The randomness, volatility, and intermittence of solar power generation make it difficult to achieve the desired accuracy of PV output-power prediction. Therefore, the time learning weight (TLW) proposed in this paper is used to improve the time correlation of the LSTM network. The Fusion Activation Function (FAF) is used to resolve gradient disappearance. Learning Factor Adaptation (LFA) and Momentum Resistance Weight Estimation (MRWE) are used to accelerate weight convergence and improve global search capabilities. Finally, this paper synthesizes the improvement and proposes the AHPA-LSTM model to stabilize the convergence domain. Using actual data verification, the δ MAPE indicator of the improved model is only 2.85% on a sunny day, 5.92% on a cloudy day, 7.71% on a rainy day, and only 5.8% on average. Therefore, the AHPA-LSTM model under full climate and climatic conditions has a good predictive effect which is generally applicable to the prediction of ultra-short-term PV power generation. INDEX TERMS Photovoltaic output power, ultra-short-term prediction, long short term memory (LSTM), time weight decoupling, adaptive hyperparameter adjustment.
The traditional photovoltaic (PV) forecasting method depends on sufficient historical data (PV power station historical power generation data and numerical weather prediction meteorological data), which is not suitable for a newly built PV power plant. In order to calculate the PV array irradiance and to predict the PV power, a physical prediction approach based on solar irradiance on inclined surfaces is proposed. This method selects three decomposition models and four transposition models to be combined into 12 combination forecasting models. Furthermore, solar spectral response, incidence angle, and soiling factor are taken into account in the modified model. The results show that the methods combining the Liu-Jordan transposition model have higher forecasting accuracy under the different weather types. Among them, the Erbs + Liu-Jordan model predictions are the most accurate. P mi measured power output P fi predicted power output C api capacity of analysed PV plants
High PV penetration into DC microgrids could bring serious stabilization challenges for power electronics engineers, as renewables are accessible to DC bus voltage oscillation, hence leading to degradation of power quality and even system collapse. This article addresses the stabilization issue for a DC microgrid with high PV penetration rate, in which a decentralized composite generalized predictive control strategy is designed. First, a disturbance observer is designed to deal with the lumped uncertainties of both PV intermittency and variation of constant power loads, which is subsequently integrated into the control design through feedforward compensation loops. Second, an offset-free generalized predictive controller is designed by employing the receding-horizon optimization process. In this regard, the new method could endow the DC microgrid system of two main distinguishable features: (1) large-signal stability for global system; (2) optimized transient-time control performance based on an offset-free voltage regulation objective. The effectiveness and performance improvement of the proposed methodology are verified by comparative experimental studies on a DC microgrid platform with 50% PV integration.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.