem that the chip temperature abnormally fluctuates due to the response lag of the cooling capacity of the existing liquid cooling system under a transient high-load scene.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for electronic devices, specifically to a method and system for intelligent temperature control and regulation of single-phase immersion liquid cooling with phase change latent heat slow release. Background Technology

[0002] In data centers and high-performance computing environments, with the continuous increase in server computing density and power consumption, traditional air-cooling systems have gradually revealed problems such as low heat dissipation efficiency, large temperature fluctuations, and high power usage effectiveness (PUE), making it difficult to meet the thermal management requirements under transient high load conditions. Liquid cooling technology, due to its high thermal conductivity and low energy consumption, has become the main development direction for improving server thermal management efficiency. In particular, single-phase immersion liquid cooling combined with the application of phase change materials shows promising prospects in terms of improving cooling capacity and optimizing energy efficiency.

[0003] For example, the invention disclosed in publication number CN119718029A is a multi-stage phase change air-cooled coupled server cooling system, belonging to the field of server cooling technology. This invention adopts a cooling and heat dissipation method that combines liquid cooling and air cooling, which can effectively reduce energy loss rate and maximize the utilization of natural cold sources. The thermal conductivity and specific heat capacity of liquids are much higher than those of air, so the liquid cooling system can transfer heat more effectively. At the same time, the liquid cooling system can provide more stable and uniform temperature control, reduce the risk of server failure due to overheating, and thus improve the overall reliability of the data center. This invention also uses phase change technology to improve the heat transfer coefficient, which allows the system to maintain a stable temperature even when operating under high load.

[0004] For example, invention CN112764496B discloses a liquid cooling system and a liquid cooling method for implementing liquid cooling, based on a coolant reservoir for housing electronic heating components and a coolant return tank. The method includes monitoring the temperature T1 of the coolant reservoir; monitoring the core temperature T2 of the heating component operating in the coolant reservoir; controlling the coolant outflow rate of the coolant reservoir and the coolant return rate of the coolant return tank when the difference T between the core temperature T2 and the temperature T1 is less than a first preset temperature and the core temperature T2 is greater than a second preset temperature; and controlling the coolant outflow rate of the coolant reservoir and / or the coolant return rate of the coolant return tank when the difference T between the core temperature T2 and the temperature T1 is greater than a third preset temperature. This avoids excessively high temperatures in the electronic heating components and reasonably controls their power consumption.

[0005] In summary, while existing technologies have improved server liquid cooling efficiency and energy consumption control, most solutions still suffer from the following limitations: a lack of quantitative identification mechanisms for transient thermal shocks, an inability to dynamically assess cooling capacity, an inability to adaptively adjust cooling strategies based on operating conditions, and a lack of closed-loop feedback optimization paths after strategy execution. These issues result in sluggish response and rigid strategies when facing complex load fluctuations, making it difficult to continuously maintain server thermal stability and cooling efficiency.

[0006] Therefore, in order to address the above problems, there is an urgent need for a method and system for intelligent temperature control and regulation of single-phase immersion liquid cooling with phase change latent heat slow release. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a method and system for intelligent temperature control and regulation of single-phase immersion liquid cooling with phase change latent heat slow release, which solves the problem of abnormal temperature fluctuations in chips caused by the lag in cooling capacity response in existing liquid cooling systems under instantaneous high load scenarios.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a method and system for intelligent temperature control of single-phase immersion liquid cooling with phase change latent heat slow release, comprising: S1, collecting temperature control data of a phase change liquid cooling system, and performing timestamp alignment, outlier cleaning, noise suppression, and scale normalization on the temperature control data to obtain pre-processed temperature control data of the phase change liquid cooling system; S2, receiving the pre-processed temperature control data of the phase change liquid cooling system, evaluating the operating status of the server, assigning a thermal response level identifier based on the evaluation results, and constructing a thermal response status dataset; S3, extracting the thermal response status dataset, evaluating the cooling slow release capability under temperature control warning and thermal shock states, generating and executing adaptive cooling adjustment strategies in a graded manner; S4, after the cooling adjustment strategy is executed, evaluating the response efficiency of the cooling strategy, generating cooling strategy optimization instructions, and realizing closed-loop adjustment of the strategy based on monitoring the changing trend of the evaluation results within a continuous adjustment cycle.

[0011] Further, phase change liquid cooling temperature control data is collected, and timestamp alignment, outlier cleaning, noise suppression, and scale normalization are performed on the preprocessed phase change liquid cooling temperature control data. The specific steps to obtain the preprocessed phase change liquid cooling temperature control data are as follows: Phase change liquid cooling temperature control data is collected, including CPU temperature, coolant inlet temperature, coolant outlet temperature, coolant flow rate, coolant density, coolant specific heat capacity, coolant flow rate, phase change material temperature, ambient temperature, and server power consumption. Time series processing of the phase change liquid cooling temperature control data is performed using a time-series correction method based on time tag consistency to remove discontinuous data caused by collection interruptions and duplicate recordings. The phase change liquid cooling temperature control data is smoothed using a sliding window averaging method to eliminate local fluctuations caused by control switching and environmental disturbances. Incremental detection of the phase change liquid cooling temperature control data is performed using a filtering method based on first-order incremental changes to identify and remove data points with abnormal change rates. Scale transformation of the phase change liquid cooling temperature control data is performed using a classification normalization method to unify different physical quantities into a standard range.

[0012] Further, the specific steps for receiving preprocessed phase change liquid cooling temperature control data and evaluating the server's operating status are as follows: The preprocessed phase change liquid cooling temperature control data is received and sequentially filled into a fixed-length sliding window according to the sampling time sequence. In each sliding window, the CPU temperature sequence, server power consumption sequence, coolant inlet temperature sequence, coolant outlet temperature sequence, and coolant flow rate sequence are extracted to evaluate the server's operating status. The CPU temperature is differentiated with respect to time to obtain the temperature change rate; the server power consumption is differentiated with respect to time to obtain the power consumption change rate; the temperature change rate and the power consumption change rate are multiplied together to obtain the heat source change. The following terms are used to calculate the thermal shock driving factor: The phase change material temperature is subtracted from the CPU temperature, the square of the difference is added to a minimum term, and this becomes the thermal buffer suppression term. The heat source change term is divided by the thermal buffer suppression term to obtain the thermal shock driving factor. The coolant outlet temperature is subtracted from the coolant inlet temperature, and the resulting temperature difference is divided by the product of the coolant flow rate, coolant density, and coolant specific heat capacity to obtain the heat dissipation capacity per unit volume term. The absolute value of the derivative of the coolant flow rate with respect to time is used as the flow rate disturbance term. The heat dissipation capacity per unit volume term and the flow rate disturbance term are added together to obtain the cooling response factor. The thermal shock driving factor and the cooling response factor are multiplied together to obtain the thermal response evaluation value.

