Server heat dissipation regulation and control method, device and equipment and medium

Abstract

The invention discloses a server heat dissipation regulation and control method, device and equipment and a medium, and relates to the technical field of server heat dissipation. According to the scheme, server heat dissipation regulation and control parameters corresponding to a plurality of proportional-integral-differential parameter combinations are obtained, and fuzzy evaluation is carried out on proportional gains, integral gains and differential gains in the parameter combinations, namely, based on actual information in the server heat dissipation regulation and control process, the server heat dissipation regulation and control parameters corresponding to the server heat dissipation regulation and control parameters are obtained. Complete and all-around evaluation of each gain parameter is realized, an evaluation system is perfected, and the method is more objective compared with a manual setting mode; meanwhile, when a target proportion-integral-differential parameter combination is finally generated, actual evaluation scores corresponding to all proportional gains, all integral gains and all differential gains are fully considered, and uncertainty of a common optimization algorithm on a proportion-integral-differential parameter matching relation is avoided; stable optimization of the proportional-integral-differential parameters can be realized in a shorter time, so that accurate heat dissipation regulation and control of the server can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of server heat dissipation, and particularly to a server heat dissipation regulation method, device, equipment and medium. Background Art

[0002] As the main technical means for server heat dissipation regulation, a Proportional-Integral-Derivative (PID) controller consists of three parts: proportional, integral, and derivative; the control of the fan speed is achieved by setting three parameters: proportional gain Kp, integral gain Ki, and derivative gain Kd. Therefore, the proportional-integral-derivative regulation performance is closely related to these three parameters. If the parameters do not match the input and output of the server system, it may lead to serious consequences such as slow system convergence, large fluctuations in the adjustment process, and even overheating of components.

[0003] Currently, the result evaluation system for proportional-integral-derivative regulation is mainly based on theoretical parameters such as overshoot, deviation, and settling time under the ideal curve state. However, this system is difficult to describe the temperature change curve in server fan regulation and lacks a unified standard. Therefore, when developing server heat dissipation strategies currently, the proportional-integral-derivative parameters can only be manually tuned. However, manually tuning the proportional-integral-derivative parameters depends on experience, takes a long time, and the results are subjective, and may not achieve the optimal regulation effect.

[0004] In view of the above, how to solve the current lack of an objective proportional-integral-derivative regulation result evaluation system, and the fact that manually tuning the proportional-integral-derivative parameters is easily affected by subjective factors and cannot achieve the optimal regulation effect is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] The present invention provides a server heat dissipation regulation method, device, equipment and medium to at least solve the problems in the related art that there is a lack of an objective proportional-integral-derivative regulation result evaluation system, and manually tuning the proportional-integral-derivative parameters is easily affected by subjective factors and cannot achieve the optimal regulation effect.

[0006] The present invention provides a server heat dissipation regulation method, and the method includes:

[0007] Obtain server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations;

[0008] According to the corresponding server heat dissipation regulation parameters, perform fuzzy evaluation on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination respectively to obtain the evaluation scores of each proportional gain, the evaluation scores of each integral gain, and the evaluation scores of each derivative gain;

[0009] Generate a target proportional-integral-derivative parameter combination that meets the preset requirements based on each proportional gain, each integral gain, each derivative gain, and the corresponding evaluation scores, so as to perform heat dissipation regulation on the server based on the target proportional-integral-derivative parameter combination.

[0010] The present invention also provides a server heat dissipation regulation device, and the device includes:

[0011] An acquisition module, configured to acquire server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations;

[0012] An evaluation module, configured to perform fuzzy evaluation on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination respectively according to the corresponding server heat dissipation regulation parameters, so as to obtain the evaluation scores of each proportional gain, the evaluation scores of each integral gain, and the evaluation scores of each derivative gain;

[0013] A generation module, configured to generate a target proportional-integral-derivative parameter combination that meets the preset requirements based on each proportional gain, each integral gain, each derivative gain, and the corresponding evaluation scores, so as to perform heat dissipation regulation on the server based on the target proportional-integral-derivative parameter combination.

[0014] The present invention also provides an electronic device, including: a memory, configured to store a computer program; a processor, configured to implement the steps of any one of the above server heat dissipation regulation methods when executing the computer program.

[0015] The present invention also provides a computer-readable storage medium, in which a computer program is stored, and wherein the computer program implements the steps of any one of the above server heat dissipation regulation methods when being executed by a processor.

[0016] The present invention also provides a computer program product, including a computer program, and the computer program implements the steps of any one of the above server heat dissipation regulation methods when being executed by a processor.

[0017] The present invention obtains server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations, and based on the corresponding server heat dissipation regulation parameters, conducts fuzzy evaluations on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination respectively. That is, based on the actual information in the server heat dissipation regulation process, a complete and all-round evaluation of the proportional-integral-derivative parameters is realized, improving the problem of the current lack of a unified standard for evaluating the proportional-integral-derivative regulation results, which is more objective than the manual tuning method. Further, when finally generating the target proportional-integral-derivative parameter combination, this solution fully considers the actual evaluation scores corresponding to each proportional gain, each integral gain, and each derivative gain, avoiding the uncertainty in the matching relationship of proportional-integral-derivative parameters in the ordinary proportional-integral-derivative optimization algorithm, and being able to achieve stable optimization of the proportional-integral-derivative parameters in a shorter time to achieve precise heat dissipation regulation of the server.

[0018] In addition, the present invention also provides a server heat dissipation regulation device, equipment, and medium with the same effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of a server heat dissipation regulation method provided by an embodiment of the present invention;

[0021] Figure 2 It is a schematic diagram of a temperature regulation curve under heat dissipation regulation test provided by an embodiment of the present invention;

[0022] Figure 3 It is a schematic diagram of a fan regulation curve under heat dissipation regulation test provided by an embodiment of the present invention;

[0023] Figure 4 It is a schematic diagram of the proportional-integral-derivative regulation temperature performance of different convergence coefficients in a normal scenario provided by an embodiment of the present invention;

[0024] Figure 5 It is a schematic diagram of the proportional-integral-derivative regulation fan speed performance of different convergence coefficients in a normal scenario provided by an embodiment of the present invention;

[0025] Figure 6 It is a flowchart of obtaining the optimal proportional-integral-derivative parameters for server heat dissipation regulation provided by an embodiment of the present invention;

[0026] Figure 7 Schematic diagram of a server heat dissipation regulation device provided by an embodiment of the present invention. Detailed implementation manners

[0027] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0028] It should be noted that in the description of the present invention, the terms "including", "comprising" or any other variation thereof are intended to cover a non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. The terms "first", "second", etc. in the present invention are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0029] In order to enable those skilled in the art of the present technology to better understand the solution of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0030] As the core technology of server heat dissipation regulation, the proportional-integral-derivative (PID) controller dynamically adjusts the fan speed to control the component temperature through the coordinated action of three parameters, Kp, Ki, and Kd. Its principle is based on the real-time response of the error signal: the proportional link directly reflects the current temperature difference, the integral link eliminates the historical cumulative error, and the derivative link predicts the future temperature change trend. However, the parameter setting and system matching degree directly determine the control effect: if Kp, Ki, and Kd are not selected properly, it will lead to temperature regulation lag, drastic fluctuations in the fan speed, and even the risk of component overheating. Currently, the industry generally relies on the manual tuning method. After initially selecting the parameter range based on empirical data, testers repeatedly debug and subjectively judge indicators such as the fan response speed and temperature stability. This mode not only consumes a large amount of human resources, but also has the defects that the results are affected by subjective experience and it is difficult to ensure the optimization of parameters.

