CN110852498A - Method for predicting data center energy consumption efficiency PUE based on GRU neural network - Google Patents

Method for predicting data center energy consumption efficiency PUE based on GRU neural network Download PDF

Info

Publication number
CN110852498A
CN110852498A CN201911052854.3A CN201911052854A CN110852498A CN 110852498 A CN110852498 A CN 110852498A CN 201911052854 A CN201911052854 A CN 201911052854A CN 110852498 A CN110852498 A CN 110852498A
Authority
CN
China
Prior art keywords
gru
energy consumption
pue
prediction model
data center
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.)
Pending
Application number
CN201911052854.3A
Other languages
Chinese (zh)
Inventor
赵鹏
康宗
杨丽娜
王佩哲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Jiaotong University
Original Assignee
Xian Jiaotong University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xian Jiaotong University filed Critical Xian Jiaotong University
Priority to CN201911052854.3A priority Critical patent/CN110852498A/en
Publication of CN110852498A publication Critical patent/CN110852498A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Software Systems (AREA)
  • Tourism & Hospitality (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • General Business, Economics & Management (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • Marketing (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Primary Health Care (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method for predicting a data center energy consumption efficiency PUE based on a GRU neural network, which comprises the following steps: 1) collecting attribute data related to energy consumption of a data center; 2) normalizing the attribute data collected in the step 1) and selecting features; 3) selecting a GRU prediction model as a prediction model, setting a loss function J (w) and an optimization method optimizer, and then training the GRU prediction model by using part of data obtained by the feature selection in the step 2); 4) and (3) evaluating the trained GRU prediction model by utilizing the residual data obtained by the feature selection in the step 2), and predicting the data center energy consumption efficiency PUE by utilizing the evaluated GRU prediction model.

