Control methods and electronic equipment for liquid cooling systems

Abstract

This application provides a control method and electronic device for a liquid cooling system, which can be applied to the field of heat dissipation technology. The control method includes: determining the load level information of the electronic device within a target time period based on performance index data; preprocessing the target data to obtain multiple sub-target data frames; performing linear prediction encoding on each of the multiple sub-target data frames to obtain multiple sub-linear prediction coefficients; matching benchmark data from a preset health benchmark library based on the load level information; comparing the multiple sub-target data frames and multiple sub-linear prediction coefficients with the matched benchmark data to determine the residual energy and coefficient distance corresponding to the target data; determining health indicators based on the residual energy and coefficient distance corresponding to the target data; generating early warning information based on the health indicators and a preset health threshold; and using the early warning information to prompt control operations corresponding to the early warning information on the liquid cooling system.

Description

Technical Field

[0001] This application relates to the field of heat dissipation technology, specifically to a control method and electronic device for a liquid cooling system. Background Technology

[0002] Existing liquid cooling systems typically employ a passive control strategy based on fixed thresholds, using parameters such as temperature and flow rate to perform simple equipment start-up, shutdown, or speed adjustment.

[0003] This approach fails to identify early-stage performance degradation in the system, only responding when a fault becomes apparent, resulting in control lag, low energy efficiency, and reliability risks. Therefore, there is an urgent need for a liquid cooling system control scheme that combines intelligent early warning with proactive control. Summary of the Invention

[0004] In view of the above problems, this application provides a control method and electronic equipment for a liquid cooling system to improve the operational reliability of the liquid cooling system.

[0005] According to a first aspect of this application, a control method for a liquid cooling system is provided, comprising: determining load level information of an electronic device within a target time period based on performance index data of the electronic device within the target time period; preprocessing the target data to obtain multiple sub-target data frames, wherein the target data includes at least one of sound data and vibration data; performing linear prediction encoding on the multiple sub-target data frames respectively to obtain multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames; matching benchmark data from a preset health benchmark library based on the load level information; comparing the multiple sub-target data frames and the multiple sub-linear prediction coefficients with the matched benchmark data respectively to determine the residual energy and coefficient distance corresponding to the target data; determining at least one health indicator based on the residual energy and coefficient distance corresponding to the target data; and generating at least one warning message based on the at least one health indicator and a preset health threshold, wherein the warning message is used to prompt control operations corresponding to the warning message to be performed on the liquid cooling system.

[0006] A second aspect of this application provides a control device for a liquid cooling system, comprising: a load information determination module, used to determine the load level information of an electronic device during a target time period based on performance index data of the electronic device during the target time period; a preprocessing module, used to preprocess target data to obtain multiple sub-target data frames, the target data including at least one of sound data and vibration data; a coefficient prediction module, used to perform linear prediction encoding on the multiple sub-target data frames respectively to obtain multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames; a benchmark matching module, used to match benchmark data from a preset health benchmark library based on the load level information; a data comparison module, used to compare the multiple sub-target data frames and the multiple sub-linear prediction coefficients with the matched benchmark data respectively to determine the residual energy and coefficient distance corresponding to the target data; and an early warning generation module, used to determine at least one health indicator based on the residual energy and coefficient distance corresponding to the target data, and generate at least one early warning message based on the at least one health indicator and a preset health threshold.

[0007] A third aspect of this application provides an electronic device, a memory configured to store instructions, and a processor connected to the memory, configured to execute instructions to perform the following operations: determining the load level information of the electronic device within a target time period based on performance index data of the electronic device within a target time period; preprocessing the target data to obtain multiple sub-target data frames, the target data including at least one of sound data and vibration data; performing linear prediction coding on the multiple sub-target data frames respectively to obtain multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames; matching benchmark data from a preset health benchmark library based on the load level information; comparing the multiple sub-target data frames and the multiple sub-linear prediction coefficients with the matched benchmark data respectively to determine the residual energy and coefficient distance corresponding to the target data; determining at least one health indicator based on the residual energy and coefficient distance corresponding to the target data; and generating at least one warning message based on the at least one health indicator and a preset health threshold, the warning message being used to prompt control operations corresponding to the warning message on the liquid cooling system.

[0008] A fourth aspect of this application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0009] In this embodiment, the load level is determined based on the performance indicators of the electronic device; sound and vibration data are preprocessed to obtain sub-target data frames; sub-linear prediction coefficients are extracted through linear predictive coding; benchmark data is matched from a health benchmark library based on the load level; residual energy and coefficient distance are calculated by comparing real-time data with benchmark data; and early warning information is generated based on residual characteristics, and corresponding control operations are executed. This achieves intelligent monitoring and active control of the liquid cooling system, significantly improving the reliability and energy efficiency of the liquid cooling system. Attached Figure Description

[0010] The above-mentioned contents, other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0011] Figure 1 This illustration schematically depicts an application scenario of a control method for a liquid cooling system and an electronic device according to embodiments of this application.

[0012] Figure 2 A flowchart illustrating a control method for a liquid cooling system according to an embodiment of this application is shown schematically.

[0013] Figure 3 A flowchart illustrating the preprocessing of target data according to an embodiment of this application is shown schematically;

[0014] Figure 4 This schematically illustrates a system architecture diagram of a control method for a liquid cooling system according to an embodiment of this application;

[0015] Figure 5 This schematic diagram illustrates the structural block diagram of a control device for a liquid cooling system according to an embodiment of this application;

[0016] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a control method for a liquid cooling system according to an embodiment of this application. Detailed Implementation

[0017] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely examples and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0018] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0019] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0020] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0021] Embodiments of this application provide a control method and electronic device for a liquid cooling system.

