CN116760033A - Real-time power demand prediction system based on artificial intelligence - Google Patents
Real-time power demand prediction system based on artificial intelligence Download PDFInfo
- Publication number
- CN116760033A CN116760033A CN202311047162.6A CN202311047162A CN116760033A CN 116760033 A CN116760033 A CN 116760033A CN 202311047162 A CN202311047162 A CN 202311047162A CN 116760033 A CN116760033 A CN 116760033A
- Authority
- CN
- China
- Prior art keywords
- data
- prediction
- model
- module
- representing
- 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.)
- Granted
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000007781 pre-processing Methods 0.000 claims abstract description 22
- 238000005457 optimization Methods 0.000 claims abstract description 21
- 238000004364 calculation method Methods 0.000 claims abstract description 20
- 230000002787 reinforcement Effects 0.000 claims abstract description 18
- 230000006870 function Effects 0.000 claims description 37
- 238000004422 calculation algorithm Methods 0.000 claims description 30
- 238000012545 processing Methods 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 12
- 125000003275 alpha amino acid group Chemical group 0.000 claims description 11
- 150000001875 compounds Chemical class 0.000 claims description 11
- 238000007405 data analysis Methods 0.000 claims description 11
- 230000014509 gene expression Effects 0.000 claims description 11
- 230000001965 increasing effect Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 238000004891 communication Methods 0.000 claims description 10
- 238000003066 decision tree Methods 0.000 claims description 10
- 230000000007 visual effect Effects 0.000 claims description 10
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 238000007726 management method Methods 0.000 claims description 8
- 230000003014 reinforcing effect Effects 0.000 claims description 6
- 208000025174 PANDAS Diseases 0.000 claims description 5
- 208000021155 Paediatric autoimmune neuropsychiatric disorders associated with streptococcal infection Diseases 0.000 claims description 5
- 240000004718 Panda Species 0.000 claims description 5
- 235000016496 Panda oleosa Nutrition 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 5
- 230000006872 improvement Effects 0.000 claims description 5
- 238000013450 outlier detection Methods 0.000 claims description 5
- 238000011217 control strategy Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 230000000694 effects Effects 0.000 claims description 3
- 230000005684 electric field Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 3
- 230000009545 invasion Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 2
- 238000003672 processing method Methods 0.000 claims description 2
- 238000010248 power generation Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000005611 electricity Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013079 data visualisation Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000012887 quadratic function Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 210000000352 storage cell Anatomy 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06315—Needs-based resource requirements planning or 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00001—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by the display of information or by user interaction, e.g. supervisory control and data acquisition systems [SCADA] or graphical user interfaces [GUI]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- Power Engineering (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Human Computer Interaction (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a real-time power demand prediction system based on artificial intelligence, which comprises a data acquisition module, a data preprocessing module, a data prediction cloud platform module, an optimization adjustment module of a model and a safety reinforcement module, wherein the output end of the data acquisition module is connected with the input end of the data preprocessing module, the output end of the data preprocessing module is connected with the input end of the data prediction cloud platform module, the output end of the data prediction cloud platform module is connected with the input end of the optimization adjustment module of the model, and the output end of the optimization adjustment module of the model is connected with the input end of the safety reinforcement module. The method solves the problems that the existing power load prediction system has low prediction accuracy, poor intelligent degree and low calculation efficiency, cannot meet the requirement of the prediction accuracy, causes low utilization rate of power generation equipment, causes a large amount of operation cost and profit loss, and even affects the economic operation of the power system.
Description
Technical Field
The invention belongs to the field of power demand prediction, and particularly relates to a real-time power demand prediction system based on artificial intelligence.
Background
The power demand prediction is an important component of power system planning and is also the basis of economic operation of a power system, is extremely important for power system planning and operation, is a series of prediction work which is carried out by taking a power load as a target, and comprises prediction of future power demand quantity, prediction of future power consumption quantity and prediction of a load curve from the viewpoint of the prediction target, and mainly aims to predict time distribution and space distribution of the future power load, so that reliable decision basis is provided for power system planning and operation, the higher the precision of the load demand is, the more beneficial to improving the utilization rate of power generation equipment and the effectiveness of economic scheduling, and conversely, when the load prediction error is larger, not only can cause a great amount of operation cost and profit loss, but also can influence the reliability of power system operation and the supply and demand balance of a power market, so that the accurate prediction of the power demand is very important.
The existing prediction method of the power load prediction system is relatively backward, the prediction precision is low, the intelligent degree is poor, the calculation efficiency is low, the requirement of the prediction precision cannot be met, the utilization rate of power generation equipment is low, a large amount of operation cost and profit loss are caused, and even the economic operation of a power system is influenced.
To this end, we propose a real-time power demand prediction system based on artificial intelligence to solve the above-mentioned problems.
Disclosure of Invention
Aiming at the defects of the technology, the invention discloses a real-time power demand prediction system based on artificial intelligence, which can establish a power demand prediction model through an XGBoost algorithm, compare and analyze a prediction result and an actual result, calculate the difference between the prediction result and the actual result, evaluate the precision of the power demand prediction model through calculation of a relative error, a root mean square error, an average absolute percentage error and an average absolute error, calculate the weight of the prediction model through a linear combination method, and dynamically adjust the weight proportion to optimize the power demand prediction system.