Description

Method for predicting data center energy consumption efficiency PUE based on GRU neural network
Technical Field
The invention belongs to the technical field of energy conservation of data centers, and relates to a method for predicting a PUE (power consumption efficiency) value of a data center based on a GRU (generalized regression) neural network.
Background
With the rapid development of technologies such as cloud computing, internet of things, artificial intelligence and the like, data centers serving as infrastructure grow rapidly in scale and quantity. The data center is a large-scale collection of electric devices, including IT devices for processing, storing and forwarding data, cooling control systems for maintaining environment at proper temperature and humidity, and power supply systems, and the like, and the power consumption of these devices is very large to ensure their normal operation. In 2014, the electricity usage of the data center in the united states was about 700 hundred million kilowatt-hours, which accounts for about 1.8% of the total electricity usage in the united states, and according to the current trend, the data center in the united states is expected to consume about 730 million kilowatt-hours by 2020. Due to the considerations of operation cost, energy, environment and the like, reducing the power consumption of the data center and improving the energy consumption efficiency of the data center are problems which need to be solved urgently at present.
Energy consumption efficiency of data centers is generally evaluated using pue (power Usage efficiency). PUE represents the ratio of the total amount of electricity supplied to the data center to the amount of electricity used to supply IT equipment only, and theoretically, the closer the PUE is to 1, the more energy efficient IT is. In the energy management process of the data center, the PUE can be used for evaluating the energy consumption efficiency of the data center and providing relevant information such as power demand and the like for the energy management of the data center. So if the PUE of the data center can be accurately predicted, effective suggestions are provided for energy consumption management of the data center. However, the energy consumption of a data center is very complex, and servers, cooling systems, power supply systems, weather and environment all affect the energy consumption, so that it is very challenging to accurately predict the PUE.
For the existing methods for predicting PUE and energy consumption, common ann (artificial neural network) is proposed to predict PUE of a data center in google, and other researches propose to predict PUE of a data center by using methods such as an Expert System with confidence (abelie Rule Based Expert System) and a polynomial linear regression model. These efforts open new ideas for energy consumption prediction in data centers, but have some disadvantages. One aspect is that the attributes considered for predicting a PUE are not strongly or fully correlated with the PUE; on the other hand, the adopted model does not have the function of considering the time sequence of the energy consumption attribute of the data center, and the adopted models are non-time sequence related machine learning algorithms, so that the characteristics of continuous changes of time sequence variables such as temperature, humidity and the like are ignored.
Therefore, a prerequisite for accurate prediction of the PUE of the data center is to take into account as many features as possible related to energy consumption and the chronological features of these features.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a method for predicting the data center energy consumption efficiency PUE based on a GRU neural network, and the method can be used for more accurately predicting the data center energy consumption efficiency PUE.
In order to achieve the above purpose, the method for predicting the data center energy consumption efficiency PUE based on the GRU neural network comprises the following steps:
1) collecting attribute data related to energy consumption of a data center;
2) normalizing the attribute data collected in the step 1) and selecting features;
3) selecting a GRU prediction model as a prediction model, setting a loss function J (w) and an optimization method optimizer, and then training the GRU prediction model by using part of data obtained by the feature selection in the step 2);
4) and (3) evaluating the trained GRU prediction model by utilizing the residual data obtained by the feature selection in the step 2), and predicting the energy consumption efficiency PUE of the data center by utilizing the evaluated GRU prediction model.
The specific operation of the step 1) is as follows: collecting attribute data related to energy consumption of a data center according to a set time interval, wherein the collected attribute data related to the energy consumption of the data center comprises attribute data related to IT equipment, attribute data related to environment, attribute data related to a cooling system and attribute data related to infrastructure, and the attribute data related to the IT equipment comprises server load, electric load of a UPS and heat gain rate; the environment-related attribute data includes temperature, humidity, and dew point temperature; cooling system related attribute data including coolingRate and CRAC power; the attribute data related to the infrastructure includes lighting and HVAC power, and the attribute data X related to the energy consumption of the data center collected at the time tt＝(x1,x2,...,xn)。
The specific operation of the step 2) is as follows:
21) collecting the obtained attribute data X related to the energy consumption of the data centert＝(x1,x2,...,xn) Normalized to [0,1 ]]In between, the normalized data is recorded as Xi＝(x1,x2,...,xn)；
22) And selecting m attribute data with the maximum correlation with the predicted PUE by using a feature selection algorithm.
The specific operation of step 22) is:
feature selection is performed by using an RFECV method, and m attribute data X with the maximum relevance to the predicted PUE are selected (X is X)1,x2,...,xm)。
In step 3), a GRU recurrent neural network is used as a PUE prediction model, wherein a reset gate r exists in a GRU neuron structure in the GRU recurrent neural networktAnd update gate ztTwo gate control units, which are used for realizing the training information generated in the training step before memorizing or forgetting, wherein the specific calculation process of the GRU unit is as follows:
zt＝σ(Wz·[ht-1，xt]+bz)
rt＝σ(Wr·[ht-1，xt]+br)
Figure BDA0002255763670000032
wherein z istFor controlling how much information h is in the previous statet-1Is written to the current state htσ (-) is a sigmoid activation function that scales values between (0,1), i.e.：
Figure BDA0002255763670000041
t is the parameter to be scaled in, where a closer value of σ () to 0 means more writes and a closer value of σ () to 1 means less writes, where ht-1Indicating the status information, x, output at time t-1tRepresenting the input at the current time, Wz、WrAnd WnAs weight parameter, bz、brAnd bhIs a bias constant;
rtfor controlling how much information from a previous state is written to a current candidate state
Figure BDA0002255763670000042
rtThe closer to 1 the value of (d) is, the more information is written indicating the previous state, rtThe closer to 0, the less information is written to the previous state;
Figure BDA0002255763670000043
representing candidate state information at the current time instant, tanh (-) activation function for scaling the value to [ -1,1]Namely:
Figure BDA0002255763670000044
the specific process of training the GRU prediction model in the step 3) is as follows:
assuming that the mean square error loss function J (w) is used for evaluating the fitting degree of the GRU prediction model to the training data during training, the expression of J (w) is as follows:
Figure BDA0002255763670000045
where N denotes the number of training data, yiThe tag representing the ith piece of data, i.e. the true PUE value,
Figure BDA0002255763670000046
PUE values predicted for GRU prediction models;
let the training optimizer Adam, and the way to update the parameter w is:
first, the gradient g of the mean square error loss function J (w) to the parameter w is calculatedtThen, the gradient g is calculatedtFirst moment m oftAnd second moment vtWherein, in the step (A),
mt＝β1·mt-1+(1-β1)·gt
vt＝β2·vt-1+(1-β2)·gt 2
wherein, β1And β2Is a hyper-parameter;
then, correcting the first moment and the second moment by the following formula;
Figure BDA0002255763670000051
Figure BDA0002255763670000052
wherein, β1 tAnd β2 tAre respectively β1And β2To the power of t.
The updating process of the parameter w is as follows:
Figure BDA0002255763670000053
where α denotes the learning rate, and ε is a hyperparameter.
The specific operation of the step 4) is as follows:
by determining the coefficient R2Evaluating the GRU prediction model, wherein a coefficient R is determined2The expression of (a) is:
Figure BDA0002255763670000054
wherein, yiIn order to test the true PUE value of the sample,
Figure BDA0002255763670000055
for the predicted value of the GRU prediction model,
Figure BDA0002255763670000056
determining the coefficient R for the mean of the PUE of the test sample2Representing the fitting degree of the GRU prediction model to the real data, and determining a coefficient R2The closer to 1, the higher the prediction accuracy of the GRU prediction model.
The invention has the following beneficial technical effects:
according to the method for predicting the PUE of the data center energy consumption efficiency value based on the GRU neural network, when the concrete operation is carried out, the GRU prediction model is selected as the prediction model, due to the fact that the GRU prediction model has the function of memorizing training information, the related attributes of energy consumption can be fully considered, such as time sequence of temperature, humidity and the like, compared with the prior art, the correlation between the PUE and energy consumption characteristics is fully mined in the training process, and the accuracy of PUE prediction is improved.
Drawings
Fig. 1 is a schematic structural diagram of a network unit of a GRU;
FIG. 2 shows the correlation coefficient R of the present invention and other prediction methods2A performance comparison chart under the evaluation index;
FIG. 3 is a plot of a fit of predicted PUE values to actual PUE values in accordance with the present invention.
Detailed Description
The invention is described in further detail below with reference to the accompanying drawings:
referring to fig. 1, the method for predicting the energy consumption efficiency value PUE of the data center based on the gru (gatedcurrrentunit) neural network according to the present invention includes the following steps:
1) collecting attribute data related to energy consumption of a data center;
the specific operation of the step 1) is as follows: collecting attribute data related to energy consumption of a data center according to a set time interval (10min), wherein the collected attribute data related to the energy consumption of the data center comprises attribute data related to IT equipment, attribute data related to environment, attribute data related to a cooling system and attribute data related to infrastructure, and the attribute data related to the IT equipment comprises server load, power load of a UPS and heat gain rate; the environment-related attribute data includes temperature, humidity, and dew point temperature; cooling system related attribute data including cooling rate and crac (computer Room Air condition) power; the attribute data related to the infrastructure includes lighting and HVAC (Heating, ventilating and air Conditioning) power, and the attribute data X related to the energy consumption of the data center collected at the time point tt＝(x1,x2,...,xn)。