[0013] Further, the specific steps for assigning thermal response level identifiers based on the evaluation results and constructing a thermal response status dataset are as follows: The thermal response evaluation value and the thermal response threshold are compared in real time. When the thermal response evaluation value is less than or equal to the Level 1 thermal response threshold, the server is determined to be in a steady-state operation, and the thermal response level identifier is assigned a value of 0. When the thermal response evaluation value is greater than the Level 1 thermal response threshold but less than the Level 2 thermal response threshold, the server is determined to be in a temperature control warning state, and the thermal response level identifier is assigned a value of 1. When the thermal response evaluation value is greater than or equal to the Level 2 thermal response threshold, the server is determined to be in a thermal shock state, and the thermal response level identifier is assigned a value of 2. The thermal response evaluation value and the thermal response level identifier are stored in a structured manner to construct the thermal response status dataset.

[0014] Further, the specific steps for extracting the thermal response state dataset and evaluating the cooling slow-release capability under temperature control warning and thermal shock conditions are as follows: Extract the thermal response state dataset and real-time phase change liquid cooling temperature control data; subtract the coolant inlet temperature from the coolant outlet temperature, divide the resulting temperature difference by the coolant flow rate to obtain the cooling temperature difference term; add one to the cooling temperature difference term and multiply by the thermal response evaluation value to obtain the thermal response enhancement term; subtract the CPU temperature from the phase change material temperature, square the resulting difference, and add a minimum term to obtain the thermal buffer suppression term; divide the ambient temperature by the coolant inlet temperature and add a minimum term, add one to the resulting ratio to obtain the environmental disturbance correction term; multiply the server power consumption, thermal buffer suppression term, and environmental disturbance correction term together to obtain the cooling load adjustment term; divide the thermal response enhancement term by the cooling load adjustment term to obtain the slow-release adaptation evaluation value.

[0015] Further, the specific steps for generating and executing the adaptive cooling regulation strategy in stages are as follows: Based on the combined thermal response level identifier and the slow-release adaptation evaluation value, an adaptive cooling regulation strategy is generated: If the thermal response level identifier is 2 and the slow-release adaptation evaluation value is greater than the slow-release regulation threshold, it is marked as high risk of slow-release hysteresis. The controller issues a liquid cooling pump acceleration command to increase the liquid cooling pump speed to the high speed range and sets the phase change material heat absorption trigger temperature to the minimum activation value; If the thermal response level identifier is 1 and the slow-release adaptation evaluation value is greater than the slow-release regulation threshold, it is marked as the slow-release capability boundary. The controller issues a liquid cooling pump acceleration command to increase the liquid cooling pump speed to the medium speed range and sets the phase change material 3 state to heat absorption preparation; In other cases, it is marked as stable cooling response, and the current cooling regulation strategy remains unchanged; Based on the cooling regulation strategy, the liquid cooling pump speed adjustment and phase change material 3 heat absorption configuration are executed to complete the cooling execution control.

[0016] Furthermore, after the cooling adjustment strategy is executed, the specific steps for evaluating the response effectiveness of the cooling strategy are as follows: Define the complete process of a cooling adjustment strategy from its initial issuance to the next strategy update as an adjustment cycle; after each cooling adjustment strategy is executed, extract the strategy issuance timestamp and the timestamp of the first collected thermal response evaluation value after adjustment to calculate the strategy response duration; extract the thermal response evaluation values ​​before and after adjustment; add one to the strategy response duration and take the logarithm to the base 10 to obtain the response lag adjustment term; subtract the thermal response evaluation value after strategy execution from the thermal response evaluation value before strategy execution, take the absolute value of the difference, and divide it by the thermal response evaluation value before strategy execution to obtain the thermal response change correction term; multiply the response lag adjustment term by the thermal response change correction term to obtain the cooling feedback quality evaluation value.

[0017] Furthermore, the specific steps for generating cooling strategy optimization instructions are as follows: Real-time comparison of the cooling feedback quality assessment value and the feedback quality threshold to generate cooling strategy optimization instructions: When the cooling feedback quality assessment value is less than the feedback quality threshold, it is determined that the current cooling strategy adjustment efficiency is insufficient, triggering a strategy enhancement mechanism to increase the operating speed of the liquid cooling pump, shorten the adjustment cycle, and increase the heat absorption activation sensitivity of the phase change material 3; When the cooling feedback quality assessment value is greater than or equal to the feedback quality threshold, it is determined that the current cooling strategy adjustment effect is good, maintaining the current strategy and extending the adjustment cycle.

[0018] Furthermore, based on the changing trend of the evaluation results within the continuous adjustment cycle, the specific steps for achieving closed-loop adjustment of the strategy are as follows: Monitor the changing trend of the cooling feedback quality evaluation value within the continuous adjustment cycle: If the continuous fixed-number adjustment cycle is lower than the feedback quality threshold, reconstruct the cooling strategy configuration, adjust the pump control logic, and reset the phase change material heat absorption trigger temperature; If the continuous fixed-number adjustment cycle is higher than the feedback quality threshold, trigger the strategy energy-saving mechanism to reduce the operating frequency of the liquid cooling pump and the activation frequency of the phase change material.

[0019] The second aspect of this invention provides a phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control system, including a temperature control data preprocessing module, a thermal shock identification and judgment module, a cooling strategy generation and execution module, and a cooling strategy closed-loop optimization module. The temperature control data preprocessing module collects phase change liquid cooling temperature control data and performs timestamp alignment, outlier cleaning, noise suppression, and scale normalization on the data to obtain preprocessed phase change liquid cooling temperature control data. The thermal shock identification and judgment module receives the preprocessed phase change liquid cooling temperature control data, evaluates the server's operating status, assigns a thermal response level identifier based on the evaluation results, and constructs a thermal response state dataset. The cooling strategy generation and execution module extracts the thermal response state dataset, evaluates the cooling slow-release capability under temperature control warning and thermal shock states, and generates and executes adaptive cooling adjustment strategies in a tiered manner. The cooling strategy closed-loop optimization module evaluates the response performance of the cooling adjustment strategy after execution, generates cooling strategy optimization instructions, and realizes closed-loop strategy adjustment based on the changing trend of the evaluation results within a continuous adjustment cycle.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) This phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control method and system, by constructing a thermal response evaluation model integrating multiple factors such as CPU temperature change rate, server power consumption change rate, coolant temperature difference and flow rate disturbance, realizes real-time quantification of thermal shock intensity and cooling response capability during server operation, which helps to accurately identify temperature control warning state and thermal shock state. The model is suitable for temperature control safety monitoring in high heat density and dynamic load scenarios, and has good real-time performance and adaptability.

