CN116383773A - Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval - Google Patents
Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval Download PDFInfo
- Publication number
- CN116383773A CN116383773A CN202310047835.1A CN202310047835A CN116383773A CN 116383773 A CN116383773 A CN 116383773A CN 202310047835 A CN202310047835 A CN 202310047835A CN 116383773 A CN116383773 A CN 116383773A
- Authority
- CN
- China
- Prior art keywords
- energy efficiency
- data
- value
- interval
- adaptive prediction
- 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
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 claims abstract description 81
- 230000002159 abnormal effect Effects 0.000 claims abstract description 51
- 230000003044 adaptive effect Effects 0.000 claims abstract description 48
- 238000005457 optimization Methods 0.000 claims abstract description 32
- 238000007781 pre-processing Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims description 26
- 230000006870 function Effects 0.000 claims description 22
- 238000005265 energy consumption Methods 0.000 claims description 19
- 238000012795 verification Methods 0.000 claims description 18
- 238000010187 selection method Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 230000008030 elimination Effects 0.000 claims description 6
- 238000003379 elimination reaction Methods 0.000 claims description 6
- 238000001816 cooling Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 238000005057 refrigeration Methods 0.000 claims description 3
- 230000008485 antagonism Effects 0.000 claims description 2
- 230000007704 transition Effects 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 3
- 230000005856 abnormality Effects 0.000 description 27
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002542 deteriorative effect Effects 0.000 description 1
- 238000012850 discrimination method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
- G06Q10/20—Administration of product repair or maintenance
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Pure & Applied Mathematics (AREA)
- Educational Administration (AREA)
- Mathematical Physics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Bioinformatics & Cheminformatics (AREA)
Abstract
The invention discloses a data center energy efficiency anomaly detection method, a system and a medium based on a self-adaptive prediction interval, wherein the method comprises the following steps: selecting an energy efficiency index, collecting and storing field historical data, and calculating an energy efficiency value; preprocessing historical data; establishing a regression prediction model through feature selection and super-parameter optimization; calculating a time-by-time energy efficiency prediction error set, calculating probability density of the time-by-time energy efficiency prediction error set, and superposing a confidence interval to obtain a self-adaptive prediction error interval; preprocessing real-time operation data, inputting the preprocessed real-time operation data into a trained regression prediction model to obtain an energy efficiency predicted value at the current moment, superposing an adaptive prediction error interval to obtain an energy efficiency adaptive prediction interval at the current moment, and comparing whether an energy efficiency actual value is in the interval to obtain an energy efficiency abnormal detection result. The invention can better cover the energy efficiency change range of the data center, effectively avoid the random error and modeling error of the system, realize the reliable detection of abnormal energy efficiency and improve the energy efficiency management level of the data center.
Description
Technical Field
The invention belongs to the technical field of data center energy efficiency anomaly detection, and particularly relates to a data center energy efficiency anomaly detection method, system and medium based on a self-adaptive prediction interval.
Background
In recent years, with the development and maturity of emerging technologies, the data center business market is rapidly developing, and the industry has been put into the category of "new infrastructure" by the country. The data center is increasingly prominent in the problems of high energy consumption and low energy efficiency level while promoting rapid development of digital economy. In the running process of the system, the energy efficiency of the data center is possibly abnormal due to the problems of equipment abnormality, operation abnormality, IT equipment energy consumption abnormality and the like, if the abnormal conditions are not found in time and countermeasures are taken, the long-term running abnormality of the system can not only cause energy waste, but also possibly cause larger abnormality and even failure. Therefore, the abnormal detection of the energy efficiency index of the data center is enhanced, and the method has important significance for improving the energy utilization efficiency and the system operation reliability and realizing the high-efficiency machine room.
At present, a machine learning method is often adopted as a method for monitoring and evaluating energy efficiency indexes of a data center. By adopting a supervised learning method, a prediction model of the relation between the operation parameters and the overall energy efficiency is established, and energy efficiency monitoring and evaluation can be realized. However, establishing an energy efficiency prediction model based on the operation parameters cannot avoid the energy efficiency abnormality caused by the abnormality of the operation parameters, and cannot determine whether the energy efficiency is abnormal at this time. Meanwhile, the anomaly detection is carried out according to the predicted value of the prediction model, the energy efficiency change range of the data center cannot be reflected, the system random error and the modeling error cannot be effectively avoided, and the reliability of the anomaly energy efficiency detection result is to be questionable.
Disclosure of Invention
The invention mainly aims to overcome the defects and shortcomings of the prior art and provide a data center energy efficiency anomaly detection method, a system and a medium based on a self-adaptive prediction interval, which predict a reasonable range of data center energy efficiency according to historical energy efficiency data, outdoor meteorological parameters and time parameters, realize reliable detection of the anomaly energy efficiency and improve the energy efficiency management level of the data center.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the invention provides a data center energy efficiency anomaly detection method based on an adaptive prediction interval, which comprises the following steps:
selecting an energy efficiency index, collecting outdoor meteorological parameters, time data and energy consumption data, and calculating energy efficiency data according to the energy consumption data;
preprocessing historical data, including abnormal value detection, rejection and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data;
adopting a feature selection method to perform feature selection on the preprocessed historical data; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
training a regression prediction model by using the data with the characteristics selected, and determining an optimal super-parameter combination by using a model super-parameter optimization method so as to obtain a trained regression prediction model;
calculating a predicted value of the time-by-time energy efficiency by using the trained regression prediction model, and subtracting an actual value from the predicted value to obtain a predicted error set;
for the prediction error set, calculating probability density of the prediction error set by adopting a kernel density estimation algorithm, and superposing a confidence interval to obtain a self-adaptive prediction error interval;
processing the data acquired in real time by using the data preprocessing method which is the same as that of the historical data, constructing an input set by adopting the characteristics selected in the characteristic selection method, inputting a trained regression prediction model to obtain the energy efficiency predicted value at the current moment, superposing the self-adaptive prediction error interval to obtain the energy efficiency self-adaptive prediction interval at the current moment, and comparing whether the actual energy efficiency value is in the energy efficiency self-adaptive prediction interval at the current moment to judge whether the energy efficiency is abnormal or not.