The invention adopts the following technical scheme:
an artificial intelligence based real-time power demand prediction system comprising:
the data acquisition module is used for acquiring power supply and demand data in real time, wherein the power supply and demand data comprise power load data, power consumption, power supply quantity, voltage, current, weather data and industry data, and the data acquisition module acquires historical power demand data, weather data and industry data through a power supplier, an energy management system and an equipment monitoring database data sharing platform and acquires real-time power equipment data through a wireless communication mode;
the data preprocessing module is used for cleaning the acquired power supply and demand data, and the data preprocessing module adopts a data processing tool pandas to process missing values, abnormal values and repeated values in the acquired data;
the data prediction cloud platform module is used for constructing a data prediction model according to historical power demand data, weather data and industry data, predicting power demand in real time, accessing, uploading and managing the data cloud of the data preprocessing module through the Internet, building the real-time power demand prediction model through an XGBoost algorithm, training the prediction model through learning sample data, comprehensively analyzing a prediction result, comparing the prediction data with an actual result, calculating the difference between the prediction result and the actual result, evaluating the precision of the prediction model, and converting data into a shape or an image by utilizing a computer graphics and image processing method to achieve data visual display;
the model optimization and adjustment module is used for carrying out improvement and optimization on the data prediction model, calculating the weight of the prediction model by a linear combination method and dynamically adjusting the weight proportion;
the security reinforcement module is used for improving the robustness and security of the system, increasing the difficulty of invasion of an attacker and improving the security level of the system, wherein the security reinforcement is a process of configuring a software system, and the security level of the system is improved by patching, reinforcing account security, reinforcing service, modifying security configuration, optimizing access control strategies and increasing security mechanisms aiming at a server operating system, a database and an application middleware software system;
the output end of the data acquisition module is connected with the input end of the data preprocessing module, the output end of the data preprocessing module is connected with the input end of the data prediction cloud platform module, the output end of the data prediction cloud platform module is connected with the input end of the optimization adjustment module of the model, and the output end of the optimization adjustment module of the model is connected with the input end of the safety reinforcement module.
As a further technical scheme of the invention, the data preprocessing module adopts a data processing tool pandas to process missing values, abnormal values and repeated values in the acquired data, and the data preprocessing module achieves format standardization, abnormal data clearing, error data correction and repeated data clearing through a smooth noise data and outlier detection method, and adopts a mahalanobis distance detection multi-element outlier.
As a further technical scheme of the invention, the data prediction cloud platform module comprises a cloud storage unit, a prediction model building unit, a data analysis processing unit and a visual display unit.
As a further technical scheme of the invention, the prediction model building unit builds a real-time power demand prediction model through an XGBoost algorithm and trains and learns the power demand model through sample data, and the method comprises the following steps:
step one: the XGBoost algorithm is a supervised integrated learning algorithm of a summation model of a plurality of decision trees, and the calculation formula is as follows:
（1）
in the formula (1), the components are as follows,representing model predictive value, +.>Indicate->Sample number->Representation about sample->K represents the number of decision trees,/-for the prediction function of (a)>An independent function representing a function space, F representing a function space, consisting of a decision tree, a given sample set having n samples, m features,/a>The independent function representing the function space has +.>Data attribute feature point, independent function with data set s, +.>Is a model complexity coefficient;
the XGBoost algorithm training data set is calculated as follows:
（2）
in the formula (2), the amino acid sequence of the compound,representation about->、/>Data set of->Indicate->Personal category labels, i.e.)>Actual value of the individual category, at the same time +.>，/>Representing a real set;
step two: the XGBoost algorithm objective function comprises a loss error function and a regularization term, wherein the loss error functionThe difference between the target predicted value and the target actual value is described; regular item->Adjusting the complexity of the control tree to avoid excessive repetition of the control tree;
step three: after the XGBoost algorithm creates a tree, feature extraction is directly completed by utilizing a feature load, two values of weight and gain are called, feature selection is completed, and a calculation expression of the gain is as follows:
（3）
in the formula (3), the amino acid sequence of the compound,for the gain expression, subscript L, R indicates the information tag identity of the left and right subtrees, respectively, after splitting, ++>Gain representing the ith data, +.>Representing left subtree gain, ">Representing right subtree gain, ">Weight representing the ith data, +.>Weight representing left subtree,/->Weight representing right subtree->Representing regularization parameters;
taking the evaluation index of the prediction model as the power prediction accuracy, the power prediction accuracy can be expressed as:
（4）
in the formula (4), the amino acid sequence of the compound,for power prediction accuracy, +.>Representing the actual output power at instant i, +.>Predicted output power representing model output, +.>The starting capacity of the integrated electric field is represented, and the output of the input formula is compared with a set threshold value.
As a further technical scheme of the invention, the data analysis processing unit is used for comparing and analyzing the predicted result and the actual result, calculating the difference between the predicted result and the actual result, inputting data into the model after constructing an algorithm model, calculating and processing data information in the model, and comparing the predicted result and the actual result to evaluate the precision of the predicted model, and the method comprises the following steps:
（5）
in the formula (5), the amino acid sequence of the compound,refers to the actual value of the electrical load, +.>And->Respectively, refers to the minimum and maximum values of the electrical load, U is the amount of index normalized data, and the accuracy of the predictive model is evaluated by using the following indices:
i. relative error:
（6）
ii. Root mean square error:
（7）
iii, average absolute percentage error:
（8）
iv, average absolute error:
（9）
in the above formulas (6) to (9),refers to the corresponding predicted load.
