Short Term Solar Power Generation Forecasting: A Novel Approach

Photovoltaics' (PV's) are widely preferred in electricity generation market in recent years. However many parameters effect solar power generation such as irradiance, temperature, humidity etc. Therefore, solar power generation forecasting is quite significant to plan and manage energy distribution. In this study, a novel methodology called Mycielski-Markov is utilized to forecast solar power generation for short term period. This novel hybrid method is developed based on two different techniques; Mycielski signal processing technique and probabilistic Markov chain. Mycielski investigates the data history and finds the recurrence of the solar radiation data. It predicts the next data due to the recurrence in a deterministic way. On the other hand, Markov produces the transition probabilities of the solar energy states and forecast new state according to these probabilities. It is obtained that, the methods in proposed hybrid hierarchy; provide a good forecasting accuracy with a 0.87 correlation of determination value.