As a preferable technical solution, the energy efficiency index includes: electric energy utilization rate, refrigeration load coefficient, power supply load coefficient and renewable energy utilization rate;
the energy consumption data comprise parameter data which are used for calculating energy efficiency data and are directly acquired through a sensor;
the energy efficiency data includes: the current time energy efficiency value, the previous 1 time energy efficiency value and the previous 24 time energy efficiency value;
the outdoor weather parameters include: outdoor temperature, outdoor humidity;
the time data includes: year, month, date, time of day, season type, and workday type; among them, season types include: a cooling season, a transition season, and a non-cooling season; workday types include: workday, non-workday.
As an optimal technical scheme, the method for detecting and rejecting the abnormal value of the energy efficiency data is 3 sigma criterion:
L1＝μ-3σ
L2＝μ+3σ
wherein X is i For a certain column of i-th data, N is the total number of samples, mu is the average value of the column of data, sigma is the standard deviation of the column of data, L1 is the lower judgment limit of an abnormality threshold, L2 is the upper judgment limit of the abnormality threshold, and the abnormality value judgment standard is that the sample value is smaller than the lower judgment limit L1 or larger than the upper judgment limit L2;
respectively eliminating abnormal values of the current time energy efficiency value, the previous 1 time energy efficiency value and the previous 24 time energy efficiency value;
the method for carrying out missing value interpolation on the energy efficiency data is an interpolation method:
X i ＝X i+24k
wherein X is i For a certain column of ith data, i.e. a certain time energy efficiency value, X i+24k For the energy efficiency value at the same time of the previous k days, when k=1, if X i+24 Taking X as normal value i+24 Assignment of values to X i If X i+24 For an outlier, when looking at k=2, X i+24*2 Whether the value is normal or not, if so, assigning the value to X i If not, look at k=3 until X i+24k Until the value is normal, the normal value X i+24k Assignment to X i ；
Performing missing value interpolation on the energy efficiency value at the previous 1 moment and the energy efficiency value at the previous 24 moment;
the method for eliminating the distortion value of the outdoor meteorological data comprises the following steps of:
wherein X is i For a certain column of ith data, the abnormal value judgment standard is that the difference value between the sample value and the data average value of the front moment and the back moment is more than 5;
the method for carrying out missing value interpolation on the outdoor meteorological data is a mean value interpolation method:
wherein X is i For the ith data of a certain column, taking the average value of the data at the front and rear moments and assigning the average value to X i 。
As an optimal technical scheme, the feature selection method adopts one of a Pearson correlation coefficient method, an embedding method, a mutual information method and a recursive feature elimination method.
As a preferable technical solution, the regression prediction model specifically includes:
the regression prediction model fitting adopts an artificial intelligent regression algorithm, and adopts one of a vector machine, a decision tree, a neural network, ensemble learning, deep learning and an antagonism network;
the regression prediction model super-parameter optimization method adopts one of random search, grid search, bayesian optimization and evolutionary algorithm.
As a preferable technical scheme, the specific process of the regression prediction model fitting is as follows:
randomly dividing a data sample into a training set and a testing set according to a certain proportion, and using the training set and the testing set for model training and verification testing;
constructing an artificial intelligent algorithm regression prediction model, and training and verifying the data set by using default parameters;
optimizing the super parameters of the regression prediction model by using a super parameter optimization method by taking the evaluation index of the regression prediction model as an optimization target, wherein the verification method is k-fold cross verification;
storing the regression prediction model optimized by the steps;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the regression prediction model is updated.
As an optimal technical scheme, the method adopts a kernel density estimation algorithm to calculate the probability density of the prediction error set, and superimposes a confidence interval to obtain a self-adaptive prediction error interval, and comprises the following specific procedures:
leading in a data sample prediction error set, carrying out bandwidth optimization on a nuclear density estimation model by adopting a super-parameter optimization method, and performing k-fold cross verification by adopting a verification method;
establishing a kernel density estimation model by adopting an optimal bandwidth h, and calculating the probability density of each prediction error value e, wherein the probability density function is as follows:
wherein K (e) is a kernel function, and one of a Gaussian kernel function, a gamma kernel function, a uniform kernel function and a triangular kernel function is adopted;
confidence interval [ f ] with confidence of 1-alpha in calculating probability density α/2 ，f 1-α/2 ]Constructing an adaptive prediction error interval [ e ] according to the functional relation between the probability density and the prediction error value 1 ，e 2 ]；
After the update time period T, the steps are carried out again by using the latest accumulated data set, and the adaptive prediction error interval is updated.