[0023] (2) The intelligent temperature control method and system for single-phase immersion liquid cooling with slow-release latent heat of phase change achieves adaptive dynamic generation and precise execution of the cooling strategy by setting a graded adjustment strategy based on thermal response level and slow-release adaptation assessment value, and jointly controlling the liquid cooling pump speed and the heat absorption state of the phase change material. This adjustment mechanism can automatically adjust the cooling intensity according to the thermal state of the system, improving the adjustment flexibility and environmental adaptability of the liquid cooling system.

[0024] (3) The phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control method and system realizes closed-loop adjustment and optimization update of the cooling strategy by setting a cooling feedback quality assessment mechanism and a continuous adjustment cycle trend monitoring method, thus ensuring the stability and continuous improvement capability of the cooling control. This closed-loop control scheme can automatically trigger the enhancement or energy-saving mode when the strategy response is insufficient or the energy efficiency deviates, thereby improving the intelligent operation level of the system.

[0025] (4) This intelligent temperature control method and system for single-phase immersion liquid cooling with slow-release latent heat of phase change achieves a reasonable reduction in the operating frequency of the liquid cooling pump and the activation frequency of the phase change material while ensuring heat dissipation safety by establishing a response mapping between the strategy feedback threshold and the control parameters of the cooling components. This effectively improves the energy efficiency ratio and operational reliability of the liquid cooling system. This method balances temperature control performance with energy-saving goals and is suitable for application scenarios with high requirements for PUE and stability, such as green data centers. Attached Figure Description

[0026] Figure 1 Flowchart of a method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change;

[0027] Figure 2 This is a structural diagram of a single-phase immersion liquid-cooled intelligent temperature control system for slow-release of latent heat of phase change.

[0028] Figure 3 This is a graph showing the trend of thermal response evaluation values ​​over a continuous sliding window.

[0029] Figure 4 This is a schematic diagram of the structure of a phase change material.

[0030] In the diagram, 1 is the TIM layer; 2 is the aluminum foil layer; 3 is the phase change material; and 4 is the outer polyimide film. Detailed Implementation

[0031] 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.

[0032] Please see Figures 1-4 This invention provides a technical solution: a method and system for intelligent temperature control of single-phase immersion liquid cooling with phase change latent heat slow release, comprising: S1, collecting temperature control data of a phase change liquid cooling system, and performing timestamp alignment, outlier cleaning, noise suppression, and scale normalization on the temperature control data to obtain pre-processed temperature control data of the phase change liquid cooling system; S2, receiving the pre-processed temperature control data of the phase change liquid cooling system, evaluating the operating status of the server, assigning a thermal response level identifier based on the evaluation results, and constructing a thermal response status dataset; S3, extracting the thermal response status dataset, evaluating the cooling slow release capability under temperature control warning and thermal shock states, generating and executing adaptive cooling adjustment strategies in a graded manner; S4, after the cooling adjustment strategy is executed, evaluating the response performance of the cooling strategy, generating cooling strategy optimization instructions, and realizing closed-loop adjustment of the strategy based on monitoring the changing trend of the evaluation results within a continuous adjustment cycle.

[0033] Specifically, phase change liquid cooling temperature control data is collected, and the collected temperature control data is sequentially subjected to timestamp alignment, outlier cleaning, noise suppression, and scale normalization operations to obtain preprocessed phase change liquid cooling temperature control data. The specific steps include: collecting phase change liquid cooling temperature control data, which includes: CPU temperature, acquired through a thermistor mounted on the processor package surface to reflect the chip's instantaneous thermal state; coolant inlet temperature and coolant outlet temperature, acquired through digital temperature sensors installed at the inlet and outlet respectively, to measure the temperature difference in the heat exchange process; coolant flow rate, collected using a flow meter to measure the change in liquid volume passing through the cooling channel per unit time; coolant density, determined by combining coolant type and current temperature conditions to ensure the accuracy of energy calculations; and coolant specific heat capacity, obtained through thermophysical parameters. The system is calibrated based on real-time operating conditions to characterize the heat absorption capacity of a unit mass of coolant. The coolant flow rate is continuously measured using a speed sensor installed inside the pipeline to obtain the flow speed of the liquid in the cooling circuit. The phase change material temperature is collected by a thermoelectric sensor installed on the surface of the phase change material 3 to monitor the heat storage and release state of the phase change medium. The ambient temperature is measured by an environmental monitoring device installed in the space where the server is located to help identify the external thermal environment in which the system is located. The server power consumption is calculated by collecting real-time voltage and current signals from the power supply end to determine the heat load level during operation. After data acquisition, the temperature control data is processed using a time-series correction method based on time-label consistency to reconstruct a complete and continuous time series, and to remove misaligned or redundant records caused by sampling delays, equipment jitter, or abnormal repetitions. Next, a sliding window averaging method is used to smooth the temperature control data. By averaging continuous data within a fixed window width, high-frequency fluctuations caused by cooling controller switching or external disturbances are eliminated. Furthermore, a first-order incremental detection method is used to identify and remove data points with abnormal rate changes. Anomalies are identified by calculating the difference between adjacent data and comparing it with a set threshold. Finally, normalization is performed on various types of temperature control data according to the category of physical quantity, converting the original values ​​into a unified dimensionless interval to ensure the scale consistency and integration of different data in subsequent calculations.

[0034] In this implementation scheme, phase change liquid cooling temperature control data is collected and rigorously preprocessed to ensure the continuity, consistency, and stability of each physical quantity data over time. Time stamp alignment eliminates interruptions and duplicate recordings during the acquisition process, improving the integrity of the timeline. A sliding window averaging method effectively suppresses local fluctuations caused by controller switching and external disturbances, enhancing the stability of the temperature control data. A filtering method based on first-order incremental changes identifies and removes abrupt data points, ensuring that the sampled data accurately reflects the system's operating status. Classification and normalization are used to standardize the scale of various physical quantities, achieving effective fusion of data with different dimensions. The synergistic effect of these techniques significantly improves the processing accuracy and input consistency of the temperature control data, providing a reliable data foundation for subsequent thermal response assessment and cooling regulation strategy formulation.

