CN117473275B - Energy consumption detection method for data center - Google Patents
Energy consumption detection method for data center Download PDFInfo
- Publication number
- CN117473275B CN117473275B CN202311818174.4A CN202311818174A CN117473275B CN 117473275 B CN117473275 B CN 117473275B CN 202311818174 A CN202311818174 A CN 202311818174A CN 117473275 B CN117473275 B CN 117473275B
- Authority
- CN
- China
- Prior art keywords
- energy consumption
- data
- time
- consumption detection
- 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.)
- Active
Links
- 238000005265 energy consumption Methods 0.000 title claims abstract description 99
- 238000001514 detection method Methods 0.000 title claims abstract description 64
- 238000000034 method Methods 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 23
- 230000009466 transformation Effects 0.000 claims abstract description 13
- 230000006870 function Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 11
- 230000004927 fusion Effects 0.000 claims description 10
- 230000009467 reduction Effects 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000010187 selection method Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 2
- 230000007613 environmental effect Effects 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 9
- 238000004458 analytical method Methods 0.000 abstract description 3
- 238000004590 computer program Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 238000003491 array Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000004378 air conditioning Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 239000000498 cooling water Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3058—Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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
-
- 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
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses an energy consumption detection method of a data center, which relates to the technical field of energy consumption detection and comprises the steps of collecting original data of the data center and fusing; based on the VTGAN model, taking LSTM as a generator and 1D CNN as a discriminator to generate an energy consumption detection model; analyzing the energy consumption sequence in the time and frequency domain by using Fourier transformation and wavelet transformation to form a comprehensive feature set; training and verifying the energy consumption detection model through the comprehensive feature set; the size of the sliding window is dynamically adjusted, the time sequence is decomposed into a plurality of subsequences, detection is carried out respectively, then the detection results are combined, and the energy consumption condition is determined according to the detection results. According to the method, the advantages of LSTM and 1D CNN are combined by introducing the VTGAN model, so that time sequence data are effectively processed, and the accuracy and efficiency of energy consumption monitoring are improved. The invention enables a comprehensive analysis of energy consumption data in the time and frequency domain by using fourier transforms and wavelet transforms.
Description
Technical Field
The invention relates to the technical field of energy consumption detection, in particular to an energy consumption detection method of a data center.
Background
In the operational management of data centers, energy consumption monitoring is a critical issue. Traditional energy consumption monitoring methods mainly rely on simple historical data analysis and basic prediction models, and often cannot accurately capture complex modes of energy consumption data, especially when facing large-scale and high-dynamic data center environments. In addition, when the traditional method processes time series data, frequency characteristics and nonlinear characteristics of the data are often ignored, so that the accuracy of energy consumption prediction and anomaly detection is insufficient. Accordingly, there is an urgent need for techniques to improve the accuracy and efficiency of data center energy consumption monitoring, particularly in terms of real-time monitoring and prediction of energy consumption.
Disclosure of Invention
The present invention has been made in view of the above-described problems occurring in the conventional energy consumption detection method of a data center.
Therefore, the problem to be solved by the present invention is how to provide a method for detecting energy consumption of a data center.