2) Normalizing the attribute data collected in the step 1) and selecting features;
the specific operation of the step 2) is as follows:
21) collecting the obtained attribute data X related to the energy consumption of the data centert＝(x1,x2,...,xn) Normalized to [0,1 ]]In between, the normalized data is recorded as Xi＝(x1,x2,...,xn)；
Let the original data be Xi＝(x1,x2,...,xn) Then, the way of normalizing the data attribute values is as follows:
Figure BDA0002255763670000071
wherein x represents the original attribute value, xminIndicates the minimum value, x, that all data center attributes havemaxRepresenting the maximum value, x, that an attribute hasnormalRepresenting the normalized attribute value with the value of 0,1]The normalized data X is processediIs marked as Xi＝(x1,x2,...,xn)；
22) The m attribute data having the greatest correlation with the predicted PUE are selected by using a Feature selection algorithm (RFECV).
The specific operation of step 22) is:
feature selection is performed by using an RFECV method, and m attribute data X with the maximum relevance to the predicted PUE are selected (X is X)1,x2,...,xm)。
3) Selecting a GRU prediction model as a prediction model, setting a loss function J (w) and an optimization method optimizer, and then training the GRU prediction model by using part of data obtained by the feature selection in the step 2);
the specific operation of the step 3) is as follows:
in the existing prediction methods for the PUE value of the data center, machine learning methods such as simple ANN (Artificial neural network) and polynomial linear regression are used, and the methods have the defects that the data center energy consumption related attributes such as: due to the time-series characteristics of temperature, humidity and the like, the invention proposes to use a GRU (gated Recurrent Unit) Recurrent neural network as a prediction model of the PUE.
Using a GRU recurrent neural network as the PUE prediction model, wherein a reset gate (reset gate) r is present in a GRU neuron structure within the GRU recurrent neural networktAnd update gate (update gate) ztTwo gate control units, which are used for realizing the training information generated in the training step before memorizing or forgetting, wherein the specific calculation process of the GRU unit is as follows:
zt＝σ(Wz·[ht-1，xt]+bz)
rt＝σ(Wr·[ht-1，xt]+br)
Figure BDA0002255763670000082
wherein z istFor controlling the previous stateHow much information h ist-1Is written to the current state htσ () is a sigmoid activation function that scales values between (0,1), i.e.:
Figure BDA0002255763670000083
t is the parameter to be scaled in, where a closer value of σ () to 0 means more writes and a closer value of σ () to 1 means less writes, where ht-1Indicating the status information, x, output at time t-1tRepresenting the input at the current time, Wz、WrAnd WhAs weight parameter, bz、brAnd bhIs a bias constant;
rtfor controlling how much information from a previous state is written to a current candidate state
Figure BDA0002255763670000084
rtThe closer to 1 the value of (d) is, the more information is written indicating the previous state, rtThe closer to 0, the less information is written to the previous state;
Figure BDA0002255763670000085
representing candidate state information at the current time instant, tanh (-) activation function for scaling the value to [ -1,1]Namely:
Figure BDA0002255763670000086
the specific process of training the GRU prediction model in the step 3) is as follows:
assuming that the mean square error loss function J (w) is used for evaluating the fitting degree of the GRU prediction model to the training data during training, the expression of J (w) is as follows:
Figure BDA0002255763670000091
wherein N represents the number of trainsNumber of data, yiThe tag representing the ith piece of data, i.e. the true PUE value,
Figure BDA0002255763670000092
PUE values predicted for GRU prediction models;
let the training optimizer Adam, and the way to update the parameter w is:
first, the gradient g of the mean square error loss function J (w) to the parameter w is calculatedtThen, the gradient g is calculatedtFirst moment m oftAnd second moment vtWherein, in the step (A),
mt＝β1·mt-1+(1-β1)·gt
Figure BDA0002255763670000093
wherein, β1And β2Is a hyper-parameter;
then, correcting the first moment and the second moment by the following formula;
Figure BDA0002255763670000094
Figure BDA0002255763670000095
wherein, β1 tAnd β2 tAre respectively β1And β2To the power of t.
The updating process of the parameter w is as follows:
Figure BDA0002255763670000096
where α denotes the learning rate, and ε is a hyperparameter.
4) And (3) evaluating the trained GRU prediction model by utilizing the residual data obtained by the feature selection in the step 2), and predicting the energy consumption efficiency PUE of the data center by utilizing the evaluated GRU prediction model.
The specific operation of the step 4) is as follows:
by determining the coefficient R2Evaluating the GRU prediction model, wherein a coefficient R is determined2The expression of (coeffient determination) is:
Figure BDA0002255763670000101
wherein, yiIn order to test the true PUE value of the sample,
Figure BDA0002255763670000102
for the predicted value of the GRU prediction model,
Figure BDA0002255763670000103
determining the coefficient R for the mean of the PUE of the test sample2Representing the fitting degree of the GRU prediction model to the real data, and determining a coefficient R2The closer to 1, the higher the prediction accuracy of the GRU prediction model.
As shown in FIG. 2, the present invention is shown at R2The prediction performance of the method is superior to prediction models such as ANN and SVR, and the time sequence characteristic of the energy consumption attribute can be fully considered due to the memory characteristic of GRU; the reason why GRU is superior to LSTM is that GRU has fewer gating units and thus fewer parameters, so training time is less, complexity is low, and prediction effect is better.