As a further technical scheme of the invention, the optimization adjustment module of the model is used for debugging the weight ratio calculated by the predicted result through the comparison analysis of the predicted data and the actual result, and the method comprises the following steps:
step one: determining the initial stageStarting weight: setting the value of the first actual result asEach prediction model->Is the predicted value of (2)Wherein i->Then each prediction model +.>Is assigned according to the first prediction effect, a well performing model will be given a high initial weight;
step two: dynamically adjusting weights
A. Let current total m predicted results and actual results, including the predicted results of the previous k-1 times except the mth time、/>、…、/>And the experimental results of the first k times>、/>、…、/>Wherein k is a time window, and default is taken to be 5; when m is<Taking k=m at k; corresponding->、/>、…、/>Weights are +.>、、…、/>；
B. Predictive modelWherein i->Correlation coefficients of the calculation results of the previous k-1 times and the actual results of the previous k times except the mth time;
C. computing a predictive modelDeviation of the mth calculation result and the actual result;
D. according to A, B result, each prediction model is adaptively solvedWeight of mth adjustment->Wherein the value of the weight is increased for a prediction model having high correlation with the actual result and small deviation from the actual result; the value of the weight is reduced for a prediction model having low correlation with the actual result and a prediction model having large deviation from the actual result.
According to the further technical scheme, the security reinforcement module realizes security reinforcement of the cloud platform through the double-layer detection firewall, the double-layer detection firewall detects abnormity of the communication request and the communication content through the inspection engine, and potential safety hazard detection and improvement risk point reinspection of the cloud platform are carried out regularly.
The invention has the positive beneficial effects of distinguishing the prior art:
according to the invention, the outlier and the missing value are detected by adopting the outlier detection method, and the outlier and the missing value of the data acquired by the data acquisition module are corrected, so that more accurate data is obtained, and the prediction result of the established model is more accurate.
According to the method, the power demand prediction model is built through the XGBoost algorithm, the prediction result and the actual result are compared and analyzed, the difference between the prediction result and the actual result is calculated, and the accuracy of the power demand prediction model is estimated through calculation of the relative error, the root mean square error, the average absolute percentage error and the average absolute error, so that the prediction result of the power demand prediction model is closer to the actual result, and the accuracy of the power load prediction model is further improved.
According to the method, the weight of the prediction model is calculated by a linear combination method, and the weight proportion is dynamically adjusted, so that the comprehensive data of the power prediction model is closer to the real result, the accuracy of power demand prediction is further improved, and the prediction accuracy is improved to the greatest extent.
Drawings
For a clearer description of an embodiment of the invention or of a technical solution in the prior art, the drawings that are necessary for the description of the embodiment or of the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, from which, without inventive faculty, other drawings are obtained for a person skilled in the art, in which:
FIG. 1 is a schematic diagram of the overall architecture of an artificial intelligence based real-time power demand prediction system of the present invention;
FIG. 2 is a schematic block diagram of the XGBoost algorithm in the real-time power demand prediction system based on artificial intelligence;
FIG. 3 is a visual display basic structure diagram of an artificial intelligence based real-time power demand prediction system of the present invention;
FIG. 4 is a block diagram of a security reinforcement module of an artificial intelligence based real-time power demand prediction system according to the present invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
As shown in fig. 1-4, an artificial intelligence based real-time power demand prediction system, comprising:
the data acquisition module is used for acquiring power supply and demand data in real time, wherein the power supply and demand data comprise power load data, power consumption, power supply quantity, voltage, current, weather data and industry data, and the data acquisition module acquires historical power demand data, weather data and industry data through a power supplier, an energy management system and an equipment monitoring database data sharing platform and acquires real-time power equipment data through a wireless communication mode;
in a specific embodiment, parameters such as temperature, humidity, wind speed, wind direction and air pressure are observed through a temperature sensor, a humidity sensor, a wind speed sensor, an air pressure sensor and the like, data arrangement is carried out on the parameters such as temperature, humidity, wind speed, wind direction and air pressure, and power equipment data are extracted through a mode of combining wireless and wired communication.
The data preprocessing module is used for cleaning the acquired power supply and demand data, and the data preprocessing module adopts a data processing tool pandas to process missing values, abnormal values and repeated values in the acquired data;
in a specific embodiment, the data relating to two or more attributes or variables are called multivariate data by means of mahalanobis distance detection of the multivariate outliers, the core idea being to convert the multivariate outlier detection task into a univariate outlier detection problem, for a set of multivariate data, the set of multivariate data is providedFor the object O in the dataset, from O to +.>The mahalanobis distance is +.>S is the covariance matrix,>is a univariate variable, which can be tested for Grubbs, if +.>If an outlier is determined, then O is also considered an outlier.
