CN114021483A - Ultra-short-term wind power prediction method based on time domain characteristics and XGboost - Google Patents

Ultra-short-term wind power prediction method based on time domain characteristics and XGboost Download PDF

Info

Publication number
CN114021483A
CN114021483A CN202111400222.9A CN202111400222A CN114021483A CN 114021483 A CN114021483 A CN 114021483A CN 202111400222 A CN202111400222 A CN 202111400222A CN 114021483 A CN114021483 A CN 114021483A
Authority
CN
China
Prior art keywords
data
wind
wind power
xgboost
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202111400222.9A
Other languages
Chinese (zh)
Inventor
蒋甲丁
王凤
朱黎
王晓宇
孔德安
徐立军
高钾
薛建雄
李瑾
王营超
刘金庭
李柯豪
朱勇
周裕松
其丽格尔
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xinjiang Institute of Engineering
Original Assignee
Xinjiang Institute of Engineering
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xinjiang Institute of Engineering filed Critical Xinjiang Institute of Engineering
Priority to CN202111400222.9A priority Critical patent/CN114021483A/en
Publication of CN114021483A publication Critical patent/CN114021483A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/06—Wind turbines or wind farms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/06—Power analysis or power optimisation

Abstract

The invention discloses an ultra-short-term wind power prediction method based on time domain features and XGboost, which is characterized in that the features are classified and marked by a machine learning method according to influence factors such as wind speed, wind speed change rate, wind direction, air pressure, temperature, humidity and the like in a wind field environment, normalized and vectorized to enable the features to have domain feature information in a wind environment space, model training is carried out on the represented features by utilizing a machine learning algorithm, a wind environment feature set with the largest influence on the wind power and an influence coefficient of the wind environment feature set on the wind power are found out, then all the features are integrated into a model by utilizing an integrated learning method to predict the ultra-short-term power of the wind power, a model which enables the root mean square error and the average absolute error of a prediction result to be the minimum is found out according to a k-fold cross validation method, and finally a prediction model is formed. Factors influencing wind power are decomposed from five dimensions and four directions, and characteristics of the influencing factors are refined and fully expanded and excavated.

