To address the issue of insufficient accuracy in ultra-short-term power prediction using historical data in the field of photovoltaic power generation
an ultra-short-term PV power prediction method is proposed
which integrates numerical weather prediction (NWP) similarity day and multi-head attention-long and short-term memory network (MHA-LSTM). The correlation coefficient method is used to analyze the inherent relationship between PV power and meteorological information
combing the meteorological factors that are highly correlated with PV power in the historical data
constructing the variable matrix of key meteorological information
and carrying out the segmentation process in order to carry out a more fine-grained analysis of similar days. A multi-head self-attention mechanism is introduced to automatically capture the complex functional relationships embedded in the NWP data
and the data information enhanced by the multi-head self-attention mechanism is processed by using a long- and short-term memory neural network (LSTM) to obtain more accurate ultrashort-term prediction results. Finally
a number of different weather types are selected to compare and analyze the prediction accuracy of the algorithm
and the experimental results show that the prediction method proposed in this paper achieves a significant improvement in prediction accuracy compared with the traditional algorithm
which verifies the effectiveness and accuracy of the method.
KAPLANI E, KAPLANIS S.A stochastic simulation model for reliable PV system sizing providing for solar radiation fluctuations[J]. Applied energy, 2012, 97: 970-981.
SARKAR S, AJJARAPU V.MW resource assessment model for a hybrid energy conversion system with wind and solar resources[J]. IEEE transactions on sustainable energy, 2011, 2(4): 383-391.
MANDAL P, MADHIRA S T S, HAQUE A U, et al. Forecasting power output of solar photovoltaic system using wavelet transform and artificial intelligence techniques[J]. Procedia computer science, 2012, 12: 332-337.
VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, California, USA,2017: 6000-6010.
LEZMI E, XU J L.Time series forecasting with transformer models and application to asset management[J]. SSRN electronic journal, 2023:1-44.