In order to solve the technical problems, the invention provides the following technical scheme: the energy consumption detection method of the data center comprises the steps of collecting original data of the data center and fusing; based on the VTGAN model, taking LSTM as a generator and 1D CNN as a discriminator to generate an energy consumption detection model; analyzing the energy consumption sequence in the time and frequency domain by using Fourier transformation and wavelet transformation to form a comprehensive feature set; training and verifying the energy consumption detection model through the comprehensive feature set; the size of the sliding window is dynamically adjusted, the time sequence is decomposed into a plurality of subsequences, detection is carried out respectively, then the detection results are combined, and the energy consumption condition is determined according to the detection results.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: collecting data center raw data and fusing the raw data, wherein the raw data comprises power consumption data, environment parameter data, equipment operation data and network use data; preprocessing the collected original data; by selecting the most contributing feature to energy consumption detection by the information gain based feature selection method, the formula is as follows,
where IG (X, Y) is the information gain of the feature X on the target variable Y, H (Y) is the entropy of the target variable Y, and H (y|x) is the conditional entropy of the target variable Y given the feature X;
the selected characteristics are subjected to dimension reduction processing based on a PCA algorithm, the characteristics after the dimension reduction processing are fused based on a self-encoder, the formula is as follows,
wherein F is the integrated feature after fusion, AE is the self-encoder, E 1 ,E 2 ,…,E n Is the feature after dimension reduction treatment;
the time series data is constructed using a fixed size sliding window method, as follows,
wherein S is t Is time series data of time t, F is integrated characteristic after fusion, and w is the size of a sliding window.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: the method comprises the steps of constructing a multi-layer LSTM network as a generator, wherein time sequence data is used as input, constructing a hidden layer by using multi-layer LSTM units, and generating output with the same dimension as target energy consumption data at an output layer by using a ReLU activation function; constructing a 1D CNN network as a discriminator, comprising the steps of receiving energy consumption data generated by a generator, extracting characteristics of the energy consumption data by using a plurality of convolution layers, reducing the dimension of the data by using a maximum pooling layer, outputting a classification result which indicates whether the input energy consumption data is real or generated, and converting the output into a probability value between 0 and 1 by using a sigmoid activation function at an output layer; the wasperstein loss will be used as an objective function, expressed as,
where L is the value of the objective function, D (x) is the output of the arbiter for the real data x, D (G (z)) is the output of the arbiter for the generated data G (z), E [. Cndot. ] represents the desired value;
training an energy consumption detection model by using an Adam optimization algorithm; the generator and the arbiter will be trained alternately until the energy consumption detection model converges.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: analyzing the energy consumption sequence in the time and frequency domains using fourier transforms and wavelet transforms to form a comprehensive feature set, including, normalizing the energy-consuming time sequence data; fourier transform is applied to the normalized time-series data X (t) to obtain frequency domain features F (ω),
in the method, in the process of the invention,frequency, i is an imaginary unit, and t is time;
wavelet transformation is applied to the time series data X (t) to obtain wavelet coefficients of different scales and positions,
where a is a scale parameter, b is a position parameter,is a wavelet function;
and fusing the frequency domain features and the wavelet coefficients to form a comprehensive feature set.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: the step of fusing the frequency domain features and the wavelet coefficients comprises the step of selecting key frequency features from Fourier transform results, wherein the key frequency features comprise main frequency components and energy distribution; selecting key features from wavelet transformation results, including wavelet coefficients and energy features of specific scales and positions; splicing the selected Fourier transform characteristic and wavelet transform characteristic to form a fusion characteristic vector; and carrying out standardization processing and dimension reduction processing on the fusion feature vector to form a comprehensive feature set.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: training and verifying the energy consumption detection model by the comprehensive feature set comprises dividing the comprehensive feature set into a training set, a verification set and a test set; training the LSTM generator using a training set to generate a simulated sample of energy consumption data; training a 1D CNN discriminator using the generated simulated samples and the real samples to distinguish between the real and generated data; evaluating the performance of the model through the verification set, and adjusting the parameters of the energy consumption detection model according to the verification result; and testing the final performance of the energy consumption detection model through a test set.
As a preferable scheme of the energy consumption detection method of the data center, the invention comprises the following steps: the step of dynamically adjusting the size of the sliding window comprises selecting an initial window size based on the performance of the historical data; at each time step, calculating a detection error Et of the energy consumption detection model; if Et is greater than the threshold ϵ, increasing the size of the sliding window; otherwise, reducing the size of the sliding window; the size of the sliding window is updated, as follows,
in which W is t+1 Is the window size of the next time step, alpha is the adjustment coefficient, W t Refers to the window size at time t.
A computer readable storage medium having stored thereon a computer program, characterized in that the computer program when executed by a processor implements the steps of a method for detecting energy consumption of a data center.
A computer device comprising a memory and a processor, said memory storing a computer program, characterized in that said processor, when executing said computer program, implements the steps of the method for detecting energy consumption of a data center.