As an preferable technical scheme, the method for obtaining the energy efficiency prediction value of the current time and the adaptive prediction error interval are overlapped to obtain the energy efficiency adaptive prediction interval of the current time, specifically:
the energy efficiency predicted value b at the current moment and the self-adaptive prediction error interval [ e ] 1 ，e 2 ]Superposition to obtain the energy efficiency self-adaptive prediction interval [ b+e ] at the current moment 1 ，b+e 2 ]；
Whether the energy efficiency is abnormal is judged by comparing whether the actual value of the energy efficiency is in the energy efficiency self-adaptive prediction interval at the current moment, specifically:
judging whether the actual value c of the energy efficiency at the current moment is in the energy efficiency self-adaptive prediction interval [ b+e ] at the current moment 1 ，b+e 2 ]If not, the energy efficiency is abnormal.
The invention also provides a data center energy efficiency abnormality detection system based on the self-adaptive prediction interval, which is applied to the data center energy efficiency abnormality detection method based on the self-adaptive prediction interval, and comprises a data acquisition module, a data preprocessing module, a feature selection module, a regression prediction model training module, a self-adaptive prediction error interval construction module and an abnormality detection module;
the data acquisition module is used for selecting energy efficiency indexes, acquiring outdoor meteorological parameters, time data and energy consumption data through the sensor, and calculating energy efficiency data according to the energy consumption data;
the data preprocessing module is used for preprocessing historical data, including abnormal value detection, elimination and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data;
the feature selection module is used for performing feature selection on the preprocessed historical data by adopting a feature selection method; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
the model training module is used for training a regression prediction model by using the data after feature selection, determining the optimal super-parameter combination by using a model super-parameter optimization method, and further obtaining a trained regression prediction model;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the regression prediction model is updated;
the self-adaptive prediction error interval construction module is used for generating a prediction error set, calculating probability density of the self-adaptive prediction error interval by adopting a kernel density estimation algorithm, and overlapping confidence intervals to obtain a self-adaptive prediction error interval;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the self-adaptive prediction error interval is updated;
the anomaly detection module is used for inputting the data to be detected into the trained regression prediction model to obtain the energy efficiency prediction value at the current moment, superposing the self-adaptive prediction error interval to obtain the energy efficiency self-adaptive prediction interval at the current moment, and comparing whether the actual energy efficiency value is in the energy efficiency self-adaptive prediction interval at the current moment to judge whether the energy efficiency is abnormal or not so as to obtain an anomaly detection result.
In another aspect of the present invention, a storage medium is further provided, where a program is stored, where the program, when executed by a processor, implements the method for detecting an energy efficiency anomaly of a data center based on an adaptive prediction interval.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) The data center energy efficiency abnormity detection method based on the regression prediction model and the self-adaptive prediction interval can rapidly and accurately evaluate the operation energy efficiency real-time condition of the complex nonlinear data center energy consumption system, is beneficial to data center operation maintenance personnel to rapidly make corresponding decisions according to the evaluation result of the operation energy efficiency real-time condition, prevents the system operation condition from further deteriorating, and improves the energy efficiency of the system;
(2) The regression prediction model superposition self-adaptive prediction interval anomaly detection method provided by the invention can better cover the energy efficiency change range of the data center, more comprehensively reflect the running state of the data center, effectively avoid the system random error and modeling error, and improve the reliability of the data center anomaly energy efficiency detection result;
(3) The self-adaptive prediction error interval constructed based on the kernel density estimation algorithm can adaptively determine the upper and lower limits of the interval according to the distribution condition of sample data, and reduces the false alarm rate of anomaly detection;
(4) According to the invention, the real-time data set is used regularly to dynamically update the existing anomaly detection model, so that the accuracy of anomaly detection is improved;
(5) The method and the system provided by the invention are generally applicable to the data center with stable operation, are applied to the energy efficiency management of the data center, and have better popularization.
Drawings
FIG. 1 is a flowchart of a data center energy efficiency anomaly detection method based on an adaptive prediction interval according to an embodiment of the present invention;
FIG. 2 is an energy efficiency data interpolation diagram according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a data center energy efficiency anomaly detection system based on an adaptive prediction interval according to an embodiment of the present invention;
fig. 4 is a schematic structural view of a storage medium according to an embodiment of the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those described herein, such as the selection of data center energy efficiency metrics is not limited to PUEs in embodiments, regression prediction models are not limited to lightgbms in embodiments, and hyper-parametric optimization methods are not limited to bayesian optimization and grid search in embodiments. Therefore, the scope of the invention is not limited by the specific embodiments disclosed below. All other embodiments, which can be made by those skilled in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.
Example 1
In this embodiment, a certain data center building in the south area is taken as an example to describe the implementation process of the present invention in detail. The data center establishes an energy consumption supervision platform, and can collect data of energy efficiency related parameters required by the case through the sensor. The present case selects historical energy efficiency related data for the data center during 2020.12-2022.07.
As shown in fig. 1, the embodiment provides a data center energy efficiency anomaly detection method based on an adaptive prediction interval, which includes the following steps:
s1, data acquisition: selecting specific energy efficiency indexes, collecting and storing outdoor meteorological parameters, time data and energy consumption data, and calculating energy efficiency data;
s11, the energy efficiency index comprises: electrical energy usage (PUE), refrigeration load factor (CLF), power Load Factor (PLF), renewable energy usage (RER). In this embodiment, the PUE is selected as the energy efficiency index of this embodiment;
s12, acquiring and storing historical energy efficiency related data through a sensor, wherein the historical energy efficiency related data comprises parameter data which is used for calculating energy efficiency data and is directly acquired through the sensor, and the historical energy efficiency related data specifically comprises the following steps: total power consumption, total power consumption of IT equipment, outdoor temperature, outdoor humidity, year, month, date, time, season type, workday type;
s13, calculating the time-by-time PUE of the energy efficiency index:
this results in the data of the previous time PUE, the previous 1 time PUE, and the previous 24 time PUE.
