CN117743965B - Data center energy efficiency optimization method and system based on machine learning - Google Patents
Data center energy efficiency optimization method and system based on machine learning Download PDFInfo
- Publication number
- CN117743965B CN117743965B CN202410189985.0A CN202410189985A CN117743965B CN 117743965 B CN117743965 B CN 117743965B CN 202410189985 A CN202410189985 A CN 202410189985A CN 117743965 B CN117743965 B CN 117743965B
- Authority
- CN
- China
- Prior art keywords
- temperature
- data
- value
- temperature data
- data segment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000005457 optimization Methods 0.000 title claims abstract description 29
- 238000010801 machine learning Methods 0.000 title claims abstract description 20
- 238000007667 floating Methods 0.000 claims abstract description 22
- 238000001816 cooling Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims description 17
- 230000035772 mutation Effects 0.000 claims description 16
- 238000009826 distribution Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 230000017525 heat dissipation Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 10
- 230000008859 change Effects 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005265 energy consumption Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Feedback Control In General (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a data center energy efficiency optimization method and system based on machine learning, wherein the method comprises the following steps: acquiring temperature data of data center service equipment, acquiring a variation coefficient of the temperature data at each moment, dividing the acquired temperature data into data segments according to the variation coefficient, acquiring a temperature floating index of each data segment, acquiring a local K value of each data segment according to the temperature floating index, acquiring a temperature contribution weight of each data segment according to the temperature floating index, acquiring a temperature contribution degree of each data segment according to the temperature contribution weight, acquiring a final K value of the acquired temperature data by combining the temperature contribution degree and the local K value, and finishing temperature prediction by using a KNN algorithm. The invention aims to improve the accuracy of temperature data prediction of the data center service equipment, adjust the cooling system in advance, save energy efficiency and finish the energy efficiency optimization of the data center.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a data center energy efficiency optimization method and system based on machine learning.
Background
A data center is a physical or virtual facility that centrally stores, processes, manages, and distributes large amounts of data and information, and typically includes a large number of servers, network devices, storage devices, and associated software and hardware facilities for supporting various computing tasks, applications, and services. The machine learning algorithm is applied to analyze and predict the temperature data monitored by the data center service equipment, so that the cooling system is conveniently optimized in advance, measures are adjusted, the cooling efficiency is improved, and the energy consumption of the data center equipment is reduced.
In the prior art, the temperature data of the data center equipment can be predicted by using a KNN algorithm, but the K value selection of the algorithm has a decisive effect on the final prediction result. Therefore, the final K value is determined by analyzing the collected temperature data of the data center, the future temperature data is predicted more accurately by using the K value, the cooling system is convenient to optimize and adjust in advance, the cooling efficiency is improved, and the energy consumption of the data center equipment is reduced.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a data center energy efficiency optimization method and system based on machine learning, and the adopted technical scheme is as follows:
In a first aspect, an embodiment of the present invention provides a machine learning-based data center energy efficiency optimization method, including the steps of:
Collecting temperature data of data center service equipment;
obtaining a variation coefficient of the temperature data at each moment according to the overall difference of the temperature data at each moment; taking temperature data with the mutation coefficient larger than a preset threshold value as a mutation value; dividing the acquired temperature data into data segments by taking the mutation value as a cutting point; obtaining the temperature floating index of each data segment according to the distribution of the temperature data in the data segment; obtaining a local K value of each data segment according to the temperature floating index of each data segment; obtaining the temperature contribution weight of each data segment according to the difference of the temperature data among the data segments; obtaining the temperature contribution degree of each data segment according to the distribution of the temperature contribution weights; combining the temperature contribution degree and the local K value to acquire final K values of all temperature data; and obtaining a temperature predicted value at the next moment according to the final K value and the acquired temperature data, and adjusting a cooling system of the data center service equipment according to the temperature predicted value to finish the energy efficiency optimization of the data center.
Preferably, the obtaining the mutation coefficient of the temperature data at each time according to the overall difference of the temperature data at each time includes:
calculating the difference value between the maximum value of the temperature data at all sampling moments and the temperature data at all moments, taking the opposite number of the difference value as an index of an exponential function based on a natural constant, presetting a time interval, calculating the sum value of all the temperature data in the time interval of the temperature data at all moments, calculating the ratio of the temperature data at all moments to the sum value, and taking the normalized value of the product of the ratio and the calculation result of the exponential function as the mutation coefficient of the temperature data at all moments.