The invention has the beneficial effects that: by introducing the VTGAN model and combining the advantages of LSTM and 1D CNN, the time sequence data is effectively processed, and the accuracy and efficiency of energy consumption monitoring are improved. By using fourier transforms and wavelet transforms, the invention enables comprehensive analysis of energy consumption data in the time and frequency domains, capturing more complex energy consumption patterns, including periodic and aperiodic variations. In addition, the method for dynamically adjusting the sliding window enables the model to adapt to dynamic changes of data, and flexibility and accuracy of detection are further improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:
FIG. 1 is a conceptual diagram of a method for detecting energy consumption of a data center;
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways other than those described herein, and persons skilled in the art will readily appreciate that the present invention is not limited to the specific embodiments disclosed below.
Further, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic can be included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
Example 1
Referring to fig. 1, a first embodiment of the present invention provides a method for detecting energy consumption of a data center, where the method for detecting energy consumption of a data center includes the following steps:
s1, collecting original data of a data center, and fusing;
s2, based on a VTGAN model, taking LSTM as a generator and taking 1D CNN as a discriminator to generate an energy consumption detection model;
s3, analyzing the energy consumption sequence in the time and frequency domain by using Fourier transformation and wavelet transformation to form a comprehensive feature set;
s4, training and verifying the energy consumption detection model through the comprehensive feature set;
s5, dynamically adjusting the size of the sliding window, decomposing the time sequence into a plurality of subsequences, detecting the subsequences respectively, merging detection results, and determining the energy consumption condition according to the detection results.
In step S1, the collecting and fusing data center raw data includes,
s11, collecting original data from a data center, wherein the original data comprise power consumption data, environment parameter data, equipment operation data and network use data, and the power consumption data comprise total power consumption, power consumption of each server and equipment and power consumption of different areas (such as a server room and a cooling system area); environmental parameters include indoor temperature and humidity, air flow and state of the air conditioning system, and temperature and flow of cooling water; the device operation data includes the operation state (such as on/off, load condition) of the server and other key devices, the operation time and the use frequency of the devices and the hardware performance index (such as CPU and memory utilization rate); the network usage data comprises data transmission quantity and energy consumption of network equipment;
s12, preprocessing operation is carried out on the collected original data, wherein the preprocessing operation comprises data cleaning and standardization;
s13, selecting the most contributing characteristic to energy consumption detection by a characteristic selection method based on information gain, wherein the formula is as follows,
where IG (X, Y) is the information gain of the feature X on the target variable Y, H (Y) is the entropy of the target variable Y, and H (y|x) is the conditional entropy of the target variable Y given the feature X;
s14, performing dimension reduction processing on the selected features based on a PCA algorithm, fusing the dimension-reduced features based on a self-encoder, and adopting the following formula,
wherein F is the integrated feature after fusion, AE is the self-encoder, E 1 ,E 2 ,…,E n Is the feature after dimension reduction treatment;
s15, constructing time series data by using a sliding window method with a fixed size, wherein the formula is as follows,
wherein S is t Is time series data of time t, F is integrated characteristic after fusion, and w is large of sliding window.
In step S2, the generating of the energy consumption detection model comprises the steps of,
s21, constructing a multi-layer LSTM network as a generator, wherein time sequence data is taken as input, a hidden layer is constructed by using multi-layer LSTM units, and an output with the same dimension as that of target energy consumption data is generated at an output layer by using a ReLU activation function;
s22, constructing a 1D CNN network as a discriminator, wherein the method comprises the steps of receiving energy consumption data generated by a generator, extracting the characteristics of the energy consumption data by using a plurality of convolution layers, reducing the dimension of the data by using a maximum pooling layer, outputting a classification result which indicates whether the input energy consumption data is real or generated, and converting the output into a probability value between 0 and 1 by using a sigmoid activation function at an output layer;
s23, using Wasserstein loss as an objective function, wherein the formula is,
where L is the value of the objective function, D (x) is the output of the arbiter for the real data x, D (G (z)) is the output of the arbiter for the generated data G (z), E [. Cndot. ] represents the desired value;
s24, training an energy consumption detection model by using an Adam optimization algorithm;
s25, training the generator and the discriminator alternately until the energy consumption detection model converges.
