Photovoltaic System Ensemble Prediction System
Date
2021Author
Lari, Ali JassimEgwebe, Augustine
Touati, Farid
Gonzales, Antonio S.
Khandakar, Amith Abdullah
Metadata
Show full item recordAbstract
Solar energy is the major renewable energy in Gulf area. The yearly solar irradiance is one of the highest in the world with more than 2000kWh/m2. Consequently, nations in the Gulf area have intended their energy investing on solar energy aggregation, especially Photovoltaics (PV). Photovoltaics (PV) power output is tremendously contingent on environmental situations. PV generation power prediction models are important to study the effects of unreliable environmental circumstances and approve solar power converters' optimum performance whilst meeting highest demand across numerous environmental circumstances. The ecological data which is examined and reviewed in this paper are air temperature, relative humidity, Photovoltaics (PV) surface temperature, irradiance, dust, wind speed, and output power. The model suggested in this paper adjusts and trains three prediction algorithms, including Artificial Neural Network (ANN), Multi-Variate (MV), and Support Vector Machine (SVM). The model exploits three well-known forecast algorithms and voting algorithm to determine the optimal likelihood of PV generation power. Additionally, the ensemble algorithm proves high forecast precision of the output power because of the ecological circumstances. The Mean Square Error (MSE) for the Artificial Neural network (ANN), Multivariate (MV), and Support Vector Machine (SVM) are 98, 81, and 82, respectively. In comparison, Mean Squared Error (MSE) of the voting algorithm is considerably lower which is just above 53. The anticipated PV power generation forecast algorithm establishes consistent result with respect to the environmental conditions in Qatar. This tool is likely to support in the design process of Photovoltaics (PV) plants design where energy generation is highly predictive using proposed voting algorithm.
Collections
- Electrical Engineering [2611 items ]