[0031] Meanwhile, although existing theoretical evaluation systems (such as overshoot, settling time, etc.) are applicable to idealized temperature curves, in the actual server heat dissipation scenario, component temperatures are interfered by multiple factors such as load fluctuations and environmental changes, showing highly non-linear characteristics. Traditional indicators cannot effectively quantify the regulation effect. The lack of evaluation criteria leads to the lack of a mathematical basis for automatic parameter tuning, further exacerbating the dependence on manual debugging. Therefore, in order to establish an objective evaluation system applicable to the dynamic scenario of server heat dissipation, break through the mathematical modeling bottleneck of parameter optimization, and achieve proportional-integral-derivative intelligent tuning, the present invention provides a server heat dissipation regulation method. It can be understood that the general application scenario of this solution is a server, and it can also be applied to devices that are also applicable to proportional-integral-derivative controllers to perform heat dissipation, without limitation in this solution.

[0032] Figure 1 The flowchart of a server heat dissipation regulation method provided by an embodiment of the present invention. As Figure 1 shown, the method includes:

[0033] S10: Obtain server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations.

[0034] To achieve an objective evaluation of the dynamic scenario of server heat dissipation and proportional-integral-derivative intelligent tuning, this solution first needs to obtain server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations, so as to evaluate the server heat dissipation regulation results of multiple proportional-integral-derivative parameter combinations in the subsequent process.

[0035] It should be noted that the proportional-integral-derivative parameter combination includes Kp, Ki, and Kd; multiple proportional-integral-derivative parameter combinations can be completely different from each other, or can include some identical gains, depending on the specific implementation situation. The server heat dissipation regulation parameters are specifically the relevant information of the server temperature and fan speed when using the proportional-integral-derivative parameter combination to perform heat dissipation on the server. In this embodiment, the specific content of the server heat dissipation regulation parameters is not limited. For example, it can include the temperature peak and its corresponding time, overshoot, fan speed, etc. information. At the same time, in this embodiment, the specific process of obtaining the server heat dissipation regulation parameters is not limited either. It can be obtained through historical data, or actual server heat dissipation regulation tests can be performed, depending on the specific implementation situation.

[0036] S11: Perform fuzzy evaluation on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination according to the corresponding server heat dissipation regulation parameters, so as to obtain the evaluation scores of each proportional gain, the evaluation scores of each integral gain, and the evaluation scores of each derivative gain.

[0037] Further, according to the corresponding server heat dissipation regulation parameters, fuzzy evaluations are respectively performed on Kp, Ki, and Kd in each proportional-integral-differential parameter combination to obtain the evaluation scores of each Kp, the evaluation scores of each Ki, and the evaluation scores of each Kd.

[0038] It should be noted that in this embodiment, based on the server heat dissipation regulation parameters, the performances of Kp, Ki, and Kd in the proportional-integral-differential regulation are separately evaluated, realizing a complete and all-round evaluation of the proportional-integral-differential parameter combination, improving the problem of the current lack of a unified standard for the evaluation system of proportional-integral-differential regulation results, and being more objective than the manual tuning method, and being able to more accurately determine its advantages and disadvantages according to the evaluation scores.

[0039] In this embodiment, there is no limitation on the specific process of separately performing fuzzy evaluations on Kp, Ki, and Kd in the proportional-integral-differential parameter combination, which depends on the specific implementation situation.

[0040] S12: Generate a target proportional-integral-differential parameter combination that meets the preset requirements according to each proportional gain, each integral gain, each differential gain, and the corresponding evaluation scores, so as to perform heat dissipation regulation on the server based on the target proportional-integral-differential parameter combination.

[0041] Finally, after obtaining the evaluation scores of each Kp, the evaluation scores of each Ki, and the evaluation scores of each Kd, in combination with the evaluation scores, select appropriate Kp, Ki, and Kd from all Kp, Ki, and Kd to generate a target proportional-integral-differential parameter combination that meets the preset requirements, so as to perform heat dissipation regulation on the server based on the target proportional-integral-differential parameter combination.

[0042] It should be noted that in this embodiment, there is no limitation on the preset requirements that Kp, Ki, and Kd need to meet. For example, the evaluation scores of Kp, Ki, and Kd may all need to be greater than a threshold, or the evaluation scores of Kp, Ki, and Kd may respectively need to be greater than the corresponding thresholds, which depends on the specific implementation situation.

[0043] In this embodiment, by obtaining the server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations, and based on the corresponding server heat dissipation regulation parameters, fuzzy evaluation is respectively performed on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination. That is, based on the actual information in the server heat dissipation regulation process, a complete and all-round evaluation of the proportional-integral-derivative parameters is realized, improving the problem of the current lack of a unified standard for the evaluation system of proportional-integral-derivative regulation results, which is more objective than the manual tuning method. Further, when finally generating the target proportional-integral-derivative parameter combination, this solution fully considers the actual evaluation scores corresponding to each proportional gain, each integral gain, and each derivative gain, avoiding the uncertainty in the matching relationship of proportional-integral-derivative parameters in the ordinary proportional-integral-derivative optimization algorithm, and can achieve stable optimization of proportional-integral-derivative parameters in a shorter time to achieve precise heat dissipation regulation of the server.

[0044] Based on the above embodiment, in some embodiments, to obtain accurate server heat dissipation regulation parameters, obtaining the server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations includes:

[0045] S101: Obtain multiple randomly generated proportional-integral-derivative parameter combinations.

[0046] S102: Import each proportional-integral-derivative parameter combination into the server respectively, and perform a heat dissipation regulation test to obtain the corresponding server heat dissipation regulation parameters.

[0047] Among them, the server heat dissipation regulation parameters include the temperature peak time, overshoot, temperature oscillation error, fan oscillation error, and convergence time.

[0048] To obtain accurate server heat dissipation regulation parameters, the best method is to actually perform a server heat dissipation regulation test using proportional-integral-derivative parameter combinations. Specifically, obtain multiple randomly generated proportional-integral-derivative parameter combinations. In this embodiment, the specific number of proportional-integral-derivative parameter combinations is not limited. To ensure a sufficient amount of test data, it is recommended to randomly generate more than 10 proportional-integral-derivative parameter combinations.

[0049] Further, import each proportional-integral-derivative parameter combination into the server system respectively to perform a heat dissipation regulation test to obtain the corresponding server heat dissipation regulation parameters. In this embodiment, the heat dissipation regulation test duration for each group of proportional-integral-derivative parameter combinations is not limited. To ensure sufficient testing, it is recommended to set the test duration for each group to more than 900s, and record the temperature of the server heat dissipation components (such as the central processing unit, graphics processing unit, or voltage regulator) and the data of the server heat dissipation fan speed in real time during the test.