Description

Ultra-short-term wind power prediction method based on time domain characteristics and XGboost
Technical Field
The invention relates to the field of new energy power generation, in particular to an ultra-short-term wind power prediction method based on time domain characteristics and XGboost.
Background
Wind power is the third largest power source in China, and is one of the most mature power generation modes with the most scale development conditions and the most commercial development prospects in the field of new energy power generation. However, the wind power is strongly related to the wind field environment, especially to the wind speed, and the output power of the wind turbine generator is also strongly fluctuating and uncertain due to the fluctuation and uncertainty of the wind speed. Once the wind power with severe fluctuation is connected and connected to the grid, great impact is generated on the power grid, and a serious challenge is brought to the reliability of the power system.
The wind power is used as an unstable power supply of the power system, and has the characteristics of low energy density, uneven distribution, randomness, intermittence, uncontrollable performance and the like. In order to reduce the risk of low wind power reliability, deal with peak load regulation and frequency modulation pressure of a power system, and realize more efficient wind power consumption, so as to solve the problems of low electric energy quality, frequent voltage control, difficult active power scheduling, low system stability and the like caused by large-scale wind power integration, and effectively improve the load prediction precision of wind power generation becomes increasingly important. Accurate wind power prediction, especially short-term and ultra-short-term power prediction, has great significance to sustainable development and utilization of wind power, and is an important premise and guarantee for development of new energy industries and guarantee of reliable operation and orderly upgrading and reconstruction of large-scale power grids.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides an integrated learning algorithm based on XBboost and wind field environment time domain mining according to actual needs of increasingly improved wind power prediction accuracy and influence factors such as wind speed, wind speed change rate, wind direction, temperature, humidity and air pressure in a wind field environment, obtains environmental factors influencing wind power most by classifying and screening time domain data, and then performs ultra-short-term wind power prediction by combining historical wind power characteristics.
An ultra-short-term wind power prediction method based on time domain characteristics and XGboost comprises the following steps:
the first step is as follows: raw data pre-processing
Classifying and marking the characteristics by a machine learning method according to the influence factors such as wind speed, wind speed change rate, wind direction, air pressure, temperature, humidity and the like in the wind field environment, and carrying out normalization and vectorization expression to ensure that the characteristics have field characteristic information in the wind environment space;
the specific steps of the raw data preprocessing include: the missing value in the original data is completed by adopting an average value method and a fixed value method; identifying and deleting repeated values and redundant values in the data; identifying and correcting error values in the data by adopting an abnormal value identification method and then inserting new values for error correction by adopting a missing value completion method; and finally, normalizing the data of different dimensions, wherein the data in a 0-1 space is represented by calculating the maximum and minimum values of the data of each dimension and performing 0-1 mapping on the data by a common method.
The second step is that: data analysis and feature mining
Model training is carried out on the expressed characteristics by utilizing a machine learning algorithm, a wind environment characteristic set which has the largest influence on the wind power and an influence coefficient of the wind environment characteristic set on the wind power are found out,