S2, data preprocessing: preprocessing historical data, including abnormal value detection, rejection and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data; the method specifically comprises the following steps:
s21, respectively detecting abnormal values of the PUE at the previous 1 moment and the PUE at the previous 24 moment, wherein the abnormal value judgment standard is 3 sigma, and calculating the mean value mu and the standard deviation sigma of each characteristic parameter in the original data:
L1＝μ-3σ
L2＝μ+3σ
wherein X is i For a certain column of i-th data, N is the total number of samples, mu is the average value of the column of data, sigma is the standard deviation of the column of data, L1 is the lower judgment limit of an abnormality threshold, L2 is the upper judgment limit of the abnormality threshold, and the abnormality value judgment standard is that the sample value is smaller than the lower judgment limit L1 or larger than the upper judgment limit L2;
if a certain characteristic is not in the [ mu-3 sigma, mu+3 sigma ] interval, judging that the characteristic is an abnormal value, and judging that the data point where the abnormal value is located is regarded as an abnormal point, and rejecting the data;
s22, as shown in FIG. 2, an energy efficiency data interpolation graph is shown, and the abnormal points removed in S21 need to be subjected to data interpolation, wherein the interpolation method is as follows:
X i ＝X i+24k
wherein X is i For a certain column of ith data, i.e. a certain time energy efficiency value, X i+24k For the energy efficiency value at the same time of the previous k days, when k=1, if X i+24 Taking X as normal value i+24 Assignment of values to X i If X i+24 For an outlier, when looking at k=2, X i+24*2 Whether or not to useIs a normal value, if so, assign it to X i If not, look at k=3 until X i+24k Until the value is normal, the normal value X i+24k Assignment to X i ；
Taking the PUE of the first 1 moment and the PUE of the first 24 moment after data cleaning and interpolation as alternative characteristic variables;
s23, detecting distortion values of outdoor temperature and outdoor humidity, wherein the detection method is a discrimination method:
wherein i is i For a certain column of ith data, the abnormal value judgment standard is that the difference value between the sample value and the data average value of the front moment and the back moment is more than 5;
if the difference value between a certain data of a certain characteristic and the data average value of the time before and after the certain characteristic is larger than 5, judging that the data is an abnormal value, and judging that the data point where the abnormal value is located is regarded as an abnormal point and rejecting the data;
s24, performing data interpolation on the outliers removed in the S23, wherein the interpolation method is an average interpolation method:
wherein X is i For the ith data of a certain column, taking the average value of the data at the front and rear moments and assigning the average value to X i ；
Taking the outdoor temperature and the outdoor humidity after data cleaning and interpolation as alternative characteristic variables;
s25, sorting all the characteristic variables and the PUE at the current moment into a table;
s26, detecting an abnormal value of the PUE at the current moment, wherein the abnormal value judgment standard is 3 sigma, and calculating the mean mu and standard deviation sigma in the original data;
if a certain data does not fall within the [ mu-3 sigma, mu+3 sigma ] interval, judging that the data is an abnormal value, and judging that the data point where the abnormal value is located is regarded as an abnormal point, removing the row of data, and finally obtaining 14513 valid data in total after preprocessing.
S3, selecting data characteristics: adopting a feature selection method to perform feature selection on the preprocessed historical data; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
the feature selection method comprises the following steps: pearson correlation coefficient method, embedding method, mutual information method and recursive feature elimination method;
the embodiment uses the LightGBM for feature selection, which belongs to an embedded method in feature selection. First, an algorithm is trained on the initial feature set, and the importance of each feature is obtained through the feature importance attribute of the LightGBM. Then, the features are pruned to obtain the final feature set as: time of day PUE 1, outdoor temperature, time of day PUE 24, time of day, outdoor humidity, month.
S4, model training: training a regression prediction model by using the data after the feature selection in the step S3, and determining an optimal super-parameter combination by using a model super-parameter optimization method so as to obtain a trained regression prediction model;
s41, randomly dividing the data sample into a training set and a test set according to a ratio of 4:1, thereby obtaining an 11610 row and 6 column input training set D train_x Output training set D of 11610 row 1 and column train_y Input test set D of 2903 rows and 6 columns test_x Output test set D of 2903 rows and 1 columns test_y Model training and verification testing in subsequent steps are facilitated;
s42, constructing an artificial intelligent algorithm regression model, wherein the model fitting adopts an artificial intelligent regression algorithm, and the method comprises the following steps: vector machine, decision tree, neural network, ensemble learning, deep learning, generating countermeasure network, in this embodiment, the LightGBM algorithm based on gradient lifting decision tree is adopted to construct regression prediction model, and default parameters are used for data set D train_x 、D train_y Training and using data set D test_x 、D test_y Performing verification test to obtain model evaluation parameters R 2 ＝0.92；
S43, optimizing the super parameters of the regression model by using a super parameter optimization method by taking the regression model evaluation index as an optimization target, wherein the verification method is k-fold cross verification;
selecting regression model evaluation index R 2 Is an optimization target;
the model super-parameter optimization method comprises the following steps: random search, grid search, bayesian optimization and evolution algorithm, the embodiment adopts the Bayesian optimization algorithm to optimize the super parameters of the LightGBM algorithm;
the verification method is k-fold cross-verification, where k=10.