[0050] It should be noted that the server heat dissipation regulation parameters obtained by performing the heat dissipation regulation test include the temperature peak time , overshoot , temperature oscillation error , fan oscillation error and convergence time . The following is a specific description:

[0051] Figure 2 is a schematic diagram of the temperature regulation curve under the heat dissipation regulation test provided by the embodiment of the present invention. As Figure 2 shown, the temperature peak time is the time when the temperature reaches the highest point. In Figure 2 , is 100 s, that is, at the 100th s during this regulation, the component temperature reaches the maximum value of 77.5 . 75 is the set temperature value, that is, it is best when the component temperature does not exceed or stabilizes at 75 , and the temperature exceeding the set value is overshoot; Figure 2 In , the overshoot is 2.5 Figure 2 (generally the maximum overshoot), and generally the smaller the overshoot, the better. After the system converges, the temperature is not constant and there will be fluctuations up and down. The temperature oscillation error in is about 1

[0052] Figure 3 is a schematic diagram of the fan regulation curve under the heat dissipation regulation test provided by the embodiment of the present invention. As Figure 3 shown, the temperature fluctuation of the heat dissipation component is mainly affected by the fan speed. When the fan speed is high, the component temperature is low, and vice versa. When the fan speed is low, the heat dissipation component temperature is high. Therefore, after convergence, the fan speed also shows a fluctuating state, and the difference between the maximum and minimum values of the fluctuation is the fan oscillation error . Finally, the convergence time is the time when the fan speed and the component temperature reach a balanced state.

[0053] In this embodiment, by obtaining multiple randomly generated proportional-integral-derivative parameter combinations, each proportional-integral-derivative parameter combination is respectively imported into the server, and the heat dissipation regulation test is performed to obtain the corresponding temperature peak time, overshoot, temperature oscillation error, fan oscillation error and convergence time. Through actual testing, the accurate acquisition of the server heat dissipation regulation parameters is realized, and it is also beneficial to obtain a complete all-round evaluation of the current proportional-integral-derivative parameters based on the actual information of the subsequent proportional-integral-derivative regulation.

[0054] Based on the above embodiments, in some embodiments, fuzzy evaluations are respectively performed on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination according to the corresponding server heat dissipation regulation parameters, including:

[0055] S111: Obtain a pre-set recommended level allocation table and a scoring matrix.

[0056] Among them, the recommended level allocation table contains the recommended levels corresponding to different server heat dissipation regulation parameters at different parameter magnitudes; the scoring levels in the scoring matrix correspond one-to-one with the recommended levels in the recommended level allocation table.

[0057] S112: Obtain a pre-set influence weight factor vector for the proportional gain, an influence weight factor vector for the integral gain, and an influence weight factor vector for the derivative gain.

[0058] Among them, the influence weight factor vector represents the influence degree of the corresponding gain on each server heat dissipation regulation parameter.

[0059] S113: Determine the evaluation score for the proportional gain, the evaluation score for the integral gain, and the evaluation score for the derivative gain according to the recommended level allocation table, the scoring matrix, each influence weight factor vector, and the corresponding server heat dissipation regulation parameters.

[0060] To implement the fuzzy evaluation of Kp, Ki, and Kd, in this embodiment, a pre-set recommended level allocation table and a scoring matrix are specifically obtained. It should be noted that the scoring matrix contains the level indicators of the regulation performance of the proportional-integral-derivative parameters. In this embodiment, there is no limitation on the division of the level indicators of the scoring matrix. For example, the regulation performance can be divided into two level indicators, three level indicators, or four level indicators according to the actual test data. Considering the rationality of the classification of level indicators, in specific implementation, it is generally recommended to allocate the regulation performance of the proportional-integral-derivative parameters into three fuzzy level indicators, namely excellent, medium, and poor. The three level indicators are scored as follows: excellent score 90, medium score 60, and poor score 30. The scoring matrix is .

[0061] Recommended level allocation table Contains the recommended levels corresponding to different server heat dissipation regulation parameters at different parameter magnitudes. In this embodiment, there is no limitation on the parameter ranges corresponding to each recommended level in the recommended level allocation table. It should be noted that the scoring levels in the scoring matrix correspond one-to-one with the recommended levels in the recommended level allocation table. For example, when the scoring matrix has three levels, there should also be corresponding three levels in the recommended level allocation table. The following gives an example of a recommended level allocation table:

[0062] Table 1 Recommended level allocation table

[0063] ;

[0064] In the server, the proportional-integral-derivative control is calculated through a discretization formula, and the discretization formula is as follows:

[0065] ;

[0066] Wherein, , , respectively represent the temperature values at the k-th moment, the (k-1)-th moment, and the (k-2)-th moment, represents the set temperature value in the current heat dissipation strategy.

[0067] According to the above discretization formula, it can be seen that the three parameters of proportional-integral-derivative (i.e., Kp, Ki, and Kd) have different degrees of influence on the 5 indicators in the above recommended level allocation table. For example: assume Kp = 100, the temperature of the heat dissipation component is 30 at the 1st second , and the fan speed is 20%; assume the temperature of the heat dissipation component rises to 31 at the 2nd second , and other parameters are all 0; at this time, the heat dissipation component only changes by 1 , and the fan speed instantly rises to 20% + 100 ×1 = 100% (the full turn is 100%). Due to the rapid heat dissipation, the temperature of the heat dissipation component will definitely drop sharply. As shown in Figure 3 the fan regulation curve will fluctuate violently up and down. Therefore, the present invention defines the degrees of influence of Kp, Ki, and Kd on the 5 indicators in the above recommended level allocation table as influence weight factors, and respectively pre-establishes influence weight factor vectors of Kp, Ki, and Kd :

[0068] ;

[0069] .

[0070] It should be noted that in this embodiment, the numerical values of the pre-set influence weight factor vectors of Kp, the influence weight factor vector of Ki, and the influence weight factor vector of Kd are not limited, and can be obtained by classification training according to actual test data. The following gives a set of influence weight factor vectors:

[0071] Influence weight factor vector of Kp: ;

[0072] Influence weight factor vector of Ki: ;

[0073] Influence weight factor vector of Kd: .

[0074] Finally, according to the recommended level distribution table, the scoring matrix, each influence weight factor vector, and the corresponding server heat dissipation regulation parameters, the evaluation scores of Kp, the evaluation score of Ki, and the evaluation score of Kd are determined respectively. In this embodiment, the specific process of determining the evaluation scores of Kp, Ki, and Kd is not limited.

[0075] In this embodiment, by discretizing the three parameters in the proportional-integral-derivative parameter combination, and separately evaluating according to the performance of the three parameters Kp, Ki, and Kd in the proportional-integral-derivative regulation, each parameter corresponds to one or two result performance characteristics, and the fuzzy comprehensive evaluation of the parameters is carried out by the quantitative data calculated from these characteristics, so as to perfect the fuzzy comprehensive evaluation and optimization method of the proportional-integral-derivative parameters, facilitating the proportional-integral-derivative adaptive regulation of selecting the best from the best.

[0076] Based on the above embodiments, in some embodiments, according to the recommended level distribution table, the scoring matrix, each influence weight factor vector, and the corresponding server heat dissipation regulation parameters, determining the evaluation score of the proportional gain, the evaluation score of the integral gain, and the evaluation score of the derivative gain includes:

[0077] S114: Determine the normalization matrix according to the recommended level distribution table and the server heat dissipation regulation parameters.