The data prediction cloud platform module is used for constructing a data prediction model according to historical power demand data, weather data and industry data, predicting power demand in real time, accessing, uploading and managing the data cloud of the data preprocessing module through the Internet, building the real-time power demand prediction model through an XGBoost algorithm, training the prediction model through learning sample data, comprehensively analyzing a prediction result, comparing the prediction data with an actual result, calculating the difference between the prediction result and the actual result, evaluating the precision of the prediction model, and converting the data into a graph or an image by utilizing computer graphics and image processing technology so as to achieve data visual display;
the model optimization and adjustment module is used for carrying out improvement and optimization on the data prediction model, calculating the weight of the prediction model by a linear combination method and dynamically adjusting the weight proportion;
the security reinforcement module is used for improving the robustness and security of the system, increasing the difficulty of invasion of an attacker and improving the security level of the system, wherein the security reinforcement is a process of configuring a software system, and the security level of the system is improved by patching, strengthening account security, reinforcing service, modifying security configuration, optimizing access control strategies and increasing security mechanisms aiming at the software systems such as a server operating system, a database and application middleware;
the output end of the data acquisition module is connected with the input end of the data preprocessing module, the output end of the data preprocessing module is connected with the input end of the data prediction cloud platform module, the output end of the data prediction cloud platform module is connected with the input end of the optimization adjustment module of the model, and the output end of the optimization adjustment module of the model is connected with the input end of the safety reinforcement module.
The data prediction cloud platform module comprises a cloud storage unit, a prediction model building unit, a data analysis processing unit and a visual display unit.
In particular embodiments, the cloud storage unit (Cloud Storage Cell) is a basic unit in a cloud storage system, typically a physical disk or a virtual disk on a file server, for storing files or data. The size and availability of the cloud storage units depends on the usage specifications and configuration of the cloud service provider. The cloud storage unit may contain multiple files or data, each of which may contain multiple copies, for synchronization and sharing among multiple devices. The cloud storage unit may also be partitioned into smaller blocks to more efficiently store and access data in a shorter time. When using a cloud storage system, a user can access and edit files and data stored in a cloud storage unit in a variety of ways. A user may access files and data in the cloud storage unit using a local computer, mobile device, or remote desktop and edit and manipulate using various file editing tools. The user may also use backup and restore services provided by the cloud storage system for data recovery in the event of data loss or corruption.
In particular embodiments, the predictive model building unit generally refers to a process of selecting features in a dataset and building a predictive model. The purpose of this process is to select features that are relevant to the problem and to build an algorithm to generate the predicted value. In a unit for building a predictive model, the following steps are generally required:
feature selection, selecting features related to the problem. The process of feature selection requires consideration of factors such as importance, mutual information, relevance, etc. of features to ensure that the selected features accurately describe the problem.
Feature fusion-selected features are fused together to build a meaningful model. The feature fusion method comprises principal component analysis, feature scaling, feature crossing and the like.
Model selection selecting an appropriate method to build the predictive model. Common models include linear regression, logistic regression, decision trees, support vector machines, and the like.
Predictive model training-training a predictive model using the selected features and the selected model. The goal of the training process is to minimize training errors and improve the generalization ability of the model.
In particular embodiments, during data analysis, a processing unit generally refers to a software or hardware system for performing data analysis tasks. These processing units may be implemented by various tools and software packages, such as statistical software, data visualization tools, database management systems, etc.
The processing unit typically includes the following functions:
1. reading data, namely reading the data in the file or the network into a processing unit.
2. And data cleaning, namely preprocessing and cleaning the read data to ensure the data quality.
3. Data conversion, namely converting the data into a data format which can be understood and processed by the processing unit.
4. Data analysis, performing statistical analysis, data mining, or other data analysis tasks to obtain useful information.
5. Visualization of data-the analysis results are presented in a visual manner for easier understanding and interpretation.
6. Data storage-storing the analysis results and data records in a database or other data storage system.
Different data analysis tasks may require different processing units to implement, such as data reading, data cleansing, data analysis, data visualization, and the like.
In a specific embodiment, the visual display unit refers to a visual device for displaying data and information, such as a display, a television, a computer screen, etc. These display units can display various charts and graphs. The result of the prediction can be seen quickly by the model.
Through the embodiment, the prediction model and the training prediction model can be built, the prediction model building unit builds a real-time power demand prediction model through the XGBoost algorithm and trains and learns the power demand model through sample data, and the method comprises the following steps of:
step one: the XGBoost algorithm is a supervised integrated learning algorithm of a summation model of a plurality of decision trees, and the calculation formula is as follows:
（1）
in the formula (1), the components are as follows,representing model predictive value, +.>Indicate->Sample number->Representation about sample->K represents the number of decision trees,/-for the prediction function of (a)>An independent function representing a function space, F representing a function space, consisting of a decision tree, a given sample set having n samples, m features,/a>The independent function representing the function space has +.>Data genusCharacteristic feature point, independent function of data set s, +.>Is a model complexity coefficient;
the function integrates different data information such as daily electricity consumption, power consumption and the like into an algorithm model through integral calculation of independent functions of a function space, so that the capability of the algorithm for processing the data information and calculating the data information is improved. Predictions of future power demand are generated by analysis of historical data and current power supply conditions. The data information is not limited to collecting historical power demand data, including daily electricity consumption, power consumption, and the like. In the context of a specific embodiment of the present invention,indicate->A sample, such as a certain amount of information or data in the power demand, is continuously selected and calculated by taking different data information. The number of decision trees indicates classification category, such as classifying data information by time, classifying data by power, daily amount, etc. In a specific embodiment, a time sequence prediction model, a regression model and other different models can be adopted to calculate the data information. The classified data information is then model trained and evaluated, more specifically, using historical data to train and evaluate the performance of the model, such as accuracy, recall, precision, etc. The function space represents a set of spatial information of the acquired data information. And displaying and feeding back the predicted result to the user, and providing user feedback, such as changes of the electricity utilization habit of the user. The invention can help power system management personnel to plan and manage the power system better so as to ensure the sustainability and reliability of the power system.
