Single-phase immersed liquid cooling temperature stratification evaluation and disturbance liquid injection balance control system

Abstract

The invention discloses a single-phase immersed liquid cooling temperature stratification evaluation and disturbance liquid injection balance control system, and relates to the technical field of liquid cooling circulation structures. The single-phase immersed liquid cooling temperature stratification evaluation and disturbance liquid injection balance control system comprises a backflow liquid multi-point liquid inlet dynamic monitoring module used for collecting liquid cooling branch state data; the immersion liquid temperature layering monitoring module is used for evaluating the non-uniform degree of heat distribution and adjusting a disturbance liquid injection strategy; the liquid flow distribution equilibrium evaluation module is used for analyzing the matching degree of liquid flow distribution and heat dissipation and optimizing a flow direction coverage strategy; the cooling path dynamic judgment and switching module is used for evaluating the cooling state and the thermal load change and regulating and controlling the circulation path; and the immersed liquid cooling cabinet local disturbance self-balancing regulation and control module is used for guiding local convection and adjusting the disturbance direction. The problem that heat dissipation efficiency of a chip is affected by local heat accumulation in the immersed liquid cooling cabinet due to serious cold and hot lamination after backflow of a cooling liquid branch is solved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of liquid cooling circulation structures, in particular to a single-phase immersion liquid cooling temperature stratification evaluation and disturbance liquid injection equalization control system. BACKGROUND

[0002] With the expansion of the deployment scale of high-density servers and the continuous improvement of the energy efficiency standards of data centers, the role of circulating cooling structures in immersion liquid cooling systems is increasingly prominent. Current single-phase immersion liquid cooling technology is widely used in the thermal management process of high-power devices. By building a closed liquid pool, setting up a main cooling circuit and configuring a liquid driving device, the orderly circulation of liquid between the heating area and the heat dissipation area is realized, so as to ensure that the heat of the internal components of the server is timely discharged and released to the external heat exchange unit.

[0003] For example, the patent for invention with the announcement number CN114423264B discloses a single-phase immersion liquid cooling system and a liquid cooling method. The liquid cooling system comprises a liquid supply main pipe and a liquid return main pipe, and the liquid supply main pipe and the liquid return main pipe form a circulating loop. A plurality of liquid cooling cabinets are connected in parallel to the circulating loop through liquid supply sub-pipes and liquid return sub-pipes. Then, a flow regulating unit is arranged on the liquid supply sub-pipe. The flow regulating unit comprises a differential pressure detection device and a flow regulating device. The differential pressure detection device and the flow regulating device are respectively signal connected to a control unit. The differential pressure detection device is used for detecting the pressure difference of the cooling liquid flowing through itself. The flow regulating device is used for regulating the flow of the cooling liquid flowing through itself. Thus, for the liquid supply sub-pipe of a certain branch, the control unit can regulate the flow of the cooling liquid flowing into the liquid supply sub-pipe according to the pressure difference of the liquid supply sub-pipe. In this way, the cooling liquid flow in each parallel branch of the circulating loop is adjusted as needed, and the adjustment precision is high, thereby avoiding the imbalance of the cooling liquid flow.

[0004] For example, the patent for invention with the announcement number CN109952003B discloses a data center liquid cooling system, relating to the technical field of data center heat dissipation devices. The data center liquid cooling system comprises a box body, a server, a cold plate, a cooling device, a liquid storage tank and a refrigerant pump. The box body is internally provided with single-phase refrigerant. The server is immersed in the single-phase refrigerant. The server is connected with the box body. The cold plate is arranged in the box body and connected with the box body. The cold plate is connected with the cooling device, the liquid storage tank and the refrigerant pump in sequence to form a circulating loop. The data center liquid cooling system realizes the combination of indirect contact liquid cooling, single-phase direct immersion liquid cooling and natural cooling. The single-phase refrigerant does not need to be circulated under the driving of the pump, but relies on the change of temperature and density for circulation. Although the refrigerant is circulated under the driving of the refrigerant pump, the viscosity of the refrigerant is smaller than that of the single-phase refrigerant, and the refrigerant is easy to flow, so as not to increase the loss of the refrigerant pump. Therefore, the loss of the pump caused by the single-phase refrigerant is reduced, the PUE of the data center is reduced, and the green degree of the data center is improved.

[0005] However, the existing single-phase immersion liquid cooling system generally has the following technical bottlenecks: the cooling path structure is mostly configured with a static main loop, lacking real-time response mechanism to fluid temperature gradient, flow distribution and local heat accumulation, and it is difficult to cope with the cooling demand fluctuation caused by dynamic changes of server heat load. When running high-load tasks, there are local dead angles or thermal stratification in liquid flow, resulting in decreased heat exchange efficiency in some areas and serious heat retention. At the same time, the existing system mostly circulates cooling liquid at a fixed frequency or constant flow rate, and cannot implement differentiated adjustment according to the heat dissipation pressure of different areas, limiting the further improvement of overall energy efficiency ratio and operation stability.

[0006] In view of the above problems, a single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system is urgently needed. SUMMARY

[0007] In view of the deficiencies of the prior art, the single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system is provided, which solves the problem of serious cold and hot layering after the return of the cooling liquid branch, resulting in local heat accumulation in the immersion liquid cooling cabinet and affecting the heat dissipation efficiency of the chip.

[0008] To achieve the above purpose, the following technical solutions are adopted: the single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system comprises a return liquid multi-point liquid injection dynamic monitoring module for collecting cooling liquid state data, liquid flow state data and path thermal state data in the current liquid cooling branch, and preprocessing the collected cooling liquid state data, liquid flow state data and path thermal state data to construct a standardized liquid cooling operation feature data set; an immersion liquid temperature stratification monitoring module for evaluating the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation feature data set, and dynamically adjusting the local disturbance injection strategy based on the evaluation result; a liquid flow distribution equalization evaluation module for analyzing the heat dissipation matching degree of the current liquid flow distribution based on the standardized liquid cooling operation feature data set, and driving the flow direction coverage strategy optimization based on the analysis result; a cooling path dynamic determination and switching module for inputting the thermal distribution unevenness evaluation result and the heat dissipation matching degree analysis result of the current cooling state and heat load change, and intelligently regulating and controlling the circulation path and heat exchange intervention mode based on the evaluation result; an immersion liquid cooling cabinet local disturbance self-balancing regulation module for judging the thermal layer structure evolution trend based on the temperature profile and liquid flow distribution state, guiding local convection through bottom disturbance injection, dynamically adjusting the disturbance direction and intensity in combination with temperature rise feedback and distribution response result, and improving the local diffusion effect of disturbance liquid by cooperating with the flow direction relaxation control.

[0009] Further, the specific steps for collecting the cooling liquid state data, liquid flow state data, and path thermal state data in the current liquid cooling branch are: structuring the collection of key parameters in the liquid cooling branch, collecting the cooling liquid state data, which includes the cooling liquid temperature of each layer in the immersion liquid cooling cabinet, the liquid level of the temperature sensor, the top cooling liquid temperature and the bottom cooling liquid temperature of the immersion liquid cooling cabinet, and recording the number of sensor layers, the temperature difference and height difference between each adjacent two layers in the current collection period; representing the high temperature limit and low temperature reference in the liquid heat distribution in the current period by the top cooling liquid temperature and the bottom cooling liquid temperature of the immersion liquid cooling cabinet, and constructing a representative temperature difference reference data set; in a plurality of continuous collection periods, arranging the time series of the top cooling liquid temperature and the bottom cooling liquid temperature, and extracting the stable interval of the temperature difference change in the near plurality of effective periods based on the processed time series, and recording the mean value of the stable interval as the temperature difference normalization reference value under the current running stage; collecting the liquid flow state data, which includes the instantaneous flow rate value, the continuous flow rate change data, and the liquid injection response time of each region, and recording the average liquid flow rate in the branch return path; constructing a flow rate fluctuation time series by the continuous flow rate change data in each chip region in the immersion liquid cooling cabinet per unit time, and comprehensively evaluating the distribution uniformity of the cooling liquid in the region by combining the liquid injection response time of the disturbance liquid injector to evaluate the fluctuation degree and disturbance response consistency of the regional liquid per unit time; collecting the path thermal state data, which includes the chip effective heat dissipation area of each region, the cooling liquid temperature in the return branch, the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet, and the chip temperature rising rate, and recording the total number of effective monitoring regions in the current period.

[0010] Further, the collected cooling liquid state data, liquid flow state data, and path thermal state data are preprocessed to construct a standardized liquid cooling operation feature data set. The specific steps are as follows: the collected cooling liquid state data, liquid flow state data, and path thermal state data are preprocessed, the coordinate standardization including the corresponding relationship between the liquid level height and the temperature sensor is completed, and the temperature sequence mutation and interlayer measurement point dislocation caused by sensor drift and short-time communication interruption are removed; for the distribution characteristics of the temperature of the cooling liquid of each layer, the actual pipe arrangement structure and the cooling liquid flow direction constraint of the immersed liquid cooling cabinet are combined to perform difference repair on the temperature gradient inversion and abnormal stationary interval, and the continuous and stable temperature profile is restored; before preprocessing, the path thermal state data is classified and mapped according to the chip region topology and the liquid injection point mapping relationship, and the heat dissipation area splitting processing is performed on the chip area with multiple region intersection cooling coverage; in order to cope with the data segment conflict caused by the parallel work of multiple liquid injection units in the liquid cooling operation process, a data window synchronization mechanism is introduced in the preprocessing stage, the region data with continuously increasing flow fluctuation amplitude is preferentially retained, and the repeated response field is removed, thereby improving the data consistency under the joint processing of multiple regions; all the preprocessed cooling liquid state data, liquid flow state data, and path thermal state data are normalized to construct a standardized liquid cooling operation feature data set.