[0078] S115: According to the influence weight factor vector of the proportional gain, the influence weight factor vector of the integral gain, the influence weight factor vector of the derivative gain, and the normalization matrix, determine the separate evaluation weight matrix of the proportional gain, the separate evaluation weight matrix of the integral gain, and the separate evaluation weight matrix of the derivative gain.

[0079] S116: According to the separate evaluation weight matrix of the proportional gain, the separate evaluation weight matrix of the integral gain, the separate evaluation weight matrix of the derivative gain, and the scoring matrix, determine the separate evaluation score matrix of the proportional gain, the separate evaluation score matrix of the integral gain, and the separate evaluation score matrix of the derivative gain.

[0080] Based on the fuzzy evaluation rules in the above embodiments, each group of proportional-integral-derivative parameter combinations is discretized, and separate evaluations are carried out on Kp, Ki, and Kd respectively. First, normalization calculation is performed according to the server heat dissipation regulation parameters and the recommended level distribution table to obtain the normalization matrix . For example, the temperature peak time extracted from a certain proportional-integral-derivative parameter test is 300 seconds, the overshoot is 3 , the temperature oscillation error is 1 , the fan oscillation error is 6%duty, and the convergence time is 500 seconds. Thus, based on the recommended level allocation table shown in Table 1, a normalized level table and a normalization matrix can be obtained. Specifically, as follows:

[0081] Table 2 Normalized level table

[0082] ;

[0083] .

[0084] Furthermore, according to the influence weight factor vectors of Kp, the influence weight factor vector of Ki, the influence weight factor vector of Kd, and the normalization matrix , the individual evaluation weight matrix of Kp, the individual evaluation weight matrix of Ki, and the individual evaluation weight matrix of Kd are determined. Specifically, as follows:

[0085] The individual evaluation weight matrix of Kp is: ;

[0086] The individual evaluation weight matrix of Ki is: ;

[0087] The individual evaluation weight matrix of Kd is: .

[0088] After the individual evaluation weights are calculated, according to the individual evaluation weight matrix of Kp, the individual evaluation weight matrix of Ki, the individual evaluation weight matrix of Kd, and the scoring matrix , the individual evaluation score matrix of Kp, the individual evaluation score matrix of Ki, and the individual evaluation score matrix of Kd are determined. Specifically, as follows:

[0089] The individual evaluation score matrix of Kp is: ;

[0090] The individual evaluation score matrix of Ki is: ;

[0091] The individual evaluation score matrix of Kd is: .

[0092] In summary, this solution extracts 5 parameter indicators in the server heat dissipation test, combines the ion swarm optimization algorithm and the fuzzy comprehensive evaluation method, and proposes a complete fuzzy comprehensive evaluation model and optimization method.

[0093] Based on the above embodiments, in some embodiments, according to each proportional gain, each integral gain, each derivative gain, and the corresponding evaluation scores, a target proportional-integral-derivative parameter combination that meets the preset requirements is generated, including:

[0094] S121: Among all the proportional gains, integral gains, and derivative gains, determine the candidate proportional gain, candidate integral gain, and candidate derivative gain with the highest corresponding evaluation scores.

[0095] S122: Based on the candidate proportional gain, candidate integral gain, and candidate derivative gain, construct a candidate proportional-integral-derivative parameter combination.

[0096] S123: Determine whether the candidate proportional-integral-derivative parameter combination is the same as the previous candidate proportional-integral-derivative parameter combination; if not, proceed to step S124; if so, proceed to step S125.

[0097] Among them, when the candidate proportional-integral-derivative parameter combination is constructed for the first time, the previous candidate proportional-integral-derivative parameter combination is an empty set.

[0098] S124: Regenerate multiple new proportional-integral-derivative parameter combinations and return to step S10.

[0099] S125: Determine the candidate proportional-integral-derivative parameter combination as the target proportional-integral-derivative parameter combination.

[0100] In order to determine the target proportional-integral-derivative parameter combination that meets the actual regulation requirements, in this embodiment, after completing the evaluation of the proportional-integral-derivative parameters, it is necessary to determine the candidate Kp, candidate Ki, and candidate Kd with the highest corresponding evaluation scores among all the Kp, Ki, and Kd participating in the evaluation, and construct a candidate proportional-integral-derivative parameter combination based on the candidate Kp, candidate Ki, and candidate Kd. Finally, directly determine the candidate proportional-integral-derivative parameter combination as the target proportional-integral-derivative parameter combination, and perform server heat dissipation regulation based on this, achieving efficient heat dissipation regulation.

[0101] It should be noted that since the candidate proportional-integral-derivative parameter combination is only the proportional-integral-derivative parameter combination with the best regulation effect among the multiple randomly generated proportional-integral-derivative parameter combinations in the above embodiment, in order to meet the requirements of efficient heat dissipation regulation, the candidate proportional-integral-derivative parameter combination is directly determined as the target proportional-integral-derivative parameter combination; however, this candidate proportional-integral-derivative parameter combination may not necessarily be the best proportional-integral-derivative parameter combination that meets the actual regulation requirements. Therefore, in some embodiments, after obtaining the candidate proportional-integral-derivative parameter combination, it is also necessary to iterate on the candidate proportional-integral-derivative parameter combination.

[0102] Specifically, it is determined whether the to-be-selected proportional-integral-derivative (PID) parameter combination is the same as the previous PID parameter combination. It can be understood that when the to-be-selected PID parameter combination is first constructed, the previous to-be-selected PID parameter combination is an empty set, that is, it does not exist. If it is confirmed that the to-be-selected PID parameter combination is not the same as the previous PID parameter combination, it means that the current to-be-selected PID parameter combination is not the optimal PID parameter combination, and multiple new PID parameter combinations need to be regenerated and returned to step S10 to obtain the server heat dissipation regulation parameters corresponding to multiple PID parameter combinations, so as to re-iterate the to-be-selected PID parameter combination. If it is confirmed that the to-be-selected PID parameter combination is the same as the previous PID parameter combination, it is considered that the current to-be-selected PID parameter combination is already the optimal PID parameter combination, and the to-be-selected PID parameter combination can be determined as the target PID parameter combination, and the server heat dissipation regulation is performed accordingly.

[0103] It should be noted that in this embodiment, there is no limitation on the specific method of regenerating multiple new PID parameter combinations during the iteration process. Multiple new PID parameter combinations can be randomly generated again, or better new PID parameter combinations can be further generated based on the already generated PID parameter combinations, depending on the specific implementation situation.

[0104] In this embodiment, the to-be-selected PID parameter combination with the highest evaluation score is constructed based on each Kp, each Ki, and each Kd participating in the evaluation, and the to-be-selected PID parameter combination is iterated to achieve the optimal PID adaptive regulation.

[0105] Based on the above embodiments, in some embodiments, regenerating multiple new PID parameter combinations includes:

[0106] S126: According to each proportional gain, each integral gain, each derivative gain, and the corresponding evaluation scores, respectively construct the functional relationships between the proportional gain, the integral gain, and the derivative gain and the corresponding evaluation scores.