Claims (7)

1. A method for predicting a data center energy consumption efficiency PUE based on a GRU neural network is characterized by comprising the following steps:
1) collecting attribute data related to energy consumption of a data center;
2) normalizing the attribute data collected in the step 1) and selecting features;
3) selecting a GRU prediction model as a prediction model, setting a loss function J (w) and an optimization method optimizer, and then training the GRU prediction model by using part of data obtained by the feature selection in the step 2);
4) and (3) evaluating the trained GRU prediction model by utilizing the residual data obtained by the feature selection in the step 2), and predicting the energy consumption efficiency PUE of the data center by utilizing the evaluated GRU prediction model.
2. The method for predicting the data center energy consumption efficiency rate PUE based on the GRU neural network as claimed in claim 1, wherein the specific operation of the step 1) is as follows: collecting attribute data related to energy consumption of a data center according to a set time interval, wherein the collected attribute data related to the energy consumption of the data center comprises attribute data related to IT equipment, attribute data related to environment, attribute data related to a cooling system and attribute data related to infrastructure, and the attribute data related to the IT equipment comprises server load, electric load of a UPS and heat gain rate; the environment-related attribute data includes temperature, humidity, and dew point temperature; cooling system related attribute data includes cooling rate and CRAC power; the attribute data related to the infrastructure includes lighting and HVAC power, and the attribute data X related to the energy consumption of the data center collected at the time tt＝(x1，x2，...，xn)。
3. The method for predicting the data center energy consumption efficiency rate PUE based on the GRU neural network as claimed in claim 2, wherein the specific operation of the step 2) is as follows:
21) collecting the obtained attribute data X related to the energy consumption of the data centert＝(x1，x2，...，xn) Normalized to [0,1 ]]In between, the normalized data is recorded as Xi＝(x1，x2，...，xn)；
22) And selecting m attribute data with the maximum correlation with the predicted PUE by using a feature selection algorithm.
4. The method for predicting a data center energy consumption efficiency rate PUE based on a GRU neural network as claimed in claim 3, wherein the specific operation of step 22) is:
feature selection is performed by using an RFECV method, and m attribute data X with the maximum relevance to the predicted PUE are selected (X is X)1，x2，...，xm)。
5. The method for predicting PUE (energy consumption efficiency) of data center based on GRU (neural network) of claim 1, wherein in the step 3), the GRU recurrent neural network is used as a PUE prediction model, wherein a reset gate r exists in a GRU neuron structure in the GRU recurrent neural networktAnd update gate ztTwo gate control units, which are used for realizing the training information generated in the training step before memorizing or forgetting, wherein the specific calculation process of the GRU unit is as follows:
zt＝σ(Wz·[ht-1，xt]+bz)
rt＝σ(Wr·[ht-1，xt]+br)
Figure FDA0002255763660000021
Figure FDA0002255763660000022
wherein z istFor controlling how much information h is in the previous statet-1Is written to the current state htσ () is a sigmoid activation function that scales values between (0,1), i.e.:
Figure FDA0002255763660000023
t is the parameter to be scaled in, where a closer value of σ () to 0 means more writes and a closer value of σ () to 1 means less writes, where ht-1Indicating the status information, x, output at time t-1tRepresenting the input at the current time, Wz、WrAnd WhAs weight parameter, bz、brAnd bhIs a bias constant;
rtfor controlling how much information from a previous state is written to a current candidate state
Figure FDA0002255763660000024
rtThe closer to 1 the value of (d) is, the more information is written indicating the previous state, rtThe closer to 0, the less information is written to the previous state;
Figure FDA0002255763660000025
representing candidate state information at the current time instant, tanh (-) activation function for scaling the value to [ -1,1]Namely:
Figure FDA0002255763660000031
6. the method for predicting the data center energy consumption efficiency PUE based on the GRU neural network as claimed in claim 5, wherein the specific process of training the GRU prediction model in the step 3) is as follows:
assuming that the mean square error loss function J (w) is used for evaluating the fitting degree of the GRU prediction model to the training data during training, the expression of J (w) is as follows:
Figure FDA0002255763660000032
where N denotes the number of training data, yiThe tag representing the ith piece of data, i.e. the true PUE value,
Figure FDA0002255763660000033
PUE values predicted for GRU prediction models;
let the training optimizer Adam, and the way to update the parameter w is:
first, the gradient g of the mean square error loss function J (w) to the parameter w is calculatedtThen, the gradient g is calculatedtFirst moment m oftAnd second moment vtWherein, in the step (A),
mt＝β1·mt-1+(1-β1)·gt
vt＝β2·vt-1+(1-β2)·gt 2
wherein, β1And β2Is a hyper-parameter;
then, correcting the first moment and the second moment by the following formula;
Figure FDA0002255763660000034
Figure FDA0002255763660000035
wherein, β1 tAnd β2 tAre respectively β1And β2To the t power;
the updating process of the parameter w is as follows:
where α denotes the learning rate, and ε is a hyperparameter.
7. The method for predicting the data center energy consumption efficiency rate PUE based on the GRU neural network as claimed in claim 6, wherein the specific operation of the step 4) is as follows:
by determining the coefficient R2Evaluating the GRU prediction model, wherein a coefficient R is determined2The expression of (a) is:
wherein, yiIn order to test the true PUE value of the sample,
Figure FDA0002255763660000043
for the predicted value of the GRU prediction model,
Figure FDA0002255763660000044
determining the coefficient R for the mean of the PUE of the test sample2Representing the fitting degree of the GRU prediction model to the real data, and determining a coefficient R2The closer to 1, the higher the prediction accuracy of the GRU prediction model.
CN201911052854.3A 2019-10-31 2019-10-31 Method for predicting data center energy consumption efficiency PUE based on GRU neural network Pending CN110852498A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911052854.3A CN110852498A (en) 2019-10-31 2019-10-31 Method for predicting data center energy consumption efficiency PUE based on GRU neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911052854.3A CN110852498A (en) 2019-10-31 2019-10-31 Method for predicting data center energy consumption efficiency PUE based on GRU neural network

