A multi-point wind speed prediction method in wind farm based on convolutional recurrent neural network

Abstract

The invention provides a multi-point wind speed prediction method in a wind farm based on a convolutional cyclic neural network, comprising the following steps: Step 1: Collecting the operation data of the wind farm, the collected data includes the measured wind speeds at the positions of multiple wind turbines and the measured wind direction; Step 2: According to the data collected in Step 1, establish a convolution module of a multi-point wind speed prediction model in a wind farm based on a convolutional recurrent neural network; Step 3: According to Step 1, establish a convolutional recurrent neural network-based The LSTM module of the multi-point wind speed prediction model in the wind farm; step 4: connect the output of the convolution module and the LSTM module, step 5: train the neural network model with the mean absolute error (MAE) loss function index. For the power grid, the invention helps to optimize the power grid scheduling and reduce the rotating reserve capacity, ensure the safe, reliable and economical operation of the power system, and reduce the fatigue load of the unit.

Description

technical field [0001] The invention belongs to the technical field of new energy power generation, and particularly relates to a multi-point wind speed prediction method in a wind farm based on a convolutional cyclic neural network. Background technique [0002] Wind power has entered the development stage of alternative energy from supplementary energy. However, the output of wind power has random fluctuations, and large-scale wind power grid connection will pose a serious threat to the safe and stable operation of the power system. High-precision ultra-short-term wind speed prediction for wind farms, on the one hand, is helpful for the power grid to optimize grid scheduling and reduce rotating reserve capacity, and ensure the safe, reliable and economical operation of the power system. On the other hand, for wind farms, due to the influence of wake effects, it is difficult for the optimal control of a single machine in a wind farm to ensure the optimal output of the enti...

AI Technical Summary

Problems solved by technology

However, in the current various wind speed prediction methods, only the wind speed time series signal of a single point is considered, and only the historical data and real-time data of the wind speed at a single point are often needed in the prediction process, so the system of the prediction method and the prediction accuracy of the model are still to be determined. Further improve and improve

Images