[0107] S126: Determine a specific evaluation score that meets the server heat dissipation regulation requirements.

[0108] S127: According to the specific evaluation score and each functional relationship, determine the specific proportional gain, the specific integral gain, and the specific derivative gain.

[0109] S128: Generate multiple new proportional-integral-derivative parameter combinations based on specific proportional gain, specific integral gain, specific derivative gain, candidate proportional gain, candidate integral gain, and candidate derivative gain.

[0110] To iteratively generate the optimal proportional-integral-derivative parameter combination, in this embodiment, when regenerating multiple new proportional-integral-derivative parameter combinations, specifically fit Kp, Ki, and Kd according to a linear relationship, and construct the functional relationships between Kp, Ki, and Kd and the corresponding evaluation scores respectively, as follows:

[0111] ;

[0112] where, is the evaluation score corresponding to the proportional-integral-derivative parameter, is Kp or Ki or Kd, is the coefficient, is the intercept. It can be understood that when are Kp, Ki, and Kd respectively, the and under the corresponding functional relationships are not necessarily the same.

[0113] Furthermore, according to the server heat dissipation regulation requirements, determine the specific evaluation score that meets the server heat dissipation regulation requirements. For example, in the above embodiment, three level indicators are set for the regulation result: excellent, medium, and poor. To meet the server heat dissipation regulation requirements, a specific evaluation score can be selected within the excellent range, that is, a specific evaluation score can be selected between 80 and 100 points. In this embodiment, the number of selected specific evaluation scores is not limited, and it can be one or multiple. For example, it is found during actual verification that when the specific evaluation scores are 100 and 95, it is easier to find the optimal proportional-integral-derivative parameter combination.

[0114] Subsequently, according to specific evaluation scores and various functional relationships, corresponding specific Kp, specific Ki, and specific Kd are determined. Finally, based on the specific Kp, specific Ki, specific Kd, candidate Kp, candidate Ki, and candidate Kd, multiple new proportional-integral-derivative parameter combinations are generated. The following is an example: When the specific evaluation scores are selected as 100 and 95, two sets of specific Kp, specific Ki, and specific Kd can be calculated according to the above functional relationships. These two sets of specific Kp, specific Ki, and specific Kd can form 8 sets of proportional-integral-derivative parameter combinations. Combining with the candidate proportional-integral-derivative parameter combinations composed of candidate Kp, candidate Ki, and candidate Kd, there are a total of 9 sets of new proportional-integral-derivative parameter combinations participating in the next iteration process. In addition, candidate Kp, candidate Ki, and candidate Kd can also be involved in the combination of the above two sets of specific Kp, specific Ki, and specific Kd. At this time, a total of 27 sets of new proportional-integral-derivative parameter combinations can be formed, depending on the specific implementation situation.

[0115] In this embodiment, the functional relationships between Kp, Ki, and Kd and the corresponding evaluation scores are respectively constructed by fitting according to a linear relationship, so as to select the proportional-integral-derivative parameter combination that meets the server heat dissipation regulation requirements, and the stable optimization of the proportional-integral-derivative parameters can be achieved in a shorter time.

[0116] In the above embodiment, the recommended grade distribution table is mainly for the normal pressurized test scenario of server products. When there are different usage requirements and standards for the server, the recommended grade distribution table may be different. For example, some usage standards require considering extreme scenarios, such as drastic changes in the computer room temperature field, abnormal overclocking of hardware devices, and abnormal temperature rise of heat dissipation devices caused by certain human factors; in these extreme scenarios, it is required that the proportional-integral-derivative still has effective regulation capabilities. In response to this requirement, the present invention also proposes the concept of a convergence coefficient to adapt to different usage scenarios. Therefore, obtaining the recommended grade distribution table includes:

[0117] S131: Obtain the initial recommended grade distribution table and determine the convergence coefficient.

[0118] S132: Determine the target recommended grade distribution table according to the initial recommended grade distribution table and the convergence coefficient.

[0119] Specifically, obtain the initial recommended grade distribution table and determine the convergence coefficient; determine the target recommended grade distribution table according to the initial recommended grade distribution table and the convergence coefficient, as follows:

[0120] ;

[0121] Among them, is the target recommended grade distribution table, is the initial recommended level allocation table, is the convergence coefficient.

[0122] It should be noted that in this embodiment, there is no limit on the magnitude of the convergence coefficient, which is determined according to the specific implementation situation. For example, by combining the actual test data analysis, the value range of the convergence coefficient in this solution is set to [1, 2].

[0123] Figure 4 is a schematic diagram of the proportional-integral-derivative regulation temperature performance of different convergence coefficients provided by the embodiment of the present invention under normal scenarios. Figure 5 is a schematic diagram of the proportional-integral-derivative regulation fan speed performance of different convergence coefficients provided by the embodiment of the present invention under normal scenarios. As Figure 4 and Figure 5 shown, based on the proportional-integral-derivative regulation performance of different convergence coefficients under normal scenarios, this solution also defines three types of convergence coefficient value standards, namely energy-saving type, balanced type, and environment-adaptive type:

[0124] (1) Energy-saving type: The fan speed and the temperature of the heat dissipation component are basically not overshot during the regulation process, and stable regulation of the heat dissipation component during the pressurization process can be achieved, while the fan power consumption is the lowest; at this time the value range of is 1 - 1.3;

[0125] (2) Balanced type: The fan speed has partial overshoot during the regulation process, and there is a certain heat dissipation safety redundancy compared with the energy-saving type; at this time the value range of is 1.3 - 1.6;

[0126] (3) Environment adaptability: It can meet the requirement that the heat dissipation component does not exceed the temperature under extreme conditions, and has the highest heat dissipation safety redundancy, but there are problems such as higher fan power consumption, speed oscillation, and high noise; at this time the value range of is 1.6 - 2.

[0127] In summary, in this embodiment, the concept of the convergence coefficient for different applicable scenarios is proposed. According to the proportional-integral-derivative regulation performance of different convergence coefficients under normal scenarios, three types of convergence coefficient value standards are defined, and the value range and specific performance are provided, which can be applied to different server heat dissipation regulation requirements.

[0128] Figure 6 is a flowchart for obtaining the optimal proportional-integral-derivative parameters for server heat dissipation regulation provided by the embodiment of the present invention. To enable those skilled in the art to better understand this solution, the overall process of this solution will be described below in combination with the attached Figure 6 to explain the overall process of this solution:

[0129] The present invention discretizes the three parameters of proportional-integral-derivative involved in server heat dissipation regulation, separately evaluates Kp, Ki, and Kd according to their actual performance in proportional-integral-derivative regulation. Each parameter corresponds to one or two result performance characteristics, and the fuzzy comprehensive evaluation of this parameter is carried out by the quantitative data calculated from these characteristics, so as to select the current optimal proportional-integral-derivative parameters. Subsequently, the current optimal Kp, Ki, and Kd parameters are recombined with the theoretically optimal proportional-integral-derivative parameters selected according to the recommended level distribution table, and then the process of obtaining the optimal proportional-integral-derivative parameters is iterated according to the combined proportional-integral-derivative parameters until the optimal proportional-integral-derivative parameter combination no longer updates, indicating that the proportional-integral-derivative parameter combination at this time is the optimal combination and can be used for actual server heat dissipation regulation.