Model fitting coefficient R obtained after Bayesian optimization in the embodiment 2 =0.93, the optimized LightGBM parameters are learning_rate=0.08, max_depth=12, n_evators=111, num_leave=41;
and storing the model optimized by the steps.
S5, generating a prediction error set: calculating a predicted value of the time-by-time energy efficiency by using the regression prediction model trained in the step S4, and subtracting an actual value from the predicted value to obtain a prediction error set errs;
input set D to 14513 row 6 column x Leading in a trained Bayes-LightGBM model to obtain a predicted value D y_pred Calculating a prediction error value e:
e＝D y -D y_pred
wherein D is y Is an actual value. The resulting prediction error set errs is the 14513 row 1 column dataset.
S6, constructing an adaptive prediction error interval: calculating probability density of the prediction error set in the step S5 by adopting a kernel density estimation algorithm, and overlapping confidence intervals to obtain a self-adaptive prediction error interval;
s61, importing a data sample prediction error set errs, performing bandwidth optimization on the kernel density estimation model by adopting a grid search method of a super-parameter optimization method, wherein the verification method is k-fold cross verification, k=5 is taken, and after optimization, the optimal bandwidth h=0.1;
s62, establishing a kernel density estimation model by adopting an optimal bandwidth h, and calculating the probability density of each prediction error value e, wherein the probability density function is as follows:
wherein, K (e) is a kernel function, and the optional kernel functions include a Gaussian kernel function, a gamma kernel function, a uniform kernel function and a triangular kernel function, and the Gaussian kernel function is selected in the embodiment;
s63, constructing an adaptive prediction error interval according to the confidence interval with the confidence coefficient of the probability density being 1-alpha, taking alpha=0.1, and finally constructing the adaptive prediction error interval as (-0.0221,0.0221).
S7, abnormality detection: processing the data acquired in real time by using the same data preprocessing method as the step S2, constructing an input set by adopting the characteristics selected in the step S3, inputting a trained regression prediction model to obtain an energy efficiency predicted value at the current moment, superposing an adaptive prediction error interval to obtain an energy efficiency adaptive prediction interval at the current moment, judging whether the actual value of the energy efficiency at the current moment is in the energy efficiency adaptive prediction interval at the current moment, if not, judging that the energy efficiency is abnormal, and outputting a detection result.
And S8, updating the model, and after the updating time period T=60×24=1440 h, using the latest accumulated data set to update the regression prediction model and the adaptive prediction error interval.
Through the technical scheme of the embodiment, the built data center energy efficiency abnormality detection result is high in reliability, the online real-time abnormality detection of the data center energy efficiency is facilitated, the abnormality is found in time and an alarm is given, management staff is reminded to take corresponding countermeasure measures in time, the energy efficiency management of the data center is facilitated, and a high-efficiency machine room is built.
It should be noted that, for the sake of simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the present invention is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present invention.
Example 2
Based on the same ideas the data center energy efficiency abnormality detection method based on the adaptive prediction interval in the above embodiment, the invention also provides a data center energy efficiency abnormality detection system based on the adaptive prediction interval, which can be used for executing the data center energy efficiency abnormality detection method based on the regression prediction model and the adaptive prediction interval. For ease of illustration, only those portions relevant to embodiments of the present invention are shown in a schematic structural diagram of an embodiment of a data center energy efficiency anomaly detection system based on a regression prediction model and an adaptive prediction interval, and it will be understood by those skilled in the art that the illustrated structure does not constitute a limitation of the apparatus, and may include more or less components than those illustrated, or may combine some components, or may have different arrangements of components.
As shown in fig. 3, in another embodiment of the present application, there is provided a data center energy efficiency anomaly detection system 100 based on an adaptive prediction interval, the system including a data acquisition module 101, a data preprocessing module 102, a feature selection module 103, a regression prediction model training module 104, an adaptive prediction error interval construction module 105, and an anomaly detection module 106;
the data acquisition module 101 is used for selecting specific energy efficiency indexes, acquiring and storing outdoor weather parameters, time data and energy consumption data through sensors, and calculating energy efficiency data according to the energy consumption data;
the data preprocessing module 102 is used for preprocessing historical data, including abnormal value detection, elimination and missing value interpolation of the multi-energy efficiency data; removing and interpolating distortion values of the obtained outdoor meteorological data;
a feature selection module 103, configured to perform feature selection on the preprocessed historical data by using a feature selection method; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
the model training module 104 is configured to train a regression prediction model using the data after feature selection, determine an optimal super-parameter combination thereof using a model super-parameter optimization method, and further obtain a trained regression prediction model;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the regression prediction model is updated;
the adaptive prediction error interval construction module 105 is configured to generate a prediction error set, calculate probability density of the prediction error set by using a kernel density estimation algorithm, and superimpose a confidence interval to obtain an adaptive prediction error interval;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the self-adaptive prediction error interval is updated;
the anomaly detection module 106 is configured to input data to be detected into the trained regression prediction model to obtain an energy efficiency prediction value at the current time, superimpose the adaptive prediction error interval to obtain an energy efficiency adaptive prediction interval at the current time, and compare whether the actual energy efficiency value is within the energy efficiency adaptive prediction interval at the current time to determine whether the energy efficiency is anomalous, so as to obtain an anomaly detection result.