the characteristic engineering of the data comprises: feature identification, feature extraction, feature representation and feature classification of the data. The data collected by the wind power plant comprise meteorological data, historical power data, state data of the fans and the like, and have the characteristics of multiple dimensions, large scale and high time sequence. Different types of data have different characteristics, the identification and extraction of the data characteristics need to be carried out according to specific information represented by the data, and then the automatic classification of the characteristics is realized by utilizing a machine learning algorithm. For example, the meteorological data are characterized and expanded from different height wind speeds and different height wind directions, and a plurality of dimensional characteristics such as different height wind speeds, average wind speeds in different height and different time ranges, different height wind speed change rates, and the influence of different height wind directions on different height wind speeds are formed. And finally, forming five dimensions of historical meteorological data, fan state data, historical power data, wind field environment data and geographic environment data, mining characteristics expanded from four directions of time sequence, variability, internal interactivity and external interactivity, performing characteristic representation on the original data, and forming a characteristic representation matrix of component types.
The third step: and constructing a wind power prediction model.
The method includes the steps that all features are integrated into a model by an integrated learning method to predict the ultra-short-term power of the wind power, and the model which enables the root mean square error and the average absolute error of a prediction result to be minimum is found out according to a k-fold cross validation method;
the factors influencing the wind power are numerous, and how to identify and screen out the factor influencing the wind power most is the key for constructing a wind power prediction model. The part is to screen and determine the main characteristics in the model, and the main contents include:
(1) and calculating correlation coefficients among the features determined in the second step to obtain a Pearson correlation coefficient matrix among the features. And analyzing the correlation among the input features through Pearson correlation coefficients to determine whether redundant features exist. The magnitude of the correlation coefficient represents 2 variables C1，C2The degree of linear correlation between the two is calculated by the formula:
Figure BDA0003370897080000031
in the formula:
Figure BDA0003370897080000032
are respectively C1，C2A standard deviation of (d); r e [ -1, 1]The closer r is to 1, the stronger the correlation among 2 input features, and the closer r is to 0, the weaker the correlation among 2 input features.
(2) And calculating the influence coefficient of each characteristic on the wind power from the five dimensions, screening out the dimension which has the largest influence on the wind power, and obtaining the influence coefficient distribution of each characteristic under each dimension.
(3) And finally, obtaining the characteristics of five dimensions by comparing data of different wind fields, and constructing an air-out power prediction model.
The fourth step: and training a wind power prediction model. The important thing in this section is to determine the weights of the individual main features in the model. The XGboost algorithm developed based on the ensemble learning method is adopted to carry out parameter training of a prediction model, a K-fold cross validation method is adopted to divide a training set and a test set, and a parameter model which enables RMSE and MAE to be minimum is calculated to screen out target parameters, so that the wind power prediction algorithm is constructed: TS _ XGB.
The XGboost algorithm is optimized and improved on the basis of GBDT, and regularization and second-order Taylor expansion are carried out on the target function. The mathematical model of XGBoost may be regarded as an additive model consisting of K classification and regression trees (CART).
Figure BDA0003370897080000041
In the formula: k is the number of trees; f is all possible CARTs; f. ofkIs a specific CART. In the regression process, the parameter θ ═ f1，f2，…，fkThen the target function of XGBoost becomes:
Figure BDA0003370897080000042
in the formula: the first part is a loss function; the second part is a regularization term, which is obtained by adding regularization terms of K trees. For the regularization term of the decision tree, each decision tree is improved through vector mapping, and the regularization term of the obtained XGboost is as follows:
Figure BDA0003370897080000043
in the formula: gamma and lambda are penalty coefficients of the model; t is the node number of the leaves; omega is the fraction of the leaves. Gradually approximating an optimization target function in steps, and adding 1 optimized CART (carrier-associated-tree) i.e. f on the basis of the existing t-1 trees in the t steptThen the objective function becomes:
Figure BDA0003370897080000044
taylor's second order expansion is carried out on the formula (6):
Figure BDA0003370897080000045
in the formula: gi is the first derivative of the loss function; hi is the second derivative of the loss function.
Figure BDA0003370897080000046
Figure BDA0003370897080000047
The fifth step: and (5) comparing and testing the wind power prediction model. The TS _ XGboost algorithm developed based on XGboost is obtained through the four steps of data preparation, feature mining and model construction, the accuracy of the algorithm is further verified through the data of different wind fields, and the algorithm is compared and tested with the algorithms of the same type, such as decision trees, SVM and the like, on the same data set, and the effectiveness of the algorithm is verified.
The invention has the technical effects and advantages that:
1) factors influencing wind power are decomposed from five dimensions and four directions, and characteristics of the influencing factors are refined and fully expanded and excavated.
2) The characteristics influencing the wind power are effectively identified and represented by the association degree and the membership degree of the characteristics.
3) The method fully combines the time-space characteristics of the influencing factors, realizes higher accuracy rate for predicting the wind power in the ultra-short term, and can greatly improve the effectiveness of the original prediction method.
4) The method can provide reference for wind power identification of wind farm management personnel, can improve the working efficiency of the working personnel, and further improves the utilization rate of wind power equipment and the wind power output rate.
Drawings
FIG. 1 is a diagram of an improved feature mining architecture used in an embodiment of the present invention.
FIG. 2 shows the TS _ XGB of the present invention: and (3) an ultra-short-term wind power prediction method flow chart based on time domain characteristics and XGboost.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
The method includes the steps of classifying and marking characteristics according to influence factors such as wind speed, wind speed change rate, wind direction, air pressure, temperature and humidity in a wind field environment through a machine learning method, carrying out normalization and vectorization expression to enable the characteristics to have domain characteristic information in a wind environment space, then carrying out model training on the expressed characteristics through a machine learning algorithm to find out a wind environment characteristic set which has the largest influence on wind power and influence coefficients of the wind environment characteristic set on the wind power, then utilizing an integrated learning method to fuse all the characteristics into a model to predict the ultra-short-term power of the wind power, finding out a model which enables the root mean square error and the average absolute error of a prediction result to be the smallest according to a k-fold cross validation method, and finally forming a prediction model.
Example two
As shown in fig. 1 and 2, an ultra-short-term wind power prediction method based on time domain features and XGBoost specifically includes the following steps:
the first step is as follows: and (4) preprocessing raw data. Because the data collected by the wind power plant are collected and transmitted by each fan signal sensor, the problems of data loss, repetition, error, redundancy and the like are inevitably caused in the forming and collecting process of the original data. The data preprocessing is a precondition for data analysis, and the specific steps comprise: the missing value in the original data is completed by adopting an average value method and a fixed value method; identifying and deleting repeated values and redundant values in the data; identifying and correcting error values in the data by adopting an abnormal value identification method and then inserting new values for error correction by adopting a missing value completion method; and finally, normalizing the data of different dimensions, wherein the data in a 0-1 space is represented by calculating the maximum and minimum values of the data of each dimension and performing 0-1 mapping on the data by a common method.
The second step is that: and (5) characteristic engineering of the data. The characteristic engineering of the data comprises: feature identification, feature extraction, feature representation and feature classification of the data. The data collected by the wind power plant comprise meteorological data, historical power data, state data of the fans and the like, and have the characteristics of multiple dimensions, large scale and high time sequence. Different types of data have different characteristics, the identification and extraction of the data characteristics need to be carried out according to specific information represented by the data, and then the automatic classification of the characteristics is realized by utilizing a machine learning algorithm. For example, the meteorological data are characterized and expanded from different height wind speeds and different height wind directions, and a plurality of dimensional characteristics such as different height wind speeds, average wind speeds in different height and different time ranges, different height wind speed change rates, and the influence of different height wind directions on different height wind speeds are formed. And finally, forming five dimensions of historical meteorological data, fan state data, historical power data, wind field environment data and geographic environment data, mining characteristics expanded from four directions of time sequence, variability, internal interactivity and external interactivity, performing characteristic representation on the original data, and forming a characteristic representation matrix of component types.
The third step: and constructing a wind power prediction model. The factors influencing the wind power are numerous, and how to identify and screen out the factor influencing the wind power most is the key for constructing a wind power prediction model. The part is to screen and determine the main characteristics in the model, and the main contents include:
(1) for the second stepThe correlation coefficient between the features is calculated by the determined features, so that a pearson correlation coefficient matrix between the features is obtained. And analyzing the correlation among the input features through Pearson correlation coefficients to determine whether redundant features exist. The magnitude of the correlation coefficient represents 2 variables C1，C2The degree of linear correlation between the two is calculated by the formula:
Figure BDA0003370897080000071
in the formula:
Figure BDA0003370897080000072
are respectively C1，C2A standard deviation of (d); r e [ -1, 1]The closer r is to 1, the stronger the correlation among 2 input features, and the closer r is to 0, the weaker the correlation among 2 input features.
(2) And calculating the influence coefficient of each characteristic on the wind power from the five dimensions, screening out the dimension which has the largest influence on the wind power, and obtaining the influence coefficient distribution of each characteristic under each dimension.
(3) And finally, obtaining the characteristics of five dimensions by comparing data of different wind fields, and constructing an air-out power prediction model.
The fourth step: and training a wind power prediction model. The important thing in this section is to determine the weights of the individual main features in the model. The XGboost algorithm developed based on the ensemble learning method is adopted to carry out parameter training of a prediction model, a K-fold cross validation method is adopted to divide a training set and a test set, and a parameter model which enables RMSE and MAE to be minimum is calculated to screen out target parameters, so that the wind power prediction algorithm is constructed: TS XGB.
The XGboost algorithm is optimized and improved on the basis of GBDT, and regularization and second-order Taylor expansion are carried out on the target function. The mathematical model of XGBoost may be regarded as an additive model consisting of K classification and regression trees (CART).
Figure BDA0003370897080000073
In the formula: k is the number of trees; f is all possible CARTs; f. ofkIs a specific CART. In the regression process, the parameter θ ═ f1，f2，…，fkThen the target function of XGBoost becomes:
Figure BDA0003370897080000074
in the formula: the first part is a loss function; the second part is a regularization term, which is obtained by adding regularization terms of K trees. For the regularization term of the decision tree, each decision tree is improved through vector mapping, and the regularization term of the obtained XGboost is as follows:
Figure BDA0003370897080000081
in the formula: gamma and lambda are penalty coefficients of the model; t is the node number of the leaves; omega is the fraction of the leaves. Gradually approximating an optimization target function in steps, and adding 1 optimized CART (carrier-associated-tree) i.e. f on the basis of the existing t-1 trees in the t steptThen the objective function becomes:
Figure BDA0003370897080000082
taylor's second order expansion is carried out on the formula (6):
Figure BDA0003370897080000083
in the formula: gi is the first derivative of the loss function; hi is the second derivative of the loss function.
Figure BDA0003370897080000084
Figure BDA0003370897080000085
The fifth step: and (5) comparing and testing the wind power prediction model. The TS _ XGboost algorithm developed based on XGboost is obtained through the four steps of data preparation, feature mining and model construction, the accuracy of the algorithm is further verified through the data of different wind fields, and the algorithm is compared and tested with the algorithms of the same type, such as decision trees, SVM and the like, on the same data set, and the effectiveness of the algorithm is verified.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that are within the spirit and principle of the present invention are intended to be included in the scope of the present invention.

