Long-Term Forecasting Method for Power Electronics-Based System Design

Failure of power electronics can significantly impact the long-term power system reliability, especially for the systems with a high penetration level of the renewable energy-based units. Therefore, the analysis of power electronics reliability impact on the optimum system design needs to be included in the long-term planning. However, current long-term forecasting methods cannot provide generation and demand profiles suitable for the power electronics reliability evaluation typically due to a low time resolution. In this paper, a long-term forecasting method for power electronics-based system design is proposed. The method employs statistical and artificial intelligence-based models to provide forecasting profiles and their probability of occurrence with a high time resolution over the whole long-term planning horizon. The case study results show that the accuracy of the forecast profiles obtained with the developed model is suitable for power electronics reliability analysis.