Calculation of solar energy potential of Sinop province with hybrid artificial intelligence model

Today, the world's demand for energy is increasing daily due to technological advancements and industrialisation. This rise has reached a point where limited resources are insufficient to meet energy needs. This situation has further widened the gap between energy production and consumption, amplifying the necessity for renewable energy sources. Energy, the cornerstone of modern life, has become more efficient and sustainable thanks to technological innovations like artificial intelligence. Within the scope of this study, a hybrid model—an innovative approach to artificial intelligence developed by combining various algorithms—has been designed. The solar energy potential of Sinop province was calculated using the hybrid Multi-Layer Perceptron (MLP)-Long Short-Term Memory (LSTM)model. The input parameters included cloud coverage, wind speed, visibility, pressure, altitude, temperature, and humidity. The developed model was trained and tested with precise data. The hybrid model's success rate was analysed compared to single-algorithm models. The absolute percentage error (MAPE) of MLP Model-1 was determined to be a 21.965% error margin, with an R² value of 0.84123. MAPE of MLP Model-2 was established at a 23.5132% margin of error, with an R² value of 0.83325. MAPE of the LSTM model was calculated at a 29.2656% margin of error, with an R² value of 0.77333. The most accurate and reliable results were obtained from the developed hybrid model, which yielded a MAPE value of 8.4888% error margin, with an R² value of 0.96935. This study has demonstrated that the hybrid model outperforms existing methods. The results obtained indicate that the model shows promise in estimating solar energy potential. The developed hybrid model may be employed to assess solar energy potential across various provinces.