[0022] Figure 1 The illustration shows an application scenario of the control method and electronic device of the liquid cooling system according to an embodiment of this application.

[0023] like Figure 1 As shown, the application scenario 100 according to this embodiment includes an interactive scenario of liquid cooling system operation monitoring and intelligent control, specifically including electronic device 101, network 102, and intelligent control platform 103. Network 102 is the communication medium between electronic device 101 and intelligent control platform 103, and may include connection types such as wired, wireless communication links, or fiber optic cables.

[0024] In this application scenario, electronic device 101 is a liquid-cooled system equipped with monitoring components such as acoustic sensors and vibration sensors. Intelligent control platform 103 acquires real-time operational data uploaded by electronic device 101 through a data acquisition interface, including performance index data, sound data, and vibration data.

[0025] The intelligent control platform 103 determines the equipment load level information based on performance index data, preprocesses the acoustic and vibration data to obtain multiple sub-target data frames, extracts sub-linear prediction coefficients through linear prediction coding, matches benchmark data from the health benchmark library based on load level information, calculates residual energy and coefficient distance by comparing real-time data with benchmark data, generates early warning information and executes corresponding control operations based on the comparison results of health indicators and thresholds, thereby realizing intelligent monitoring and active control of the liquid cooling system.

[0026] The following will be based on Figure 1 The described scene, through Figures 2-4 The control method of the liquid cooling system according to the disclosed embodiments will be described in detail.

[0027] Figure 2 A flowchart illustrating a control method for a liquid cooling system according to an embodiment of this application is shown.

[0028] like Figure 2 As shown, the control method of the liquid cooling system in this embodiment includes operations S210 to S260.

[0029] In operation S210, the load level information of the electronic device during the target time period is determined based on the performance index data of the electronic device during the target time period.

[0030] In this embodiment of the application, the electronic device can be a high-performance computing device using a liquid cooling medium. The target time period is a preset time window for load status assessment. The performance index data includes at least one quantitative parameter reflecting the operating status of the electronic device, such as CPU utilization, memory utilization, and input / output utilization. Based on the performance index data within the target time period, load level information characterizing the workload of the electronic device is obtained by dividing the data according to preset rules.

[0031] In operation S220, the target data is preprocessed to obtain multiple sub-target data frames.

[0032] In this embodiment, the target data includes at least one of sound data collected by an acoustic sensor and vibration data acquired by a vibration sensor; the preprocessing process performs format unification, noise filtering and segmentation optimization on the original monitoring data to generate sub-target data frames that meet the requirements of short-term stationarity; each sub-target data frame contains multiple data samples arranged according to the acquisition time, and the data samples are digital values ​​obtained by sampling continuous signals at discrete time points, which characterize the instantaneous amplitude of the target data at a specific acquisition time.

[0033] In operation S230, linear prediction coding is performed on multiple sub-target data frames to obtain multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames.

[0034] In this embodiment, linear prediction coding represents the current value of the signal as a linear combination of a finite number of historical values ​​by establishing a linear prediction model; the sub-linear prediction coefficients are a set of prediction weight parameters obtained after encoding each sub-target data frame, reflecting the short-term characteristics of the target data frame signal.

[0035] When operating S240, benchmark data is matched from a preset health benchmark library based on load level information.

[0036] In this embodiment, the preset health benchmark library stores multiple candidate benchmark data corresponding to various load scenarios defined by preset temperature data groups and preset flow data groups under different load levels. The load level information refers to the classification result of server workload based on performance index data such as CPU utilization, memory utilization and input / output utilization, calculated through preset weight relationships.

[0037] The preset health baseline library can refer to a database that is pre-established under the normal operating conditions of the liquid cooling system. It stores a set of reference data corresponding to various load scenarios defined by preset temperature data sets and preset flow data sets under different load levels. The baseline data can refer to the standardized reference values ​​stored in the preset health baseline library that characterize the various monitoring parameters of electronic equipment under the health state, including but not limited to the baseline data sequence and the baseline linear prediction coefficient.

[0038] In operation S250, multiple sub-target data frames and multiple sub-linear prediction coefficients are compared with multiple matched benchmark data to determine the residual energy and coefficient distance corresponding to the target data.

[0039] In the embodiments of this application, the residual energy is characterized by the degree of difference between each sub-target data frame and the corresponding reference data sequence; the coefficient distance is characterized by the deviation of model parameters obtained by measuring the degree of difference between the sub-linear prediction coefficients and the reference linear prediction coefficients.

[0040] In operation S260, based on the residual energy and coefficient distance corresponding to the target data, at least one health indicator is determined. Based on the at least one health indicator and a preset health threshold, at least one warning message is generated. The warning message is used to prompt the liquid cooling system to perform control operations corresponding to the warning message.

[0041] In this embodiment, the health index refers to a quantitative parameter used to characterize the health status of an electronic device, generated by fusing residual energy and coefficient distance; the preset health threshold refers to a boundary value set during the system debugging phase based on historical operating data to determine the health status level; and the warning information refers to a graded alarm signal generated when the health index exceeds the preset health threshold to indicate potential fault risks.

[0042] In this embodiment, the load level is determined based on the performance indicators of the electronic device; the acoustic and vibration data are preprocessed to obtain sub-target data frames; sub-linear prediction coefficients are extracted through linear predictive coding; benchmark data is matched from a health benchmark library based on the load level; residual energy and coefficient distance are calculated by comparing real-time data with benchmark data; and early warning information is generated based on residual characteristics, and corresponding control operations are executed. Through linear predictive coding and adaptive load matching, intelligent monitoring and active control of the liquid cooling system are achieved, significantly improving the system's reliability and energy efficiency.