In yet a further embodiment, a given sample set hasSamples, & ->A feature, wherein the training data information calculation expression is:
（2）
in the formula (2), the amino acid sequence of the compound,representation about->、/>Data set of->Indicate->Personal category labels, i.e.)>Actual value of the individual category, at the same time +.>，/>Representing a real set;
step two: the XGBoost algorithm objective function comprises a loss error function and a regularization term, wherein the loss error functionThe difference between the target predicted value and the target actual value is described; regular item->Adjusting the complexity of the control tree to avoid excessive repetition of the control tree;
in yet a further embodiment, the Loss error Function (Loss Function) is a Function that measures the difference between the model predicted and actual results. Are typically used to evaluate the predictive ability of the network to real data. The goal of the loss error function is to minimize or maximize the value of the function to minimize or maximize the variance of the model output. This may encourage the model to learn a more robust prediction mode. In particular embodiments, the loss error function includes mean square error (Mean Squared Error, MSE), cross Entropy (Cross Entropy), and the like.
A regular term refers to a set of predefined grammatical rules and conditional statements used in programming to define and verify the logical structure and code behavior of a program. In still further specific embodiments, the data information computing power is improved by setting regular expressions. Such as the start position of the matching string, the end position of the matching string, the grouping expression with parameters, for matching variable names, matching character sequences, including letters, numbers, and underlining, etc. These regular expressions can be used to match a variety of different types of data, such as text, numbers, symbols, list items, and so forth. When programming, regular expressions may be used to verify that the input data conforms to the intended format. In particular embodiments, the settings are made according to particular needs.
Step three: after the XGBoost algorithm creates a tree, feature extraction is directly completed by utilizing a feature load, two values of weight and gain are called, feature selection is completed, and a calculation expression of the gain is as follows:
（3）
in the formula (3), the amino acid sequence of the compound,for the gain expression, subscript L, R indicates the information tag identity of the left and right subtrees, respectively, after splitting, ++>Gain representing the ith data, +.>Representing left subtree gain，/>Representing right subtree gain, ">Weight representing the ith data, +.>Weight representing left subtree,/->Weight representing right subtree->Representing regularization parameters;
taking the evaluation index of the prediction model as the power prediction accuracy, the power prediction accuracy can be expressed as:
（4）
in the formula (4), the amino acid sequence of the compound,for power prediction accuracy, +.>Representing the actual output power at instant i, +.>Predicted output power representing model output, +.>The starting capacity of the integrated electric field is represented, and the output of the input formula is compared with a set threshold value.
In the prior art, gain refers to a technique of increasing the strength of a signal or amplifying a signal during signal transmission. Gain is typically used in amplifiers to amplify an input signal to a greater amplitude or to increase the strength of the signal. In audio processing, gain may also be used to convert analog signals to digital signals, or to amplify digital signals to a greater amplitude. In the invention, the predicted data information is amplified, for example, after the characteristics of the data information are extracted, the characteristic information is converted into a mode which can be identified by a user, for example, a plurality of different modes such as a graph, data information calculation and the like.
Weights refer to a magnitude that is used to describe the relative importance between two or more quantities. Generally, the weight is a coefficient of a quadratic function that defines the relative contribution between the quantities. By this arrangement, the data information calculation capability can be improved.
In a specific embodiment, a real-time power demand prediction model is established through an XGBoost algorithm, and the prediction results of training and learning of the power demand model are compared with the actual results through sample data, as shown in table 1:
TABLE 1 comparison of predicted and actual results
Through the embodiment, the invention calculates the difference between the predicted result and the actual result by comparing and analyzing the predicted result and the actual result, inputs data into the model after constructing the algorithm model, calculates and processes the data information in the model, and compares the predicted result with the actual result to evaluate the precision of the predicted model, and the method comprises the following steps:
（5）
in the formula (5), the amino acid sequence of the compound,refers to the actual value of the electrical load, +.>And->Respectively refers to the minimum and the maximum of the electrical loadThe large value, U, is the index normalized data amount, the accuracy of the prediction model is evaluated by using the following index:
i. relative error:
（6）
ii. Root mean square error:
（7）
iii, average absolute percentage error:
（8）
iv, average absolute error:
（9）
in the above formulas (6) to (9),which means that the corresponding predicted load is to be applied,
the smaller these evaluation indexes are, the higher the prediction accuracy is.
In a specific embodiment, the difference between the predicted result and the actual result is calculated by comparing the predicted data with the actual result, and the accuracy of the prediction model is evaluated by calculating the relative error, the root mean square error, the mean absolute percentage error and the mean absolute error of multiple experiments.