In step S3, the energy consumption sequence is analyzed in the time and frequency domain using fourier transforms and wavelet transforms, forming a comprehensive feature set, including,
s31, carrying out standardization processing on the time-consuming sequence data;
s32, applying Fourier transform to the normalized time series data X (t) to obtain frequency domain characteristics F (omega),
in the method, in the process of the invention,frequency, i is an imaginary unit, and t is time;
s33, wavelet transformation is applied to the time sequence data X (t) to obtain wavelet coefficients of different scales and positions,
where a is a scale parameter, b is a position parameter,is a wavelet function;
s34, fusing the frequency domain features and the wavelet coefficients to form a comprehensive feature set.
In step S4, training and validating the energy consumption detection model by integrating the feature set includes,
s41, dividing the comprehensive feature set into a training set, a verification set and a test set;
s42, training the LSTM generator by using a training set to generate a simulation sample of the energy consumption data;
s43, training a 1D CNN discriminator by using the generated simulation sample and the real sample to distinguish real and generated data;
s44, evaluating the performance of the model through the verification set, and adjusting the parameters of the energy consumption detection model according to the verification result;
s45, testing the final performance of the energy consumption detection model through a test set.
In step S5, the dynamically adjusting the size of the sliding window includes,
s51, selecting an initial window size based on historical data performance;
s52, calculating a detection error Et of the energy consumption detection model at each time step;
s53, if Et is larger than a threshold ϵ, increasing the size of the sliding window; otherwise, reducing the size of the sliding window;
s54, updating the size of the sliding window, wherein the formula is as follows,
in which W is t+1 Is the window size of the next time step, alpha is the adjustment coefficient, W t Refers to the window size at time t.
By introducing an advanced VTGAN model and combining the advantages of LSTM and 1D CNN, the invention effectively processes time sequence data and improves the accuracy and efficiency of energy consumption monitoring. By using fourier transforms and wavelet transforms, the present invention enables comprehensive analysis of energy consumption data in the time and frequency domains, capturing more complex energy consumption patterns, including periodic and aperiodic variations. In addition, the method for dynamically adjusting the sliding window enables the model to adapt to dynamic changes of data, and flexibility and accuracy of detection are further improved. Therefore, the invention not only solves the defects of the traditional energy consumption monitoring method in terms of accuracy and real-time, but also provides a more efficient and intelligent energy consumption monitoring and management solution for the data center.
Example 2
A second embodiment of the invention, which is different from the previous embodiment, is:
the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
Logic and/or steps represented in the flowcharts or otherwise described herein, e.g., a ordered listing of executable instructions for implementing logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program is printed, as the program may be electronically captured, via, for instance, optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner, if necessary, and then stored in a computer memory.
It is to be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
Claims (5)
1. The energy consumption detection method for the data center is characterized by comprising the following steps of: comprising the steps of (a) a step of,
collecting original data of a data center, and fusing;
based on the VTGAN model, taking LSTM as a generator and 1D CNN as a discriminator to generate an energy consumption detection model;
analyzing the energy consumption sequence in the time and frequency domain by using Fourier transformation and wavelet transformation to form a comprehensive feature set;
training and verifying the energy consumption detection model through the comprehensive feature set;
dynamically adjusting the size of a sliding window, decomposing the time sequence into a plurality of subsequences, respectively detecting, then merging detection results, and determining the energy consumption condition according to the detection results;
the collecting and fusing data center raw data includes,
collecting raw data from a data center, the raw data including power consumption data, environmental parameter data, equipment operation data, and network usage data;
preprocessing the collected original data;