It should be noted that, the data center energy efficiency abnormality detection system based on the adaptive prediction interval and the data center energy efficiency abnormality detection method based on the adaptive prediction interval according to the present invention are in one-to-one correspondence, and the technical features and the beneficial effects described in the embodiments of the data center energy efficiency abnormality detection method based on the adaptive prediction interval are applicable to the embodiments of the data center energy efficiency abnormality detection system based on the adaptive prediction interval, and specific content may be referred to the description in the embodiments of the method according to the present invention, which is not repeated herein.
It should be noted that, in the system provided in the foregoing embodiment, only the division of the foregoing functional modules is illustrated, in practical application, the foregoing functional allocation may be performed by different functional modules according to needs, that is, the internal structure is divided into different functional modules to perform all or part of the functions described above, and the system is a data center energy efficiency anomaly detection method based on an adaptive prediction interval applied to the foregoing embodiment.
Example 3
As shown in fig. 4, in another embodiment of the present application, there is further provided a storage medium storing a program, where the program, when executed by a processor, implements a data center energy efficiency anomaly detection method based on an adaptive prediction interval, specifically:
selecting an energy efficiency index, collecting outdoor meteorological parameters, time data and energy consumption data, and calculating energy efficiency data according to the energy consumption data;
preprocessing historical data, including abnormal value detection, rejection and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data;
adopting a feature selection method to perform feature selection on the preprocessed historical data; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
training a regression prediction model by using the data with the characteristics selected, and determining an optimal super-parameter combination by using a model super-parameter optimization method so as to obtain a trained regression prediction model;
calculating a predicted value of the time-by-time energy efficiency by using the trained regression prediction model, and subtracting an actual value from the predicted value to obtain a predicted error set;
for the prediction error set, calculating probability density of the prediction error set by adopting a kernel density estimation algorithm, and superposing a confidence interval to obtain a self-adaptive prediction error interval;
processing the data acquired in real time by using the data preprocessing method which is the same as that of the historical data, constructing an input set by adopting the characteristics selected in the characteristic selection method, inputting a trained regression prediction model to obtain the energy efficiency predicted value at the current moment, superposing the self-adaptive prediction error interval to obtain the energy efficiency self-adaptive prediction interval at the current moment, and comparing whether the actual energy efficiency value is in the energy efficiency self-adaptive prediction interval at the current moment to judge whether the energy efficiency is abnormal or not.
It is to be understood that portions of the present application 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.
The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.
Claims (10)
1. The data center energy efficiency anomaly detection method based on the self-adaptive prediction interval is characterized by comprising the following steps of:
selecting an energy efficiency index, collecting outdoor meteorological parameters, time data and energy consumption data, and calculating energy efficiency data according to the energy consumption data;
preprocessing historical data, including abnormal value detection, rejection and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data;
adopting a feature selection method to perform feature selection on the preprocessed historical data; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
training a regression prediction model by using the data with the characteristics selected, and determining an optimal super-parameter combination by using a model super-parameter optimization method so as to obtain a trained regression prediction model;
calculating a predicted value of the time-by-time energy efficiency by using the trained regression prediction model, and subtracting an actual value from the predicted value to obtain a predicted error set;
for the prediction error set, calculating probability density of the prediction error set by adopting a kernel density estimation algorithm, and superposing a confidence interval to obtain a self-adaptive prediction error interval;
processing the data acquired in real time by using the data preprocessing method which is the same as that of the historical data, constructing an input set by adopting the characteristics selected in the characteristic selection method, inputting a trained regression prediction model to obtain the energy efficiency predicted value at the current moment, superposing the self-adaptive prediction error interval to obtain the energy efficiency self-adaptive prediction interval at the current moment, and comparing whether the actual energy efficiency value is in the energy efficiency self-adaptive prediction interval at the current moment to judge whether the energy efficiency is abnormal or not.
2. The data center energy efficiency anomaly detection method based on the adaptive prediction interval according to claim 1, wherein the energy efficiency index comprises: electric energy utilization rate, refrigeration load coefficient, power supply load coefficient and renewable energy utilization rate;
the energy consumption data comprise parameter data which are used for calculating energy efficiency data and are directly acquired through a sensor;
the energy efficiency data includes: the current time energy efficiency value, the previous 1 time energy efficiency value and the previous 24 time energy efficiency value;
the outdoor weather parameters include: outdoor temperature, outdoor humidity;
the time data includes: year, month, date, time of day, season type, and workday type; among them, season types include: a cooling season, a transition season, and a non-cooling season; workday types include: workday, non-workday.
3. The method for detecting abnormal data center energy efficiency based on the adaptive prediction interval according to claim 1, wherein the method for detecting and rejecting abnormal values of energy efficiency data is a 3σ criterion:
L1＝μ-3σ
L2＝μ+3σ
wherein X is i For a column of ith data, N is the total number of samples, μ isThe average value sigma of the column data is the standard deviation of the column data, L1 is the lower judgment limit of an abnormal threshold value, L2 is the upper judgment limit of the abnormal threshold value, and the abnormal value judgment standard is that the sample value is smaller than the lower judgment limit L1 or larger than the upper judgment limit L2;
respectively eliminating abnormal values of the current time energy efficiency value, the previous 1 time energy efficiency value and the previous 24 time energy efficiency value;
the method for carrying out missing value interpolation on the energy efficiency data is an interpolation method:
X i ＝X i+24k
wherein X is i For a certain column of ith data, i.e. a certain time energy efficiency value, X i+24k For the energy efficiency value at the same time of the previous k days, when k=1, if X i+24 Taking X as normal value i+24 Assignment of values to X i If X i+24 For an outlier, when looking at k=2, X i+24*2 Whether the value is normal or not, if so, assigning the value to X i If not, look at k=3 until X i+24k Until the value is normal, the normal value X i+24k Assignment to X i ；
Performing missing value interpolation on the energy efficiency value at the previous 1 moment and the energy efficiency value at the previous 24 moment;
the method for eliminating the distortion value of the outdoor meteorological data comprises the following steps of:
wherein X is i For a certain column of ith data, the abnormal value judgment standard is that the difference value between the sample value and the data average value of the front moment and the back moment is more than 5;
the method for carrying out missing value interpolation on the outdoor meteorological data is a mean value interpolation method:
wherein X is i For the ith data of a certain column, taking the average value of the data at the front and rear moments and assigning the average value to X i 。
4. The method for detecting abnormal data center energy efficiency based on adaptive prediction intervals according to claim 1, wherein the feature selection method adopts one of a Pearson correlation coefficient method, an embedding method, a mutual information method and a recursive feature elimination method.