Preferably, the temperature floating index of each data segment is obtained according to the distribution of the temperature data in the data segment, and the expression is:
In the method, in the process of the invention, Represents the/>Temperature float index of individual data segment,/>Represents the/>Standard deviation of all temperature data of each data segment,/>Represents the/>Average value of all temperature data of each data segment,/>Represents the/>J-th temperature data in data segment,/>Represents the/>Temperature data number of individual data segments,/>Representing the normalization function.
Preferably, the obtaining the local K value of each data segment according to the temperature floating index of each data segment includes:
and presetting a K value super parameter for each data segment, and taking an upward rounding value of the product of the K value super parameter and the temperature floating index as a local K value of each data segment.
Preferably, the temperature contribution weight of each data segment is obtained according to the difference of temperature data among the data segments, and the expression is:
In the method, in the process of the invention, Represents the/>Temperature contribution weight of individual data segments,/>Represents the/>The temperature floating index of the individual data segments,Represents the/>Temperature float index of individual data segment,/>Represents the/>Corresponding time of last temperature data in each data segment,/>Representing the corresponding moment of acquiring the last temperature data in all the temperature data,/>Representation/>And/>Time interval between,/>Represents the/>Average value of all temperature data of each data segment,/>Representing the mean value of all temperature data acquired,/>Represents the/>Temperature data number of individual data segments,/>Represents the number of all temperature data collected, e represents a natural constant,Representing the normalization function.
Preferably, the obtaining the temperature contribution degree of each data segment according to the distribution of the temperature contribution weights includes:
and calculating the sum value of the temperature contribution weights of all the data segments, and taking the ratio of the temperature contribution weight of each data segment to the sum value as the temperature contribution degree of each data segment.
Preferably, the combining the temperature contribution degree and the local K value to acquire final K values of all temperature data includes:
And calculating the product of the temperature contribution degree of each data segment and the local K value, and taking the upward rounding value of the sum value of the products of all the data segments as the final K value for collecting all the temperature data.
Preferably, the temperature predicted value at the next moment is obtained according to the final K value and the collected temperature data, and the expression is:
In the method, in the process of the invention, First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature reference weights of the individual temperature data,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representing the final K value of all temperature data acquired,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representation/>Predicted value of temperature at time/>First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature data,/>Representing a prediction error, wherein the prediction error is the average value of the difference value between any final K temperature predicted values and the real temperature value.
Preferably, the adjusting the cooling system of the service equipment of the data center according to the temperature predicted value completes the optimization of the energy efficiency of the data center, and the method comprises the following steps:
calculating the absolute value of the difference between the current time temperature value and the predicted value of the next time temperature, calculating the ratio of the absolute value of the difference to the current time temperature value, and marking as If the temperature predicted value at the next moment is smaller than the temperature value at the current moment, reducing the/>, of the heat dissipation power of the cooling system of the data center service equipmentOtherwise, the/>, the heat dissipation power of the cooling system of the data center service equipment is improved。
In a second aspect, an embodiment of the present invention further provides a data center energy efficiency optimization system based on machine learning, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of any one of the methods described above when the processor executes the computer program.
The invention has at least the following beneficial effects:
According to the invention, the temperature data of the service equipment of the data center are acquired, all acquired temperature data are divided into data segments for accurately analyzing the change condition of the temperature data in a short time, the temperature contribution degree of each data segment is obtained according to the change degree of the temperature data in each data segment and the influence degree among the temperature data, and the final K value when the temperature data are predicted by using a KNN algorithm is obtained according to the temperature contribution degree, so that the K value is selected and combined with the distribution characteristic of the temperature data, the accuracy and the reliability of K value determination are improved, the accuracy of the temperature data prediction is further improved, and the energy efficiency optimization efficiency of the data center is effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of steps of a machine learning based data center energy efficiency optimization method according to one embodiment of the present invention;
FIG. 2 is a flow chart for data center energy efficiency optimization index acquisition.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description refers to the specific implementation, structure, characteristics and effects of the data center energy efficiency optimization method and system based on machine learning according to the invention by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the machine learning-based data center energy efficiency optimization method and system provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart illustrating steps of a machine learning-based data center energy efficiency optimization method according to an embodiment of the present invention is shown, the method includes the steps of:
And S001, acquiring temperature data through a temperature sensor arranged inside the data center service equipment.