[0043] The control methods for the liquid cooling system, including operations S210 to S260, are described in detail below.

[0044] In this embodiment of the application, the above operation 210 may further include: calculating a comprehensive load coefficient based on computing resource utilization, storage resource utilization, and data transmission resource utilization according to a predetermined weighting relationship; determining the load level information as heavy load when the comprehensive load coefficient is greater than or equal to a heavy load threshold; determining the load level information as medium load when the comprehensive load coefficient is less than a heavy load threshold but greater than or equal to a medium load threshold; and determining the load level information as light load when the comprehensive load coefficient is less than a medium load threshold.

[0045] In this embodiment, the computing resource utilization, storage resource utilization, and data transmission resource utilization can be the CPU utilization, memory utilization, and I / O utilization of the electronic device, respectively. A comprehensive load coefficient is calculated according to a predetermined weighting relationship, where the CPU and I / O utilization have higher weights, and the memory utilization has a relatively lower weight. The load level is divided according to the numerical range of the comprehensive load coefficient. When the coefficient reaches the heavy load threshold, it is determined as heavy load; when it is in the medium load threshold range, it is determined as medium load; and when it is below the medium load threshold, it is determined as light load.

[0046] For example, CPU utilization It can reflect the server's computing load and memory utilization. This reflects the server's data caching and program execution memory usage; input / output utilization. It reflects the level of activity in data reading, writing, and data transmission.

[0047] Calculate the load level information using the following formula. :

[0048]

[0049] Among them, load level information The range is 0-10, when The range is When the load level information is determined to be heavy load; when the range of K is within When the load level information is determined to be moderate load; when K ranges within... At that time, the load level information was determined to be light load.

[0050] In this embodiment, by assigning different weights to computing resource utilization, storage resource utilization, and data transmission resource utilization, a comprehensive load coefficient is calculated, and threshold ranges for heavy, medium, and light loads are set, thereby achieving a refined classification of the operating load of electronic devices. This solves the problem that a single indicator in the existing technical solution cannot fully reflect the true load status of electronic devices, and provides an accurate basis for matching health benchmark models under different load scenarios, thereby improving the adaptability of the liquid cooling system to complex operating conditions.

[0051] Figure 3 A flowchart illustrating the preprocessing of target data according to an embodiment of this application is shown.

[0052] like Figure 3 As shown, the above operation S220 may also include operations S301 to S303.

[0053] In operation S301, in response to the instantaneous amplitude of the target data being greater than the preset amplitude range, the length and position of the preset segmentation window are adjusted to segment the target data into multiple initial data frames.

[0054] For example, when the instantaneous amplitude of the target data is found to be greater than the preset amplitude range, the dynamic window length is adjusted based on the short-time Fourier transform. The original signal is scanned by sliding a Hamming window, with an initial window length of 20ms. The initial window length corresponds to the fundamental frequency period of the main mechanical components of the electronic device, such as 20ms for a 50Hz cold pump. The spectral entropy of each frame of signal is calculated. The spectral entropy reflects the dispersion of frequency distribution. When the fluctuation of the spectral entropy value is greater than a threshold, such as 5%, the initial window length is shortened to 10ms until the spectral entropy value stabilizes. The overlap rate of adjacent frames is always kept at 50% of the initial window length. For example, with an initial window length of 20ms, there is a 10ms overlap, and with an initial window length of 10ms, there is a 5ms overlap.

[0055] In operation S302, multiple initial data frames are decomposed into components in different frequency bands, and the components within the preset frequency band range are amplitude compressed to obtain multiple denoised data frames.

[0056] For example, wavelet transform is used to perform multi-level decomposition on the initial data frame obtained from the segmentation, and wavelet basis functions are used to perform three-level decomposition processing on the signal. For each frequency band component obtained after decomposition, soft thresholding suppression is performed on the noise components with frequencies higher than 2kHz. The components subjected to soft thresholding suppression mainly include turbulence noise components generated by electronic equipment during operation.

[0057] In operation S303, the variance ratio of the first half to the second half of multiple denoised data frames is calculated respectively, and the data frames whose variance ratio is in the preset stable range are determined as sub-target data frames.

[0058] For example, a variance ratio test mechanism can be used to calculate the variance ratio of the first half to the second half of multiple denoised data frames. If the variance ratio is within a certain range... If the frame is within the specified range, it is considered a stable frame; otherwise, it is marked as an abnormal frame and reframed.

[0059] In this embodiment, the preprocessing procedure for non-stationary signals includes three steps: adaptive framing, denoising, and variance ratio stationarity verification. This process converts the original signal into sub-data frames that meet short-time stationarity requirements, effectively overcoming the interference of the non-stationary characteristics of the acoustic and vibration signals corresponding to the liquid cooling device on linear prediction modeling. This provides high-quality input for subsequent linear prediction coding and improves the accuracy of fault feature extraction.

[0060] In this embodiment of the application, the above operation 230 may further include: for any sub-target data frame, performing the following operations: establishing a linear prediction model based on the statistical correlation between each data sample point in the sub-target data frame, wherein the linear prediction model is used to represent each data sample point in the sub-target data frame as a linear combination of a finite number of preceding historical data samples; determining the weights corresponding to each preceding historical data sample point in the linear combination by solving the regular equation of the linear prediction model, and using the weights as the sub-linear prediction coefficients corresponding to the sub-target data frame.

[0061] In this embodiment, a forward linear prediction model is established for a sub-target data frame containing n data samples. The forward linear prediction model represents the nth data sample s(n) as a linear combination of the previous p historical samples:

[0062]

[0063] in, arrive To find the linear prediction coefficients, the optimal prediction coefficients are solved using the least squares method. Specifically, the autocorrelation function of the signal is calculated, and a system of linear equations based on the autocorrelation function is constructed. The autocorrelation recursive algorithm is used to solve the system of linear equations, and the linear prediction coefficients are calculated and determined sequentially.