by selecting the most contributing feature to energy consumption detection by the information gain based feature selection method, the formula is as follows,
where IG (X, Y) is the information gain of the feature X on the target variable Y, H (Y) is the entropy of the target variable Y, and H (y|x) is the conditional entropy of the target variable Y given the feature X;
the selected characteristics are subjected to dimension reduction processing based on a PCA algorithm, the characteristics after the dimension reduction processing are fused based on a self-encoder, the formula is as follows,
wherein F is the integrated feature after fusion, AE is the self-encoder, E 1 ,E 2 ,…,E n Is the feature after dimension reduction treatment;
the time series data is constructed using a fixed size sliding window method, as follows,
wherein S is t Is time series data of time t, F is integrated characteristic after fusion, and w is the size of a sliding window;
the generating of the energy consumption detection model comprises the steps of,
constructing a multi-layer LSTM network as a generator, wherein the multi-layer LSTM network comprises the steps of taking time sequence data as input, constructing a hidden layer by using multi-layer LSTM units, and generating output with the same dimension as target energy consumption data at an output layer by using a ReLU activation function;
constructing a 1D CNN network as a discriminator, comprising the steps of receiving energy consumption data generated by a generator, extracting characteristics of the energy consumption data by using a plurality of convolution layers, reducing the dimension of the data by using a maximum pooling layer, outputting a classification result which indicates whether the input energy consumption data is real or generated, and converting the output into a probability value between 0 and 1 by using a sigmoid activation function at an output layer;
the wasperstein loss will be used as an objective function, expressed as,
where L is the value of the objective function, D (x) is the output of the arbiter for the real data x, D (G (z)) is the output of the arbiter for the generated data G (z), E [. Cndot. ] represents the desired value;
training an energy consumption detection model by using an Adam optimization algorithm;
the generator and the arbiter will be trained alternately until the energy consumption detection model converges.
2. The method for detecting energy consumption of a data center according to claim 1, wherein: the energy consumption sequences are analyzed in the time and frequency domain using fourier transforms and wavelet transforms, forming a comprehensive feature set comprising,
carrying out standardization processing on the energy-consuming time sequence data;
fourier transform is applied to the normalized time-series data X (t) to obtain frequency domain features F (ω),
in the method, in the process of the invention,frequency, i is an imaginary unit, and t is time;
wavelet transformation is applied to the time series data X (t) to obtain wavelet coefficients of different scales and positions,
where a is a scale parameter, b is a position parameter,is a wavelet function;
and fusing the frequency domain features and the wavelet coefficients to form a comprehensive feature set.
3. The method for detecting energy consumption of a data center according to claim 2, wherein: said fusing the frequency domain features and the wavelet coefficients includes,
selecting key frequency features from the fourier transform result, including a main frequency component and an energy distribution;
selecting key features from wavelet transformation results, including wavelet coefficients and energy features of specific scales and positions;
splicing the selected Fourier transform characteristic and wavelet transform characteristic to form a fusion characteristic vector;
and carrying out standardization processing and dimension reduction processing on the fusion feature vector to form a comprehensive feature set.
4. A method of energy consumption detection for a data center as claimed in claim 3, wherein: training and validating the energy consumption detection model by integrating the feature set includes,
dividing the comprehensive feature set into a training set, a verification set and a test set;
training the LSTM generator using a training set to generate a simulated sample of energy consumption data;
training a 1D CNN discriminator using the generated simulated samples and the real samples to distinguish between the real and generated data;
evaluating the performance of the model through the verification set, and adjusting the parameters of the energy consumption detection model according to the verification result;
and testing the final performance of the energy consumption detection model through a test set.
