<![CDATA[Dual-axis sun trackers improve photovoltaic energy harvesting but require relatively high operational energy for continuous actuator movement, which reduces the net energy gain and overall system efficiency. This study proposes an Adaptive Neuro-Fuzzy Inference System (ANFIS)-based model to predict the proportional energy generated by a dual-axis sun-tracker controller. The proposed approach consists of evaluating the performance of the dual-axis sun tracker under different threshold settings and developing an ANFIS prediction model. The system performance was analyzed by limiting servo motor movement based on differences in LDR sensor readings using threshold values of 50, 100, and 150. Subsequently, 24 ANFIS membership function configurations were systematically evaluated to identify the optimal model structure. Experimental results show that a threshold value of 100 provides the best proportional energy efficiency. Among the evaluated configurations, the Generalized Bell (gbellmf) membership function with a 9×9 structure achieved the lowest training error of 0.0116. Validation using an independent testing dataset yielded an RMSE of 0.1128 and an MSE of 0.0127, confirming that the selected gbellmf (9×9) configuration provides the best prediction performance among the evaluated ANFIS configurations. Furthermore, implementing the ANFIS-based control strategy increased the average proportional energy efficiency from 90% to 93%, representing an improvement of approximately 3%. These results demonstrate that the proposed ANFIS model effectively predicts proportional energy, while improving the proportional energy efficiency of the dual-axis sun tracker.]]>