When monitoring and analyzing the temperature of the data center service equipment, the temperature sensor is installed at the CPU position of the data center service equipment to acquire temperature data, and the temperature data is transmitted to an analysis system by using a wired transmission technology. The frequency of collecting temperature data in this embodiment is 1 time per second, the collection time length is 10 minutes, and the collection frequency and the collection time length can be set by the practitioner according to the actual situation, which is not limited in this embodiment.
Step S002, the K value in the KNN algorithm is determined in a self-adaptive mode according to the distribution characteristics of the acquired temperature data, the final K value of all the acquired temperature data is obtained, and the temperature predicted value at the next moment is obtained according to the final K value.
Specifically, in this embodiment, first, temperature data of a service device of a data center is collected, a variation coefficient of the temperature data at each moment is obtained, the collected temperature data is divided into data segments according to the variation coefficient, a temperature floating index of each data segment is obtained, a local K value of each data segment is obtained according to the temperature floating index, a temperature contribution weight of each data segment is obtained according to the temperature floating index, a temperature contribution degree of each data segment is obtained according to the temperature contribution weight, a final K value of the collected temperature data is obtained by combining the temperature contribution degree and the local K value, temperature prediction is completed by using a KNN algorithm, and a data center energy efficiency optimization index obtaining flow chart is shown in fig. 2. The construction process of the final K value for collecting all temperature data comprises the following steps:
When the KNN algorithm is used for predicting the temperature data, K adjacent temperature data are selected according to the temperature data to be predicted to be analyzed and calculated to obtain a predicted value, and the value of K has a large influence on the accuracy of a predicted result. The data center service equipment processes the task amount of the data in different time periods, and the load of the internal CPU position also changes in real time according to the task amount, so that the temperature generated by the CPU position also changes continuously along with time. For the temperature data at each time, a time interval is constructed, in this embodiment, the length of the time interval is 20, and the first 20 temperature data of the temperature data at each time can be set by an implementer according to the actual situation, which is not limited in this embodiment. In order to select a proper K value when predicting temperature data, so that accuracy of temperature value prediction is improved, the acquired temperature data needs to be segmented, and for segmenting the temperature data, a mutation coefficient of the temperature data at each moment is constructed, wherein the expression is as follows:
In the method, in the process of the invention, A coefficient of variation representing temperature data at the i-th time,/>Temperature data representing the i-th time,/>The ith temperature data in the time zone representing the ith temperature data, n represents the number of temperature data included in the time zone,Representing the maximum value of all the temperature data collected,/>Representing an exponential function based on natural constants,/>Representing the normalization function.
The larger the value is, the higher the value of the temperature data at the i-th time is, and the higher the possibility that the temperature data at the i-th time is the mutation coefficient is. The smaller the difference between the temperature value at time i and the temperature maximum, i.e./>The smaller the temperature value at the i-th time, the more likely the temperature value is a local variation maximum. Thus,/>The larger the value, the greater the likelihood that the temperature data at the i-th time is a coefficient of variation.
After obtaining the mutation coefficients of the temperature data at each moment, dividing all the acquired temperature data by using the mutation coefficients, specifically, dividing the mutation coefficients to be larger than a preset threshold valueAs a variant value; dividing the acquired temperature data into data segments by taking the mutation value as a cutting point, wherein/>, in the embodimentThe implementation can be set by the implementation personnel according to the actual situation, and the embodiment is not limited to this.
When the K value is selected, the larger K value is selected when the temperature data in the data segment changes more, because the larger K value can average more neighbor samples, the influence of noise is reduced, the calculation of the predicted value is more accurate, and for the data segment with smaller temperature data change, the smaller K value is selected to avoid the influence of other temperature data, and the accuracy of the predicted temperature is improved. Therefore, the temperature floating index of each data segment is constructed according to the distribution of the temperature data in the data segment, and the expression is:
In the method, in the process of the invention, Represents the/>Temperature float index of individual data segment,/>Represents the/>Standard deviation of all temperature data of each data segment,/>Represents the/>Average value of all temperature data of each data segment,/>Represents the/>J-th temperature data in data segment,/>Represents the/>Temperature data number of individual data segments,/>Representing the normalization function.
