CN-115526429-B - Wind power prediction error decoupling analysis method, processor and storage medium

CN115526429BCN 115526429 BCN115526429 BCN 115526429BCN-115526429-B

Abstract

The invention relates to a decoupling analysis method for wind power prediction errors, and belongs to the field of wind power prediction. The wind power prediction method comprises the steps of obtaining data required by wind power prediction error decoupling based on wind power plant historical data, combining an actual wind power plant prediction flow, dividing wind power prediction into three links including a numerical weather forecast, a wind-electricity conversion model and a prediction result correction, calculating errors caused by correction of the prediction result in consideration of weak coupling relation between the correction link and other two links, fitting the wind-electricity conversion model on the basis, fitting wind-electricity conversion model output under real meteorological conditions, comparing the wind-electricity conversion model output with the output under the numerical weather forecast condition, calculating error change of the wind-electricity conversion model under different meteorological conditions, and finally, building a wind power error decoupling equation by integrating the links, solving the equation and calculating error ratio. The method can reliably identify key influencing factors causing wind power prediction errors.

Inventors

Hao Luqian
SHEN FEIFAN
HE LIFU
WANG XIAOYUAN
WANG DING
SHEN YANGWU
HU YUHAN
GUO HU
YE JIANXING
ZENG XIANDONG
ZHU LIPENG
LI JIAYONG

Assignees

国网湖南省电力有限公司电力科学研究院
国网湖南省电力有限公司
国家电网有限公司

Dates

Publication Date: 20260512
Application Date: 20221028

Claims (11)

1. The wind power prediction error decoupling analysis method is characterized in that the wind power prediction process comprises a numerical weather prediction link, a wind-electricity conversion model link and a prediction result correction link, and comprises the following steps of: Acquiring data required by wind power prediction error analysis from historical operation data of a wind power plant, wherein the data comprise actual measurement weather data, numerical weather forecast data, installed capacity, real-time starting capacity, planned starting capacity, wind power prediction data and actual output power of the wind power plant; Determining a first target error caused by the prediction result correction link; Fitting a wind-electricity conversion model of the wind power plant according to the required data to establish a wind-electricity conversion fitting model; inputting the actually measured weather data of the wind power plant to the wind power conversion fitting model to obtain model output under real meteorological conditions; determining wind power prediction power under real meteorological conditions according to the installed capacity, the planned startup capacity and model output under the real meteorological conditions; determining a prediction result correction error under the real meteorological conditions according to the real-time starting capacity, the planned starting capacity and the wind power prediction power under the real meteorological conditions; determining a wind-electricity conversion model error under the real meteorological conditions according to the wind power predicted power under the real meteorological conditions, the actual output power of the wind power plant and the predicted result correction error under the real meteorological conditions; obtaining model output under numerical weather forecast corresponding to model output under real meteorological conditions; Determining a second target error caused by a wind-electricity conversion model link according to the model output under the real meteorological condition, the model output under the numerical weather forecast and the wind-electricity conversion model error under the real meteorological condition; Determining a third target error caused by the numerical weather forecast link according to the wind power prediction data, the actual output power of the wind power plant, the first target error and the second target error, and Determining the duty ratio of each target error according to the first target error, the second target error and the third target error; the determining a second target error caused by the wind-electricity conversion model link according to the model output under the real meteorological condition, the model output under the numerical weather forecast and the wind-electricity conversion model error under the real meteorological condition comprises the following steps: calculating the second target error according to the following formula: Wherein, the Is the second target error of the first set, Is the model output under the numerical weather forecast, Is the wind-electricity conversion model error under the real meteorological condition, and The model output under the real meteorological conditions; the model output under the numerical weather forecast can be calculated according to the following formula: wherein f m is the output of the wind-electricity conversion model under the condition of numerical weather forecast, For the installed capacity of the container, In order to plan the capacity of the power on, And predicting data for wind power.
2. The wind power prediction error decoupling analysis method of claim 1, further comprising: And analyzing links causing main factors of errors in each link in the wind power prediction process according to the duty ratio of each target error.
3. The wind power prediction error decoupling analysis method according to claim 1, wherein the determining the first target error caused by the prediction result correction step includes: determining an equivalent wind power predicted value according to the wind power predicted data, the real-time starting-up capacity and the planned starting-up capacity; subtracting the equivalent wind power predicted value from the wind power predicted data to obtain the first target error.
4. The method for decoupling a wind power prediction error according to claim 1, wherein the fitting a wind-to-electrical conversion model of a wind farm according to the required data to establish a wind-to-electrical conversion fitted model comprises: Determining wind-electricity conversion model output according to the installed capacity, the planned starting capacity and wind power prediction data; inputting the numerical weather forecast data as a wind-electricity conversion model, and outputting the numerical weather forecast data and the wind-electricity conversion model to form a training set and a testing set; Training the training set by XGBoost to fit a wind-electricity conversion model; after training is completed, judging whether the fitted wind-electricity conversion model meets the precision requirement or not by using the test set; initial parameters of XGBoost are modified, and training is conducted on the training set repeatedly until accuracy requirements are met; And taking the fitted wind-electricity conversion model meeting the precision requirement as the wind-electricity conversion fitting model.
5. The wind power prediction error decoupling analysis method of claim 4, wherein the wind-to-electrical conversion model comprises one of: A wind-electricity conversion model based on BP neural network, a wind-electricity conversion model based on LSTM cyclic neural network and a wind-electricity conversion model based on CNN-LSTM hybrid neural network.
6. The wind power prediction error decoupling analysis method of claim 1, wherein determining wind power prediction power under real weather conditions from the installed capacity, the planned startup capacity, and model output under real weather conditions comprises: Calculating wind power prediction power under real meteorological conditions according to the following formula: Wherein, the Is the wind power predicted power under the real meteorological condition, Is the planned starting capacity, Is the installed capacity of the machine in question, Is the model output under real meteorological conditions.
7. The method for decoupling wind power prediction error analysis according to claim 6, wherein determining the prediction result correction error under real weather conditions based on the real-time start-up capacity, the planned start-up capacity, and the wind power prediction power under real weather conditions comprises: Calculating a prediction result correction error under real meteorological conditions according to the following formula: Wherein, the Is the prediction result correction error under the real meteorological condition, Is the real-time power-on capacity, Is the planned power-on capacity of the device, The wind power prediction power under the real meteorological conditions.
8. The method for decoupling a wind power prediction error according to claim 7, wherein determining a wind-electricity conversion model error under real weather conditions based on the wind power prediction under real weather conditions, the wind farm actual output power, and the prediction result correction error under real weather conditions comprises: And calculating the wind-electricity conversion model error under the real meteorological condition according to the following formula: Wherein, the Is the wind power predicted power under the real meteorological condition, Is the actual output power of the wind farm, Is the wind-electricity conversion model error under the real meteorological condition, and Is the prediction result correction error under the real meteorological condition.
9. The method for decoupling a wind power prediction error according to claim 1, wherein determining a third target error caused by the numerical weather forecast link according to the wind power prediction data, the wind farm actual output power, the first target error, and the second target error comprises: calculating the third target error according to the following formula: Wherein, the Is the wind power prediction data, Is the first target error of the first set, Is the second target error, and Is the third target error of the first and second target errors, Is the actual output power of the wind farm.
10. A processor configured to perform the wind power prediction error decoupling analysis method of any one of claims 1 to 9.
11. A machine-readable storage medium having instructions stored thereon, which when executed by a processor, cause the processor to implement a wind power prediction error decoupling analysis method according to any one of claims 1 to 9.

Description

Wind power prediction error decoupling analysis method, processor and storage medium Technical Field The invention relates to the field of wind power prediction, in particular to a wind power prediction error decoupling analysis method, a processor and a storage medium. Background Wind energy is an important renewable clean energy source, has abundant reserves and wide distribution, and therefore wind power generation has become an important research direction for the development and utilization of renewable energy sources. However, the wind resources in the nature have the characteristics of randomness, intermittence and the like, and the random wind speed and wind cause the output power of the wind turbine to show obvious fluctuation, so that the stability of the voltage and the frequency of the power system is greatly threatened. When the wind power in the power system is connected in excess of a certain proportion, the safety and stability of the whole system face a great challenge. In order to effectively absorb new energy sources such as wind power with outstanding intermittence and randomness, the power system needs to be added with additional spare capacity, so that the total operation cost of the power system is increased. In order to reduce the configuration cost of the spare capacity of the power system and improve the safety and reliability of the overall operation of the power system, how to accurately predict the wind power is an important difficult problem to be broken through in the field of operation and control of the power system. The existing wind power prediction technology is difficult to accurately describe and characterize the inherent characteristics of wind power generation, and the prediction result inevitably has errors. The method has the advantages that objective and comprehensive analysis and evaluation are carried out on the error of wind power prediction, key influence factors and weak links in the wind power prediction process are facilitated to be identified, further the existing wind power prediction technology is guided to be further improved and optimized, and the method has important significance in improving the wind power prediction precision. However, the existing wind power prediction error analysis method does not consider the change of errors of all links of wind power prediction after the conditions such as meteorological data and the like are changed. Under the condition of neglecting the error change characteristic, the reliability of the existing wind power prediction error analysis method is greatly reduced. Disclosure of Invention The embodiment of the invention aims to provide a wind power prediction error decoupling analysis method, a processor and a storage medium. In order to achieve the above objective, a first aspect of the present invention provides a wind power prediction error decoupling analysis method, where a wind power prediction process includes a numerical weather prediction link, a wind-electricity conversion model link, and a prediction result correction link, and the wind power prediction error decoupling analysis method includes: acquiring data required by wind power prediction error analysis from historical operation data of a wind power plant, wherein the data comprise actual measurement weather data, numerical weather forecast data, installed capacity, real-time starting capacity, planned starting capacity, wind power prediction data and actual output power of the wind power plant; Determining a first target error caused by a prediction result correction link; Fitting a wind-electricity conversion model of the wind power plant according to the required data to establish a wind-electricity conversion fitting model; inputting the actually measured weather data of the wind power plant into a wind power conversion fitting model to obtain model output under real meteorological conditions; Determining wind power prediction power under real meteorological conditions according to the installed capacity, the planned starting capacity and the model output under the real meteorological conditions; determining a prediction result correction error under the real meteorological conditions according to the real-time starting capacity, the planned starting capacity and the wind power prediction power under the real meteorological conditions; determining a wind-electricity conversion model error under the real meteorological conditions according to the wind power predicted power under the real meteorological conditions, the wind power plant actual output power and the predicted result correction error under the real meteorological conditions; obtaining model output under numerical weather forecast corresponding to model output under real meteorological conditions; Determining a second target error caused by a wind-electricity conversion model link according to the model output under the real meteorological condition, the model output under the numerica