5. A method of energy consumption detection for a data center as claimed in claim 3, wherein: the dynamically adjusting the size of the sliding window includes,
selecting an initial window size based on the performance of the historical data;
at each time step, calculating a detection error Et of the energy consumption detection model;
if Et is greater than the threshold valueIncreasing the size of the sliding window; otherwise, reducing the size of the sliding window;
the size of the sliding window is updated, as follows,
in which W is t+1 Is the window size of the next time step, alpha is the adjustment coefficient, W t Refers to the window size at time t.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311818174.4A CN117473275B (en) | 2023-12-27 | 2023-12-27 | Energy consumption detection method for data center |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311818174.4A CN117473275B (en) | 2023-12-27 | 2023-12-27 | Energy consumption detection method for data center |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117473275A CN117473275A (en) | 2024-01-30 |
| CN117473275B true CN117473275B (en) | 2024-03-26 |
Family
ID=89624140
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311818174.4A Active CN117473275B (en) | 2023-12-27 | 2023-12-27 | Energy consumption detection method for data center |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117473275B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598851A (en) * | 2019-08-29 | 2019-12-20 | 北京航空航天大学合肥创新研究院 | Time series data abnormity detection method fusing LSTM and GAN |
| CN115878992A (en) * | 2021-09-28 | 2023-03-31 | 比亚迪股份有限公司 | Monitoring method and monitoring system for comprehensive pipe rack power supply system |
| CN116318830A (en) * | 2023-02-02 | 2023-06-23 | 北京邮电大学 | Log intrusion detection system based on generation of countermeasure network |
| CN116976472A (en) * | 2022-04-13 | 2023-10-31 | 北京国双科技有限公司 | Method, system, equipment and storage medium for predicting residential electric energy consumption |
-
2023
- 2023-12-27 CN CN202311818174.4A patent/CN117473275B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598851A (en) * | 2019-08-29 | 2019-12-20 | 北京航空航天大学合肥创新研究院 | Time series data abnormity detection method fusing LSTM and GAN |
| CN115878992A (en) * | 2021-09-28 | 2023-03-31 | 比亚迪股份有限公司 | Monitoring method and monitoring system for comprehensive pipe rack power supply system |
| CN116976472A (en) * | 2022-04-13 | 2023-10-31 | 北京国双科技有限公司 | Method, system, equipment and storage medium for predicting residential electric energy consumption |
| CN116318830A (en) * | 2023-02-02 | 2023-06-23 | 北京邮电大学 | Log intrusion detection system based on generation of countermeasure network |
Non-Patent Citations (1)
| Title |
|---|
| 基于LSTM—GAN 的加油时序数据异常检测;赵飏;计算机应用与软件;20220712;第39卷(第第13-19页期);第13-19页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117473275A (en) | 2024-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xie et al. | Time series prediction based on NARX neural networks: An advanced approach | |
| Liu et al. | An enhanced encoder–decoder framework for bearing remaining useful life prediction | |
| CN113344295B (en) | Method, system and medium for predicting residual life of equipment based on industrial big data | |
| CN111382906B (en) | Power load prediction method, system, equipment and computer readable storage medium | |
| CN107292446B (en) | Hybrid wind speed prediction method based on component relevance wavelet decomposition | |
| KR102149495B1 (en) | Optimization apparatus for training conditions of environmental prediction model and operating thereof | |
| CN111950805B (en) | Medium-and-long-term power load prediction method and device, computer equipment and storage medium | |
| CN114970774B (en) | Intelligent transformer fault prediction method and device | |
| CN113392732B (en) | Partial discharge ultrasonic signal anti-interference method and system | |
| CN110879927A (en) | Sea clutter amplitude statistical distribution field modeling method for sea target detection | |
| CN111310963A (en) | Power generation data prediction method and device for power station, computer equipment and storage medium | |
| CN112085111A (en) | Load identification method and device | |
| Oh et al. | Investigating the effects of daily travel time patterns on short-term prediction | |
| Yan et al. | Forecasting uranium resource price prediction by extreme learning machine with empirical mode decomposition and phase space reconstruction | |
| Chen et al. | Multi-scale split dual calibration network with periodic information for interpretable fault diagnosis of rotating machinery | |
| CN116668198B (en) | Flow playback test method, device, equipment and medium based on deep learning | |
| CN117369418A (en) | Fault diagnosis method, system, storage medium and equipment for feeding system of numerical control machine tool | |
| CN117473275B (en) | Energy consumption detection method for data center | |
| CN117216522A (en) | Bearing fault type diagnosis method, device and equipment based on machine learning | |
| Feng et al. | Spatiotemporal prediction based on feature classification for multivariate floating-point time series lossy compression | |
| Mishra et al. | Power quality disturbances classification with deep learning approach | |
| CN116561569A (en) | Industrial power load identification method based on EO feature selection and AdaBoost algorithm | |
| CN116502672A (en) | Neural network quantitative deployment method, system, equipment and medium | |
| CN116758922A (en) | Voiceprint monitoring and diagnosing method for transformer | |
| Dang et al. | seq2graph: discovering dynamic dependencies from multivariate time series with multi-level attention |
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 |