Can reflect the/>Degree of discretization of individual data segments, when/>The greater the degree of dispersion of the data segments, the more dispersed the data segment changes, and the greater the temperature float index of the data segment. /(I)Reflecting the/>Overall degree of deviation of all temperature data in each data segment relative to temperature data mean value,/>The larger the description of the/>The greater the overall degree of fluctuation of the individual data segments, the greater the corresponding temperature float index.
Constructing a local K value of each data segment based on the temperature floating index of each data segment, wherein the expression is as follows:
In the method, in the process of the invention, Represents the/>Local K value of data segment,/>Represents the/>K-value superparameters of individual data segments, in this embodiment/>The implementer can set according to the actual situation, and the embodiment does not limit the above, i.e./>Represents the/>Temperature float index of individual data segment,/>Representing rounding up symbols.
The local K values of the data segments are obtained, the influence degree of the temperature floating indexes of different data segments on the temperature predicted value is different, the more the data segments are similar to the change degree of the temperature predicted value, the higher the contribution degree of the corresponding data segments is, and when the KNN algorithm is used for carrying out temperature prediction on the next time, the closer the data segments are to the current time point, the higher the possibility of the reference degree is, and the higher the contribution degree of the data segments is. Thus, in summary of the above analysis, a temperature contribution weight for each data segment is constructed, expressed as:
In the method, in the process of the invention, Represents the/>Temperature contribution weight of individual data segments,/>Represents the/>The temperature floating index of the individual data segments,Represents the/>Temperature float index of individual data segment,/>Represents the/>Corresponding time of last temperature data in each data segment,/>Representing the corresponding moment of acquiring the last temperature data in all the temperature data,/>Representation/>And/>Time interval between,/>Represents the/>Average value of all temperature data of each data segment,/>Representing the mean value of all temperature data acquired,/>Represents the/>Temperature data number of individual data segments,/>Represents the number of all temperature data collected, e represents a natural constant,Representing the normalization function.
The smaller the instruction of the first/>The smaller the difference between the degree of change of the individual data segment and the degree of change of the adjacent data segment, the more similar the degree of change of the first data segment and the degree of change of the adjacent data segment, the/>The greater the contribution of the individual data segments; meanwhile, when/>The closer the data segment is to the last temperature data time of the whole data segment, the greater the reference degree, i.e./>, when temperature prediction is carried outThe larger the value of (2) is, the description of (1) >The greater the contribution of the individual data segments; /(I)The size of (3) illustrates the/>Differences in the overall temperature value and the temperature mean value of the individual data segments,/>The smaller the instruction of the first/>The closer the temperature values of the individual data segments are, the more/>, is statedThe greater the contribution of the individual data segments to the temperature prediction; /(I)The larger the description of the first/>The larger the proportion of each data segment in all temperature data, the more/>The larger the contribution of individual data segments, i.e./>The smaller the description of the first/>The larger the contribution of the individual data segments.
And constructing the temperature contribution degree of each data segment based on the temperature contribution weight of each data segment, wherein the expression is as follows:
In the method, in the process of the invention, Represents the/>Temperature contribution of individual data segments; /(I)Represents the/>Temperature contribution weights for the individual data segments; n represents the number of segments of acquired temperature data segments.
And constructing a final K value for collecting all temperature data based on the temperature contribution degree of each data segment, wherein the expression is as follows:
In the method, in the process of the invention, Representing the final K value of all temperature data acquired,/>Represents the/>Temperature contribution of individual data segments,/>Represents the/>The local K value of each data segment, n represents the number of data segments divided by the acquired temperature data,/>Representing rounding up symbols.
The final K value is obtained through the steps, the temperature value at the next moment is predicted according to the final K value, and the expression is:
In the method, in the process of the invention, First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature reference weights of the individual temperature data,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representing the final K value of all temperature data acquired,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representation/>Predicted value of temperature at time/>First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature data,/>Representing a prediction error, wherein the prediction error is the average value of the difference value between any final K temperature predicted values and the real temperature value.
If pair is toThe temperature at the moment is predicted, and analysis/>The first K temperature data of a moment, the distance/>, for the first K temperature dataThe closer the moment is, the greater the reference value, i.e. the greater the weight, and vice versa, the smaller the/>For the r-th temperature data pair/>Reference degree of time-of-day temperature prediction,/>Represents the/>Weights of individual temperature data at the time of temperature prediction,/>I.e. the temperature predictions integrate/>Temperature data influence and error influence before the moment make the temperature predictive value more accurate.