[0035] Specifically, the steps for receiving preprocessed phase change liquid cooling temperature control data and evaluating the server's operating status are as follows: The phase change liquid cooling temperature control data, after timestamp alignment, outlier cleaning, noise suppression, and scale normalization, is received and sequentially filled into a fixed-length sliding window according to the sampling time order, ensuring the data sequence within each window has complete continuity in the time dimension. Within each sliding window, the CPU temperature sequence, server power consumption sequence, coolant inlet temperature sequence, coolant outlet temperature sequence, and coolant flow rate sequence are extracted as the input basis for thermal response analysis. The CPU temperature change rate is calculated by dividing the difference between adjacent CPU temperature values ​​by the sampling interval, reflecting the processor's heating trend. Similarly, the server power consumption change rate is calculated by dividing the difference between adjacent server power consumption values ​​by the sampling interval, reflecting dynamic load characteristics. Multiplying the temperature change rate and power consumption change rate one by one generates a heat source change term, characterizing the severity of heat generation. The difference between the phase change material temperature and the CPU temperature is squared and then superimposed with a minima term to calculate a heat buffer suppression term, reflecting the phase change material's ability to mitigate heat rise. The minima term consists of very small but non-zero positive real numbers used to avoid numerical instability caused by division by zero during calculations; its value range is [insert range here]. arrive Unless otherwise specified, all subsequent minima will use the definitions and ranges provided herein. The thermal shock driving factor is obtained by dividing the heat source change term by the thermal buffer suppression term, characterizing the intensity of the current transient thermal disturbance. The temperature difference is obtained by subtracting the coolant inlet temperature from the coolant outlet temperature. This temperature difference is then divided by the product of the coolant flow rate, coolant density, and coolant specific heat capacity to calculate the heat dissipation capacity per unit volume, serving as a quantitative indicator of the instantaneous heat exchange capacity of the liquid cooling system. The absolute value of the difference in coolant flow rate between adjacent moments is divided by the sampling interval to calculate the coolant flow rate disturbance term, reflecting the flow instability caused by pump speed changes. The heat dissipation capacity per unit volume and the coolant flow rate disturbance term are added one by one to form the cooling response factor, used to evaluate the real-time response efficiency of the current cooling path to thermal disturbances. Finally, the thermal shock driving factor and the cooling response factor are multiplied to obtain the thermal response evaluation value, providing a quantitative basis for the accurate determination of subsequent cooling adjustment strategies.

[0036] The specific formula for calculating the thermal response evaluation value is as follows:

[0037] ;

[0038] In the formula, This represents the thermal response assessment value. Indicates CPU temperature. Indicates the temperature of the phase change material. Indicates server power consumption. Indicates the coolant inlet temperature. Indicates the coolant outlet temperature. Indicates coolant flow rate. Indicates the density of the coolant. This indicates the specific heat capacity of the coolant. Indicates the coolant flow rate. Indicates a minus term.

[0039] In this embodiment, Table 1 is a thermal response evaluation value data table, which shows the temperature control parameters and thermal response evaluation results of the five sliding windows in the current control cycle, covering the CPU temperature change rate, server power consumption change rate, phase change material temperature, coolant inlet and outlet temperatures, coolant flow rate, coolant density, coolant specific heat capacity, coolant flow rate, and the finally calculated thermal response evaluation value. The specific details are as follows: In sliding window 1, the CPU temperature change rate is 0.85, the server power consumption change rate is 200, the phase change material temperature is 43, the coolant inlet temperature is 35, the outlet temperature is 47, the coolant flow rate is 0.58, the coolant density is 1000, the coolant specific heat capacity is 4186, and the coolant flow rate is 2.5, resulting in a final calculated thermal response evaluation value of 0.26; In sliding window 2, the CPU temperature change rate is 0.81, the server power consumption change rate is 180, the phase change material temperature is 41, the coolant inlet temperature is 35, the outlet temperature is 46, the coolant flow rate is 0.60, the coolant density is 1000, the coolant specific heat capacity is 4186, and the coolant flow rate is 2.1, resulting in a final calculated thermal response evaluation value of 0.18; In sliding window 3, the CPU temperature change rate is 0.92, the server power consumption change rate is 210, the phase change material temperature is 42, the coolant inlet temperature is 36, and the outlet temperature is... In sliding window 4, with a CPU temperature change rate of 1.00%, a server power consumption change rate of 190%, a phase change material temperature of 43%, a coolant inlet temperature of 37%, an outlet temperature of 48%, a coolant flow rate of 0.57%, a coolant density of 1000%, a coolant specific heat capacity of 4186%, and a coolant flow rate of 2.6%, the final calculated thermal response evaluation value is 0.29. In sliding window 5, with a CPU temperature change rate of 1.10%, a server power consumption change rate of 200%, a phase change material temperature of 44%, a coolant inlet temperature of 38%, an outlet temperature of 49%, a coolant flow rate of 0.62%, a coolant density of 1000%, a coolant specific heat capacity of 4186%, and a coolant flow rate of 2.8%, the final calculated thermal response evaluation value is 0.31.

[0040] Table 1. Thermal Response Evaluation Values ​​Data Table

[0041]

[0042] like Figure 3The figure shows the trend of thermal response evaluation values ​​calculated over five consecutive sliding windows. By calculating multi-dimensional indicators such as CPU temperature change rate, server power consumption change rate, coolant flow rate change, and temperature difference, the obtained thermal response evaluation values ​​can be used to measure the current thermal shock intensity and cooling response capability of the system. As can be seen from the figure, the thermal response evaluation value is highest in window 1, indicating a significant thermal disturbance during this period; it then decreases in windows 2 and 3, indicating that the cooling strategy has mitigated the thermal shock to some extent; while it rises again in windows 4 and 5, suggesting the need for dynamic monitoring and optimization of the current cooling strategy. Figure 3 This model is used to verify the real-time performance and sensitivity of the thermal response evaluation model in this invention, and can provide a basis for the intelligent generation and closed-loop optimization of subsequent cooling regulation strategies.

[0043] In this implementation scheme, the preprocessed phase change liquid cooling temperature control data is segmented using a sliding window method. The CPU temperature sequence, server power consumption sequence, coolant inlet temperature sequence, coolant outlet temperature sequence, and coolant flow rate sequence are extracted sequentially. The temperature change rate and power consumption change rate are calculated using the time derivative. A heat source change term, a thermal buffer suppression term, a thermal shock driving factor, a unit volume heat dissipation capacity term, a flow rate disturbance term, and a cooling response factor are constructed to ultimately obtain a thermal response evaluation value. This value accurately characterizes the matching relationship between the thermal shock intensity and cooling capacity under the current operating state of the server. This invention enables high-precision dynamic evaluation of the thermal response process on a time-series scale, improving the real-time performance of thermal disturbance identification, the accuracy of cooling adjustment, and the targeted nature of temperature control strategy adaptation.