5. The data center energy efficiency anomaly detection method based on the adaptive prediction interval according to claim 1, wherein the regression prediction model is specifically:
the regression prediction model fitting adopts an artificial intelligent regression algorithm, and adopts one of a vector machine, a decision tree, a neural network, ensemble learning, deep learning and an antagonism network;
the regression prediction model super-parameter optimization method adopts one of random search, grid search, bayesian optimization and evolutionary algorithm.
6. The data center energy efficiency anomaly detection method based on the adaptive prediction interval according to claim 5, wherein the specific process of regression prediction model fitting is as follows:
randomly dividing a data sample into a training set and a testing set according to a certain proportion, and using the training set and the testing set for model training and verification testing;
constructing an artificial intelligent algorithm regression prediction model, and training and verifying the data set by using default parameters;
optimizing the super parameters of the regression prediction model by using a super parameter optimization method by taking the evaluation index of the regression prediction model as an optimization target, wherein the verification method is k-fold cross verification;
storing the regression prediction model optimized by the steps;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the regression prediction model is updated.
7. The method for detecting abnormal energy efficiency of a data center based on an adaptive prediction interval according to claim 1, wherein the calculating the probability density of the prediction error set by using a kernel density estimation algorithm and superposing the confidence interval to obtain the adaptive prediction error interval comprises the following specific procedures:
leading in a data sample prediction error set, carrying out bandwidth optimization on a nuclear density estimation model by adopting a super-parameter optimization method, and performing k-fold cross verification by adopting a verification method;
establishing a kernel density estimation model by adopting an optimal bandwidth h, and calculating the probability density of each prediction error value e, wherein the probability density function is as follows:
wherein K (e) is a kernel function, and one of a Gaussian kernel function, a gamma kernel function, a uniform kernel function and a triangular kernel function is adopted;
confidence interval [ f ] with confidence of 1-alpha in calculating probability density α/2 ，f 1-α/2 ]Constructing an adaptive prediction error interval [ e ] according to the functional relation between the probability density and the prediction error value 1 ，e 2 ]；
After the update time period T, the steps are carried out again by using the latest accumulated data set, and the adaptive prediction error interval is updated.
8. The method for detecting abnormal energy efficiency of a data center based on an adaptive prediction interval according to claim 1, wherein the obtaining the predicted value of the energy efficiency at the current time, and superimposing the adaptive prediction error interval, obtains the adaptive prediction interval of the energy efficiency at the current time, specifically:
the energy efficiency predicted value b at the current moment and the self-adaptive prediction error interval [ e ] 1 ，e 2 ]Superposition to obtain the energy efficiency self-adaptive prediction interval [ b+e ] at the current moment 1 ，b+e 2 ]；
Whether the energy efficiency is abnormal is judged by comparing whether the actual value of the energy efficiency is in the energy efficiency self-adaptive prediction interval at the current moment, specifically:
judging whether the actual value c of the energy efficiency at the current moment is in the energy efficiency self-adaptive prediction interval [ b+e ] at the current moment 1 ，b+e 2 ]If not, the energy efficiency is abnormal.
9. The data center energy efficiency anomaly detection system based on the self-adaptive prediction interval is characterized by being applied to the data center energy efficiency anomaly detection method based on the self-adaptive prediction interval, which is disclosed in any one of claims 1-8, and comprises a data acquisition module, a data preprocessing module, a feature selection module, a regression prediction model training module, a self-adaptive prediction error interval construction module and an anomaly detection module;
the data acquisition module is used for selecting energy efficiency indexes, acquiring outdoor meteorological parameters, time data and energy consumption data through the sensor, and calculating energy efficiency data according to the energy consumption data;
the data preprocessing module is used for preprocessing historical data, including abnormal value detection, elimination and missing value interpolation of energy efficiency data; removing and interpolating distortion values of outdoor meteorological data;
the feature selection module is used for performing feature selection on the preprocessed historical data by adopting a feature selection method; the historical data comprises outdoor weather parameters, time data and energy efficiency data;
the model training module is used for training a regression prediction model by using the data after feature selection, determining the optimal super-parameter combination by using a model super-parameter optimization method, and further obtaining a trained regression prediction model;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the regression prediction model is updated;
the self-adaptive prediction error interval construction module is used for generating a prediction error set, calculating probability density of the self-adaptive prediction error interval by adopting a kernel density estimation algorithm, and overlapping confidence intervals to obtain a self-adaptive prediction error interval;
after the updating time period T, the steps are carried out again by using the latest accumulated data set, and the self-adaptive prediction error interval is updated;
the anomaly detection module is used for inputting the data to be detected into the trained regression prediction model to obtain the energy efficiency prediction value at the current moment, superposing the self-adaptive prediction error interval to obtain the energy efficiency self-adaptive prediction interval at the current moment, and comparing whether the actual energy efficiency value is in the energy efficiency self-adaptive prediction interval at the current moment to judge whether the energy efficiency is abnormal or not so as to obtain an anomaly detection result.