Publications (1)

Publication Number Publication Date
CN110852498A true CN110852498A (en) 2020-02-28

Family

ID=69598737

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911052854.3A Pending CN110852498A (en) 2019-10-31 2019-10-31 Method for predicting data center energy consumption efficiency PUE based on GRU neural network

Country Status (1)

Country Link
CN (1) CN110852498A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111693667A (en) * 2020-05-06 2020-09-22 杭州电子科技大学 Water quality detection system and method based on gated recursive array
CN112734097A (en) * 2020-12-31 2021-04-30 中南大学 Unmanned train energy consumption prediction method, system and storage medium
CN112991331A (en) * 2021-04-19 2021-06-18 广州大一互联网络科技有限公司 Operation and maintenance method and device of data center using insulating cooling liquid
CN113837480A (en) * 2021-09-29 2021-12-24 河北工业大学 Impact load prediction method based on improved GRU and differential error compensation
CN114692942A (en) * 2022-01-17 2022-07-01 陕西智引科技有限公司 Data center energy consumption prediction system based on GRU neural network model
CN114723303A (en) * 2022-04-15 2022-07-08 中国电信股份有限公司 Method, device and equipment for determining energy-saving space of machine room and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105204978A (en) * 2015-06-23 2015-12-30 北京百度网讯科技有限公司 Data center operation data analysis system based on machine learning
CN108448610A (en) * 2018-03-12 2018-08-24 华南理工大学 A kind of short-term wind power prediction method based on deep learning
CN109325624A (en) * 2018-09-28 2019-02-12 国网福建省电力有限公司 A kind of monthly electric power demand forecasting method based on deep learning

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105204978A (en) * 2015-06-23 2015-12-30 北京百度网讯科技有限公司 Data center operation data analysis system based on machine learning
CN108448610A (en) * 2018-03-12 2018-08-24 华南理工大学 A kind of short-term wind power prediction method based on deep learning
CN109325624A (en) * 2018-09-28 2019-02-12 国网福建省电力有限公司 A kind of monthly electric power demand forecasting method based on deep learning