[0064] In this embodiment of the application, by establishing a linear prediction model based on statistical correlation, the current data sample is represented as a linear combination of a finite number of historical samples, and the optimal weight coefficient is solved using a regular equation to quantify and extract the intrinsic correlation features of the signal. The generated predictive coding coefficient can be used as a key feature parameter characterizing the health status of electronic devices.

[0065] In this embodiment of the application, the above operation 240 may further include: locating the corresponding load level partition in the preset health benchmark library according to the load level information, wherein the load level partition has multiple candidate benchmark data stored in advance; performing weighted processing on the multiple candidate benchmark data to determine the benchmark data that matches the load level information.

[0066] In this embodiment, the corresponding load level partition in the health benchmark library is located based on the real-time load level information. The load level partition has multiple candidate benchmark data stored in advance. A weighted statistical algorithm is used to fuse the candidate benchmark data to generate the benchmark data that best matches the current load state.

[0067] For example, when the current load level is identified as heavy load, the system automatically locates the data partition marked as heavy load in the health benchmark database. This data partition stores approximately 10 sets of benchmark data collected under historical health conditions. The weights are determined by calculating the correlation coefficient between each set of benchmark data and the current operating characteristics. A weighted average algorithm is then used to generate the benchmark linear prediction coefficients, where the weight coefficients are dynamically adjusted based on the data collection time and data quality.

[0068] In this embodiment, by establishing a mapping relationship between load levels and benchmark data, the corresponding partition is located based on the real-time load level, and multiple candidate benchmark data are fused using a weighted algorithm to generate benchmark data that best matches the current load state. This achieves dynamic optimization and matching of benchmark data, effectively improving the accuracy and adaptability of health status assessment, and providing a precise reference benchmark for the intelligent control of liquid cooling systems.

[0069] In this embodiment of the application, the above operation S260 may further include: for any sub-target residual sequence, sub-linear prediction coefficients and matching reference data, performing the following operations: processing the sub-target data frame based on the reference linear prediction coefficients in the reference data to obtain prediction residuals; calculating the coefficient difference between the sub-linear prediction coefficients and the reference linear prediction coefficients; dividing the prediction residuals by the coefficient difference to obtain relative residuals; summing the squares of the relative residuals corresponding to all sub-target data frames to obtain residual energy; averaging the coefficient difference of all sub-target data frames to obtain coefficient distance.

[0070] In this embodiment, the prediction residual refers to the difference between the predicted value calculated using the baseline linear prediction coefficient and the first n-1 values ​​in the sub-target data frame and the target value acquired in the nth sub-target data frame; the coefficient difference can be obtained by calculating the cosine distance or Euclidean distance between the sub-linear prediction coefficient and the baseline linear prediction coefficient.

[0071] It should be noted that the preset health benchmark library includes a benchmark sound comparison table and a benchmark vibration comparison table, as shown in Table 1 and Table 2:

[0072] Table 1

[0073]

[0074] Table 2

[0075]

[0076] In the above operation steps, the corresponding baseline linear prediction coefficients and baseline data sequences, including sound data and vibration data, can be selected from Table 1 and Table 2, as well as multiple sets of preset temperature data and preset flow data that need to be compared.

[0077] It should be noted that the letters in the table only represent the type of data. The numerical values ​​of the data can be the same or different. In Tables 1 and 2, the t corresponding to the preset temperature data and preset flow rate data represents the number of preset temperature data and preset flow rate data. The number depends on the number of sensors, and can also be the same as the number of sampling points in the vibration data frame and the sound data frame, which is n.

[0078] In this embodiment, by comparing the difference between the real-time data frame and the reference data sequence, and the distance between the real-time predicted coding coefficient and the reference coefficient, the residual energy and the average coefficient distance are obtained by summing and averaging respectively. The system health status is quantified from two dimensions: signal reconstruction error and model parameter deviation, providing a fusion judgment basis for fault early warning.

[0079] In this embodiment, the above operation S260 may further include: when the target data includes sound data and vibration data, weightedly fusing the residual energy and coefficient distance corresponding to the sound data to obtain an acoustic health index; weightedly fusing the residual energy and coefficient distance corresponding to the vibration data to obtain a vibration health index; when the acoustic health index is greater than a preset acoustic threshold, generating a minor fault warning message, which is used to prompt the execution of a primary alarm strategy, the primary alarm strategy including: generating a log event; when the vibration health index is greater than a preset vibration threshold, generating a medium-level fault warning message, which is used to prompt the execution of a medium-level alarm strategy, the medium-level alarm strategy including triggering the indicator light of the electronic device to flash, sending an email or SMS at least one; when the acoustic health index is greater than a preset acoustic threshold and the vibration health index is greater than a preset vibration threshold, generating a severe fault warning message, which is used to prompt the execution of an active protection strategy, the active protection strategy including: migrating the computing load to other electronic devices, reducing the power consumption of the electronic devices, and safely shutting down the electronic devices at least one.

[0080] In this embodiment, the residual features of the acoustic channel and the vibration channel are weighted and fused to generate acoustic health indicators and vibration health indicators. A graded response is executed based on the comparison results of each indicator with a preset threshold: when only the acoustic indicator exceeds the limit, a primary alarm is triggered by logging; when only the vibration indicator exceeds the limit, a medium-level alarm including audible and visual prompts is initiated; when both acoustic and vibration indicators exceed the limit simultaneously, an active protection strategy including load migration, power consumption limitation, and safe shutdown is executed.