[0011] Further, the specific steps for evaluating the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation feature data set are as follows: real-time measurement data of temperature sensors of each layer inside the immersed liquid cooling cabinet are collected to construct a vertical temperature profile sequence in the current collection period, the temperature gradient of the current position is represented by the ratio of the temperature difference and the height difference between each adjacent two layers, and the local maximum gradient of the profile is extracted as the absolute value of the temperature gradient between all adjacent measurement points; at the same time, the upper limit value of the acceptable gradient fluctuation range is obtained by calculating the interquartile range of all temperature gradients, which is recorded as the profile fluctuation reference value; the cooling liquid temperature difference between each adjacent two layers is calculated in turn and divided by the liquid level height difference of the temperature sensors of the adjacent two layers, the ratio of the cooling liquid temperature difference and the liquid level height difference of the temperature sensors is taken as the absolute value to obtain the interlayer temperature gradient value; the average interlayer temperature value is obtained by adding the interlayer temperature gradient values of the sensors of each layer and dividing by the number of sensor layers; the upper and lower temperature difference normalized ratio is obtained by subtracting the bottom cooling liquid temperature from the top cooling liquid temperature of the immersed liquid cooling cabinet and then dividing by the temperature difference normalization reference value; the gradient offset ratio is obtained by dividing the local maximum gradient of the profile by the profile fluctuation reference value; the temperature layer difference risk value is obtained by multiplying the average interlayer temperature value, the upper and lower temperature difference normalized ratio, and the gradient offset ratio.

[0012] Further, the dynamic adjustment of the local disturbance liquid injection strategy based on the evaluation result comprises the following steps: comparing the current temperature layer difference risk value with a temperature layer difference risk threshold value in real time; when the temperature layer difference risk value is less than or equal to the temperature layer difference risk threshold value, maintaining the current branch backflow state unchanged, keeping the existing liquid injection point configuration and flow direction, continuing to perform the multi-point liquid injection operation, monitoring the change amplitude of the temperature collection value of each layer, and updating the data collection and calculation of each layer temperature measuring point at a fixed period; when the temperature layer difference risk value is greater than the temperature layer difference risk threshold value, immediately closing the branch pipeline valve, switching the main circulation cooling path, synchronously suspending the current liquid injection point liquid injection operation, activating the bottom micro-disturbance liquid injector to perform periodic low-speed disturbance injection, and reconstructing the internal liquid injection path mapping relationship of the immersion liquid cooling cabinet.

[0013] Further, the step of analyzing the matching degree of the current liquid flow distribution and heat dissipation based on the standardized liquid cooling operation feature data set comprises the following steps: calculating the ratio of the continuous flow rate change data and the liquid injection response time of each effective monitoring area in the current period to construct a dynamic index set of cooling liquid distribution uniformity; then, in the dynamic index set of cooling liquid distribution uniformity, a median extraction method is used to obtain a normalized reference value of the distribution consistency level in the current period liquid cooling branch, which is denoted as the cooling liquid distribution uniformity reference value; the instantaneous flow rate value of the current area, the effective heat dissipation area of the chip, and the corresponding cooling liquid density correction factor are multiplied to obtain the regional liquid flow kinetic energy value of the current area; the ratio of the distribution uniformity of each area cooling liquid to the cooling liquid distribution uniformity reference value is subtracted and then multiplied by a distribution deviation weight factor to obtain a distribution uniformity adjustment value; the regional liquid flow kinetic energy value is divided by the corresponding distribution uniformity adjustment value to obtain a regional balanced effective liquid flow value, and the regional balanced effective liquid flow values of each effective monitoring area in the current period are accumulated to obtain a liquid flow balanced distribution evaluation value of the current period.

[0014] Further, the driving flow coverage strategy based on the analysis results comprises the following specific steps: based on the calculation result of the liquid flow balance distribution evaluation value, the flow control strategy is dynamically adjusted according to the fluctuation trend of the real-time liquid flow balance distribution evaluation value: when the liquid flow balance distribution evaluation value shows an upward trend in three or more consecutive periods, and the upward amplitude does not exceed the fluctuation amplitude threshold, it indicates that the cooling liquid flow structure tends to be stable and has high matching degree with the heat load, the current flow distribution structure is frozen, the path reconstruction is suspended, and an observation enhancement stage is entered: the sampling density of the micro disturbance area around the liquid inlet point is expanded, the micro low flow rate disturbance injection test is regularly performed, and it is verified whether the current stable state is a real stable state to avoid local false stable state failure; when the liquid flow balance distribution evaluation value shows an upward amplitude exceeding the fluctuation amplitude threshold in two periods, it is determined that the liquid flow coverage structure is suddenly unbalanced, all the liquid injection points are interrupted, the main circulation path is reset, the cooling liquid self-balancing starting mechanism is activated, and the gradient recovery is performed according to the flow rate average value of each liquid injection point, while the delay boost strategy is enabled to avoid sudden pressure impact, and the spatial distribution and convergence speed of the chip temperature difference are monitored in real time during the recovery process, and when the disorder area is found, the liquid inlet angle and flow path are forcibly rearranged.

[0015] Further, the comprehensive evaluation of the current cooling state and the heat load change based on the heat distribution unevenness evaluation result and the liquid flow distribution and heat dissipation matching degree analysis result comprises the following specific steps: the temperature layer difference risk value and the liquid flow balance distribution evaluation value are obtained, the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the liquid outlet of the immersion liquid cooling cabinet is calculated, the temperature layer difference risk value, the liquid flow balance distribution evaluation value and the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the liquid outlet of the immersion liquid cooling cabinet are multiplied to obtain a temperature difference intensification load term; the chip temperature rising rate is added to the reciprocal of the average liquid flow rate in the current branch return path and then added by one to obtain a liquid flow decay penalty value; the temperature difference intensification load term is divided by the liquid flow decay penalty value and then logarithmized to obtain the path switching judgment value at the current time.

[0016] Further, the intelligent regulation of the circulation path and the intervention mode of the heat exchange based on the evaluation results comprises the following specific steps: real-time comparison of the current path switching decision value and the path switching threshold value, the path switching threshold value comprising a first switching threshold value and a second switching threshold value; when the path switching decision value is less than or equal to the second switching threshold value, the current branch backflow path operating state is maintained unchanged, the flow distribution structure and the flow direction of each liquid inlet point are maintained, the path switching execution is suspended, only the backflow liquid temperature and the chip temperature difference are periodically sampled at low frequency, the diffusion trend of the liquid on the cooling surface is recorded, and the next cycle is executed for judgment; when the path switching decision value is greater than the second switching threshold value and less than or equal to the first switching threshold value, the branch pipeline valve opening degree is adjusted downward and the main circulation channel pressure is gradually increased, the two cooling paths are kept in short-time parallel operation, the flow field buffer area is established while ensuring the stability of the chip heat exchange efficiency, and the dynamic path transfer recording mechanism is started, whether to switch to the main circulation channel completely is judged according to the path switching decision value change trend of the next cycle; when the path switching decision value is greater than the first switching threshold value, the path switching instruction is immediately executed, the branch backflow channel is closed, the main circulation path is switched for forced cooling, the backflow liquid direction and the liquid inlet point weight configuration are reset, all disturbance liquid injection operations are suspended, and the main channel liquid velocity equalization constraint logic is activated, thereby ensuring large-flow cooling while limiting the concentrated injection to the overload area, improving the heat exchange efficiency and preventing the local temperature rise of the chip edge caused by high flow rate impact.

[0017] Further, the steps for judging the evolution trend of the thermal layer structure based on the temperature profile and the liquid flow distribution state, and guiding local convection by bottom disturbance injection, dynamically adjusting the disturbance direction and intensity through temperature rise feedback and distribution response results, and improving the local diffusion effect of the disturbance liquid by combining with the flow direction relaxation control are as follows: during the opening period of the branch return path, the temperature profile data inside the immersion liquid cooling cabinet collected by the vertical temperature sensor array is continuously monitored, and the liquid flow equalization distribution evaluation value and the path switching judgment value are jointly analyzed to judge whether there is a trend that the thermal layer structure tends to be stable, when it is monitored that the top liquid temperature of the immersion liquid cooling cabinet is higher than the bottom for three consecutive periods, and the multi-layer temperature difference changes converge, the micro jet and the lateral directional channel arranged at the bottom of the immersion liquid cooling cabinet are automatically used to intermittently inject disturbance cooling liquid at the minimum flow rate without changing the main cooling path and the return path structure, to cause slow local liquid convection, so that the cold liquid floats up and the hot liquid shifts sideways; after each round of disturbance injection, the chip heat zone temperature rise rate and the immersion liquid cooling cabinet multi-layer temperature difference change trend are analyzed, when the temperature difference converges significantly and the liquid flow distribution index shows an upward trend, the current disturbance frequency and direction are maintained; when the liquid flow distribution index does not change, the structure self-adaptive adjustment state is entered, the disturbance jet angle is rotated and the injection position is replaced to build a new disturbance path, and the liquid inlet guide direction is adjusted for short-term fine adjustment; at the same time, the response of the main channel flow rate change is temporarily delayed during the disturbance process, so that the disturbance liquid can complete the initial rising diffusion in the local area.