10. A storage medium storing a program, characterized in that: the program, when executed by a processor, implements the data center energy efficiency anomaly detection method based on the adaptive prediction interval of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310047835.1A CN116383773A (en) | 2023-01-31 | 2023-01-31 | Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310047835.1A CN116383773A (en) | 2023-01-31 | 2023-01-31 | Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116383773A true CN116383773A (en) | 2023-07-04 |
Family
ID=86971927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310047835.1A Pending CN116383773A (en) | 2023-01-31 | 2023-01-31 | Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116383773A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116663747A (en) * | 2023-07-19 | 2023-08-29 | 广东云下汇金科技有限公司 | Intelligent early warning method and system based on data center infrastructure |
| CN116992221A (en) * | 2023-07-31 | 2023-11-03 | 武汉天翌数据科技发展有限公司 | Fault detection method, device and equipment of operation and maintenance platform and storage medium |
| CN117269751A (en) * | 2023-11-22 | 2023-12-22 | 国网江西省电力有限公司电力科学研究院 | GIS isolating switch switching position confirmation method |
| CN117559633A (en) * | 2023-11-03 | 2024-02-13 | 华能国际电力股份有限公司河北清洁能源分公司 | Power operation and maintenance regulation and control system based on energy efficiency abnormality |
-
2023
- 2023-01-31 CN CN202310047835.1A patent/CN116383773A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116663747A (en) * | 2023-07-19 | 2023-08-29 | 广东云下汇金科技有限公司 | Intelligent early warning method and system based on data center infrastructure |
| CN116663747B (en) * | 2023-07-19 | 2024-04-12 | 广东云下汇金科技有限公司 | Intelligent early warning method and system based on data center infrastructure |
| CN116992221A (en) * | 2023-07-31 | 2023-11-03 | 武汉天翌数据科技发展有限公司 | Fault detection method, device and equipment of operation and maintenance platform and storage medium |
| CN116992221B (en) * | 2023-07-31 | 2024-03-26 | 武汉天翌数据科技发展有限公司 | Fault detection method, device and equipment of operation and maintenance platform and storage medium |
| CN117559633A (en) * | 2023-11-03 | 2024-02-13 | 华能国际电力股份有限公司河北清洁能源分公司 | Power operation and maintenance regulation and control system based on energy efficiency abnormality |
| CN117269751A (en) * | 2023-11-22 | 2023-12-22 | 国网江西省电力有限公司电力科学研究院 | GIS isolating switch switching position confirmation method |
| CN117269751B (en) * | 2023-11-22 | 2024-04-02 | 国网江西省电力有限公司电力科学研究院 | GIS isolating switch switching position confirmation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116383773A (en) | Data center energy efficiency anomaly detection method, system and medium based on self-adaptive prediction interval | |
| CN109359698B (en) | Leakage identification method based on long-time memory neural network model | |
| CN113298297B (en) | Wind power output power prediction method based on isolated forest and WGAN network | |
| CN112801388B (en) | Power load prediction method and system based on nonlinear time series algorithm | |
| CN111582542B (en) | Power load prediction method and system based on anomaly repair | |
| CN113408659A (en) | Building energy consumption integrated analysis method based on data mining | |
| CN114023399A (en) | Air particulate matter analysis early warning method and device based on artificial intelligence | |
| CN111814964A (en) | Air pollution treatment method based on air quality condition prediction and storage medium | |
| CN113449919B (en) | Power consumption prediction method and system based on feature and trend perception | |
| CN115034485A (en) | Wind power interval prediction method and device based on data space | |
| CN118051827A (en) | Power grid fault prediction method based on deep learning | |
| CN113988398A (en) | Wind turbine generator power prediction method and device, electronic equipment and storage medium | |
| CN116706992A (en) | Self-adaptive power prediction method, device and equipment for distributed photovoltaic cluster | |
| CN107706938B (en) | A kind of wind power waving interval analysis method returned based on quantile | |
| US20220243347A1 (en) | Determination method and determination apparatus for conversion efficiency of hydrogen production by wind-solar hybrid electrolysis of water | |
| CN115829145A (en) | Photovoltaic power generation capacity prediction system and method | |
| CN117408394B (en) | Carbon emission factor prediction method and device for electric power system and electronic equipment | |
| CN117077870B (en) | Water resource digital management method based on artificial intelligence | |
| CN113869590A (en) | EEMD-LSTM-based regional energy internet load prediction method and system | |
| CN116591911A (en) | Intelligent detection operation and maintenance system and method facing to offshore wind turbine generator set | |
| CN117272184A (en) | ISSA-RF algorithm-based photovoltaic power generation system fault classification method | |
| CN116720983A (en) | Power supply equipment abnormality detection method and system based on big data analysis | |
| CN113723670B (en) | Photovoltaic power generation power short-term prediction method with variable time window | |
| Hu et al. | The early warning model of dust concentration in smart construction sites based on long short term memory network | |
| CN117113267B (en) | Prediction model training method based on big data and photovoltaic power generation performance detection method |
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 |