Through the embodiment, the prediction result optimization adjustment module of the invention adjusts the weight ratio calculated by the prediction result by comparing and analyzing the prediction data with the actual result, and comprises the following steps:
step one: determining initial weights: setting the value of the first actual result asEach prediction model->Is the predicted value of (2)Wherein i->Then each prediction model +.>The initial weight of the model has higher initial weight for meeting the requirement of good first prediction effect;
step two: dynamically adjusting weights
A. Let current total m predicted results and actual results, including the predicted results of the previous k-1 times except the mth time、/>、…、/>And the experimental results of the first k times>、/>、…、/>Wherein k is a time window, and default is taken to be 5; when m is<Taking k=m at k; corresponding->、/>、…、/>Weights are +.>、、…、/>；
B. Predictive modelWherein i->Correlation coefficients of the calculation results of the previous k-1 times and the actual results of the previous k times except the mth time;
C. computing a predictive modelDeviation of the mth calculation result and the actual result;
D. according to A, B result, each prediction model is adaptively solvedWeight of mth adjustment->Wherein the value of the weight is increased for a prediction model having high correlation with the actual result and small deviation from the actual result; and reducing the value of the weight for the prediction model with low correlation with the actual result and the prediction model with larger deviation from the actual result.
In a specific embodiment, the weight of the prediction model is calculated by a linear combination method, the weight proportion is dynamically adjusted, the influence of external factors on the prediction result is reduced, and the accuracy of the real-time power prediction model is higher. Prediction model accuracy and optimization model accuracy pair such as table 2:
table 2 prediction model accuracy and optimization model accuracy vs
Through the embodiment, the security reinforcement of the cloud platform is realized through the double-layer detection firewall, the double-layer detection firewall detects the communication request and the communication content abnormally through the inspection engine, and the potential safety hazard detection and the improvement of the risk point rechecking of the cloud platform are carried out regularly.
In a specific embodiment, a network security policy is formulated first, a specification and a standard of network use are defined, and a reasonable security policy is formulated, including aspects of a north-south firewall policy, a east-west firewall policy, a vulnerability management policy, a user authority management policy, a password security policy and the like. The process of configuring the software system aims at the software systems such as a server operating system, a database, application middleware and the like, improves the security protection level of the system by patching, enhancing account security, reinforcing service, modifying security configuration, optimizing access control strategy and increasing security mechanism, and realizes security reinforcement of the system by security equipment deployment, hardware equipment encryption, network topology optimization, vulnerability management and repair.
Claims (7)
1. An artificial intelligence-based real-time power demand prediction system, which is characterized in that: comprising:
the data acquisition module is used for acquiring power supply and demand data in real time, wherein the power supply and demand data comprise power load data, power consumption, power supply quantity, voltage, current, weather data and industry data, and the data acquisition module acquires historical power demand data, weather data and industry data through a power supplier, an energy management system and an equipment monitoring database data sharing platform and acquires real-time power equipment data through a wireless communication mode;
the data preprocessing module is used for cleaning the acquired power supply and demand data, and the data preprocessing module adopts a data processing tool pandas to process missing values, abnormal values and repeated values in the acquired data;
the data prediction cloud platform module is used for constructing a data prediction model according to historical power demand data, weather data and industry data, predicting power demand in real time, accessing, uploading and managing the data cloud of the data preprocessing module through the Internet, building the real-time power demand prediction model through an XGBoost algorithm, training the prediction model through learning sample data, comprehensively analyzing a prediction result, comparing the prediction data with an actual result, calculating the difference between the prediction result and the actual result, evaluating the precision of the prediction model, and converting data into a shape or an image by utilizing a computer graphics and image processing method to achieve data visual display;
the model optimization and adjustment module is used for carrying out improvement and optimization on the data prediction model, calculating the weight of the prediction model by a linear combination method and dynamically adjusting the weight proportion;
the security reinforcement module is used for improving the robustness and security of the system, increasing the difficulty of invasion of an attacker and improving the security level of the system, wherein the security reinforcement is a process of configuring a software system, and the security level of the system is improved by patching, reinforcing account security, reinforcing service, modifying security configuration, optimizing access control strategies and increasing security mechanisms aiming at a server operating system, a database and an application middleware software system;
the output end of the data acquisition module is connected with the input end of the data preprocessing module, the output end of the data preprocessing module is connected with the input end of the data prediction cloud platform module, the output end of the data prediction cloud platform module is connected with the input end of the optimization adjustment module of the model, and the output end of the optimization adjustment module of the model is connected with the input end of the safety reinforcement module.
2. An artificial intelligence based real time power demand prediction system according to claim 1 wherein: the data preprocessing module adopts a data processing tool pandas to process missing values, abnormal values and repeated values in the acquired data, and achieves format standardization, abnormal data removal, error data correction and repeated data removal through a smooth noise data and outlier detection method, and adopts a mahalanobis distance to detect multiple outliers.
3. An artificial intelligence based real time power demand prediction system according to claim 1 wherein: the data prediction cloud platform module comprises a cloud storage unit, a prediction model building unit, a data analysis processing unit and a visual display unit.