And step S003, the cooling system of the data center service equipment is adjusted in advance according to the temperature predicted value, and the energy efficiency optimization of the data center is completed.
Predicting temperature data at the next moment according to the current moment and the previous K pieces of temperature data to obtain a temperature predicted value at the next moment, calculating the absolute value of the difference between the temperature predicted value at the next moment and the temperature value at the current moment, calculating the ratio of the absolute value of the difference to the temperature value at the current moment, and marking asIf the temperature predicted value at the next moment is smaller than the temperature value at the current moment, reducing the/>, of the heat dissipation power of the cooling system of the data center service equipmentOtherwise, the/>, the heat dissipation power of the cooling system of the data center service equipment is improved。
Based on the same inventive concept as the above method, the embodiment of the present invention further provides a data center energy efficiency optimization system based on machine learning, which includes a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor executes the computer program to implement the steps of any one of the above data center energy efficiency optimization methods based on machine learning.
In summary, the embodiment of the invention solves the problem of low temperature prediction precision caused by error in K value determination in the KNN algorithm, and the K value is determined in a self-adaptive manner by analyzing the distribution characteristics of the temperature data, so that the accuracy and reliability of K value determination are improved, the precision of temperature data prediction is further improved, and the energy efficiency optimization efficiency of a data center is effectively improved.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (6)
1. The data center energy efficiency optimization method based on machine learning is characterized by comprising the following steps of:
Collecting temperature data of data center service equipment;
Obtaining a variation coefficient of the temperature data at each moment according to the overall difference of the temperature data at each moment; taking temperature data with the mutation coefficient larger than a preset threshold value as a mutation value; dividing the acquired temperature data into data segments by taking the mutation value as a cutting point; obtaining the temperature floating index of each data segment according to the distribution of the temperature data in the data segment; obtaining a local K value of each data segment according to the temperature floating index of each data segment; obtaining the temperature contribution weight of each data segment according to the difference of the temperature data among the data segments; obtaining the temperature contribution degree of each data segment according to the distribution of the temperature contribution weights; combining the temperature contribution degree and the local K value to acquire final K values of all temperature data; obtaining a temperature predicted value of the next moment according to the final K value and the acquired temperature data, and adjusting a cooling system of data center service equipment according to the temperature predicted value to finish data center energy efficiency optimization;
the obtaining the mutation coefficient of the temperature data at each moment according to the overall difference of the temperature data at each moment comprises the following steps:
Calculating the difference value between the maximum value of the temperature data at all sampling moments and the temperature data at all moments, taking the opposite number of the difference value as an index of an exponential function based on a natural constant, presetting a time interval, calculating the sum value of all the temperature data in the time interval of the temperature data at all moments, calculating the ratio of the temperature data at all moments to the sum value, and taking the normalized value of the product of the ratio and the calculation result of the exponential function as the mutation coefficient of the temperature data at all moments;
And obtaining the temperature floating index of each data segment according to the distribution of the temperature data in the data segment, wherein the expression is as follows:
In the method, in the process of the invention, Represents the/>Temperature float index of individual data segment,/>Represents the/>Standard deviation of all temperature data of each data segment,/>Represents the/>Average value of all temperature data of each data segment,/>Represents the/>J-th temperature data in data segment,/>Represents the/>Temperature data number of individual data segments,/>Representing a normalization function;
And obtaining the temperature contribution weight of each data segment according to the difference of temperature data among the data segments, wherein the expression is as follows:
In the method, in the process of the invention, Represents the/>Temperature contribution weight of individual data segments,/>Represents the/>The temperature floating index of the individual data segments,Represents the/>Temperature float index of individual data segment,/>Represents the/>Corresponding time of last temperature data in each data segment,/>Representing the corresponding moment of acquiring the last temperature data in all the temperature data,/>Representation/>And (3) withTime interval between,/>Represents the/>Average value of all temperature data of each data segment,/>Representing the mean value of all temperature data acquired,/>Represents the/>Temperature data number of individual data segments,/>Represents the number of all temperature data collected, e represents a natural constant,/>Representing a normalization function;
And obtaining a temperature predicted value at the next moment according to the final K value and the acquired temperature data, wherein the expression is as follows:
In the method, in the process of the invention, First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature reference weights of the individual temperature data,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representing the final K value of all temperature data acquired,/>Represents the/>Time interval of corresponding time of each temperature data and temperature predicted value,/>Representation/>Predicted value of temperature at time/>First/>, within the first final K-value temperature data representing the temperature predicted valueTemperature data,/>Representing a prediction error, wherein the prediction error is the average value of the difference value between any final K temperature predicted values and the real temperature value.