[0018] The present application has the following beneficial effects: (1) The single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system can realize dynamic switching and fine control of the liquid flow direction and the injection path by constructing the linkage regulation mechanism of the main circulation path and the branch return path, effectively improve the response speed of the cooling liquid in the high heat load area, and avoid local heat accumulation.

[0019] (2) The single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system can realize real-time sensing of the evolution trend of the liquid hot and cold layer structure by introducing the coupling analysis method of the immersion liquid cooling cabinet temperature layer difference risk value and the liquid flow equalization distribution evaluation value, and identify the thermal inertia section caused by the return path lag in the immersion liquid cooling cabinet in advance, to ensure the stable circulation of temperature and flow.

[0020] (3) The single-phase immersion liquid cooling temperature stratification evaluation and disturbance injection equalization control system can maintain the consistency and plasticity of the flow direction inside the channel by introducing a periodic low-speed flushing process under stable circulation, effectively inhibit the thermal inertia accumulation caused by long-time stable operation, and improve the heat dissipation stability of the server under high load continuous operation.

[0021] (4) This single-phase immersion liquid cooling temperature stratification assessment and disturbance injection balance control system can automatically identify the cold liquid floating hysteresis zone and hot liquid accumulation zone during the cooling cycle by introducing a joint constraint mechanism of temperature profile fluctuation reference value and local temperature difference threshold, drive the fine-tuning of the disturbance structure, and realize the micro-scale redistribution of the circulation flow direction.

[0022] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a structural diagram of the single-phase immersion liquid cooling temperature stratification assessment and disturbance injection balance control system of the present invention; Figure 2 A line graph of the path switching determination value involved in the present invention; Figure 3 This is a flow chart of the single-phase immersion liquid cooling temperature stratification assessment and disturbance injection balance control system of the present invention; Figure 4 This is a schematic diagram of the single-phase immersion liquid cooling temperature stratification assessment and disturbance injection balance control system of the present invention.

[0024] In the figure, 1. Immersion liquid cooling cabinet; 2. Branch pipe valve; 3. Main pipe valve; 4. Main liquid outlet pipe; 5. Pump; 6. Heat exchanger; 7. Branch pipe. DETAILED DESCRIPTION

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0026] See also Figures 1-4The embodiment of the application provides a technical scheme: a single-phase immersion liquid cooling temperature stratification evaluation and disturbance liquid injection equalization control system, comprising a reflux liquid multi-point liquid inlet dynamic monitoring module, which is used for collecting cooling liquid state data, liquid flow state data and path thermal state data in a current liquid cooling branch, and pre-processing the collected cooling liquid state data, liquid flow state data and path thermal state data to construct a standardized liquid cooling operation feature data set; an immersion liquid temperature stratification monitoring module, which is used for evaluating the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation feature data set, and dynamically adjusting a local disturbance liquid injection strategy based on the evaluation result; a liquid flow distribution equalization evaluation module, which is used for analyzing the liquid flow distribution and heat dissipation matching degree based on the standardized liquid cooling operation feature data set, and driving a flow direction coverage strategy optimization based on the analysis result; a cooling path dynamic determination and switching module, which is used for comprehensively evaluating the current cooling state and thermal load change by taking the thermal distribution unevenness evaluation result and the liquid flow distribution and heat dissipation matching degree analysis result as inputs, and intelligently regulating and controlling the circulation path and heat exchange intervention mode based on the evaluation result; and an immersion liquid cooling cabinet local disturbance self-balancing regulation and control module, which is used for judging the thermal layer structure evolution trend based on the temperature profile and liquid flow distribution state, guiding local convection through bottom disturbance liquid injection, dynamically adjusting the disturbance direction and intensity in combination with the temperature rise feedback and distribution response result, and improving the local diffusion effect of the disturbance liquid by cooperating with the flow direction relaxation control.

[0027] Specifically, the specific steps for collecting the cooling liquid state data, the liquid flow state data and the path thermal state data in the current liquid cooling branch are as follows: The key parameters in the liquid cooling branch are collected and structured, the cooling liquid state data is collected, and the cooling liquid state data includes the cooling liquid temperature of each layer in the immersion liquid cooling cabinet 1, the liquid level height of the temperature sensor, the top cooling liquid temperature and the bottom cooling liquid temperature of the immersion liquid cooling cabinet 1, and the number of sensor layers in the current collection period, the temperature difference and the height difference between each adjacent two layers are recorded, The top cooling liquid temperature and the bottom cooling liquid temperature of the immersion liquid cooling cabinet 1 are used to represent the high temperature limit and the low temperature reference in the liquid heat distribution in the current period to form a representative temperature difference reference data set; in a plurality of continuous collection periods, the top cooling liquid temperature and the bottom cooling liquid temperature are time series arranged, and based on the processed time series, the stable interval of the temperature difference change in the plurality of effective periods is extracted, and the mean value of the stable interval is taken as the temperature difference normalization reference value under the current operation stage; The liquid flow state data is collected, and the liquid flow state data includes the instantaneous flow rate value, the continuous flow rate change data and the liquid injection response time of each region, and the average liquid flow rate in the branch reflux path is recorded; The flow velocity fluctuation time series is constructed by using the continuous flow velocity change data per unit time in each chip area in the immersion liquid cooling cabinet 1. At the same time, the fluctuation degree of the regional liquid per unit time and the consistency of the disturbance response are comprehensively evaluated in combination with the injection response time of the disturbance injector to obtain the distribution uniformity of the coolant in the area. Collect path thermal status data, including the effective heat dissipation area of ​​the chip in each area, the coolant temperature in the return branch, the coolant temperature at the outlet of the immersion liquid cooling cabinet 1, and the chip temperature rise rate. At the same time, record the total number of valid monitoring areas in the current cycle.

[0028] In this implementation plan, a comprehensive perception of the three core parameters of the heat flow path in the liquid cooling cycle is achieved, a representative temperature difference reference data set is constructed based on the coolant temperature gradient, and the thermal distribution characteristics are extracted; the coolant distribution uniformity is calculated jointly by the flow velocity fluctuation and the consistency of the injection response to characterize the flow stability in the local area; at the same time, the heat transfer efficiency between the paths is evaluated by the chip heat dissipation area and the liquid temperature rise difference, providing a high-precision input basis and dynamic response support for the subsequent temperature layer risk identification, liquid flow optimization and control, and path switching strategy.

[0029] Specifically, the collected coolant status data, liquid flow status data, and path thermal status data are preprocessed to construct a standardized liquid cooling operation feature data set. The specific steps are as follows: preprocess the collected coolant status data, liquid flow status data, and path thermal status data, introduce a hierarchical identification index on the basis of retaining the original physical meaning, complete the coordinate standardization processing including the correspondence between the liquid level height and the temperature sensor, and construct a hierarchical thermal data mapping structure; at the same time, perform a rapid window sliding comparison on the temperature mutation points that appear in the acquisition, eliminate the temperature series mutations and inter-layer measurement point misalignment data caused by sensor drift and short-term communication interruption, and ensure the continuity and credibility of the multi-cycle temperature trend.

[0030] Based on the distribution characteristics of the coolant temperature in each layer, combined with the actual piping structure of the immersion liquid cooling cabinet 1 and the coolant flow direction constraints, a distribution constraint matrix is ​​established to assist in determining the temperature anomaly area. A difference repair operation is performed on the detected temperature gradient reversal and abnormal static intervals, prioritizing the preservation of structural continuity and consistency of heat conduction direction, and restoring a continuous and stable temperature profile reflecting the true flow-thermal coupling state.

[0031] Before preprocessing, path thermal status data is classified and mapped according to the chip area topology and injection point mapping relationship, and regional heat flux analysis is used to assist in verifying coverage validity. For chip areas with multi-zone cross-cooling coverage, heat dissipation area splitting is performed to ensure that the energy attribution of each hot zone in the subsequent processing flow is clear and the response path is independent.

[0032] To cope with the data segment conflict caused by the parallel work of multiple liquid injection units in the liquid cooling process, a sliding data window synchronization mechanism based on sampling timing is introduced in the preprocessing stage, combined with the response time difference in the region to analyze the conflict, and the region data with continuously increasing flow fluctuation amplitude is preferentially retained and the repeated response field is removed, thereby improving the data consistency and response feature integrity under the joint processing of multiple regions.

[0033] Finally, the cooling liquid state data, liquid flow state data, and path thermal state data after preprocessing are normalized to unify the numerical scale and distribution form, and a standardized liquid cooling operation feature data set is constructed.

[0034] In this embodiment, the collected cooling liquid state data, liquid flow state data, and path thermal state data are standardized and preprocessed to construct a liquid cooling operation feature data set with continuous structure, accurate distribution, and strong comparability, thereby improving the temperature profile restoration degree, flow response consistency, and thermal load attribution accuracy, and providing high-quality data support for subsequent thermal layer structure identification, local disturbance control, and circulation path intelligent adjustment.