4. A real-time power demand prediction system based on artificial intelligence according to claim 3, wherein: the prediction model building unit builds a real-time power demand prediction model through an XGBoost algorithm and trains and learns the power demand model through sample data, and the method comprises the following steps of:
step one: the XGBoost algorithm is a supervised integrated learning algorithm of a summation model of a plurality of decision trees, and the calculation formula is as follows:
（1）
in the formula (1), the components are as follows,representing model predictive value, +.>Indicate->Sample number->Representation about sample->And k represents a decision treeNumber of (1),. About.>An independent function representing a function space, F representing a function space, consisting of a decision tree, a given sample set having n samples, m features,/a>The independent function representing the function space has +.>Data attribute feature point, independent function with data set s, +.>Is a model complexity coefficient;
the XGBoost algorithm training data set is calculated as follows:
（2）
in the formula (2), D represents、/>Data set of->Indicate->Personal category labels, i.e.)>Actual values of individual classes at the same time，/>Representing a real set;
step two: the XGBoost algorithm objective function comprises a loss error function and a regularization term, wherein the loss error functionThe difference between the target predicted value and the target actual value is described; regular item->Adjusting the complexity of the control tree to avoid excessive repetition of the control tree;
step three: after the XGBoost algorithm creates a tree, feature extraction is directly completed by utilizing a feature load, two values of weight and gain are called, feature selection is completed, and a calculation expression of the gain is as follows:
（3）
in the formula (3), the amino acid sequence of the compound,for the gain expression, subscript L, R indicates the information tag identity of the left and right subtrees, respectively, after splitting, ++>Gain representing the ith data, +.>Representing left subtree gain, ">Representing right subtree gain, ">Weight representing the ith data, +.>Weight representing left subtree,/->Weight representing right subtree->Representing regularization parameters;
taking the evaluation index of the prediction model as the power prediction accuracy, the power prediction accuracy can be expressed as:
（4）
in the formula (4), the amino acid sequence of the compound,for power prediction accuracy, +.>Representing the actual output power at instant i, +.>Predicted output power representing model output, +.>The starting capacity of the integrated electric field is represented, and the output of the input formula is compared with a set threshold value.
5. A real-time power demand prediction system based on artificial intelligence according to claim 3, wherein: the data analysis processing unit is used for comparing and analyzing the predicted result and the actual result, calculating the difference between the predicted result and the actual result, inputting data into the model after constructing an algorithm model, calculating and processing data information in the model, and comparing the predicted result and the actual result to evaluate the precision of the predicted model, and the method comprises the following steps:
（5）
in the formula (5), the amino acid sequence of the compound,refers to the actual value of the electrical load, +.>And->Respectively, refers to the minimum and maximum values of the electrical load, U is the amount of index normalized data, and the accuracy of the predictive model is evaluated by using the following indices:
i. relative error:
（6）
ii. Root mean square error:
（7）
iii, average absolute percentage error:
（8）
iv, average absolute error:
（9）
in the above formulas (6) to (9),refers to the corresponding predicted load.
6. An artificial intelligence based real time power demand prediction system according to claim 1 wherein: the optimization adjustment module of the model is used for debugging the weight ratio calculated by the prediction result through comparing and analyzing the prediction data with the actual result, and the method comprises the following steps of:
step one: determining initial weights: setting the value of the first actual result asEach prediction model->Is +.>Wherein i->Then each prediction model +.>Is assigned according to the first prediction effect, a well performing model will be given a high initial weight;
step two: dynamically adjusting weights
A. Let current total m predicted results and actual results, including the predicted results of the previous k-1 times except the mth time、、…、/>And the experimental results of the first k times>、/>、…、/>Wherein k is a time window, and default is taken to be 5; when m is<Taking k=m at k; corresponding->、/>、…、/>Weights are +.>、/>、…、；
B. Predictive modelWherein i->Correlation coefficients of the calculation results of the previous k-1 times and the actual results of the previous k times except the mth time;
C. computing a predictive modelDeviation of the mth calculation result and the actual result;
D. according to A, B result, each prediction model is adaptively solvedWeight of mth adjustment->Wherein the value of the weight is increased for a prediction model having high correlation with the actual result and small deviation from the actual result; the value of the weight is reduced for a prediction model having low correlation with the actual result and a prediction model having large deviation from the actual result.