Claims (5)

1. An ultrashort-term wind power prediction method based on time domain characteristics and XGboost is characterized by comprising the following steps:
the first step is as follows: raw data pre-processing
Classifying and marking the features by a machine learning method according to the influence factors in the wind field environment, and carrying out normalization and vectorization expression to ensure that the features have domain feature information in the wind environment space;
the second step is that: data analysis and feature mining
Performing model training on the expressed characteristics by using a machine learning algorithm, and finding out a wind environment characteristic set which has the greatest influence on the wind power and an influence coefficient of the wind environment characteristic set on the wind power;
different types of data have different characteristics, the identification and extraction of the data characteristics need to be carried out according to specific information represented by the data, and then the automatic classification of the characteristics is realized by utilizing a machine learning algorithm; finally, five dimensions of historical meteorological data, fan state data, historical power data, wind field environment data and geographic environment data are formed, characteristic mining is carried out on the original data from the characteristics expanded from four directions of time sequence, variability, internal interactivity and external interactivity, and a characteristic representation matrix of component types is formed;
the third step: wind power prediction model construction
The method includes the steps that all features are integrated into a model by an integrated learning method to predict the ultra-short-term power of the wind power, and the model which enables the root mean square error and the average absolute error of a prediction result to be minimum is found out according to a k-fold cross validation method;
the factors influencing the wind power are numerous, and the key for constructing a wind power prediction model is to identify and screen out the factor influencing the wind power most; the step is to screen and determine the main characteristics in the model, and the main contents comprise:
(1) and calculating correlation coefficients among the features determined in the second step to obtain a Pearson correlation coefficient matrix among the features. Analyzing the correlation among the input features through a Pearson correlation coefficient to determine whether redundant features exist or not; the magnitude of the correlation coefficient represents 2 variables C1，C2The degree of linear correlation between the two is calculated by the formula:
Figure FDA0003370897070000021
in the formula:
Figure FDA0003370897070000022
are respectively C1，C2A standard deviation of (d); r e [ -1, 1]The closer r is to 1, the stronger the correlation among the 2 input features, and the closer r is to 0, the weaker the correlation among the 2 input features;
(2) calculating the influence coefficient of each characteristic on the wind power from five dimensions, screening out the dimension which has the largest influence on the wind power, and obtaining the influence coefficient distribution of each characteristic under each dimension;
(3) obtaining characteristics of the final five dimensions by comparing data of different wind fields, and constructing an air-out power prediction model;
the fourth step: wind power prediction model training
This part is to determine the weight of the main features in the model; the XGboost algorithm developed based on the ensemble learning method is adopted to carry out parameter training of a prediction model, a K-fold cross validation method is adopted to divide a training set and a test set, and a parameter model which enables RMSE and MAE to be minimum is calculated to screen out target parameters, so that the wind power prediction algorithm is constructed: TS _ XGB;
the XGboost algorithm is optimized and improved on the basis of GBDT, and regularization and second-order Taylor expansion are carried out on a target function; the mathematical model of XGBoost can be regarded as an addition model composed of K classification and regression trees (CART);
Figure FDA0003370897070000023
in the formula: k is the number of trees; f is all possible CARTs; f. ofkIs a specific CART; in the regression process, the parameter θ ═ f1，f2，…，fkThen the target function of XGBoost becomes:
Figure FDA0003370897070000024
in the formula: the first part is a loss function; the second part is a regular term and is obtained by adding the regular terms of the K trees; for the regularization term of the decision tree, each decision tree is improved through vector mapping, and the regularization term of the obtained XGboost is as follows:
Figure FDA0003370897070000025
in the formula: gamma and lambda are penalty coefficients of the model; t is the node number of the leaves; omega is the fraction of the leaves; gradually approximating an optimization target function in steps, and adding 1 optimized CART (carrier-associated-tree) i.e. f on the basis of the existing t-1 trees in the t steptThen, thenThe objective function becomes:
Figure FDA0003370897070000031
taylor's second order expansion is carried out on the formula (6):
Figure FDA0003370897070000032
in the formula: gi is the first derivative of the loss function; hi is the second derivative of the loss function;
Figure FDA0003370897070000033
Figure FDA0003370897070000034
the fifth step: wind power prediction model comparison test
The TS _ XGboost algorithm developed based on XGboost is obtained through the four previous steps of data preparation, feature mining and model construction.
2. The ultra-short-term wind power prediction method based on the time domain feature and the XGboost as claimed in claim 1, wherein: in the first step, the influencing factors in the wind farm environment include, but are not limited to: wind speed, rate of change of wind speed, wind direction, air pressure, temperature, humidity.
3. The ultra-short-term wind power prediction method based on the time domain feature and the XGboost as claimed in claim 1, wherein: in the first step, the specific steps of raw data preprocessing include: the missing value in the original data is completed by adopting an average value method and a fixed value method; identifying and deleting repeated values and redundant values in the data; identifying and correcting error values in the data by adopting an abnormal value identification method and then inserting new values for error correction by adopting a missing value completion method; and finally, normalizing the data of different dimensions, wherein the data in a 0-1 space is represented by calculating the maximum and minimum values of the data of each dimension and performing 0-1 mapping on the data by a common method.
4. The ultra-short-term wind power prediction method based on the time domain feature and the XGboost as claimed in claim 1, wherein: in the second step, the characteristic engineering of the data comprises: data feature identification, feature extraction, feature representation and feature classification; the data collected by the wind power plant comprise meteorological data, historical power data, state data of the fans and the like, and have the characteristics of multiple dimensions, large scale and high time sequence.
5. The ultra-short-term wind power prediction method based on the time domain feature and the XGboost as claimed in claim 1, wherein: in the fourth step, the accuracy of the algorithm is further verified, and the algorithm is similar to the same type of algorithm, including but not limited to: decision tree DecisionTree, SVM.
CN202111400222.9A 2021-11-24 2021-11-24 Ultra-short-term wind power prediction method based on time domain characteristics and XGboost Pending CN114021483A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111400222.9A CN114021483A (en) 2021-11-24 2021-11-24 Ultra-short-term wind power prediction method based on time domain characteristics and XGboost