[0035] Specifically, to evaluate the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation feature data set, the specific steps are as follows: collecting real-time measurement data of temperature sensors at each layer inside the immersion liquid cooling cabinet 1, constructing a vertical temperature profile sequence in the current collection period, representing the temperature gradient at the current position with the ratio of the temperature difference to the height difference between each adjacent two layers, and extracting the one with the maximum absolute value among all adjacent measurement points as the local maximum gradient of the profile; at the same time, the upper limit value of the acceptable gradient fluctuation range is obtained by calculating the interquartile range of all temperature gradients, which is denoted as the profile fluctuation reference value; the cooling liquid temperature difference between each adjacent two layers is calculated in turn and divided by the height difference of the liquid surface where the temperature sensors are located, and the ratio of the cooling liquid temperature difference to the height difference of the liquid surface where the temperature sensors are located is taken as the absolute value to obtain the interlayer temperature gradient value; the interlayer temperature gradient values of each layer in the sensor are added and divided by the number of sensor layers to obtain the average interlayer temperature value; the upper and lower temperature difference of the immersion liquid cooling cabinet 1 is obtained by subtracting the bottom cooling liquid temperature from the top cooling liquid temperature and then dividing by the temperature difference reference value; the gradient offset ratio is obtained by dividing the local maximum gradient of the profile by the profile fluctuation reference value; the average interlayer temperature value, the upper and lower temperature difference normalization ratio, and the gradient offset ratio are multiplied to obtain the temperature layer difference risk value.

[0036] The temperature layer difference risk value calculation formula is: ; In the formula: P represents the temperature layer difference risk value. Sensor layer number, representing the total number of temperature collection points vertically arranged in the immersion liquid cooling cabinet 1, used for segmented calculation of temperature gradient distribution, derived from the structure design of the immersion liquid cooling cabinet 1 and the arrangement of the probe; representing the temperature of the cooling liquid of the first layer, used for calculating the temperature difference between adjacent heights, and is the basic data for constructing the vertical temperature profile, derived from the real-time collection value of the temperature sensor of the first layer; representing the temperature of the cooling liquid of the first layer, used for calculating the liquid level distance between each two layers, and then deriving the temperature gradient distribution, derived from the static height parameter table of the internal structure of the immersion liquid cooling cabinet 1; representing the temperature of the cooling liquid at the top, used for representing the temperature state of the uppermost layer of liquid, derived from the real-time measurement value of the temperature sensor at the top region of the immersion liquid cooling cabinet 1; representing the temperature of the cooling liquid at the bottom, used for representing the temperature state of the bottommost layer of liquid, derived from the real-time measurement value of the temperature sensor at the bottom region of the immersion liquid cooling cabinet 1; representing the temperature difference normalization reference value, used for normalizing the temperature difference amplitude between the top and the bottom, so that the risk values under different working conditions are comparable, derived from the mean value of the cooling liquid temperature difference in the steady state interval at the initial stage of the immersion liquid cooling cabinet 1; representing the local maximum gradient of the profile, used for quantifying the temperature mutation phenomenon occurring at a certain height section of the immersion liquid cooling cabinet 1, and is an important basis for judging the position of the thermal stratification space; representing the profile fluctuation reference value, used for measuring the maximum temperature change gradient acceptable in the immersion liquid cooling cabinet 1 within the same monitoring period, and is a relative reference quantity for judging whether the temperature disturbance exceeds the normal interlayer transition range, derived from the statistical fluctuation interval automatic calculation result of the temperature gradient value set of all adjacent measuring points within the monitoring period.

[0037] In this embodiment, by comprehensively analyzing the temperature difference distribution between different vertical layers in the immersion liquid cooling cabinet 1, the profile gradient change and its fluctuation stability over time, it is identified whether the liquid thermal stratification structure is in different evolution states of continuous aggravation, slow convergence and critical balance, and then it is judged whether the cooling liquid is in the case of top high-temperature liquid retention and bottom cooling liquid compensation deficiency due to insufficient backflow path structure and local disturbance.

[0038] Specifically, and based on the evaluation results, the specific steps of dynamically adjusting the local disturbance liquid injection strategy are as follows: Real-time comparison of the current temperature layer difference risk value and the temperature layer difference risk threshold value is used to dynamically identify the thermal distribution evolution trend in the immersion liquid cooling cabinet 1 and drive the adaptive regulation and control of the liquid injection strategy accordingly: When the temperature layer difference risk value is less than or equal to the temperature layer difference risk threshold, the current thermal stratification structure is determined to be within the controllable range. The current branch reflux state is maintained unchanged, the existing liquid inlet point configuration and diversion direction are maintained, and multi-point liquid injection is continued to maintain liquid flow uniformity. During this process, the data collection and calculation tasks of the temperature measurement points on each layer of the immersion liquid cooling cabinet are updated at a fixed period, and the change range of the temperature collection values ​​of each layer is monitored in real time to dynamically evaluate whether the temperature difference micro-fluctuation has an upward trend. When the temperature layer difference risk value is greater than the temperature layer difference risk threshold, it indicates that a heat retention structure has formed inside the immersion liquid cooling cabinet 1. The branch pipe valve 2 is immediately closed, and the main circulation cooling path is switched to prevent high-temperature liquid that has not been treated by the heat exchanger from entering the main cooling circuit in reverse; the current liquid inlet point injection operation is synchronously suspended to block the potential heat conduction reflux path, and then the micro-disturbance liquid injector arranged at the bottom of the immersion liquid cooling cabinet 1 is activated to perform periodic low-speed disturbance injection without affecting the main cooling channel, guiding the cold liquid to float and the local hot liquid to move sideways; at the same time, the state of all injection channels in the immersion liquid cooling cabinet 1 is reconstructed, and the injection path mapping relationship is reconstructed based on the temperature layer distribution, and the disturbance channels are preferentially allocated to the temperature difference concentrated area, thereby improving the efficiency of breaking the thermal layer structure.

[0039] In this implementation, the temperature stratification state in the immersion liquid cooling cabinet 1 is dynamically monitored and the evolution trend of the thermal layer structure is determined in real time. Combined with the comparison result of the temperature layer difference risk value and the preset threshold, the cooling path switching and the disturbance injection strategy are driven to adaptively adjust to ensure balanced heat distribution in the liquid cooling environment and prevent the formation of high-temperature retention areas, thereby improving the thermal control response efficiency and stability in different operating stages.

[0040] Specifically, the specific steps for analyzing the current liquid flow distribution and heat dissipation matching degree based on the standardized liquid cooling operation characteristic data set are as follows: by calculating the ratio of the continuous flow velocity change data of each effective monitoring area in the current cycle to the liquid injection response time, a dynamic indicator set of coolant distribution uniformity is constructed; then, in the dynamic indicator set of coolant distribution uniformity, the median extraction method is used to obtain the normalized reference value of the distribution consistency level in the liquid cooling branch of the current cycle, which is recorded as the coolant distribution uniformity reference value; the instantaneous flow velocity value of the current area, the effective heat dissipation area of ​​the chip and the corresponding coolant density correction factor are multiplied to obtain the regional liquid kinetic energy value of the current area; the distribution uniformity adjustment value is obtained by subtracting the ratio of the distribution uniformity of the coolant in each area to the coolant distribution uniformity reference value and then multiplying it by the distribution deviation weight factor; the regional liquid kinetic energy value is divided by the corresponding distribution uniformity adjustment value to obtain the regional balanced effective liquid flow value, and the regional balanced effective liquid flow values ​​of each effective monitoring area in the current cycle are accumulated to obtain the liquid flow balanced distribution evaluation value of the current cycle.

[0041] The calculation formula for the liquid flow balance distribution evaluation value is: ; wherein: C represents the liquid flow distribution evaluation value; n represents the total number of effective monitoring areas; represents the instantaneous flow rate value of the jth area, used to characterize the variation of the flow speed of the cooling liquid in the area, is the basis parameter for analyzing the local fluid energy dynamics, and is derived from the multi-point micro flow rate sensor arranged in the immersion liquid cooling cabinet 1; represents the effective heat dissipation area of the chip in the jth area, used to estimate the required cooling liquid contact range of the area, is an input parameter for evaluating the distribution characteristics of the cooling demand side, and is derived from the real-time heat flux response of the chip surface and the projection analysis result of the physical geometry; represents the distribution uniformity of the cooling liquid in the jth area, used to reflect the flow rate fluctuation and coverage diffusion of the liquid in the area per unit time, is a reference quantity for identifying whether there is a flow dead angle and local accumulation; represents the cooling liquid distribution uniformity reference value, used to characterize the consistency level of the overall liquid flow distribution in the current period, is a normalized reference for measuring the deviation degree of a single area, and is derived from the ratio of the continuous flow rate change data of each effective monitoring area in the current period to the liquid injection response time, to construct a dynamic index set of cooling liquid distribution uniformity; then, in the dynamic index set of cooling liquid distribution uniformity, the median extraction method is used to obtain the normalized reference value of the distribution consistency level in the current period of the liquid cooling branch; represents the cooling liquid density correction factor of the jth area, with a value range of 0.85 to 1.15, used to correct the difference in inertial response and pressure drop performance of different cooling liquids under the same flow structure, and is derived from the ratio relationship between the instantaneous density value of the target cooling medium under the current temperature and pressure condition and the standard density reference value. In specific calculation, firstly, the local temperature and pressure data of the cooling liquid in each monitoring area are collected, and the actual density value of the area is calculated by combining the refrigeration working medium density state equation; then, the standard density reference value of the medium under the rated operating condition is extracted by calling the material performance database, and the ratio of the two is the cooling liquid density correction factor. The cooling liquid density correction factor reflects the density drift degree of the cooling liquid in the current operating environment compared with the standard working condition. When the actual density is much higher than the standard value, it means that there is local retention of the cooling liquid, and the value of the cooling liquid density correction factor needs to be increased; otherwise, it indicates that the medium diffusion is enhanced, and the value of the cooling liquid density correction factor needs to be reduced to improve the overall distribution flexibility; A distribution deviation weight factor, with a value range of 0.5 to 2, is used to adjust the weight of the spatial consistency and multi-point distribution deviation of the liquid flow in the monitoring area, and is derived from the structural difference analysis results between the velocity vectors in each area in the current monitoring period. In specific implementation, first, the instantaneous flow velocity vector is extracted in each area, and it is mapped to a unified coordinate reference frame, the spatial vector clustering grouping operation is performed, and the flow direction concentration and deviation direction trend are identified; then the velocity center deviation distance, azimuth angle difference and velocity amplitude difference between each cluster are calculated, and the velocity distribution deviation tensor is formed accordingly; finally, the deviation concentration index of the overall velocity field is aggregated and calculated based on the deviation weight distribution between each area and the main flow direction, which is denoted as the distribution deviation weight factor. When the liquid flow distribution is seriously uneven, the flow velocity difference between areas is large, and the cold and hot spots exist obvious imbalance trend, the distribution deviation weight factor takes a larger value, otherwise the liquid flow distribution is basically uniform, the local small difference is acceptable, and the distribution deviation weight factor takes a smaller value, reducing the hypersensitive reaction to slight deviation, maintaining the stability of regulation and control, and avoiding the disturbance of flow field caused by excessive adjustment.

[0042] In the embodiment, the distribution state of the cooling liquid inside the liquid-cooled server between different chip regions is quantitatively evaluated to identify the adaptation degree and flow field balance level between the current liquid flow structure and the heat load. By fusing the instantaneous flow velocity change, disturbance response delay and flow velocity fluctuation standard deviation of each area per unit time, the coverage consistency, disturbance response stability and flow balance of the cooling liquid are dynamically described, and then a comprehensive index reflecting the global liquid flow coordination is constructed. The liquid flow balance evaluation value is used as the core judgment basis for cooling path dynamic optimization, liquid injection strategy adjustment and abnormal state identification, which assists in judging whether there is an uneven phenomenon such as flow velocity deviation, insufficient cold zone coverage and local blockage in the current liquid cooling structure, and provides accurate data support for subsequent disturbance guidance, path switching and cooling reconstruction strategy, thereby improving the self-adaptive regulation and control capability and overall cooling efficiency under complex heat load changes.

[0043] Specifically, and based on the analysis results, the flow direction coverage strategy optimization specific steps are as follows: Based on the calculation results of the liquid flow balance distribution evaluation value, the flow direction control strategy in the liquid cooling structure is dynamically adjusted according to the fluctuation trend of the liquid flow balance distribution evaluation value in the time sequence. The complete process from data acquisition to preprocessing to liquid flow balance distribution evaluation value calculation to result caching in a fixed time window is called a period.

[0044] When the liquid flow equalization distribution evaluation value is steadily rising in three or more consecutive periods, and the growth rate of each period does not exceed the set fluctuation threshold, it indicates that the distribution state of the cooling liquid inside the structure tends to be balanced, the flow path covers the heat source area fully, and the flow field disturbance is low, the thermal load response matching degree is high. At this time, freeze the current flow distribution structure, suspend path reconstruction and valve control adjustment, enter the observation enhancement stage: expand the sampling density of the micro disturbance area around the liquid inlet point, obtain more detailed boundary flow disturbance information through high-frequency sampling, and periodically perform micro low flow disturbance injection test to verify whether there is hidden false steady state risk in the surface stable state, and ensure that the local flow is not hidden by dead water retention and thermal inertia response.

[0045] When the liquid flow equalization distribution evaluation value jumps in two consecutive periods, and the rising amplitude exceeds the fluctuation amplitude threshold, it is determined that the liquid flow covering structure has mutated imbalance, all existing liquid injection points are immediately interrupted, the main circulation cooling path is reconstructed, and the cooling liquid self-balancing starting mechanism is activated. The cooling liquid self-balancing starting mechanism is a rapid recovery process based on flow speed self-adaptive adjustment and liquid injection channel reconstruction, which specifically includes three stages: the initial stage closes all liquid injection modules and branch backflow paths to ensure zero liquid injection stable state; the transition stage opens the main channel circulation at a safe bottom flow rate, and sorts the liquid injection points according to the flow speed average in the running period, and preferentially activates the liquid injection point closest to the target average, to realize gradient recovery; the last stage introduces a delayed boost strategy: after each liquid injection point is restarted, the corresponding flow rate is increased after a delay of one period, to avoid local flow field impact and pipeline fluctuation rebound caused by sudden pressure. At the same time, the spatial distribution and convergence speed of the chip temperature difference are monitored in real time, and once the temperature difference in a certain region is detected to be continuously fluctuating and the flow speed is abnormal, the liquid injection angle and flow path of the region are immediately rearranged to ensure that the newly recovered structure has good dynamic adaptability and heat exchange stability.

[0046] When the liquid flow equalization distribution evaluation value continuously decreases in five consecutive periods, and the temperature of each chip core region synchronously decreases, it indicates that the current liquid cooling structure has reached a relatively optimal balanced cooling state, the trigger frequency of all disturbance liquid injection nozzles is reduced, the holding time of each branch pipeline valve 2 in the main cooling channel is prolonged to reduce the flow field disturbance caused by frequent switching, and a micro pressure difference weak disturbance mechanism is introduced: the boundary heat exchange surface activity and micro circulation ability are maintained through slight periodic pressure fluctuations to prevent the formation of thermal inertia layer due to excessive stability. To prevent the cooling liquid flow from being fixed and causing internal thermal energy accumulation during the stable state process, a flexible flushing process is automatically triggered once every ten periods, the liquid cooling path structure is kept unchanged, a large volume of liquid is injected for a short time to push the cooling liquid to slowly redistribute as a whole, thereby realizing flow direction adjustment and balanced performance reactivation, and ensuring the continuous cooling efficiency and structural flexibility of the liquid cooling structure during operation.

[0047] In this embodiment, based on the dynamic change trend of the liquid flow balance distribution evaluation value, an adaptive liquid cooling flow control mechanism is constructed, so that the circulating cooling process can timely respond to the stability change of the cooling liquid distribution structure, the flow field mutation and the local imbalance, thereby realizing the precise regulation of the heat load matching degree and the timing judgment of the cooling path reconstruction. When the liquid flow balance distribution evaluation value presents a smooth rising trend, by freezing the current flow distribution structure, enhancing the perturbation observation density and executing low-intensity perturbation verification, it is helpful to identify the local pseudo-steady state and stabilize the overall heat dissipation structure; when the liquid flow balance distribution evaluation value suddenly increases and breaks through the threshold, by the self-balancing starting mechanism, the abnormal channel is quickly interrupted and the liquid injection path is reconstructed, which effectively avoids the sudden drop of cooling efficiency and the risk of chip heat aggregation; and in the case that the liquid flow balance distribution evaluation value continuously decreases and is accompanied by synchronous temperature decrease, the steady-state maintenance mode is entered, the micro-pressure difference perturbation and flexible flushing process are introduced, which not only maintains the boundary heat exchange activity, but also prevents the accumulation of thermal inertia, and comprehensively improves the robustness and stability of the liquid cooling circulation structure in the dynamic heat dissipation scene.

[0048] Specifically, for taking the heat distribution unevenness evaluation result and the liquid flow distribution and heat dissipation matching degree analysis result as inputs, the specific steps for comprehensively evaluating the current cooling state and heat load change are as follows: obtaining a temperature layer difference risk value and a liquid flow balance distribution evaluation value, calculating the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet 1, multiplying the temperature layer difference risk value, the liquid flow balance distribution evaluation value and the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet 1 to obtain a temperature difference intensification load term; adding the reciprocal of the average liquid flow rate in the current branch return path to the chip temperature rise rate and then adding one to obtain a liquid flow decay penalty value; taking the logarithm of the temperature difference intensification load term divided by the liquid flow decay penalty value to obtain a path switching judgment value at the current time.

[0049] The path switching judgment value calculation formula is: ; In the formula, represents the path switching judgment value; represents the temperature layer difference risk value; represents the liquid flow balance distribution evaluation value; represents the cooling liquid temperature in the current return branch, which is used to evaluate the remaining heat absorption capacity of the liquid in the path and is a key variable for judging whether there is secondary cooling potential, and is derived from the collection value of the in-branch temperature probe; represents the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet 1, which is used to form a difference with the return liquid temperature and reflects the temperature transfer strength between the cooling paths, and is derived from the measurement result of the real-time temperature sensor at the outlet of the immersion liquid cooling cabinet 1; represents the average liquid flow rate in the current branch backflow path, is used to reflect the liquid patency of the backflow channel, and is an important factor for judging the backflow efficiency, and is derived from the flow rate sensor arranged in the branch pipeline 7 to collect a sequence; represents the chip temperature rising rate, is used to quantify the intensity of the current thermal load change of the chip, is an important parameter for judging whether there is a short-term heat dissipation lag trend, and is derived from the time difference value of the chip surface temperature sensor in two consecutive periods.