[0044] Specifically, the steps for assigning thermal response level labels based on the evaluation results and constructing a thermal response state dataset are as follows: real-time comparison of thermal response evaluation values ​​with thermal response thresholds, including primary thermal response thresholds and secondary thermal response thresholds. The thermal response evaluation value is obtained by multiplying the thermal shock driving factor and the cooling response factor, which is used to quantitatively reflect the degree of matching between thermal disturbance intensity and cooling capacity. When the thermal response assessment value is less than or equal to the Level 1 thermal response threshold, the server is determined to be in a steady-state operation, indicating that the current temperature control is stable and the cooling regulation capacity is sufficient, and the thermal response level identifier is assigned a value of 0. When the thermal response assessment value is greater than the Level 1 thermal response threshold but less than the Level 2 thermal response threshold, the server is determined to be in a temperature control warning state, indicating that the thermal disturbance signal is gradually increasing and the cooling regulation capacity is facing boundary pressure, and the thermal response level identifier is assigned a value of 1. When the thermal response assessment value is greater than or equal to the Level 2 thermal response threshold, the server is determined to be in a thermal shock state, indicating that the current temperature control capacity is unable to effectively cope with the continuous changes in heat load and there is a significant risk of temperature rise, and the thermal response level identifier is assigned a value of 2. The thermal response assessment value and the thermal response level identifier are stored in a structured manner to construct a thermal response status dataset, providing basic data support for the dynamic adjustment of subsequent cooling regulation strategies.

[0045] In this implementation plan, by comparing the thermal response assessment value and the thermal response threshold in real time, and combining this with the assignment of thermal response level identifiers and the construction of a thermal response status dataset, the current temperature control status of the server can be accurately determined. This process uses the thermal response assessment value as the basic indicator, clearly distinguishing between steady-state operation, temperature control warning, and thermal shock states, and recording the thermal response level identifier and thermal response assessment value in a structured form. This enhances the targeting and real-time performance of temperature control adjustments, effectively improves the dynamic adaptability of cooling adjustment strategies, and provides accurate data support and judgment criteria for subsequent cooling response control.

[0046] Specifically, the steps for extracting the thermal response state dataset and evaluating the cooling slow-release capability under temperature control warning and thermal shock conditions are as follows: Extract the thermal response state dataset and real-time phase change liquid cooling temperature control data, including coolant outlet temperature, coolant inlet temperature, coolant flow rate, phase change material temperature, CPU temperature, ambient temperature, and server power consumption; subtract the coolant inlet temperature from the coolant outlet temperature, calculate the temperature difference, and then divide it by the corresponding coolant flow rate to obtain the cooling temperature difference term formed by the combined change in coolant temperature and flow rate per unit time; increment the cooling temperature difference term by one and multiply it by the current thermal response evaluation value to form the final evaluation value. The following calculations are performed: A thermal response enhancement term is considered, taking into account the coupling relationship between temperature difference and thermal response. The CPU temperature is subtracted from the phase change material temperature, the difference is squared, and a minimum term is added to obtain a buffering capacity index for CPU temperature changes during the phase change process, serving as a thermal buffer suppression term. The ambient temperature is divided by the coolant inlet temperature, a minimum term is added, and then one is added to the ratio to form a temperature disturbance correction ratio, serving as an environmental disturbance correction term. Server power consumption is multiplied by both the thermal buffer suppression term and the environmental disturbance correction term to obtain a cooling load adjustment term reflecting the cooling load pressure under the current operating conditions. The thermal response enhancement term is used as the numerator, divided by the cooling load adjustment term as the denominator, to finally calculate the sustained-release adaptation evaluation value.

[0047] The specific formula for the sustained-release adaptation assessment value is as follows:

[0048] ;

[0049] In the formula, J represents the sustained-release adaptation assessment value, G represents the thermal response level identifier, and R represents the thermal response assessment value. Indicates the coolant outlet temperature. Indicates the coolant inlet temperature. Indicates the coolant flow rate. Indicates server power consumption. Indicates the temperature of the phase change material. Indicates CPU temperature. Indicates minterms, Indicates ambient temperature.

[0050] In this implementation scheme, a multi-parameter calculation model is constructed by jointly extracting data from thermal response state datasets and real-time phase change liquid cooling temperature control data. This model includes parameters such as coolant outlet temperature, coolant inlet temperature, coolant flow rate, phase change material temperature, CPU temperature, ambient temperature, and server power consumption. By combining and analyzing cooling temperature difference terms, thermal response enhancement terms, thermal buffer suppression terms, environmental disturbance correction terms, and cooling load adjustment terms, a slow-release adaptation evaluation value is quantified. This enables accurate evaluation of the cooling slow-release capability under temperature control warning and thermal shock conditions. This helps to enhance the response sensitivity and adaptive adjustment accuracy of the temperature control strategy to complex thermal disturbance scenarios, thereby improving the dynamic stability and energy efficiency control level of the cooling system.

[0051] Specifically, the steps for generating and executing an adaptive cooling regulation strategy in stages are as follows: An adaptive cooling regulation strategy is generated by combining the thermal response level identifier and the slow-release adaptation assessment value to ensure that the cooling response behavior is real-time and targeted. If the thermal response level identifier is 2 and the slow-release adaptation assessment value is greater than the slow-release regulation threshold, it is determined that the current cooling slow-release capacity has a risk of lag, and it is marked as a high-risk state of slow-release lag. The controller issues a liquid cooling pump acceleration command to increase the liquid cooling pump speed to the high-speed range within the regulation cycle, thereby improving the coolant flow rate and heat exchange efficiency. At the same time, the phase change material heat absorption trigger temperature is set as the minimum activation value to activate the latent heat absorption behavior of the phase change to improve the thermal shock suppression capability. The high-speed range refers to the speed range in which the liquid cooling pump operates under high load adaptation conditions, possessing the ability to quickly increase the coolant circulation rate to enhance heat exchange performance. If the thermal response level identifier is 1 and the slow-release adaptation assessment value is greater than the slow-release regulation threshold, it is determined that the current cooling capacity is close to the slow-release edge. The boundary is marked as the boundary state of the slow-release capacity. The controller issues an acceleration command to the liquid cooling pump, increasing the speed of the liquid cooling pump to the medium speed range to maintain the heat flux transfer efficiency. The phase change material 3 is set to the heat absorption preparation state, entering the latent heat response preparation stage in advance. The medium speed range refers to the stable operating range of the liquid cooling pump in the cooling enhancement state, which is used to improve the cooling flow efficiency without wasting resources. The other states are marked as the cooling response stable state, maintaining the current cooling adjustment strategy unchanged and avoiding resource fluctuations caused by ineffective adjustment. Based on the generated cooling adjustment strategy, the liquid cooling pump speed adjustment and phase change material 3 heat absorption configuration are executed in the control logic sequence to complete the cooling execution control operation and realize the dynamic and stable adjustment of the server's thermal management.