7. An artificial intelligence based real time power demand prediction system according to claim 1 wherein: the security reinforcement module realizes security reinforcement of the cloud platform through a double-layer detection firewall, and the double-layer detection firewall performs abnormal detection on the communication request and the communication content through an inspection engine and periodically performs potential safety hazard detection and improved risk point rechecking of the cloud platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311047162.6A CN116760033B (en) | 2023-08-21 | 2023-08-21 | Real-time power demand prediction system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311047162.6A CN116760033B (en) | 2023-08-21 | 2023-08-21 | Real-time power demand prediction system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116760033A true CN116760033A (en) | 2023-09-15 |
| CN116760033B CN116760033B (en) | 2024-04-12 |
Family
ID=87957601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311047162.6A Active CN116760033B (en) | 2023-08-21 | 2023-08-21 | Real-time power demand prediction system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116760033B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103399867A (en) * | 2013-07-05 | 2013-11-20 | 西安交通大学 | Self-adaptive adjusting method for linear combination prediction model weight |
| CN109858674A (en) * | 2018-12-27 | 2019-06-07 | 国网浙江省电力有限公司 | Monthly load forecasting method based on XGBoost algorithm |
| CN111340273A (en) * | 2020-02-17 | 2020-06-26 | 南京邮电大学 | Short-term load prediction method for power system based on GEP parameter optimization XGboost |
| CN112330050A (en) * | 2020-11-20 | 2021-02-05 | 国网辽宁省电力有限公司营口供电公司 | Power system load prediction method considering multiple features based on double-layer XGboost |
| CN114169595A (en) * | 2021-11-25 | 2022-03-11 | 中国海洋大学 | Learning situation prediction method based on big data |
| CN115130741A (en) * | 2022-06-20 | 2022-09-30 | 北京工业大学 | Multi-model fusion based multi-factor power demand medium and short term prediction method |
| CN115392340A (en) * | 2022-07-11 | 2022-11-25 | 江苏科技大学 | Power load prediction system and prediction method for multi-energy electric propulsion ship |
| CN115409292A (en) * | 2022-10-31 | 2022-11-29 | 广东电网有限责任公司佛山供电局 | Short-term load prediction method for power system and related device |
| US20230244197A1 (en) * | 2022-02-01 | 2023-08-03 | Enerallies, Inc. | Machine-learning-enhanced distributed energy resource management system |
-
2023
- 2023-08-21 CN CN202311047162.6A patent/CN116760033B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103399867A (en) * | 2013-07-05 | 2013-11-20 | 西安交通大学 | Self-adaptive adjusting method for linear combination prediction model weight |
| CN109858674A (en) * | 2018-12-27 | 2019-06-07 | 国网浙江省电力有限公司 | Monthly load forecasting method based on XGBoost algorithm |
| CN111340273A (en) * | 2020-02-17 | 2020-06-26 | 南京邮电大学 | Short-term load prediction method for power system based on GEP parameter optimization XGboost |
| CN112330050A (en) * | 2020-11-20 | 2021-02-05 | 国网辽宁省电力有限公司营口供电公司 | Power system load prediction method considering multiple features based on double-layer XGboost |
| CN114169595A (en) * | 2021-11-25 | 2022-03-11 | 中国海洋大学 | Learning situation prediction method based on big data |
| US20230244197A1 (en) * | 2022-02-01 | 2023-08-03 | Enerallies, Inc. | Machine-learning-enhanced distributed energy resource management system |
| CN115130741A (en) * | 2022-06-20 | 2022-09-30 | 北京工业大学 | Multi-model fusion based multi-factor power demand medium and short term prediction method |
| CN115392340A (en) * | 2022-07-11 | 2022-11-25 | 江苏科技大学 | Power load prediction system and prediction method for multi-energy electric propulsion ship |
| CN115409292A (en) * | 2022-10-31 | 2022-11-29 | 广东电网有限责任公司佛山供电局 | Short-term load prediction method for power system and related device |
Non-Patent Citations (2)
| Title |
|---|
| 唐德栋: ""基于人工智能的短期电力负荷预测方法研究"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, 15 September 2020 (2020-09-15), pages 18 - 19 * |
| 高盼: ""基于Hadoop的电力大数据可视化研究与实现"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, 15 January 2020 (2020-01-15), pages 8 - 26 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116760033B (en) | 2024-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109816221B (en) | Project risk decision method, apparatus, computer device and storage medium | |
| CN112734128B (en) | 7-day power load peak prediction method based on optimized RBF | |
| CN114723285B (en) | Power grid equipment safety evaluation prediction method | |
| CN114048870A (en) | Power system abnormity monitoring method based on log characteristic intelligent mining | |
| CN115013859B (en) | User portrait-based household regulation and control method for heat supply secondary network | |
| CN113868953B (en) | Multi-unit operation optimization method, device and system in industrial system and storage medium | |
| CN115718861A (en) | Method and system for classifying power users and monitoring abnormal behaviors in high-energy-consumption industry | |
| CN117113159A (en) | Deep learning-based power consumer side load classification method and system | |
| CN116466672B (en) | Data center machine room parameter regulation and control method based on artificial intelligence and related device | |
| Ahmed et al. | Enhancing stock portfolios for enterprise management and investment in energy industry | |
| CN116760033B (en) | Real-time power demand prediction system based on artificial intelligence | |
| CN117113202A (en) | Power loop energy consumption detection method and equipment based on joint error stacking model | |
| CN113076217B (en) | Disk fault prediction method based on domestic platform | |
| CN115330050A (en) | Building load prediction method based on hybrid model | |
| CN114168409A (en) | Service system running state monitoring and early warning method and system | |
| Singh et al. | Enhancing wind power forecasting from meteorological parameters using machine learning models | |
| CN116956174B (en) | Classification model for cold head state classification detection and life prediction and generation method of prediction model | |
| Zhang et al. | Research on Sales Forecast of Automobile Spare Parts Based on LightGBM and Feature Engineering | |
| CN117272170B (en) | Knowledge graph-based IT operation and maintenance fault root cause analysis method | |
| Yuan et al. | Identification and Calibration Method of Deviation of Main Transformer Online Monitoring Date Groups | |
| CN116826733A (en) | Photovoltaic power prediction method and system | |
| Jayasri et al. | Evaluation of Weather Forecasting Models and Handling Anomalies in Short-Term Wind Speed Data | |
| CN118017502A (en) | Digital twinning-based power distribution calculation power prediction method, system and medium | |
| CN118316036A (en) | Load prediction method, device and equipment of power system | |
| CN117977536A (en) | Smart power grid load identification 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |