MULTILAYER MICRO STRIP PATCH ANTENNA FOR BROADBAND APPLICATIONS

The performance prediction of solar thermal systems plays a crucial role in optimizing renewable energy utilization.This study presents a soft computing-based approach for predicting the thermal efficiency of a Flat Plate Solar Air Heater (SAH) using Artificial Neural Networks (ANNs).Four distinct training algorithms were employed to evaluate and compare the predictive capability of the ANN model: the default Neural Model, Levenberg-Marquardt (TrainLM), Scaled Conjugate Gradient (TrainSCG), and Conjugate Gradient with Powell-Beale Restarts (TrainCGP).Experimental data collected under varying environmental and operating conditions-including solar radiation, ambient temperature, inlet air temperature, and air mass flow rateserved as the input features, with thermal efficiency as the target output.Each model was trained and tested using a standardized dataset split and evaluated using key statistical metrics: Mean Squared Error (MSE), Correlation Coefficient (R), and Coefficient of Determination (R).The comparative analysis of these models highlights the effectiveness of soft computing techniques in providing accurate, data-driven predictions without the need for complex physical modeling.Among the evaluated models, the training algorithm achieving the lowest MSE and highest R and R values was identified as the most efficient for thermal performance forecasting of the SAH.The results demonstrate the potential of neural network-based soft computing approaches as reliable tools in the domain of solar energy system analysis and optimization.