[0130] On this basis, in order to achieve a better regulation effect, feedforward compensation can also be added to the server heat dissipation regulation process. Feedforward compensation is an active control strategy that adjusts the control quantity in advance before the interference or system change takes effect, reducing the response delay. Its core process is divided into three steps:

[0131] First, monitor the disturbance or model the system. Specifically, it monitors in real time the interference sources related to heat dissipation (such as processor load, ambient temperature), and obtains data through sensors. For example, when the server load suddenly increases, this change is immediately identified. Establish a mathematical model of the heat dissipation system to describe the relationship between fan speed, load change, and temperature, which is used to predict the future state (such as how long the temperature will rise after the load suddenly increases).

[0132] Further generate the compensation signal. Specifically, according to experience or experimental data, quantify the impact of the disturbance. For example, when the load increases by 10%, the fan speed needs to be increased by 5%. Calculate the compensation value directly according to the proportion and adjust it in advance. Use the inverse model of the system to calculate the control quantity. For example, if it is known that a sudden increase in load will cause the temperature to rise by 2 degrees Celsius after 5 seconds, then calculate the required increase in fan speed according to the model 5 seconds in advance.

[0133] Finally, superimpose the feedforward compensation signal with the output of feedback control such as the proportional-integral-derivative parameter combination in the above embodiment to form a composite control. For example, the feedforward immediately increases the fan speed to cope with the sudden increase in load, and the proportional-integral-derivative parameter combination then fine-tunes according to the real-time temperature to avoid overshoot or oscillation.

[0134] In summary, by adding feedforward compensation, the lag dependence on error feedback can be reduced, and the dynamic response can be significantly accelerated; at the same time, it is applicable to measurable disturbances (such as load fluctuations) or scenarios with known system delays, and complements the proportional-integral-derivative parameters to improve the overall stability.

[0135] In addition, by combining feedforward compensation and proportional-integral-derivative parameter regulation, the server heat dissipation regulation effect can be further improved through the following methods:

[0136] (1)Redundant design of server heat dissipation;

[0137] Specifically, deploy redundant cooling paths for the server's heat dissipation devices, such as configuring an air-cooling and a liquid-cooling system simultaneously, or paralleling multiple groups of fans. During server operation, monitor the operating status of the server's heat dissipation devices, such as monitoring the rotational speed deviation of the heat dissipation fans, current anomalies, data overlimits, etc. Further, determine whether the heat dissipation devices are faulty based on the operating status of the heat dissipation devices; for example, when the fan rotational speed is lower than 50% of the set value and lasts for 10 seconds, it is determined as a fault. If it is confirmed that the heat dissipation device is faulty, terminate the operation of the faulty heat dissipation device and put the redundant device of the heat dissipation device into operation, thus ensuring that a single fault does not affect the overall heat dissipation.

[0138] (2)Predictive maintenance of heat dissipation devices;

[0139] Specifically, collect the historical operating data of the server's heat dissipation devices, monitor the changes in the operating parameters of the server's heat dissipation devices, such as recording the changes in parameters such as fan rotational speed, temperature, and load over time. Further, establish a degradation model, and predict the remaining life of the heat dissipation devices based on the changes in operating parameters and machine learning algorithms (such as regression analysis). When the remaining life is lower than the preset life, output a prompt message indicating the replacement of the heat dissipation device and notify the operation and maintenance personnel to replace the component.

[0140] (3)Server dynamic power management;

[0141] Specifically, read the power consumption value and temperature value of the processors (central processing unit and / or graphics processing unit) in the server in real time. Determine whether both the power consumption value and the temperature value are greater than the corresponding thresholds. When it is confirmed that both the power consumption value and the temperature value are greater than the corresponding thresholds, reduce the chip frequency through dynamic voltage and frequency adjustment to reduce heat generation, and increase the rotational speed of the heat dissipation fans.

[0142] (4)Long-term planning for server heat dissipation;

[0143] Specifically, establish a global thermal model of the server, and determine the predicted temperature value of the server after a preset time period based on the server cluster layout, load distribution where the server is located, and the environmental conditions where the server is located, such as the temperature change within the next 1 hour. Obtain the energy consumption strategy and temperature threshold of the server, and dynamically adjust the current temperature value of the server based on the energy consumption strategy, temperature threshold, and predicted temperature value, such as allowing the server to have a higher temperature at night, so as to achieve the purpose of reducing power consumption.

[0144] In summary, through the above steps, the robustness, energy efficiency, and intelligence level of the server heat dissipation system can be systematically improved.

[0145] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0146] Figure 7 It is a schematic diagram of a server heat dissipation regulation device provided by an embodiment of the present invention. As Figure 7 shown, the device includes:

[0147] An acquisition module 10, configured to acquire server heat dissipation regulation parameters corresponding to multiple proportional-integral-derivative parameter combinations.

[0148] An evaluation module 11, configured to perform fuzzy evaluation on the proportional gain, integral gain, and derivative gain in each proportional-integral-derivative parameter combination respectively according to the corresponding server heat dissipation regulation parameters, so as to obtain the evaluation scores of each proportional gain, the evaluation scores of each integral gain, and the evaluation scores of each derivative gain.

[0149] A generation module 12, configured to generate a target proportional-integral-derivative parameter combination that meets preset requirements based on each proportional gain, each integral gain, each derivative gain, and the corresponding evaluation scores, so as to perform heat dissipation regulation on the server based on the target proportional-integral-derivative parameter combination.

[0150] In some embodiments, the acquisition module 10 includes:

[0151] A first acquisition sub-module, configured to acquire multiple randomly generated proportional-integral-derivative parameter combinations;

[0152] A test sub-module, configured to import each proportional-integral-derivative parameter combination into the server respectively, perform a heat dissipation regulation test, so as to obtain the corresponding server heat dissipation regulation parameters;

[0153] Wherein, the server heat dissipation regulation parameters include temperature peak time, overshoot, temperature oscillation error, fan oscillation error, and convergence time.

[0154] In some embodiments, the evaluation module 11 includes:

[0155] A second acquisition sub-module, configured to acquire a pre-set recommended level distribution table and a scoring matrix; wherein, the recommended level distribution table includes the recommended levels corresponding to different server heat dissipation regulation parameters at different parameter magnitudes; the scoring levels in the scoring matrix correspond one-to-one with the recommended levels in the recommended level distribution table;

[0156] A third acquisition sub-module, configured to acquire a vector of influence weight factors for the proportional gain, a vector of influence weight factors for the integral gain, and a vector of influence weight factors for the derivative gain that are preset; wherein, the vector of influence weight factors characterizes the influence degree of the corresponding gain on each server heat dissipation regulation parameter;

[0157] A first determination sub-module, configured to determine an evaluation score for the proportional gain, an evaluation score for the integral gain, and an evaluation score for the derivative gain according to a recommended level allocation table, a scoring matrix, each vector of influence weight factors, and the corresponding server heat dissipation regulation parameter.

[0158] In some embodiments, the first determination sub-module includes:

[0159] A second determination sub-module, configured to determine a normalization matrix according to the recommended level allocation table and the server heat dissipation regulation parameter;

[0160] A third determination sub-module, configured to determine a separate evaluation weight matrix for the proportional gain, a separate evaluation weight matrix for the integral gain, and a separate evaluation weight matrix for the derivative gain according to the vector of influence weight factors for the proportional gain, the vector of influence weight factors for the integral gain, the vector of influence weight factors for the derivative gain, and the normalization matrix;

[0161] A fourth determination sub-module, configured to determine a separate evaluation score matrix for the proportional gain, a separate evaluation score matrix for the integral gain, and a separate evaluation score matrix for the derivative gain according to the separate evaluation weight matrix for the proportional gain, the separate evaluation weight matrix for the integral gain, the separate evaluation weight matrix for the derivative gain, and the scoring matrix.

[0162] In some embodiments, the generation module 12 includes:

[0163] A gain determination sub-module, configured to determine a candidate proportional gain, a candidate integral gain, and a candidate derivative gain with the highest corresponding evaluation score among each proportional gain, each integral gain, and each derivative gain;

[0164] A parameter combination construction sub-module, configured to construct a candidate proportional-integral-derivative parameter combination based on the candidate proportional gain, the candidate integral gain, and the candidate derivative gain;

[0165] A parameter combination determination sub-module, configured to determine the candidate proportional-integral-derivative parameter combination as the target proportional-integral-derivative parameter combination.

[0166] In some embodiments, the generation module 12 includes:

[0167] A fifth determination sub-module, configured to determine a candidate proportional gain, a candidate integral gain, and a candidate derivative gain with the highest corresponding evaluation score among each proportional gain, each integral gain, and each derivative gain;

[0168] The first construction sub-module is used to construct a candidate proportional-integral-derivative parameter combination based on the candidate proportional gain, candidate integral gain, and candidate derivative gain;

[0169] The judgment sub-module is used to judge whether the candidate proportional-integral-derivative parameter combination is the same as the previous candidate proportional-integral-derivative parameter combination; wherein, when the candidate proportional-integral-derivative parameter combination is constructed for the first time, the previous candidate proportional-integral-derivative parameter combination is an empty set; if not, the first generation sub-module is triggered; if so, the sixth determination sub-module is triggered;

[0170] The first generation sub-module is used to regenerate a plurality of new proportional-integral-derivative parameter combinations and trigger the acquisition module 10;

[0171] The sixth determination sub-module is used to determine the candidate proportional-integral-derivative parameter combination as the target proportional-integral-derivative parameter combination.

[0172] In some embodiments, the generation sub-module includes:

[0173] The second construction sub-module is used to respectively construct the functional relationships between the proportional gain, integral gain, derivative gain and the corresponding evaluation scores according to each proportional gain, each integral gain, each derivative gain and the corresponding evaluation scores;

[0174] The seventh determination sub-module is used to determine a specific evaluation score that meets the server heat dissipation regulation requirements;

[0175] The eighth determination sub-module is used to determine a specific proportional gain, a specific integral gain and a specific derivative gain according to the specific evaluation score and each functional relationship;

[0176] The second generation sub-module is used to generate a plurality of new proportional-integral-derivative parameter combinations according to the specific proportional gain, the specific integral gain, the specific derivative gain, the candidate proportional gain, the candidate integral gain and the candidate derivative gain.

[0177] In some embodiments, the second acquisition sub-module includes:

[0178] The fourth acquisition sub-module is used to acquire an initial recommended level allocation table and determine a convergence coefficient;

[0179] The ninth determination sub-module is used to determine a target recommended level allocation table according to the initial recommended level allocation table and the convergence coefficient.

[0180] In some embodiments, it further includes:

[0181] The first monitoring module is used to monitor the operating state of the heat dissipation device of the server;

[0182] The first judgment module is used to judge whether the heat dissipation device fails according to the operating state of the heat dissipation device; if so, trigger the redundant module;

[0183] The redundant module is used to terminate the operation of the faulty heat dissipation device and put the redundant device of the heat dissipation device into operation.

[0184] In some embodiments, it further includes:

[0185] The second monitoring module is used to monitor the change of the operating parameters of the heat dissipation device of the server;

[0186] The prediction module is used to predict the remaining life of the heat dissipation device according to the change of the operating parameters and the machine learning algorithm;

[0187] The second judgment module is used to judge whether the remaining life is lower than the preset life; if so, trigger the alarm module;

[0188] The alarm module is used to output a prompt message indicating the replacement of the heat dissipation device.

[0189] In some embodiments, it further includes:

[0190] The reading module is used to read the power consumption value and temperature value of the processor in the server;

[0191] The third judgment module is used to judge whether both the power consumption value and the temperature value are greater than the corresponding thresholds; if so, trigger the processing module;

[0192] The processing module is used to reduce the voltage frequency of the processor and increase the rotation speed of the heat dissipation fan.

[0193] In some embodiments, it further includes:

[0194] The temperature prediction module is used to determine the predicted temperature value of the server after a preset time period according to the server cluster layout, load distribution where the server is located, and the environmental conditions where the server is located;

[0195] The policy acquisition module is used to acquire the energy consumption policy and temperature threshold of the server;

[0196] The adjustment module is used to adjust the current temperature value of the server according to the energy consumption policy, temperature threshold, and predicted temperature value.

[0197] For the description of the features in the corresponding embodiments of the server heat dissipation regulation device, reference can be made to the relevant descriptions in the corresponding embodiments of the server heat dissipation regulation method, which will not be elaborated here one by one.

[0198] An embodiment of the present invention further provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any of the above embodiments of the server heat dissipation regulation method.

[0199] An embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps in any of the above embodiments of the server heat dissipation regulation method when running.

[0200] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to: various media such as a USB flash drive, a read-only memory (ROM for short), a random access memory (RAM for short), a mobile hard disk, a magnetic disk, or an optical disc that can store a computer program.

[0201] An embodiment of the present invention further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps in any of the above embodiments of the server heat dissipation regulation method are implemented.

[0202] An embodiment of the present invention further provides another computer program product, including a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above embodiments of the server heat dissipation regulation method are implemented.

[0203] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0204] The above has introduced in detail a server heat dissipation regulation method, device, equipment, and medium provided by the present invention. Specific examples are used herein to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the principle of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the protection scope of the present invention.

Claims

1. A server heat dissipation control method, characterized in that: include: Obtain server heat dissipation control parameters corresponding to multiple proportional-integral-derivative parameter combinations; According to the corresponding server heat dissipation control parameters, fuzzy evaluation is performed on the proportional gain, integral gain and differential gain in each proportional-integral-differential parameter combination to obtain an evaluation score of each proportional gain, an evaluation score of each integral gain and an evaluation score of each differential gain; According to each of the proportional gains, each of the integral gains, each of the differential gains and the corresponding evaluation scores, a target proportional-integral-differential parameter combination that meets preset requirements is generated, so as to perform heat dissipation control on the server based on the target proportional-integral-differential parameter combination.

2. The server heat dissipation control method according to claim 1, characterized in that: Get server heat dissipation control parameters corresponding to multiple proportional-integral-derivative parameter combinations, including: Obtaining a plurality of randomly generated proportional-integral-derivative parameter combinations; Importing each of the proportional-integral-derivative parameter combinations into the server respectively, and performing a heat dissipation control test to obtain the corresponding heat dissipation control parameters of the server; The server heat dissipation control parameters include temperature peak time, overshoot, temperature oscillation error, fan oscillation error and convergence time.

