An Attention LSTM-based photovoltaic power generation power prediction model and a construction method thereof

Abstract

The invention belongs to the technical field of prediction or optimization. The invention discloses an Attention LSTM-based photovoltaic power generation power prediction model and a construction method thereof, the structure of the prediction model is composed of an LSTM, an attention mechanism and a full connection layer, and attention vectors are converted into a data format capable of being processed by the full connection layer through expansion and fusion operations. According to the invention, an Attention LSTM model based on an attention mechanism and a long and short term memory network is realized; feature extraction is performed on the input time sequence by the LSTM to realize information abstraction from the time sequence to high-level features; the attention mechanism automatically pays attention to the output vector of the LSTM hidden layer, and a larger weight is given to the characteristic quantity remarkably related to the current output quantity; the output vector ofthe attention mechanism is processed into a one-dimensional vector to be input into a full connection layer through unfolding and fusion operation, and the full connection layer directly outputs a photovoltaic power generation power prediction value at the next moment. The method is quick in modeling, high in transportability and high in prediction precision.

Description

technical field [0001] The invention belongs to the technical field of prediction or optimization, and in particular relates to an Attention LSTM-based photovoltaic power prediction model and a construction method thereof. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: [0003] Photovoltaic power generation is affected by factors such as solar irradiance, atmospheric temperature, relative humidity, and ventilation conditions, and has the characteristics of randomness, volatility, and intermittency. , The user's electricity safety poses a serious threat. Photovoltaic power forecasting is one of the key technologies to eliminate threats and maintain grid security and stability. It can be used for grid scheduling, smooth control, and fault detection. It can balance the supply and demand of electric energy, improve the economics of grid operation, and promote the intelligentization of photovoltaic power stations...

AI Technical Summary

Problems solved by technology

[0023] To sum up, the problems existing in the existing technology are: deep learning is an emerging hot technology in the field of artificial intelligence, which has broken through the technical bottlenecks of many industries. At present, traditional machine learning has been relatively mature in the field of photovoltaic power generation prediction, while The application of deep learning is still in its infancy. Long short-term memory network LSTM is a classic neural network model in deep learning, and it is rarely used in the field of photovoltaic power generation prediction.

Attention mechanism Attentionmechanism has been widely used in the fields of image processing, video analysis and natural language processing, but less in the field of photovoltaic power generation prediction

Images