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111400222.9A CN114021483A (en) 2021-11-24 2021-11-24 Ultra-short-term wind power prediction method based on time domain characteristics and XGboost

Publications (1)

Publication Number Publication Date
CN114021483A true CN114021483A (en) 2022-02-08

Family

ID=80065897

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111400222.9A Pending CN114021483A (en) 2021-11-24 2021-11-24 Ultra-short-term wind power prediction method based on time domain characteristics and XGboost

Country Status (1)

Country Link
CN (1) CN114021483A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114692515A (en) * 2022-06-01 2022-07-01 中材邦业(杭州)智能技术有限公司 Soft measurement method for clinker free calcium content based on time lag XGBOOST model
CN115357065A (en) * 2022-08-27 2022-11-18 华能新能源股份有限公司 Remote intelligent dehumidification control system and method for offshore wind turbine
CN116739172A (en) * 2023-06-19 2023-09-12 河海大学 Method and device for ultra-short-term prediction of offshore wind power based on climbing identification
CN117055512A (en) * 2023-10-10 2023-11-14 安徽中科维德数字科技有限公司 Intelligent model control system of continuous stirring reaction kettle and control method thereof

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114692515A (en) * 2022-06-01 2022-07-01 中材邦业(杭州)智能技术有限公司 Soft measurement method for clinker free calcium content based on time lag XGBOOST model
CN115357065A (en) * 2022-08-27 2022-11-18 华能新能源股份有限公司 Remote intelligent dehumidification control system and method for offshore wind turbine
CN115357065B (en) * 2022-08-27 2024-04-12 华能新能源股份有限公司 Remote intelligent dehumidification control system and method for offshore wind turbine
CN116739172A (en) * 2023-06-19 2023-09-12 河海大学 Method and device for ultra-short-term prediction of offshore wind power based on climbing identification
CN116739172B (en) * 2023-06-19 2023-12-08 河海大学 Method and device for ultra-short-term prediction of offshore wind power based on climbing identification
CN117055512A (en) * 2023-10-10 2023-11-14 安徽中科维德数字科技有限公司 Intelligent model control system of continuous stirring reaction kettle and control method thereof

Similar Documents

Publication Publication Date Title
CN114021483A (en) Ultra-short-term wind power prediction method based on time domain characteristics and XGboost
CN111915092B (en) Ultra-short-term wind power prediction method based on long-short-term memory neural network
CN112818604A (en) Wind turbine generator risk degree assessment method based on wind power prediction
CN110942205A (en) Short-term photovoltaic power generation power prediction method based on HIMVO-SVM
CN110717610A (en) Wind power prediction method based on data mining
Kaplan et al. A novel method based on Weibull distribution for short-term wind speed prediction
CN110991737A (en) Ultra-short-term wind power prediction method based on deep belief network
CN113822418A (en) Wind power plant power prediction method, system, device and storage medium
Chen et al. Research on wind power prediction method based on convolutional neural network and genetic algorithm
CN112288157A (en) Wind power plant power prediction method based on fuzzy clustering and deep reinforcement learning
CN112819238A (en) Short-term wind power prediction method based on chaotic chicken flock optimization algorithm
CN115995810A (en) Wind power prediction method and system considering weather fluctuation self-adaptive matching
Jiading et al. TS_XGB: Ultra-short-term wind power forecasting method based on fusion of time-spatial data and XGBoost algorithm
CN110738363A (en) photovoltaic power generation power prediction model and construction method and application thereof
CN113627674A (en) Distributed photovoltaic power station output prediction method and device and storage medium
Niu et al. Knowledge mining based on environmental simulation applied to wind farm power forecasting
CN116896093A (en) Online analysis and optimization method for grid-connected oscillation stability of wind farm
CN116470491A (en) Photovoltaic power probability prediction method and system based on copula function
CN116128211A (en) Wind-light-water combined short-term optimization scheduling method based on wind-light uncertainty prediction scene
CN115796327A (en) Wind power interval prediction method based on VMD (vertical vector decomposition) and IWOA-F-GRU (empirical mode decomposition) -based models
Guan et al. Scenario generation of wind power based on longitudinal-horizontal clustering strategy
CN113095547A (en) Short-term wind power prediction method based on GRA-LSTM-ICE model
Liu et al. Ultra-short-term wind power forecasting based on stacking model
Ren et al. A robust short-term wind power forecasting algorithm based on LSTM-XGBoost model
Cho et al. Application of Parallel ANN-PSO to Hourly Solar PV Estimation

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