3. The server heat dissipation control method according to claim 2, characterized in that: According to the corresponding server heat dissipation control parameters, fuzzy evaluation is performed on the proportional gain, integral gain and differential gain in each proportional-integral-differential parameter combination, including: Obtaining a preset recommendation level allocation table and a scoring matrix; wherein the recommendation level allocation table includes the recommended levels corresponding to different server heat dissipation control parameters at different parameter sizes; and the scoring levels in the scoring matrix correspond one-to-one to the recommended levels in the recommendation level allocation table; Obtaining the preset influence weight factor vector of the proportional gain, the influence weight factor vector of the integral gain, and the influence weight factor vector of the differential gain; wherein the influence weight factor vector represents the influence degree of the corresponding gain on the heat dissipation control parameters of each server; According to the recommendation level allocation table, the scoring matrix, each influencing weight factor vector and the corresponding server heat dissipation control parameter, an evaluation score of the proportional gain, an evaluation score of the integral gain and an evaluation score of the differential gain are determined.

4. The server heat dissipation control method according to claim 3, characterized in that: Determining the evaluation score of the proportional gain, the evaluation score of the integral gain, and the evaluation score of the differential gain according to the recommendation level allocation table, the scoring matrix, each influencing weight factor vector, and the corresponding server heat dissipation control parameter, including: Determine a normalized matrix according to the recommended level allocation table and the server heat dissipation control parameters; Determine a separate evaluation weight matrix for the proportional gain, a separate evaluation weight matrix for the integral gain, and a separate evaluation weight matrix for the differential gain according to the influence weight factor vector of the proportional gain, the influence weight factor vector of the integral gain, the influence weight factor vector of the differential gain, and the normalized matrix; The separate evaluation score matrix of the proportional gain, the separate evaluation score matrix of the integral gain and the separate evaluation score matrix of the differential gain are determined according to the separate evaluation weight matrix of the proportional gain, the separate evaluation weight matrix of the integral gain, the separate evaluation weight matrix of the differential gain and the scoring matrix.

5. The server heat dissipation control method according to any one of claims 1 to 4, characterized in that: According to each of the proportional gains, each of the integral gains, each of the differential gains and the corresponding evaluation scores, a target proportional-integral-differential parameter combination that meets the preset requirements is generated, including: Determine, among the proportional gains, the integral gains and the differential gains, a candidate proportional gain, a candidate integral gain and a candidate differential gain having the highest corresponding evaluation score; Based on the selected proportional gain, the selected integral gain and the selected differential gain, construct a selected proportional-integral-differential parameter combination; The candidate proportional-integral-derivative parameter combination is determined as the target proportional-integral-derivative parameter combination.

6. The server heat dissipation control method according to any one of claims 1 to 4, characterized in that: According to each of the proportional gains, each of the integral gains, each of the differential gains and the corresponding evaluation scores, a target proportional-integral-differential parameter combination that meets the preset requirements is generated, including: Determine, among the proportional gains, the integral gains and the differential gains, a candidate proportional gain, a candidate integral gain and a candidate differential gain having the highest corresponding evaluation score; Based on the selected proportional gain, the selected integral gain and the selected differential gain, construct a selected proportional-integral-differential parameter combination; Determine whether the to-be-selected proportional-integral-differential parameter combination is the same as the last to-be-selected proportional-integral-differential parameter combination; wherein, when the to-be-selected proportional-integral-differential parameter combination is constructed for the first time, the last to-be-selected proportional-integral-differential parameter combination is an empty set; If not, regenerate a plurality of new proportional-integral-differential parameter combinations, and return to the step of obtaining server heat dissipation control parameters corresponding to the plurality of proportional-integral-differential parameter combinations; If so, the candidate proportional-integral-derivative parameter combination is determined as the target proportional-integral-derivative parameter combination.

7. The server heat dissipation control method according to claim 6, characterized in that: Regenerate a plurality of new proportional-integral-derivative parameter combinations, including: According to each of the proportional gain, each of the integral gain, each of the differential gain and each corresponding evaluation score, constructing a functional relationship between the proportional gain, the integral gain and the differential gain and the corresponding evaluation score respectively; Determine a specific evaluation score that meets the server thermal regulation requirements; Determining a specific proportional gain, a specific integral gain, and a specific differential gain according to the specific evaluation score and each of the functional relationships; A plurality of new proportional-integral-derivative parameter combinations are generated according to the specific proportional gain, the specific integral gain, the specific differential gain, the to-be-selected proportional gain, the to-be-selected integral gain and the to-be-selected differential gain.

8. The server heat dissipation control method according to claim 3, characterized in that: Obtaining the recommended level allocation table includes: Obtain an initial recommended grade allocation table and determine the convergence coefficient; A target recommendation level allocation table is determined according to the initial recommendation level allocation table and the convergence coefficient.

9. The server heat dissipation control method according to claim 1, characterized in that: Also includes: Monitoring the operating status of the heat dissipation device of the server; Determining whether the heat dissipation device fails according to the operating status of the heat dissipation device; If so, the operation of the failed heat dissipation device is terminated, and the redundant device of the heat dissipation device is put into operation.

10. The server heat dissipation control method according to claim 1, characterized in that: Also includes: Monitoring the changes in operating parameters of the heat dissipation device of the server; Predicting the remaining life of the heat dissipation device based on the change of the operating parameters and the machine learning algorithm; Determining whether the remaining life is less than a preset life; If so, a prompt message indicating replacement of the heat dissipation device is output.

11. The server heat dissipation control method according to claim 1, characterized in that: Also includes: Reading the power consumption and temperature of the processor in the server; Determine whether the power consumption value and the temperature value are both greater than corresponding thresholds; If so, the voltage frequency of the processor is reduced, and the speed of the cooling fan is increased.

12. The server heat dissipation control method according to claim 1, characterized in that: Also includes: Determine a predicted temperature value of the server after a preset time period according to the server cluster layout, load distribution and environmental conditions of the server; Obtaining the energy consumption strategy and temperature threshold of the server; The current temperature value of the server is adjusted according to the energy consumption strategy, the temperature threshold and the predicted temperature value.

13. A server heat dissipation control device, characterized in that: include: An acquisition module, used to acquire server heat dissipation control parameters corresponding to multiple proportional-integral-differential parameter combinations; An evaluation module, configured to perform fuzzy evaluation on the proportional gain, integral gain and differential gain in each proportional-integral-differential parameter combination according to the corresponding server heat dissipation control parameter, so as to obtain an evaluation score of each proportional gain, an evaluation score of each integral gain and an evaluation score of each differential gain; A generation module is used to generate a target proportional-integral-differential parameter combination that meets preset requirements based on each of the proportional gains, each of the integral gains, each of the differential gains and the corresponding evaluation scores, so as to perform heat dissipation control on the server based on the target proportional-integral-differential parameter combination.

14. An electronic device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the server heat dissipation control method as claimed in any one of claims 1 to 12 when executing the computer program.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the server heat dissipation control method according to any one of claims 1 to 12.

Images