[0050] In this embodiment, the temperature layer difference risk value of example 1 is set to 0.36, the liquid flow balance distribution evaluation value is 0.68, the cooling liquid temperature in the backflow branch is 43.5, the cooling liquid temperature of the liquid outlet is 26.1, the average liquid flow rate is 0.94, the chip temperature rising rate is 0.11, and the path switching judgment value is 2.057; The temperature layer difference risk value of example 2 is set to 0.42, the liquid flow balance distribution evaluation value is 0.75, the cooling liquid temperature in the backflow branch is 44.2, the cooling liquid temperature of the liquid outlet is 25.7, the average liquid flow rate is 1.01, the chip temperature rising rate is 0.09, and the path switching judgment value is 2.324; The temperature layer difference risk value of example 3 is set to 0.39, the liquid flow balance distribution evaluation value is 0.72, the cooling liquid temperature in the backflow branch is 41.8, the cooling liquid temperature of the liquid outlet is 26.4, the average liquid flow rate is 0.98, the chip temperature rising rate is 0.10, and the path switching judgment value is 2.025; The temperature layer difference risk value of example 4 is set to 0.33, the liquid flow balance distribution evaluation value is 0.65, the cooling liquid temperature in the backflow branch is 42.6, the cooling liquid temperature of the liquid outlet is 25.9, the average liquid flow rate is 0.93, the chip temperature rising rate is 0.12, and the path switching judgment value is 1.981; The temperature layer difference risk value of example 5 is set to 0.45, the liquid flow balance distribution evaluation value is 0.79, the cooling liquid temperature in the backflow branch is 45.1, the cooling liquid temperature of the liquid outlet is 26.0, the average liquid flow rate is 1.07, the chip temperature rising rate is 0.08, and the path switching judgment value is 2.446; The temperature layer difference risk value of example 6 is set to 0.37, the liquid flow balance distribution evaluation value is 0.70, the cooling liquid temperature in the backflow branch is 43.0, the cooling liquid temperature of the liquid outlet is 25.8, the average liquid flow rate is 0.96, the chip temperature rising rate is 0.11, and the path switching judgment value is 2.094; For Example 7, the temperature gradient risk value was set to 0.40, the flow balance distribution assessment value was 0.67, the coolant temperature in the return branch was 42.2°C, the coolant temperature at the outlet was 25.5°C, the average liquid flow rate was 0.95°C, the chip temperature rise rate was 0.10, and the path switching decision value was 2.146. The path switching decision values ​​for each example were calculated, as shown in Table 1.

[0051] Table 1 Path switching judgment values

[0052] like Figure 2 As shown in Table 1 and Figure 2 It can be seen that the highest path switching judgment value is in case 5, indicating that the temperature layer difference risk of its immersion liquid cooling cabinet 1 is high, the liquid flow is well balanced, the chip temperature rise rate is low, and the temperature difference between hot and cold liquids is large, indicating that there is a significant matching deviation between the coolant flow structure and the heat load in this case, which is prone to imbalance in the distribution of hot and cold spots and reflux delay. It is necessary to prioritize adjusting the injection path and flow ratio and activate the active path reconstruction mechanism to restore dynamic balance; the lowest path switching judgment value is in case 4. Although its chip temperature rise rate is slightly higher, the overall liquid flow distribution stability is good, the temperature gradient of the immersion liquid cooling cabinet 1 is relatively gentle, and the temperature difference between the coolant inlet and outlet is moderate, indicating that the current flow path structure is relatively reasonable, with good heat load carrying capacity and flow stability. Path adjustment can be temporarily postponed, and the existing cooling layout can be maintained to reduce disturbances. The line graph of the path switching judgment value intuitively reflects the dynamic change characteristics of liquid flow stability, thermal coupling coordination and path adaptability in different instances during the cooling process. The higher the value, the greater the path control pressure, and the more priority is needed for dynamic reconstruction and control optimization to improve the overall cooling efficiency and stability.

[0053] Specifically, based on the evaluation results, intelligently controlling the circulation path and heat exchange intervention method involves the following steps: A real-time comparison of the current path switching determination value with the path switching threshold. The path switching thresholds include a first switching threshold and a second switching threshold, which are used to differentiate the flow control level under different heat load pressures. The current path switching determination value is a combination of the temperature layer difference risk value of the immersion liquid cooling cabinet 1, the liquid flow balance distribution assessment value, the chip temperature rise rate, and the temperature difference between the hot and cold liquids. This value reflects the combined strength of the cooling structure imbalance and heat transfer pressure.

[0054] When the path switching judgment value is less than or equal to the second switching threshold, it indicates that the current cooling state is still within the thermal load tolerance range. The current branch return path operation state remains unchanged, the flow distribution structure and diversion direction settings of each liquid inlet point are maintained, the path switching execution is suspended, and only the difference between the return liquid temperature and the chip temperature is periodically sampled at a low frequency, while the diffusion trend of the liquid on the cooling surface is recorded. When the path switching decision value is greater than the second switching threshold and less than or equal to the first switching threshold, it indicates that the current cooling flow field has a certain degree of instability trend, but has not broken through the heat load control limit. At this time, the branch pipeline valve 2 opening is reduced and the main circulating channel pressure is gradually increased, so that the branch flow is suppressed and the cooling capacity of the main channel is improved, the two cooling paths are kept in parallel for a short time, the transitional cooling redundancy structure is constructed, the flow field buffer area is formed, the temperature impact caused by cooling mutation is avoided. And start the dynamic path transfer recording mechanism, continuously collect the flow direction change of each liquid injection point and the temperature difference change of the chip surface, and after the fluctuation trend of the path switching decision value in the next period is clear, it is judged whether to completely switch to the main circulating channel.

[0055] When the path switching decision value is greater than the first switching threshold, it is judged that the cooling imbalance has reached a dangerous level, and the path switching instruction is immediately executed to close the branch backflow channel and completely switch to the main circulating path to preferentially restore the cooling stability of the chip area; At the same time, reset the backflow liquid direction and the liquid injection point weight configuration, and readjust the liquid injection proportion of each area. Suspend all perturbation liquid injection operations to prevent local disturbance and flow disturbance with the main circulation, and activate the main channel liquid velocity equalization constraint logic to dynamically constrain the liquid velocity impact behavior of high-risk areas, ensure high-flow cooling, and avoid abnormal temperature rise at the chip edge through flow direction restriction, further improve the heat exchange efficiency and prevent local hot spot out of control.

[0056] In this embodiment, the cooling path dynamic adjustment process is managed based on the path switching decision value, and an adaptive switching control mechanism is constructed considering the cooling efficiency, flow stability and chip thermal safety. By comparing the path switching decision value with the two thresholds in real time, it is dynamically identified whether there is a heat conduction imbalance and abnormal trend of liquid flow in the cooling structure, and different levels of cooling path intervention operations are triggered accordingly. When the cooling state is stable, maintain the original path and reduce the disturbance frequency; When there is moderate structural fluctuation, construct a parallel buffer path and delay the decision; When the cooling imbalance is serious, immediately execute forced path switching and flow direction reconstruction to ensure the heat exchange stability and cooling efficiency of the chip under high heat load, effectively prevent performance degradation and hardware risks caused by local temperature out of control.

[0057] Specifically, the thermal layer structure evolution trend is judged based on the temperature profile and liquid flow distribution state, and the local convection is guided by the bottom disturbance liquid injection. The disturbance direction and intensity are dynamically adjusted by combining temperature rise feedback and distribution response results, and the local diffusion effect of the disturbance liquid is improved by cooperating with the flow direction relaxation control. The specific steps are as follows: During the opening of the branch backflow path, the temperature profile data inside the immersion liquid cooling cabinet 1 collected by the vertical temperature sensor array is continuously monitored, and the liquid flow uniform distribution evaluation value and the path switching judgment value are jointly analyzed to determine whether there is a trend that the thermal layer structure tends to be stable. When the monitoring results show that the liquid temperature at the top of the immersion liquid cooling cabinet 1 is continuously higher than that at the bottom in multiple sampling periods, and the change trend of the temperature difference of the multiple layers along the vertical direction gradually converges, it is determined that the liquid heat distribution has a potential stable state, and then the disturbance liquid injection response mechanism is started: through the micro jet and the lateral directional channel arranged at the bottom of the immersion liquid cooling cabinet 1, intermittent disturbance cooling liquid injection is implemented at the minimum flow rate without reconstructing the main cooling path and the branch backflow structure, the local liquid slow convection is caused, the cold liquid is slowly raised and the hot liquid is moved laterally, so that the natural decomposition and redistribution of the local thermal layer structure are promoted.

[0058] After each round of disturbance injection is completed, the change of the temperature rise rate of the chip heat area and the dynamic evolution trend of the temperature difference of the multiple layers of the immersion liquid cooling cabinet 1 are analyzed synchronously: when the temperature difference convergence trend is obvious and the liquid flow distribution index is in a continuously rising state, it is determined that the current disturbance strategy is effective, and the disturbance frequency, liquid injection direction and injection rhythm are maintained; otherwise, when the temperature rise changes slightly after disturbance and the liquid flow distribution index does not improve obviously, the structure self-adaptive adjustment state is entered, the angle and liquid injection position of the disturbance jet are actively adjusted, and the flow direction of the main liquid inlet is linked to execute short-time fine adjustment operation, so as to build a new disturbance path to improve the local heat exchange efficiency. At the same time, the flow direction relaxation control strategy is introduced during the disturbance implementation process, the flow rate response change of the main channel is temporarily delayed, so that the disturbance liquid has enough initial diffusion time in the local area, and the intervention ability and adjustment amplitude of the disturbance liquid to the stable state of the thermal layer structure are ensured.