2. The machine learning based data center energy efficiency optimization method of claim 1, wherein the obtaining the local K value for each data segment based on the temperature floating index for each data segment comprises:
and presetting a K value super parameter for each data segment, and taking an upward rounding value of the product of the K value super parameter and the temperature floating index as a local K value of each data segment.
3. The machine learning based data center energy efficiency optimization method of claim 1, wherein the obtaining the temperature contribution degree of each data segment according to the distribution of the temperature contribution weights includes:
and calculating the sum value of the temperature contribution weights of all the data segments, and taking the ratio of the temperature contribution weight of each data segment to the sum value as the temperature contribution degree of each data segment.
4. The machine learning based data center energy efficiency optimization method of claim 1, wherein the combining the temperature contribution and the local K values to collect final K values for all temperature data comprises:
And calculating the product of the temperature contribution degree of each data segment and the local K value, and taking the upward rounding value of the sum value of the products of all the data segments as the final K value for collecting all the temperature data.
5. The machine learning based data center energy efficiency optimization method of claim 1, wherein the adjusting the cooling system of the data center service equipment according to the temperature prediction value completes the data center energy efficiency optimization, comprising:
calculating the absolute value of the difference between the current time temperature value and the predicted value of the next time temperature, calculating the ratio of the absolute value of the difference to the current time temperature value, and marking as If the temperature predicted value at the next moment is smaller than the temperature value at the current moment, reducing the/>, of the heat dissipation power of the cooling system of the data center service equipmentOtherwise, the/>, the heat dissipation power of the cooling system of the data center service equipment is improved。
6. A machine learning based data center energy efficiency optimization system comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any one of claims 1-5 when the computer program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410189985.0A CN117743965B (en) | 2024-02-21 | 2024-02-21 | Data center energy efficiency optimization method and system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410189985.0A CN117743965B (en) | 2024-02-21 | 2024-02-21 | Data center energy efficiency optimization method and system based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117743965A CN117743965A (en) | 2024-03-22 |
| CN117743965B true CN117743965B (en) | 2024-05-07 |
Family
ID=90261346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410189985.0A Active CN117743965B (en) | 2024-02-21 | 2024-02-21 | Data center energy efficiency optimization method and system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117743965B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109458703A (en) * | 2018-12-21 | 2019-03-12 | 杭州睿兴栋宇建筑科技有限公司 | A kind of wisdom building air conditioning control system based on KNN algorithm and BIM technology |
| CN112393390A (en) * | 2020-07-15 | 2021-02-23 | 上海有孚智数云创数字科技有限公司 | Cloud computing data center precision air conditioner energy-saving control method based on data analysis |
| CN113495487A (en) * | 2020-03-18 | 2021-10-12 | 海信集团有限公司 | Terminal and method for adjusting operation parameters of target equipment |
| CN115309603A (en) * | 2021-11-30 | 2022-11-08 | 宁波亮控信息科技有限公司 | Data center energy consumption prediction optimization method, system, medium and computing device |
| CN115668097A (en) * | 2020-05-20 | 2023-01-31 | 华为技术有限公司 | Chip control method and control device |
| CN116860551A (en) * | 2023-06-30 | 2023-10-10 | 鹏城实验室 | Abnormality monitoring method, device, equipment and storage medium of server |
| CN116955091A (en) * | 2023-09-20 | 2023-10-27 | 深圳市互盟科技股份有限公司 | Data center fault detection system based on machine learning |
| CN117193502A (en) * | 2023-11-06 | 2023-12-08 | 惠州市鑫晖源科技有限公司 | E-bidding host power state analysis method based on machine learning |
| CN117269655A (en) * | 2023-11-17 | 2023-12-22 | 国网山东省电力公司东营供电公司 | Transformer substation power equipment temperature anomaly monitoring method, system, terminal and medium |
| CN117520103A (en) * | 2024-01-08 | 2024-02-06 | 广州大一互联网络科技有限公司 | Temperature monitoring method for data center machine room based on thermal imaging |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190195525A1 (en) * | 2017-12-21 | 2019-06-27 | At&T Intellectual Property I, L.P. | Method and apparatus for operating heating and cooling equipment via a network |