[0081] For example, the acoustic health index is calculated by fusing the acoustic residual energy and the acoustic coefficient distance with weights of 0.6 and 0.4 respectively. The vibration health index is calculated by fusing the vibration residual energy and the vibration coefficient distance with weights of 0.5 each. When the acoustic health index exceeds the threshold of 0.7, a minor fault warning is generated and a primary alarm strategy is executed, recording an "acoustic anomaly" event in the system log. When the vibration health index exceeds the threshold of 0.8, a medium-level fault warning is generated and a medium-level alarm strategy is executed, triggering the yellow indicator light on the front panel of the device to flash at a frequency of 1Hz, and simultaneously sending an alarm email to the administrator via the Simple Mail Transfer Protocol (SMTP). When both the acoustic and vibration health indices exceed the thresholds, a severe fault warning is generated and an active protection strategy is executed. The computing load is migrated to other nodes in the cluster via the resource manager, and the power consumption of the electronic device is limited to 70% of the rated value via the interface. If the warning is not cleared within 10 minutes, an orderly shutdown process is initiated.

[0082] It should be noted that the above weight values, threshold settings, and corresponding early warning strategies are all illustrative examples. In practical applications, they can be flexibly configured and adjusted according to specific equipment types, operating environments, and reliability requirements.

[0083] In this embodiment, a hierarchical early warning mechanism is established by fusing the residual characteristics of acoustic and vibration signals. Based on the threshold exceedance of acoustic health indicators and vibration health indicators, primary log recording, intermediate audible and visual alarms, and active protection strategies are triggered respectively, realizing closed-loop control from anomaly detection to active protection, which significantly improves the timeliness of fault handling and system reliability.

[0084] In this embodiment, after obtaining multiple sub-linear prediction coefficients corresponding to multiple sub-target data frames in operation S230, the method further includes determining multiple sets of target temperature data and multiple sets of target flow data corresponding to the collection time of each data sample point in the multiple sub-target data frames from multiple temperature data and multiple flow data of the electronic device within the target time period. Operation S240 may further include determining target benchmark data from multiple candidate benchmark data in a preset health benchmark library based on load level information, multiple sets of target temperature data and multiple sets of target flow data. The preset health benchmark library stores multiple candidate benchmark data corresponding to multiple load scenarios defined by preset temperature data sets and preset flow data sets under different load levels.

[0085] In this embodiment, multiple temperature data and multiple flow data that share the same time identifier as each data sample point in multiple sub-target data frames are identified from multiple temperature data and multiple flow data. Multiple temperature data values ​​from multiple sensors with the same time identifier are aggregated into a temperature data group, and multiple flow data values ​​from multiple sensors with the same time identifier are aggregated into a flow data group. The aggregated multiple temperature data groups are arranged according to the chronological order of their time identifiers to form multiple sets of temperature data aligned with the acquisition time of each data sample point in the sub-target data frame. The aggregated multiple flow data groups are also arranged according to the chronological order of their time identifiers to form multiple sets of flow data aligned with the acquisition time of each data sample point in the sub-target data frame.

[0086] In this embodiment, a timestamp matching mechanism is used to achieve precise alignment of multi-source data. For example, a millisecond-level time stamp is established for each acoustic vibration data sample point. Completely synchronized monitoring data is extracted from the data streams collected by temperature and flow sensors through timestamp comparison. Temperature data sets are aggregated at the same time for the inlet and outlet temperatures of the coolant pipe and the surface temperature of the chip on the electronic device. Flow data sets are aggregated at the same time for the main circulation pipe and branch flow data. Temperature data sets at time 1, 2, 3, and up to time 4 are arranged in chronological order of acquisition to form temperature data corresponding one-to-one with the acoustic vibration signal data samples. Flow data sets at time 1, 2, 3, and up to time 5 are arranged in chronological order of acquisition to form flow data completely synchronized with the acoustic vibration signal data samples.

[0087] In this embodiment, by identifying temperature and flow data with the same time marker, the values ​​from sensors at different locations are aggregated into groups and arranged in chronological order to ensure accurate correspondence with the data samples of the acoustic and vibration signals. This establishes a unified time reference, enables collaborative analysis of multimodal data, and lays the data foundation for subsequent scene matching based on multi-parameter fusion.

[0088] In this embodiment of the application, multiple temperature data and multiple flow data that have the same time identifier as each data sample point in multiple sub-target data frames are identified from multiple temperature data and multiple flow data; multiple temperature data values ​​belonging to different time identifiers but from the same sensor location are aggregated into multiple temperature data groups, and multiple flow data values ​​belonging to different time identifiers but from the same sensor location are aggregated into multiple flow data groups.

[0089] It should be noted that by installing high-precision sensors, such as temperature sensors, flow sensors, acoustic or vibration sensors, on various key components of electronic devices, real-time monitoring and acquisition of this data can be achieved. For example, installing temperature sensors at the inlet and outlet of coolant pipes and on the surface of server chips can accurately measure the temperature changes of the coolant and chips; installing current sensors on the motor of a water pump can monitor the motor's operating current in real time, thereby reflecting the pump's operating status; and deploying vibration and acoustic sensors at the pumps, pipes, cold plates, and connections of the liquid cooling system converts the collected analog signals into digital signals, which are then transmitted to the data processing unit through a data transmission interface.

[0090] In this embodiment, historical data from a single monitoring location are aggregated longitudinally to construct a sequence of operational characteristics reflecting the changes of that location over time. For example, for a specific monitoring location such as the coolant pipe inlet, temperature readings at multiple consecutive acquisition times are aggregated into a temperature data set to form a temperature change trend sequence for that location; similarly, flow readings from the main circulation pipeline at multiple consecutive acquisition times are aggregated into a flow data set to form a flow change characteristic sequence for that location.

