Numerical weather forecast total radiation system error classification calculation method

Abstract

The invention discloses a numerical weather forecast total radiation system error classification calculation method. The method comprises steps of preprocessing historical power sample data of a photovoltaic power station, including power sequence missing measurement processing and power sequence smoothing processing; feature extraction being carried out from the photovoltaic daily output amplitude and the fluctuation condition, and corresponding daily power sequence features being constructed; constructing an input vector matrix based on a self-organizing neural network, and performing normalization processing; determining a forecast weather type according to the forecast day numerical forecast data; and calculating a numerical weather forecast total radiation system error. Compared withthe prior art, changes of all meteorological data are integrated from the perspective of power characteristics, and the classification result is more suitable for photovoltaic power station historicaldata classification compared with multi-meteorological-factor clustering and has higher accuracy; the influence of NWP misinformation on system error calculation is avoided, and the obtained NWP total radiation system error is more accurate.

Description

technical field [0001] The invention belongs to the field of photovoltaic power generation, and in particular relates to a numerical weather forecast total radiation system error calculation method based on the classification of historical data of photovoltaic power stations. Background technique [0002] The output power of photovoltaic power generation system is obviously affected by environmental factors, and its output has volatility, uncertainty and indirectness. Accurate prediction of photovoltaic power generation power can help the grid to formulate fine dispatching plans and reduce the impact of photovoltaic power generation instability on the grid. shock. The premise of accurate photovoltaic power forecasting is accurate weather forecast data. Numerical weather forecast is often used as a data source for photovoltaic power forecasting. Therefore, establishing an effective numerical weather forecast error correction model is of great significance to improve the forec...

AI Technical Summary

Problems solved by technology

In the field of photovoltaic power generation forecasting, most scholars use the data of meteorological elements such as irradiance, temperature, and humidity provided by numerical weather forecasting service agencies as input variables of the power forecasting model without error correction, and the power forecasting accuracy cannot be guaranteed. During the seasonal change period, the accuracy of numerical weather prediction is poor, and the accuracy of photovoltaic power prediction is also reduced, and the prediction results cannot meet the assessment standards

[0004] Relevant literatures apply the model output statistics (MOS) method in meteorological operational forecasting to the field of power prediction in the case of measured meteorological data. Three dynamic MOS methods of classification median, classification regression and classification clustering are used to establish power and raw values. The relationship between the weather forecast and the three methods are given appropriate weights to realize the forecast of photovoltaic power combination; at the same time, some literature proposes to use the long-term average difference between the numerical weather forecast prediction field and the actual measurement field of the weather station as the systematic error of numerical weather prediction. To a certain extent, the calculation of the numerical weather prediction system error has been realized, but it treats all samples in the same way, and the obtained systematic error cannot accurately correct the numerical weather prediction, and its accuracy needs to be further improved

Images