Short-term photovoltaic output probability prediction method based on simplest gated neural network

Abstract

The invention relates to a short-term photovoltaic output probability prediction method based on a simplest gated neural network, and the method comprises the following steps: 1), carrying out the normalization of original data containing a plurality of to-be-selected weather variables, and carrying out the reduction of the dimension of the original data through employing a maximum information coefficient MIC; 2) dividing the reduced feature data set into a training data set and a test data set, and respectively dividing the training data set and the test data set into four weather type data of sunny days, cloudy days, cloudy days and rainy days by adopting a K-means algorithm; 3) constructing a neural network quantile regression model and performing training by adopting the training dataset; and 4) performing prediction by adopting the trained neural network quantile regression model to obtain quantiles under various conditions, and obtaining an approximately complete probability density function through kernel density estimation. Compared with the prior art, the invention has the advantages that the prediction reliability and precision are improved, the prediction interval is narrower, the coverage rate is higher, and the method is simple and rapid.

Description

technical field [0001] The invention relates to the field of photovoltaic power generation big data processing, in particular to a short-term photovoltaic output probability prediction method based on maximum information coefficient feature selection and the simplest gating neural network. Background technique [0002] As people pay more and more attention to energy shortage and environmental problems, photovoltaic power generation, as one of the most potential utilization technologies of solar energy, has been rapidly developed. Power generation has the characteristics of randomness, intermittency and volatility. With the increasing proportion of photovoltaic power generation installed capacity in the power system, it brings challenges to the safe and stable operation of traditional power grids. Reliable and effective prediction of photovoltaic power output power It is of great significance to optimize the grid configuration, reduce the operating cost of the grid, and ensur...

AI Technical Summary

Problems solved by technology

However, when the sequence length is too long, RNN will face long-term dependency problems. Long-term short-term memory network (LSTM) was proposed to solve this problem. Compared with RNN, the long-term and short-term memory ability of LSTM model helps it in time In the field of sequence prediction, the prediction accuracy has been significantly improved. However, there are a large number of weight and offset parameters that need to be optimized in the LSTM model, and the optimization may be slow under large data sets, especially when the span of dependencies that need to be captured is large, that is, the input When the time series length is long, it is difficult for LSTM to converge successfully

In addition, LSTM also faces the possible "overfitting" problem, that is, during training, the network's generalization ability is weak due to overemphasis on the prediction accuracy of the training set, and the prediction accuracy is reduced in actual prediction.

Images