[0091] In this embodiment, the acoustic vibration signal and temperature and flow data are accurately matched based on the time stamp to establish a unified time reference. Then, the historical data of each independent monitoring point are longitudinally aggregated, and the monitoring values ​​of the same sensor at different times are integrated into a feature sequence reflecting the temporal changes. Time alignment ensures the synchronization of multi-source data, provides a time-consistent data foundation for subsequent correlation analysis, highlights the operating trend characteristics of a single monitoring point, and is conducive to identifying the performance degradation pattern of the equipment.

[0092] In this embodiment, determining the target benchmark data from multiple candidate benchmark data in a preset health benchmark library based on load level information, multiple sets of target temperature data, and multiple sets of target flow data may further include: for any set of target temperature data and target flow data, performing the following operations: using load level information as the first-level index, locating the corresponding load level partition in the preset health benchmark library; determining the temperature difference degree based on the difference between the temperature values ​​of each sampling point in the target temperature data of the current group and the temperature values ​​of the corresponding sampling points in the preset temperature data; determining the flow difference degree based on the difference between the flow values ​​of each sampling point in the target flow data of the current group and the flow values ​​of the corresponding sampling points in the preset flow data; performing a weighted summation of the temperature difference degree and the flow difference degree, and using the summation result as the scene similarity degree, where a smaller scene similarity value indicates a higher degree of similarity; and determining the target benchmark data from multiple candidate benchmark data of the corresponding load level partition based on all scene similarities.

[0093] In this embodiment, after extracting the acoustic and vibration signal features, multi-dimensional operating condition data such as temperature and flow rate are further combined. A dual matching mechanism of load level and operating parameters is used to select benchmark data from the health benchmark library. This achieves accurate matching of operating scenarios, effectively improves the adaptability of benchmark data to real-time status, and significantly enhances the reliability of fault early warning.

[0094] In this embodiment, based on real-time acquired load level information, data partitioning and positioning are performed in a preset health benchmark database. By using load level information as the primary retrieval condition, the load level partition corresponding to the current operating load state of the electronic device is determined within the benchmark database architecture containing multiple load level partitions. The currently collected target temperature data set and target flow data set are compared with multiple preset temperature data sets and preset flow data sets stored in the load level target partition using multi-dimensional features. Through a set similarity calculation model, the matching degree between the current operating scenario and each benchmark scenario in terms of temperature and flow parameters is quantitatively analyzed to determine the scenario similarity. Based on the calculated scenario similarities, the benchmark scenario with the highest similarity is selected from the current load level partition as the target load scenario, and the benchmark data corresponding to the target load scenario is used as a benchmark reference. The benchmark data includes benchmark linear prediction coefficients and benchmark data sequences.

[0095] For example, scene similarity can be calculated using the following formula:

[0096] S= ( )+ ( )

[0097] Where S is a smaller value, it indicates a higher degree of scene similarity; t is the number of temperature sensors; and c is the number of flow sensors. It should be noted that the values ​​of t and c can also be equal to the number of prediction coefficients. Characterizes the temperature value of the sampling point at different times i in the target temperature data of the current group. Characterizes the temperature value of the sampling point at different times i in the preset temperature data. Characterizes the flow value of the sampling point at different times i in the target flow data of the current group. It represents the flow value of the sampling point in the preset flow data at different times i, where i can represent any time when the data is collected.

[0098] In this embodiment, the corresponding partition in the health benchmark library is located according to the load level, and then the similarity between the current temperature and flow data and the preset scenario data in the partition is calculated to determine the matching target load scenario and its benchmark data, which improves the efficiency and accuracy of benchmark data retrieval and ensures the reliability of subsequent residual calculation and status assessment.

[0099] Figure 4 The diagram illustrates the system architecture of a control method for a liquid cooling system according to an embodiment of this application.

[0100] In the embodiments of this application, such as Figure 4As shown, a control system for a liquid cooling system is provided, which mainly includes the following components:

[0101] The health status benchmark construction module operates under the premise of equipment health. It includes a target data acquisition system, a preset health benchmark library establishment system, and a predictive coding feature extraction system. The final output is a linear prediction system reflecting the correspondence between sound and vibration data under various load scenarios, serving as the health benchmark. The actual operation monitoring and control module operates during routine equipment operation. It includes a target data acquisition system, a non-stationary signal matrix stabilization processing system (whose output is multiple sub-target data frames), and a linear prediction system feature extraction system. The final output is the residual energy and coefficient distance corresponding to the target data under the actual operating scenario. The health status assessment and early warning control module assesses the equipment health status by calculating the residual energy and coefficient distance between the actual operating data and the benchmark data. Based on the assessment results, it outputs tiered early warning control signals, ultimately achieving closed-loop control of the equipment's operating status through actuators.

[0102] During operation, the system first establishes a baseline for multiple load scenarios under healthy conditions using a health status baseline construction module, outputting a linear prediction system reflecting the correspondence between sound and vibration data. In actual operation, the actual operation monitoring and control module acquires real-time data, which is then processed by a non-stationary signal matrix stabilization system to output multiple sub-target data frames. These frames are further processed by the linear prediction system feature extraction system to ultimately obtain the residual energy and coefficient distance under the actual operating scenario. The health status assessment and early warning control module compares these features with the health baseline, generates corresponding control commands based on the magnitude of the deviation, and uses actuators to adjust and control the equipment's operating status, forming a complete monitoring, assessment, and control closed loop.

[0103] It should be noted that the coordinated efforts of the various systems not only realize the complete process from data acquisition and signal processing to feature extraction, but also directly drive control execution through health assessment results, achieving closed-loop management from status monitoring to early warning control.

