This article presents a research approach to enhancing the quality of short-term power output forecasting models for photovoltaic plants using a Long Short-Term Memory (LSTM) recurrent neural network. Typically, time-related indicators are used as inputs for forecasting models of PV generators. However, this study proposes replacing the time-related inputs with clear sky solar irradiance at the specific location of the power plant. This feature represents the maximum potential solar radiation that can be received at that particular location on Earth. The Ineichen/Perez model is then employed to calculate the solar irradiance. To evaluate the effectiveness of this approach, the forecasting model incorporating this new input was trained and the results were compared with those obtained from previously published models. The results show a reduction in the Mean Absolute Percentage Error (MAPE) from 3.491% to 2.766%, indicating a 24% improvement. Additionally, the Root Mean Square Error (RMSE) decreased by approximately 0.991 MW, resulting in a 45% improvement. These results demonstrate that this approach is an effective solution for enhancing the accuracy of solar power output forecasting while reducing the number of input variables.

The paper presents the study of the 500-kV Vietnamese power system. The oscillatory response of the system is analyzed both with a modal analysis and a time domain analysis. The 500-kV system of Vietnam is modeled in details, including all the power plants with the corresponding regulators. The model is developed in collaboration with the Institute of Energy Science IES-VAST of Vietnam and it is validated with the data provided by the National Load Dispatch Centre (NLDC) of Vietnam. The simulation results reveal a clear identification of potential inter-area oscillations between North and South of the Country. A worsening of the observed phenomenon should be considered as possible, especially in the perspective of the new installation of a significant amount of power from renewable sources into the Vietnamese system.

In recent years, distributed energy resources have grown rapidly in Vietnamese Power System, especially the rooftop Photovoltaic sources. The operational characteristics of these sources such as lack of real-time measuring data and controllable capability have caused significant difficulties in operation and management for power system operator at both transmission and distribution level, to maintain the stability and security of the power system. To cope with this situation, the development of an applications aimed to calculate and estimation the potential rooftop PV generation in near real-time, is crucial for operational monitoring and scheduling power systems for Vietnamese power system operator. In this study, an application is implemented for estimating and forecasting active power of rooftop PV systems based on satellite imagery data source. In addition, to increase the forecast accuracy, a feedback mechanism is proposed utilizing Linear Least Squares algorithm. The predicted outcomes indicate the superiority of the proposed procedure compared to the original forecast.

The increase in power generation from uncertain renewable energy sources combined with the aleatory nature of loads is significantly affecting electrical network management, which aims to find a constant equilibrium between generated and consumed power. These problems are even more significant in geographical islands' networks disconnected from the main grid, where loads suffer a further variability due to the seasonal tourist influx. For these reasons an accurate forecasting of power generation is an essential tool to improve network's management. In this framework, the paper proposes a direct method for forecasting photovoltaic power generation. The method uses Long Short-Term Memory (LSTM) neural networks applied to real power and meteorological data of a PV power plant installed in an Italian small island, named Ustica.