[0059] As shown in Figure 3 The flow chart of the single-phase immersion liquid cooling temperature layering evaluation and disturbance liquid equalization control system of the application is shown in the figure, which shows a cooling path switching control process under the driving of chip heat load. Starting from chip heat generation, the chip continuously releases heat during operation, and the temperature of the liquid cooling medium is monitored in real time after absorbing heat. The system judges whether the temperature of the cooling liquid after absorbing heat meets the standard. If it meets the standard, it enters the branch, directly flows back to the panel and is cooled again, realizes local rapid heat removal, and avoids excessive energy consumption and redundant circulation; if it does not meet the standard, it enters the main path, the cooling liquid passes through the post-circulation link and flows to the heat exchanger for further cooling to ensure that the temperature of the cooling liquid returns to a reasonable range. The liquid after heat exchange enters the monitoring link again to form a complete closed-loop control.

[0060] As shown in Figure 4The single-phase immersion liquid cooling temperature stratification evaluation and disturbance liquid injection equalization control system principle diagram of the application is shown, which shows the working principle structure, and is mainly used for heat exchange management in high heat load scenes including chips and servers, and realizes efficient cooling in an immersion liquid cooling mode. The immersion liquid cooling cabinet 1 internally bears the equipment to be cooled and is completely covered by the cooling liquid, realizing uniform heat absorption; the branch pipeline valve 2 controls the cooling liquid to return to the cooling circuit through the branch pipeline 7 backflow path from the cabinet; the main pipeline valve 3 is used for switching whether the cooling liquid enters the main cooling path, realizing control switching of the branch and the main line; the main liquid outlet pipeline 4 is the main channel for the cooling liquid to flow out of the cabinet and enter the cooling circuit; the pump 5 provides liquid circulation driving force to push the cooling liquid to flow in the pipeline; the heat exchanger 6 is used for releasing the absorbed heat to the external environment, realizing rapid export of heat; the branch pipeline 7 constitutes an auxiliary path for liquid cooling liquid backflow, which can be used for local circulation and secondary cooling. The cooling liquid is driven to flow between the immersion liquid cooling cabinet 1 and the heat exchanger 6 by the pump, and the flow direction is adjusted by the valve, realizing multi-path regulation and control and dynamic cooling strategy switching, and guaranteeing that the equipment still maintains stable heat management capability in a high heat density operation state.

[0061] In the embodiment, based on the joint analysis of the temperature profile change trend and the liquid flow equalization characteristics during the operation of the branch backflow path, it is dynamically identified whether there is a potential risk section in which the thermal layer structure in the immersion liquid cooling cabinet 1 tends to be stable but is not completely decoupled. When it is monitored that the top liquid temperature is higher than the bottom liquid temperature for three consecutive periods and the temperature difference distribution tends to converge, intermittent liquid injection intervention is implemented by the bottom micro-jet and the lateral directional channel with the minimum disturbance, without affecting the main circulation path and the backflow channel structure, triggering local liquid slow convection, realizing cold liquid floating and hot liquid horizontal diffusion, and thus actively breaking the false steady-state thermal layer accumulation phenomenon. The disturbance effect is judged by the real-time feedback of the chip thermal zone temperature rise rate and the multi-layer temperature difference change trend of the immersion liquid cooling cabinet 1 after the disturbance liquid injection, the local heat exchange efficiency is improved, the continuous accumulation of the high-temperature retention area is inhibited, and the dynamic steady-state regulation and control capability of the overall liquid cooling structure is guaranteed.

[0062] It should be noted that, in this document, the terms such as first and second are merely used to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0063] The preferred embodiments of the application disclosed above are only to facilitate the elucidation of the application. The preferred embodiments do not describe all the details of the application and limit the application to the specific embodiments described. Obviously, many modifications and variations can be made in light of the teachings above. The description is chosen and described in order to best explain the principles of the application and its practical application to thereby enable others skilled in the art to best utilize the application and get the best results from the application. The application is only limited by the claims and their full scope and equivalents.

Claims

1. A single-phase immersion liquid cooling temperature stratification assessment and perturbation liquid injection equalization control system, characterized by: The application relates to a liquid cooling system and a liquid cooling method. The application comprises: A backflow liquid multi-point injection dynamic monitoring module is used for collecting cooling liquid state data, liquid flow state data and path heat state data in a current liquid cooling branch, pre-processing the collected cooling liquid state data, liquid flow state data and path heat state data, and constructing a standardized liquid cooling operation characteristic data set; An immersed liquid temperature stratification monitoring module is used for evaluating the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation characteristic data set, and dynamically adjusting a local disturbance liquid injection strategy based on the evaluation result; A liquid flow distribution balance evaluation module is used for analyzing the cooling liquid flow distribution and heat dissipation matching degree based on the standardized liquid cooling operation characteristic data set, and driving a flow direction coverage strategy optimization based on the analysis result; A cooling path dynamic determination and switching module is used for comprehensively evaluating the current cooling state and heat load change by taking the thermal distribution unevenness evaluation result and the cooling liquid flow distribution and heat dissipation matching degree analysis result as inputs, and intelligently regulating and controlling the circulation path and heat exchange intervention mode based on the evaluation result; 2. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: An immersed liquid cooling cabinet local disturbance self-balancing regulation and control module is used for judging the thermal layer structure evolution trend based on the temperature profile and liquid flow distribution state, guiding local convection through bottom disturbance liquid injection, dynamically adjusting the disturbance direction and intensity by combining temperature rise feedback and distribution response results, and improving the local diffusion effect of the disturbance liquid by cooperating with the flow direction relaxation control. The specific steps for collecting the cooling liquid state data, liquid flow state data and path heat state data in the current liquid cooling branch are as follows: Key parameters in the liquid cooling branch are collected to obtain the cooling liquid state data, including the cooling liquid temperature of each layer in the immersed liquid cooling cabinet, the liquid level height of the temperature sensor, the top cooling liquid temperature and the bottom cooling liquid temperature of the immersed liquid cooling cabinet, and the number of sensor layers, the temperature difference and the height difference between each adjacent two layers in the current collection period are recorded; The top cooling liquid temperature and the bottom cooling liquid temperature of the immersed liquid cooling cabinet are used to represent the high-temperature limit and the low-temperature reference in the liquid heat distribution in the current period to form a representative temperature difference reference data set; In a plurality of continuous collection periods, the top cooling liquid temperature and the bottom cooling liquid temperature are time series arranged, and the stable interval of the temperature difference change in the plurality of effective periods is extracted based on the processed time series, and the mean value of the stable interval is taken as the temperature difference normalization reference value in the current operation stage; The liquid flow state data is collected, including the instantaneous flow rate value, the continuous flow rate change data and the liquid injection response time of each region, and the average liquid flow rate in the branch backflow path is recorded; The continuous flow rate change data of each chip region in the immersed liquid cooling cabinet in unit time is used to construct a flow rate fluctuation time series, and the fluctuation degree and disturbance response consistency of the regional liquid in unit time are comprehensively evaluated to obtain the distribution uniformity of the cooling liquid in the region. Collect path thermal state data, the path thermal state data includes: each region of the chip effective heat dissipation area, the backflow branch internal cooling liquid temperature, the immersion liquid cooling cabinet outlet liquid temperature and the chip temperature rising rate, record the total number of effective monitoring regions in the current period.

3. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The collected cooling liquid state data, liquid flow state data and path thermal state data are preprocessed, and a standardized liquid cooling operation feature data set is constructed. The collected cooling liquid state data, liquid flow state data and path thermal state data are preprocessed, and the coordinate standardization including the correspondence relationship between the liquid level and the temperature sensor is completed, and the temperature sequence mutation and interlayer measurement point dislocation caused by sensor drift and short-time communication interruption are eliminated; for the distribution characteristics of the cooling liquid temperature of each layer, combined with the actual pipe arrangement structure of the immersion liquid cooling cabinet and the cooling liquid flow direction constraint, the temperature gradient inversion and abnormal static interval are differentially repaired to restore the continuous and stable temperature profile; The path thermal state data is classified and mapped according to the chip region topology and the liquid injection point mapping relationship before preprocessing, and the heat dissipation area splitting processing is performed on the chip area with multiple region intersection cooling coverage; In order to cope with the data segment conflict caused by the parallel work of multiple liquid injection units in the liquid cooling process, a data window synchronization mechanism is introduced in the preprocessing stage, and the region data with continuously increasing flow fluctuation amplitude is preferentially retained and the repeated response field is eliminated; All the preprocessed cooling liquid state data, liquid flow state data and path thermal state data are normalized to construct a standardized liquid cooling operation feature data set.

4. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The steps for evaluating the thermal distribution unevenness of the cooling liquid in the vertical direction based on the standardized liquid cooling operation feature data set are as follows: Collect real-time measurement data of temperature sensors of each layer inside the immersion liquid cooling cabinet, construct a vertical temperature profile sequence in the current collection period, express the temperature gradient of the current position by the ratio of the temperature difference value and the height difference value between each adjacent two layers, and extract the maximum absolute value among all adjacent measurement points as the local maximum gradient of the profile; at the same time, by calculating the interquartile range of all temperature gradients, the upper limit value of the acceptable gradient fluctuation range is obtained, which is recorded as the profile fluctuation reference value; The cooling liquid temperature difference between each adjacent two layers is calculated in turn and divided by the liquid level height difference of the temperature sensors in the adjacent two layers, and the ratio of the cooling liquid temperature difference and the liquid level height difference of the temperature sensors is taken as the absolute value to obtain the interlayer temperature gradient value; The interlayer temperature gradient values of each layer of the sensor are added and divided by the number of sensor layers to obtain the average interlayer temperature value; The top cooling liquid temperature of the immersion liquid cooling cabinet is subtracted from the bottom cooling liquid temperature and then divided by the temperature difference normalization reference value to obtain the up-down temperature difference normalization ratio; The profile local maximum gradient is divided by the profile fluctuation reference value to obtain the gradient offset ratio; The average interlayer temperature value, the up-down temperature difference normalization ratio and the gradient offset ratio are multiplied to obtain the temperature layer difference risk value.

5. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The specific steps for dynamically adjusting the local disturbance liquid injection strategy based on the evaluation result are as follows: Real-time comparison of the current temperature layer difference risk value and the temperature layer difference risk threshold value: When the temperature layer difference risk value is less than or equal to the temperature layer difference risk threshold value, the current branch backflow state is maintained unchanged, the existing liquid inlet point configuration and flow direction are kept, the multi-point liquid injection operation is continued to be performed, the change range of the temperature acquisition value of each layer is monitored, and the data acquisition and calculation of each layer temperature measuring point are updated at a fixed period; When the temperature layer difference risk value is greater than the temperature layer difference risk threshold value, the branch pipeline valve is immediately closed, the main circulating cooling path is switched, the current liquid inlet point liquid injection operation is synchronously paused, the bottom micro-disturbance liquid injector is activated to perform periodic low-speed disturbance injection, and the internal liquid injection path mapping relationship of the immersion liquid cooling cabinet is reconstructed.

6. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The specific steps for analyzing the matching degree of the current liquid flow distribution and heat dissipation based on the standardized liquid cooling operation characteristic data set are as follows: A cooling liquid distribution uniformity dynamic index set is constructed by calculating the ratio of the continuous flow rate change data of each effective monitoring area in the current period to the liquid injection response time. Then, in the cooling liquid distribution uniformity dynamic index set, a median extraction method is used to obtain a normalized reference value of the distribution consistency level in the current period liquid cooling branch, which is denoted as the cooling liquid distribution uniformity reference value. The instantaneous flow rate value of the current area, the effective heat dissipation area of the chip and the corresponding cooling liquid density correction factor are multiplied to obtain the regional liquid flow kinetic energy value of the current area. The distribution uniformity adjustment value is obtained by multiplying the distribution uniformity of each area by the cooling liquid distribution uniformity reference value and then multiplying the distribution deviation weight factor. The regional balanced effective liquid flow value is obtained by dividing the regional liquid flow kinetic energy value by the corresponding distribution uniformity adjustment value. The regional balanced effective liquid flow values of each effective monitoring area in the current period are accumulated to obtain the liquid flow balanced distribution evaluation value of the current period.

7. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The specific steps for driving the flow coverage strategy optimization based on the analysis result are as follows: Based on the calculation result of the liquid flow balanced distribution evaluation value, the flow control strategy is dynamically adjusted according to the fluctuation trend of the real-time liquid flow balanced distribution evaluation value: When the liquid flow balanced distribution evaluation value shows an upward trend in three or more consecutive periods and the upward amplitude does not exceed the fluctuation amplitude threshold value, it indicates that the cooling liquid flow structure tends to be stable and has a high matching degree with the heat load. The current flow distribution structure is frozen and the path reconstruction is paused, and the observation enhancement stage is entered: the sampling density of the micro-disturbance area around the liquid inlet point is expanded, the micro low-flow disturbance liquid injection test is regularly performed, and the current stable state is verified to be a real stable state to avoid local false stable state failure; When the upward amplitude of the liquid flow balanced distribution evaluation value exceeds the fluctuation amplitude threshold value in two periods, it is determined that the liquid flow coverage structure is suddenly unbalanced. The current all liquid injection points are interrupted, the main circulating path is reset, the cooling liquid self-balancing starting mechanism is activated, the flow rate average value of each liquid injection point is sorted for gradient recovery, the delayed boost strategy is enabled to avoid sudden pressure impact, and the spatial distribution and convergence speed of the chip temperature difference are monitored in real time during the recovery process. When a disorder area is found, the liquid inlet angle and flow path are forcibly rearranged. When the liquid flow uniform distribution evaluation value shows a downward trend in five consecutive cycles, and the chip temperature decreases synchronously, it is determined that the current cooling liquid layout has reached an optimal uniform structure, the disturbance nozzle frequency is reduced, the total pipe valve holding time is extended, and a micro-pressure-difference weak disturbance mechanism is introduced to maintain the boundary heat exchange activity, while a flexible flushing is automatically performed every ten cycles to achieve flow redistribution through short-time large-volume liquid injection to prevent thermal inertia accumulation.

8. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The specific steps for comprehensively evaluating the current cooling state and thermal load change based on the heat distribution unevenness evaluation result and the liquid flow distribution and heat dissipation matching degree analysis result are as follows: Obtain the temperature layer difference risk value and the liquid flow uniform distribution evaluation value, calculate the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet, multiply the temperature layer difference risk value, the liquid flow uniform distribution evaluation value, and the absolute value of the difference between the cooling liquid temperature in the current return branch and the cooling liquid temperature at the outlet of the immersion liquid cooling cabinet to obtain a temperature difference intensification load term; Add the chip temperature rise rate to the reciprocal of the average liquid flow rate in the current branch return path and then add one to obtain a liquid flow decay penalty value; Divide the temperature difference intensification load term by the liquid flow decay penalty value and take the logarithm to obtain the path switching decision value at the current time.

9. The single phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill equalization control system of claim 1, wherein: The specific steps for intelligently regulating the circulation path and heat exchange intervention mode based on the evaluation result are as follows: Real-time comparison of the current path switching decision value and the path switching threshold value, the path switching threshold value includes a first switching threshold value and a second switching threshold value: When the path switching decision value is less than or equal to the second switching threshold value, the current branch return path operating state is maintained unchanged, the flow distribution structure and flow direction setting of each liquid inlet point are maintained, the path switching execution is suspended, only the return liquid temperature and chip temperature difference are periodically sampled at low frequency, and the liquid diffusion trend on the cooling surface is recorded, and the next cycle is executed for judgment; When the path switching decision value is greater than the second switching threshold value and less than or equal to the first switching threshold value, the branch pipe valve opening is adjusted and the main circulation channel pressure is gradually increased, the two cooling paths are operated in parallel for a short time, the flow field buffer area is established while ensuring the stability of the chip heat exchange efficiency, and the dynamic path transfer recording mechanism is started to determine whether to switch to the main circulation path completely according to the path switching decision value change trend of the next cycle; When the path switching decision value is greater than the first switching threshold value, the path switching instruction is immediately executed, the branch return channel is closed, the main circulation path is switched to forced cooling, the return liquid direction and liquid inlet point weight configuration are reset, all disturbance liquid injection operations are suspended, and the main channel liquid velocity uniformity constraint logic is activated to ensure high flow cooling while limiting the concentrated injection to the overloaded area, thereby improving the heat exchange efficiency and preventing local temperature rise at the chip edge caused by high flow rate impact.

10. The single-phase immersion liquid cooling temperature stratification assessment and perturbation liquid fill-equalization control system of claim 1, wherein: The specific steps for judging the thermal layer structure evolution trend based on the temperature profile and liquid flow distribution state, guiding local convection through bottom disturbance injection, dynamically adjusting the disturbance direction and intensity based on the temperature rise feedback and distribution response result, and improving the local diffusion effect of the disturbance liquid by flow direction relaxation control are as follows: During the opening of the branch backflow path, the temperature profile data inside the immersion liquid cooling cabinet collected by the vertical temperature sensor array is continuously monitored, and the trend of the thermal stratification structure tends to be stable is analyzed and judged in combination with the liquid flow uniform distribution evaluation value and the path switching judgment value; When it is monitored that the top liquid temperature of the immersion liquid cooling cabinet is higher than the bottom for three consecutive periods, and the multi-layer temperature difference changes converge, a micro jet and a lateral directional channel arranged at the bottom of the immersion liquid cooling cabinet are used to intermittently inject disturbance cooling liquid at the minimum flow rate without changing the main cooling path and the backflow path structure, to cause local liquid slow convection, make the cold liquid float up and the hot liquid move sideways. After each round of disturbance injection, the chip thermal zone temperature rise rate and the multi-layer temperature difference change trend of the immersion liquid cooling cabinet are analyzed, when the temperature difference converges significantly and the liquid flow distribution index shows an upward trend, the current disturbance frequency and direction are maintained; when the liquid flow distribution index does not change, the structure adaptive adjustment state is entered, the disturbance jet angle is rotated and the liquid injection position is changed to build a new disturbance path, and the liquid inlet flow direction is adjusted for a short time; at the same time, the response of the main channel flow rate change is temporarily delayed during the disturbance process to ensure that the disturbance liquid can complete the initial upward diffusion in the local area.

Images