[0104] Based on the above-described control method for a liquid cooling system, this application also provides a control method apparatus for a liquid cooling system. The following will be combined with... Figure 5 The device is described in detail.

[0105] Figure 5 A schematic block diagram of the control device for a liquid cooling system according to an embodiment of this application is shown.

[0106] like Figure 5 As shown, the control method apparatus 500 for the liquid cooling system in this embodiment includes a load information determination module 510, a preprocessing module 520, a coefficient prediction module 530, a benchmark matching module 540, a data comparison module 550, and an early warning generation module 560.

[0107] The load information determination module 510 is used to determine the load level information of the electronic device during the target time period based on the performance index data of the electronic device during the target time period. In one embodiment, the load information determination module 510 can be used to perform the operation S210 described above, which will not be repeated here.

[0108] The preprocessing module 520 is used to preprocess the target data to obtain multiple sub-target data frames. The target data includes at least one of sound data and vibration data. In one embodiment, the preprocessing module 520 can be used to perform the operation S220 described above, which will not be repeated here.

[0109] The coefficient prediction module 530 is used to perform linear prediction coding on multiple sub-target data frames respectively, to obtain multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames. In one embodiment, the coefficient prediction module 530 can be used to perform the operation S230 described above, which will not be repeated here.

[0110] The benchmark matching module 540 is used to match benchmark data from a preset health benchmark library based on load level information. In one embodiment, the benchmark matching module 540 can be used to perform the operation S240 described above, which will not be repeated here.

[0111] The data comparison module 550 is used to compare multiple sub-target data frames and multiple sub-linear prediction coefficients with the matched benchmark data respectively, and determine the residual energy and coefficient distance corresponding to the target data. In one embodiment, the data comparison module 550 can be used to perform the operation S250 described above, which will not be repeated here.

[0112] The early warning generation module 560 is used to determine at least one health indicator based on the residual energy and coefficient distance corresponding to the target data, and to generate at least one early warning message based on the at least one health indicator and a preset health threshold. In one embodiment, the early warning generation module 560 can be used to perform the operation S260 described above, which will not be repeated here.

[0113] According to embodiments of this application, any multiple modules among the load information determination module 510, preprocessing module 520, coefficient prediction module 530, benchmark matching module 540, data comparison module 550, and early warning generation module 560 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. According to embodiments of this application, at least one of the load information determination module 510, preprocessing module 520, coefficient prediction module 530, benchmark matching module 540, data comparison module 550, and early warning generation module 560 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), programmable logic array (PLA), system-on-a-chip, system-on-a-substrate, system-on-package, application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, and firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the load information determination module 510, preprocessing module 520, coefficient prediction module 530, benchmark matching module 540, data comparison module 550, and early warning generation module 560 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0114] Figure 6 A block diagram schematically illustrates an electronic device suitable for implementing a control method for a liquid cooling system according to an embodiment of this application.

[0115] like Figure 6 As shown, an electronic device according to an embodiment of this application includes a processor 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage portion 608 into a random access memory (RAM) 603. The processor 601 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 601 may also include onboard memory for caching purposes. The processor 601 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this application.

[0116] RAM 603 stores various programs and data required for the operation of the electronic device. Processor 601, ROM 602, and RAM 603 are interconnected via bus 604. Processor 601 executes various operations of the method flow according to embodiments of this application by executing programs in ROM 602 and / or RAM 603. It should be noted that programs may also be stored in one or more memories other than ROM 602 and RAM 603. Processor 601 may also execute various operations of the method flow according to embodiments of this application by executing programs stored in one or more memories.

[0117] According to embodiments of this application, the electronic device may further include an input / output (I / O) interface 605, which is also connected to a bus 604. The electronic device may also include one or more of the following components connected to the input / output (I / O) interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the input / output (I / O) interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 610 as needed so that computer programs read from it can be installed into the storage section 608 as needed.

[0118] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0119] According to embodiments of this application, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this application, the computer-readable storage medium may include ROM 602 and / or RAM 603 and / or one or more memories other than ROM 602 and RAM 603 described above.

[0120] Embodiments of this application also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this application.

[0121] When the computer program is executed by the processor 601, it performs the functions defined in the system / apparatus of this application embodiment. According to the embodiments of this application, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0122] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and downloaded and installed via the communication section 609, and / or installed from the removable medium 611. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0123] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from the removable medium 611. When the computer program is executed by the processor 601, it performs the functions defined in the system of this application embodiment. According to the embodiments of this application, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0124] According to embodiments of this application, program code for executing the computer programs provided in the embodiments of this application can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0125] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0126] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

[0127] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A control method for a liquid cooling system, the liquid cooling system being used in electronic equipment, the method comprising: Based on the performance index data of the electronic device during the target time period, determine the load level information of the electronic device during the target time period; The target data is preprocessed to obtain multiple sub-target data frames, wherein the target data includes at least one of sound data and vibration data; For any of the sub-target data frames, the following operations are performed: a linear prediction model is established based on the statistical correlation between each data sample point within the sub-target data frame, wherein the linear prediction model is used to represent each data sample point within the sub-target data frame as a linear combination of previous historical data samples; by solving the regularization equation of the linear prediction model, the weights corresponding to each previous historical data sample point in the linear combination are determined, and the weights are used as the sub-linear prediction coefficients corresponding to the sub-target data frame, wherein the sub-linear prediction coefficients characterize the linear prediction relationship between data samples within the sub-target data frame; Based on the load level information, benchmark data is matched from a preset health benchmark library; The multiple sub-target data frames and multiple sub-linear prediction coefficients are compared with the matched benchmark data to determine the residual energy and coefficient distance corresponding to the target data. Based on the residual energy and coefficient distance corresponding to the target data, at least one health indicator is determined. Based on the at least one health indicator and a preset health threshold, at least one early warning message is generated. The early warning message is used to prompt the liquid cooling system to perform control operations corresponding to the early warning message.