[0052] In this implementation scheme, precise identification and graded response of cooling status are achieved by combining thermal response level indicators with sustained-release adaptation assessment values. Under high-risk conditions of sustained-release lag, the cooling response speed is improved by increasing the liquid cooling pump speed and setting the phase change material endothermic trigger temperature as the minimum activation value. Under conditions at the sustained-release capacity boundary, endothermic preparation is activated in advance to enhance the thermal shock sustained-release capability. Under stable cooling response conditions, the current strategy is maintained to reduce energy consumption. These strategies achieve sensitive response and dynamic control to changes in heat load, enhancing the adaptability and stability of temperature control.

[0053] Specifically, after the cooling adjustment strategy is executed, the specific steps for evaluating the response performance of the cooling strategy are as follows: The complete process from the start of a cooling adjustment strategy issuance by the controller to the next cooling adjustment strategy update is defined as an independent adjustment cycle, used to clarify the time boundary between each strategy activation and feedback; after each cooling adjustment strategy is executed, the strategy issuance timestamp and the timestamp of the first thermal response evaluation value collected after adjustment are extracted to calculate the strategy response duration, used to reflect the time sensitivity of the cooling system response; and the thermal response evaluation values ​​before and after adjustment are extracted and quantitatively compared to measure the thermal shock performance before and after cooling adjustment. The thermal response change amplitude is formed by the combined effect of the dynamic factor and the cooling response factor; the logarithm to base 10 is taken after adding one to the strategy response time to eliminate the calculation anomaly when the response time is zero, resulting in a response lag adjustment term, which is used to measure the weight of the cooling strategy response speed; the thermal response change correction term is calculated by subtracting the thermal response evaluation value after the strategy execution from the thermal response evaluation value before the strategy execution, taking the absolute value of the difference, and dividing it by the thermal response evaluation value before the strategy execution, which is used to measure the degree of improvement of the adjustment effect; the response lag adjustment term is multiplied by the thermal response change correction term to obtain the cooling feedback quality evaluation value, which is used as the basis for generating subsequent cooling strategy optimization instructions.

[0054] The specific formula for calculating the cooling feedback quality assessment value is as follows:

[0055] ;

[0056] In the formula, This indicates the cooling feedback quality assessment value. Indicates the strategy response time. This represents the thermal response evaluation value before strategy execution. This represents the thermal response evaluation value after the strategy is executed. Indicates a minus term.

[0057] In this implementation scheme, by introducing the joint calculation of regulation cycle division, strategy response duration, thermal response evaluation value change, response lag adjustment term, and thermal response change correction term after the cooling regulation strategy is executed, a cooling feedback quality evaluation value is constructed, enabling quantitative analysis and dynamic evaluation of the cooling regulation strategy response effectiveness. Under the coupled effect of thermal shock driving factors and cooling response factors, this evaluation mechanism can accurately reflect the regulation effect of the cooling strategy under different thermal response level labels and slow-release adaptation evaluation value conditions, significantly improving the accuracy and timeliness of cooling feedback regulation, and providing reliable support for subsequent cooling strategy optimization.

[0058] Specifically, the steps for generating cooling strategy optimization instructions are as follows: Real-time comparison of the cooling feedback quality assessment value and the feedback quality threshold is used to generate cooling strategy optimization instructions, which include the following: When the cooling feedback quality assessment value is less than the feedback quality threshold, based on the deviation between the cooling feedback quality assessment value and the feedback quality threshold, it is determined that the current cooling adjustment strategy's adjustment efficiency within the current adjustment cycle is insufficient. This triggers a strategy enhancement mechanism, and the control logic outputs instructions to increase the liquid cooling pump's operating speed, switching the liquid cooling pump from the current speed range to a higher speed range, shortening the adjustment cycle of the current cooling adjustment strategy, and simultaneously increasing the heat absorption activation sensitivity of the phase change material. This improves the thermal buffer response rate by lowering the heat absorption trigger temperature. When the cooling feedback quality assessment value is greater than or equal to the feedback quality threshold, it is determined that the current cooling adjustment strategy has a good adjustment effect. The control logic maintains the current cooling adjustment strategy within the original speed range and extends the adjustment cycle to reduce cooling resource consumption and thermal management redundancy.

[0059] In this implementation scheme, the operating level of the liquid cooling pump and the endothermic activation sensitivity of the phase change material are adaptively adjusted by dynamically comparing the cooling feedback quality assessment value with the feedback quality threshold. When the cooling feedback quality assessment value is lower than the feedback quality threshold, the liquid cooling pump is automatically upgraded to the high response range, the adjustment cycle is shortened, and the endothermic response capability is enhanced, thereby improving cooling efficiency. When the cooling feedback quality assessment value is not lower than the feedback quality threshold, the current strategy is maintained and the adjustment cycle is extended to improve cooling stability and energy efficiency, thus achieving closed-loop optimization and efficient control of the cooling adjustment strategy.

[0060] Specifically, based on the changing trend of the evaluation results within the continuous adjustment cycle, the specific steps for achieving closed-loop adjustment of the strategy are as follows: Monitor the changing trend of the cooling feedback quality evaluation value within the continuous adjustment cycle. When the cooling feedback quality evaluation value is less than the feedback quality threshold within a fixed number of consecutive adjustment cycles, it is determined that the current cooling strategy is continuously failing. The cooling strategy configuration is reconstructed, the liquid cooling pump control gear setting parameters are updated, and the activation boundary value of the phase change material heat absorption trigger temperature is reset to improve the timeliness and accuracy of cooling control. When the cooling feedback quality evaluation value is greater than the feedback quality threshold within a fixed number of consecutive adjustment cycles, it is determined that the current cooling strategy is continuously effective. The strategy energy-saving mechanism is triggered. While maintaining the cooling capacity, the operating frequency of the liquid cooling pump is reduced to reduce energy consumption, and the activation frequency of phase change material heat absorption is reduced simultaneously to extend the material's service life and optimize the system power consumption level, thereby achieving adaptive closed-loop dynamic adjustment of the cooling strategy.