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
NING LIU ET AL.: "data center power management for regulation service using neural network-based power prediction", 《2017 18TH INTERNATIONAL SYMPOSIUM ON QUALITY ELECTRONIC DESIGN (ISQED)》 *
王增平 等: "基于GRU-NN模型的短期负荷预测方法", 《电力系统自动化》 *
秦育: "大型公共建筑能耗预测模型与监管系统研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111693667A (en) * 2020-05-06 2020-09-22 杭州电子科技大学 Water quality detection system and method based on gated recursive array
CN112734097A (en) * 2020-12-31 2021-04-30 中南大学 Unmanned train energy consumption prediction method, system and storage medium
CN112734097B (en) * 2020-12-31 2023-09-05 中南大学 Unmanned train energy consumption prediction method, system and storage medium
CN112991331A (en) * 2021-04-19 2021-06-18 广州大一互联网络科技有限公司 Operation and maintenance method and device of data center using insulating cooling liquid
CN112991331B (en) * 2021-04-19 2021-10-26 广州大一互联网络科技有限公司 Operation and maintenance method and device of data center using insulating cooling liquid
CN113837480A (en) * 2021-09-29 2021-12-24 河北工业大学 Impact load prediction method based on improved GRU and differential error compensation
CN113837480B (en) * 2021-09-29 2023-11-07 河北工业大学 Impact load prediction method based on improved GRU and differential error compensation
CN114692942A (en) * 2022-01-17 2022-07-01 陕西智引科技有限公司 Data center energy consumption prediction system based on GRU neural network model
CN114723303A (en) * 2022-04-15 2022-07-08 中国电信股份有限公司 Method, device and equipment for determining energy-saving space of machine room and storage medium
CN114723303B (en) * 2022-04-15 2023-10-31 中国电信股份有限公司 Method, device, equipment and storage medium for determining energy-saving space of machine room

Similar Documents

Publication Publication Date Title
CN110852498A (en) Method for predicting data center energy consumption efficiency PUE based on GRU neural network
Tan et al. Ultra-short-term wind power prediction by salp swarm algorithm-based optimizing extreme learning machine
CN114066071B (en) Power parameter optimization method based on energy consumption, terminal equipment and storage medium
Almalaq et al. Parallel building: a complex system approach for smart building energy management
CN110866592B (en) Model training method, device, energy efficiency prediction method, device and storage medium
CN112434787B (en) Terminal space energy consumption prediction method, medium and equipment based on total energy consumption of building
CN112413831A (en) Energy-saving control system and method for central air conditioner
CN107644297B (en) Energy-saving calculation and verification method for motor system
CN116307652A (en) Artificial intelligent resource allocation method for intelligent power grid
CN111241755A (en) Power load prediction method
CN112418495A (en) Building energy consumption prediction method based on longicorn stigma optimization algorithm and neural network
Barzola-Monteses et al. Energy consumption of a building by using long short-term memory network: a forecasting study
CN114119273A (en) Park comprehensive energy system non-invasive load decomposition method and system
CN113449919B (en) Power consumption prediction method and system based on feature and trend perception
CN116167531A (en) Photovoltaic power generation prediction method based on digital twin
CN115310727B (en) Building cooling, heating and power load prediction method and system based on transfer learning
CN113821903A (en) Temperature control method and device, modular data center and storage medium
CN113128666A (en) Mo-S-LSTMs model-based time series multi-step prediction method
CN114384931B (en) Multi-target optimal control method and equipment for unmanned aerial vehicle based on strategy gradient
CN116976488A (en) Air conditioner energy consumption prediction method for optimizing BP neural network based on sparrow search algorithm
CN111709585A (en) Air conditioner load prediction method and device and storage medium
CN114240687A (en) Energy hosting efficiency analysis method suitable for comprehensive energy system
Li et al. Practice and application of LSTM in temperature prediction of HVAC system
CN115879190A (en) Model construction method and device and building load prediction method and device
Kreider et al. Expert systems, neural networks and artificial intelligence applications in commercial building HVAC operations

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200228