2. The method of claim 1, wherein, The performance metrics data include computing resource utilization, storage resource utilization, and data transmission resources; determining the load level information of the electronic device within the target time period based on the performance metrics data of the electronic device within the target time period includes: Based on the computing resource utilization, storage resource utilization, and data transmission resource utilization, a comprehensive load coefficient is calculated according to a predetermined weighting relationship. If the overall load factor is greater than or equal to the heavy load threshold, the load level information is determined to be a heavy load. If the overall load factor is less than the heavy load threshold and greater than or equal to the medium load threshold, the load level information is determined to be a medium load. If the overall load factor is less than the moderate load threshold, the load level information is determined to be a light load.

3. The method of claim 1, wherein, The preprocessing of the target data to obtain multiple sub-target data frames includes: In response to the instantaneous amplitude of the target data being greater than a preset amplitude range, the length and position of the preset segmentation window are adjusted to segment the target data into multiple initial data frames; The multiple initial data frames are decomposed into components in different frequency bands, and the components within a preset frequency band range are amplitude compressed to obtain multiple denoised data frames. The variance ratio of the first half to the second half of each of the plurality of denoised data frames is calculated, and the data frames whose variance ratio is within a preset stable range are determined as the sub-target data frames.

4. The method of claim 1, wherein, The matching of benchmark data from a preset health benchmark library based on the load level information includes: Based on the load level information, locate the corresponding load level partition in the preset health benchmark library, wherein multiple candidate benchmark data are pre-stored in the load level partition; The multiple candidate benchmark data are weighted to determine the benchmark data that matches the load level information.

5. The method of claim 1, wherein, The step of comparing the multiple sub-target data frames and multiple sub-linear prediction coefficients with the matched benchmark data to determine the residual energy and coefficient distance corresponding to the target data includes: For any sub-target data frame, sub-linear prediction coefficients, and matching baseline data, perform the following operations: The sub-target data frame is processed based on the baseline linear prediction coefficients in the baseline data to obtain the prediction residual; Calculate the coefficient difference between the sub-linear prediction coefficient and the baseline linear prediction coefficient; divide the prediction residual by the coefficient difference to obtain the relative residual; The residual energy is obtained by summing the squares of the relative residuals corresponding to all sub-target data frames. The coefficient distance is obtained by averaging the coefficient differences corresponding to all sub-target data frames.

6. The method of claim 1, wherein, The step of determining at least one health indicator based on the residual energy and coefficient distance corresponding to the target data, and generating at least one early warning message based on the at least one health indicator and a preset health threshold, includes: When the target data includes sound data and vibration data, the residual energy and coefficient distance corresponding to the sound data are weighted and fused to obtain the acoustic health index. The vibration health index is obtained by weighted fusion of the residual energy and coefficient distance corresponding to the vibration data. If the acoustic health index exceeds a preset acoustic threshold, a minor fault warning message is generated. The minor fault warning message is used to prompt the execution of a primary alarm strategy. The primary alarm strategy includes generating a log event. When the vibration health index exceeds a preset vibration threshold, a medium-level fault warning message is generated. The medium-level fault warning message is used to prompt the execution of a medium-level alarm strategy. The medium-level alarm strategy includes triggering the indicator light of the electronic device to flash, sending an email or a text message, or at least one of these actions. If the acoustic health index is greater than a preset acoustic threshold and the vibration health index is greater than a preset vibration threshold, a serious fault warning message is generated. The serious fault warning message is used to prompt the execution of an active protection strategy. The active protection strategy includes at least one of the following: migrating the computing load to other electronic devices, reducing the power consumption of the electronic devices, and safely shutting down the electronic devices.

7. The method of claim 1, wherein, The benchmark data includes candidate benchmark data and target benchmark data. After obtaining multiple sub-linear prediction coefficients corresponding to the multiple sub-target data frames, the method further includes: From multiple temperature data and multiple flow data of the electronic device within the target time period, determine multiple sets of target temperature data and multiple sets of target flow data corresponding to the collection time of each data sample point in the multiple sub-target data frames; The matching of benchmark data from a preset health benchmark library based on the load level information includes: Based on the load level information, multiple sets of target temperature data, and multiple sets of target flow data, target benchmark data is determined from multiple candidate benchmark data in a preset health benchmark library; the preset health benchmark library stores multiple candidate benchmark data corresponding to various load scenarios defined by preset temperature data sets and preset flow data sets under different load levels.

8. The method of claim 7, wherein, The step of determining the target benchmark data from multiple candidate benchmark data in a preset health benchmark library based on the load level information, multiple sets of target temperature data, and multiple sets of target flow data includes: For any set of target temperature data and target flow rate data, perform the following operations: Using the load level information as the first-level index, locate the corresponding load level partition in the preset health benchmark library; The temperature difference is determined by comparing the temperature values ​​of each sampling point in the target temperature data of the current group with the temperature values ​​of the corresponding sampling points in the preset temperature data. The degree of flow difference is determined by comparing the flow values ​​of each sampling point in the target flow data of the current group with the flow values ​​of the corresponding sampling points in the preset flow data. The temperature difference and flow rate difference are weighted and summed, and the sum is used as the scene similarity. The smaller the scene similarity value, the higher the degree of similarity. Based on the similarity of all the scenarios, target benchmark data is determined from multiple candidate benchmark data corresponding to the load level partition.

9. An electronic device, comprising: include: At least one processor; as well as, A memory that is communicatively connected to the at least one processor; The memory stores instructions that may be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method according to any one of claims 1-8.

Images