[0061] In this implementation plan, the cooling regulation strategy is optimized in a closed loop by monitoring the changing trend of the cooling feedback quality assessment value within a continuous adjustment cycle. When the cooling feedback quality assessment value is continuously lower than the feedback quality threshold, the cooling strategy configuration is reconfigured in a timely manner, and the control level of the liquid cooling pump and the heat absorption trigger temperature of the phase change material are adjusted. When the cooling feedback quality assessment value is continuously higher than the feedback quality threshold, the strategy energy-saving mechanism is triggered to reduce the operating frequency of the liquid cooling pump and the heat absorption activation frequency of the phase change material, thereby improving the stability and energy efficiency control level of the cooling regulation strategy.

[0062] like Figure 2 As shown, the second aspect of this invention provides a phase change latent heat slow-release single-phase immersion liquid cooling intelligent temperature control system, including a temperature control data preprocessing module, a thermal shock identification and judgment module, a cooling strategy generation and execution module, and a cooling strategy closed-loop optimization module. The temperature control data preprocessing module collects phase change liquid cooling temperature control data and performs timestamp alignment, outlier cleaning, noise suppression, and scale normalization on the data to obtain preprocessed phase change liquid cooling temperature control data. The thermal shock identification and judgment module receives the preprocessed phase change liquid cooling temperature control data, evaluates the server's operating status, assigns a thermal response level identifier based on the evaluation results, and constructs a thermal response status dataset. The cooling strategy generation and execution module extracts the thermal response status dataset, evaluates the cooling slow-release capability under temperature control warning and thermal shock conditions, and generates and executes adaptive cooling adjustment strategies in a tiered manner. The cooling strategy closed-loop optimization module evaluates the response performance of the cooling adjustment strategy after execution, generates cooling strategy optimization instructions, and realizes closed-loop strategy adjustment based on the changing trend of the evaluation results within a continuous adjustment cycle.

[0063] like Figure 4As shown, the phase change material 3 comprises the following structure: a TIM layer 1, disposed on the heat source contact surface to improve interface thermal conductivity; an aluminum foil layer 2, disposed on the outer surface of the phase change material 3 as a thermally conductive covering material; the phase change material 3, activated after the cooling regulation strategy triggers the heat absorption configuration, achieving thermal shock mitigation through latent heat absorption; and an outer polyimide film 4, used to isolate the thermally conductive components from the coolant, maintaining material structural stability and thermal conductivity continuity. This structure supports the heat absorption configuration control process of the phase change material 3 in this invention. When the thermal response level is marked as 1 or 2, the heat absorption activation parameters are set according to the cooling regulation strategy, effectively suppressing abnormal CPU temperature fluctuations by utilizing the latent heat of phase change, thereby improving the thermal mitigation capability and system response stability.

[0064] In this implementation scheme, by setting up a temperature control data preprocessing module, a thermal shock identification and judgment module, a cooling strategy generation and execution module, and a cooling strategy closed-loop optimization module, the system can sequentially complete the cleaning and normalization of phase change liquid cooling temperature control data, accurate identification of server operating status, dynamic evaluation of cooling slow-release capability, adaptive execution of cooling regulation strategy, and closed-loop optimization of cooling regulation strategy. This ensures efficient linkage between thermal response level identification, thermal response status dataset, slow-release adaptation evaluation value, and cooling feedback quality evaluation value, thereby improving the timeliness, stability, and accuracy of energy efficiency control of cooling regulation response, and ultimately achieving intelligent thermal management regulation and control for complex operating conditions.

[0065] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0066] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change, characterized in that, Includes the following steps: S1, collect phase change liquid cooling temperature control data, and perform timestamp alignment, outlier cleaning, noise suppression and scale normalization on the phase change liquid cooling temperature control data to obtain preprocessed phase change liquid cooling temperature control data; S2 receives the preprocessed phase change liquid cooling temperature control data, evaluates the server's operating status, assigns a thermal response level identifier based on the evaluation results, and constructs a thermal response status dataset. The specific steps for receiving the preprocessed phase change liquid cooling temperature control data and evaluating the server's operating status are as follows: The preprocessed phase change liquid cooling temperature control data is received and sequentially filled into a fixed-length sliding window according to the sampling time sequence. Within each sliding window, the CPU temperature sequence, server power consumption sequence, coolant inlet temperature sequence, coolant outlet temperature sequence, and coolant flow rate sequence are extracted to evaluate the server's operating status: the CPU temperature is differentiated with respect to time to obtain the temperature change rate; the server power consumption is differentiated with respect to time to obtain the power consumption change rate; the temperature change rate and the power consumption change rate are multiplied together to obtain the heat source change term; the phase change material temperature is subtracted from the CPU temperature to obtain the difference. The squared value is added to a minimum term to obtain the thermal buffer suppression term; the heat source change term is divided by the thermal buffer suppression term to obtain the thermal shock driving factor; the temperature difference obtained by subtracting the coolant inlet temperature from the coolant outlet temperature is divided by the product of the coolant flow rate, coolant density, and coolant specific heat capacity to obtain the heat dissipation capacity per unit volume term; the absolute value of the derivative of the coolant flow rate with respect to time is obtained as the flow rate disturbance term; the heat dissipation capacity per unit volume term and the flow rate disturbance term are added to obtain the cooling response factor; the thermal shock driving factor and the cooling response factor are multiplied to obtain the thermal response evaluation value. S3, extract the thermal response state dataset, evaluate the cooling slow release capability under temperature control warning state and thermal shock state, and generate and execute adaptive cooling regulation strategies in a hierarchical manner. S4. After the cooling adjustment strategy is executed, the response performance of the cooling strategy is evaluated, cooling strategy optimization instructions are generated, and closed-loop adjustment of the strategy is achieved based on the changing trend of the evaluation results within the continuous adjustment cycle.

2. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for acquiring phase change liquid cooling temperature control data and performing timestamp alignment, outlier cleaning, noise suppression, and scale normalization on the phase change liquid cooling temperature control data to obtain preprocessed phase change liquid cooling temperature control data are as follows: Collect phase change liquid cooling temperature control data, including CPU temperature, coolant inlet temperature, coolant outlet temperature, coolant flow rate, coolant density, coolant specific heat capacity, coolant flow rate, phase change material temperature, ambient temperature, and server power consumption. The phase change liquid cooling temperature control data is processed using a time-series correction method based on time label consistency to remove discontinuous data caused by data acquisition interruptions and duplicate recordings. A sliding window averaging method is used to smooth the data, eliminating local fluctuations caused by control switching and environmental disturbances. A first-order incremental change-based filtering method is used to perform incremental detection on the data, identifying and removing data points with abnormal change rates. Finally, a classification normalization method is used to scale the data, unifying different physical quantities into a standard range.

3. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for assigning thermal response level identifiers based on the evaluation results and constructing a thermal response state dataset are as follows: The thermal response assessment value and the thermal response threshold are compared in real time. When the thermal response assessment value is less than or equal to the first-level thermal response threshold, the server is determined to be in a steady-state operation and the thermal response level flag is assigned a value of 0. When the thermal response assessment value is greater than the first-level thermal response threshold and less than the second-level thermal response threshold, the server is determined to be in a temperature control warning state, and the thermal response level flag is assigned a value of 1. When the thermal response assessment value is greater than or equal to the level 2 thermal response threshold, the server is determined to be in a thermal shock state, and the thermal response level identifier is assigned a value of 2. The thermal response assessment values ​​and thermal response level identifiers are stored in a structured manner to construct a thermal response status dataset.

4. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for extracting the thermal response state dataset and evaluating the cooling slow-release capability under temperature control warning and thermal shock conditions are as follows: Extract the thermal response state dataset and real-time phase change liquid cooling temperature control data. Subtract the coolant inlet temperature from the coolant outlet temperature, divide the resulting temperature difference by the coolant flow rate to obtain the cooling temperature difference term. Add one to the cooling temperature difference term and multiply by the thermal response evaluation value to obtain the thermal response enhancement term. Subtract the CPU temperature from the phase change material temperature, square the difference, and add a minimum term to obtain the thermal buffer suppression term. Divide the ambient temperature by the coolant inlet temperature and add a minimum term, then add one to the resulting ratio to obtain the environmental disturbance correction term. Multiply the server power consumption, thermal buffer suppression term, and environmental disturbance correction term together to obtain the cooling load adjustment term. Divide the thermal response enhancement term by the cooling load adjustment term to obtain the slow-release adaptation evaluation value.

5. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for generating and implementing the hierarchical adaptive cooling adjustment strategy are as follows: Based on the combined thermal response level identifier and the slow-release adaptation assessment value, an adaptive cooling adjustment strategy is generated: if the thermal response level identifier is 2 and the slow-release adaptation assessment value is greater than the slow-release adjustment threshold, it is marked as a high risk of slow-release hysteresis. The controller issues a liquid cooling pump acceleration command to increase the speed of the liquid cooling pump to the high speed range and sets the heat absorption trigger temperature of the phase change material to the minimum activation value. If the thermal response level is marked as 1 and the slow-release adaptation assessment value is greater than the slow-release adjustment threshold, it is marked as the slow-release capacity boundary. The controller issues a liquid cooling pump acceleration command to increase the speed of the liquid cooling pump to the medium speed range and sets the phase change material (3) state to heat absorption preparation. In other cases, it is marked as cooling response stable and the current cooling adjustment strategy remains unchanged. Based on the cooling regulation strategy, the liquid cooling pump speed regulation and phase change material (3) heat absorption configuration are executed to complete the cooling execution control.

6. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: After the cooling regulation strategy is implemented, the specific steps for evaluating the response performance of the cooling strategy are as follows: The entire process from the issuance of a cooling adjustment strategy to the next strategy update is defined as an adjustment cycle. After each cooling adjustment strategy is executed, the strategy issuance timestamp and the timestamp of the first thermal response evaluation value collected after adjustment are extracted to calculate the strategy response duration. The thermal response evaluation values ​​before and after adjustment are extracted. The strategy response duration is added by one and the logarithm to the base 10 is taken to obtain the response delay adjustment item. Subtracting the thermal response evaluation value before strategy execution from the thermal response evaluation value after strategy execution, and then dividing the absolute value of the difference by the thermal response evaluation value before strategy execution, yields the thermal response change correction term; multiplying the response lag adjustment term by the thermal response change correction term yields the cooling feedback quality evaluation value.

7. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for generating the cooling strategy optimization instruction are as follows: The system compares the cooling feedback quality assessment value with the feedback quality threshold in real time and generates cooling strategy optimization instructions: when the cooling feedback quality assessment value is less than the feedback quality threshold, it determines that the current cooling strategy adjustment efficiency is insufficient, triggers the strategy enhancement mechanism, increases the operating level of the liquid cooling pump, shortens the adjustment cycle, and improves the heat absorption activation sensitivity of the phase change material; when the cooling feedback quality assessment value is greater than or equal to the feedback quality threshold, it determines that the current cooling strategy adjustment effect is good, maintains the current strategy, and extends the adjustment cycle.

8. The method for intelligent temperature control and regulation of single-phase immersion liquid cooling with slow-release latent heat of phase change according to claim 1, characterized in that: The specific steps for implementing closed-loop policy adjustment based on the changing trend of evaluation results within the continuous adjustment period are as follows: Monitor the trend of cooling feedback quality assessment value changes within a continuous adjustment cycle: if the continuous fixed number of adjustment cycles is lower than the feedback quality threshold, reconfigure the cooling strategy, adjust the pump control logic, and reset the phase change material heat absorption trigger temperature; if the continuous fixed number of adjustment cycles is higher than the feedback quality threshold, trigger the strategy energy-saving mechanism to reduce the operating frequency of the liquid cooling pump and the activation frequency of the phase change material.

9. A single-phase immersion liquid cooling intelligent temperature control system for phase change latent heat slow release, employing the single-phase immersion liquid cooling intelligent temperature control method for phase change latent heat slow release as described in any one of claims 1-8, characterized in that: It includes a temperature control data preprocessing module, a thermal shock identification and judgment module, a cooling strategy generation and execution module, and a cooling strategy closed-loop optimization module, wherein: The temperature control data preprocessing module is used to collect phase change liquid cooling temperature control data and perform timestamp alignment, outlier cleaning, noise suppression and scale normalization on the phase change liquid cooling temperature control data to obtain preprocessed phase change liquid cooling temperature control data. The thermal shock identification and determination module is used to receive preprocessed phase change liquid cooling temperature control data, evaluate the server's operating status, assign a thermal response level identifier based on the evaluation results, and construct a thermal response status dataset. The cooling strategy generation and execution module is used to extract thermal response state dataset, evaluate the cooling slow release capability under temperature control warning state and thermal shock state, and generate and execute adaptive cooling adjustment strategies in a hierarchical manner. The cooling strategy closed-loop optimization module is used to evaluate the response performance of the cooling strategy after the cooling adjustment strategy is executed, generate cooling strategy optimization instructions, and realize closed-loop adjustment of the strategy based on the changing trend of the evaluation results within the continuous adjustment cycle.

Images