Global Solar Radiation Estimation and Climatic Variability Analysis Using Extreme Learning Machine Based Predictive Model

摘要: Sustainable utilization of the freely available solar radiation as renewable energy source requires accurate predictive models to quantitatively evaluate future potentials. In this research, an evaluation preciseness extreme learning machine (ELM) model a fast and efficient framework for estimating global incident (G) is undertaken. Daily meteorological datasets suitable G estimation belongs northern parts Cheliff Basin in Northwest Algeria, used construct model. Cross-correlation functions are applied between inputs target variable (i.e., G) where several climatological information’s predictors surface level estimation. The most significant determined accordance with highest cross-correlations considering covariance dataset. Subsequently, seven ELM unique neuronal architectures terms their input-hidden-output neurons developed appropriate input combinations. prescribed model’s performance over testing phase evaluated against multiple linear regressions (MLR), autoregressive integrated moving average (ARIMA) well-established literature studies. This done statistical score metrics. quantitative terms, root mean square error (RMSE) absolute (MAE) dramatically lower optimal RMSE MAE = 3.28 2.32 Wm−2 compared 4.24 3.24 (MLR) 8.33 5.37 (ARIMA).