-
2024
- 2024-02-21 CN CN202410189985.0A patent/CN117743965B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109458703A (en) * | 2018-12-21 | 2019-03-12 | 杭州睿兴栋宇建筑科技有限公司 | A kind of wisdom building air conditioning control system based on KNN algorithm and BIM technology |
| CN113495487A (en) * | 2020-03-18 | 2021-10-12 | 海信集团有限公司 | Terminal and method for adjusting operation parameters of target equipment |
| CN115668097A (en) * | 2020-05-20 | 2023-01-31 | 华为技术有限公司 | Chip control method and control device |
| CN112393390A (en) * | 2020-07-15 | 2021-02-23 | 上海有孚智数云创数字科技有限公司 | Cloud computing data center precision air conditioner energy-saving control method based on data analysis |
| CN115309603A (en) * | 2021-11-30 | 2022-11-08 | 宁波亮控信息科技有限公司 | Data center energy consumption prediction optimization method, system, medium and computing device |
| CN116860551A (en) * | 2023-06-30 | 2023-10-10 | 鹏城实验室 | Abnormality monitoring method, device, equipment and storage medium of server |
| CN116955091A (en) * | 2023-09-20 | 2023-10-27 | 深圳市互盟科技股份有限公司 | Data center fault detection system based on machine learning |
| CN117193502A (en) * | 2023-11-06 | 2023-12-08 | 惠州市鑫晖源科技有限公司 | E-bidding host power state analysis method based on machine learning |
| CN117269655A (en) * | 2023-11-17 | 2023-12-22 | 国网山东省电力公司东营供电公司 | Transformer substation power equipment temperature anomaly monitoring method, system, terminal and medium |
| CN117520103A (en) * | 2024-01-08 | 2024-02-06 | 广州大一互联网络科技有限公司 | Temperature monitoring method for data center machine room based on thermal imaging |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117743965A (en) | 2024-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108595301B (en) | Server energy consumption prediction method and system based on machine learning | |
| CN106533750B (en) | The forecasting system and method for non-stationary type application user concurrent amount under a kind of cloud environment | |
| CN113554148A (en) | BiLSTM voltage deviation prediction method based on Bayesian optimization | |
| CN113039506A (en) | Data center infrastructure optimization method based on causal learning | |
| CN112396234A (en) | User side load probability prediction method based on time domain convolutional neural network | |
| CN111028100A (en) | Refined short-term load prediction method, device and medium considering meteorological factors | |
| CN109784528A (en) | Water quality prediction method and device based on time series and support vector regression | |
| CN112613542B (en) | Bidirectional LSTM-based load identification method for enterprise decontamination equipment | |
| CN112183826A (en) | Building energy consumption prediction method based on deep cascade generation countermeasure network and related product | |
| CN112180739A (en) | Parameter optimization method for superheated steam temperature control system | |
| CN116307215A (en) | Load prediction method, device, equipment and storage medium of power system | |
| CN117234301A (en) | Server thermal management method based on artificial intelligence | |
| CN105933138B (en) | Space-time dimension combined cloud service credibility situation assessment and prediction method | |
| CN117743965B (en) | Data center energy efficiency optimization method and system based on machine learning | |
| CN111091223A (en) | Distribution transformer short-term load prediction method based on Internet of things intelligent sensing technology | |
| CN112862060B (en) | Content caching method based on deep learning | |
| CN116629843B (en) | Remote early warning and maintenance decision support system of intelligent diesel generator set | |
| CN111091243A (en) | PCA-GM-based power load prediction method, system, computer-readable storage medium, and computing device | |
| CN116561569A (en) | Industrial power load identification method based on EO feature selection and AdaBoost algorithm | |
| CN116401561A (en) | Time-associated clustering method of equipment-level running state sequence based on local time domain features | |
| Xin et al. | Short-term load forecasting for electric vehicle charging stations based on time series distance measuring | |
| CN116187506A (en) | Short-term wind power combination probability prediction method and system considering meteorological classification | |
| CN112926801B (en) | Load curve combined prediction method and device based on quantile regression | |
| Bi et al. | Adaptive prediction of resources and workloads for cloud computing systems with attention-based and hybrid LSTM | |
| CN113535527A (en) | Load shedding method and system for real-time flow data predictive analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |