Ultra-short-term wind power prediction method for offshore wind farms considering output fluctuation process

Abstract

The invention relates to an ultra-short-term wind power forecasting method for an offshore wind farm considering the output fluctuation process, comprising the following steps: 1) constructing a time-series forecasting model of offshore wind power considering meteorological variables and meteorological fluctuation variables; 2) constructing an improvement considering meteorological factors Long-term cyclic convolutional neural network; 3) Random initialization improves the weight of the long-term cyclic convolutional neural network; 4) Constructs a multi-convolution channel with a residual structure to obtain the time series characteristics of meteorological factors; 5) Optimizes and improves the long-term cyclic convolutional neural network The weight of the network; 6) if the maximum number of iterations is reached, the iteration is terminated, and the training of the improved long-term cyclic convolutional neural network is completed; otherwise, k=k+1 is set, and step 4) is returned; 7) the improvement completed according to the training The long-term cyclic convolutional neural network is used to solve the offshore wind power time series prediction model; 8) The XGboost algorithm is used to correct the error according to the optimal characteristics. Compared with the prior art, the present invention has the advantages of high prediction accuracy and comprehensive consideration.

Description

technical field [0001] The invention relates to the field of ultra-short-term wind power prediction of wind farms, in particular to a method for ultra-short-term wind power prediction of offshore wind farms that takes into account output fluctuation process. Background technique [0002] With the continuous expansion of the grid-connected scale of offshore wind power, the impact of the random fluctuation of offshore wind power output on the safe and stable operation of the power system has become increasingly evident. Accurate ultra-short-term offshore wind power forecasting (WPP) technology is an important means to solve the dispatching and stable operation of power systems. At present, the ultra-short-term WPP technology mostly takes onshore wind power as the research object, and the ultra-short-term offshore WPP technology is still in its infancy, and its prediction results are difficult to meet the actual engineering needs. Therefore, the new prediction method is of gre...

AI Technical Summary

Problems solved by technology

However, considering the maintenance difficulty of offshore units and the laying cost of transmission cables, offshore wind farms are usually distributed in a "high concentration", so that their output fluctuations may reach a very significant level

Moreover, under different meteorological conditions, the hourly offshore wind power shows completely different fluctuation characteristics, which seriously affects the overall performance of the offshore WPP model. Therefore, how to accurately predict the wind power under various fluctuation characteristics is the main difficulty of ultra-short-term offshore WPP.

[0004] The existing technology that combines principal component analysis (PCA) with system clustering to divide wind power output scenarios, and uses BP neural network to establish different prediction models can improve the prediction accuracy to a certain extent. In changing scenarios, the prediction accuracy still needs to be improved; using direct heuristic dynamic programming to continuously correct the model parameters under different fan output conditions can improve the prediction accuracy when the fan output changes frequently. However, the model does not have the ability to identify the fan output conditions. The ability of categories can only be used for offline correction; based on the multifractal theory, the fluctuation types are divided, and the ability of online matching prediction models is realized through the variable-scale time window algorithm. However, this method can only determine the type of wind turbine output fluctuations based on wind speed fluctuations. Considering the impact of wind direction, humidity and other meteorological factors on wind turbine output fluctuations, it does not have good generalization ability

Images