University of Tehran Journal of Solar Energy Research 2588-3097 11 1 2026 01 01 RL-based MPPT and Active Disturbance Control using Dynamic surface and RL Induction Motor control for Photovoltaic Water Pumping System 2903 2922 106004 10.22059/jser.2026.408032.1686 EN Boubacar Kirgni Hamza Laboratory of Energy, Electronics, Electrotechnics, Automatics and Industrial Computing, Abdou Moumouni University (UAM), Niamey, Niger Noma Talibi Soumaïla Laboratory of Energy, Electronics, Electrotechnics, Automatics and Industrial Computing, Abdou Moumouni University (UAM), Niamey, Niger Attoumane Kosso Moustapha Laboratory for Research in Engineering and Environmental Technical Sciences (STIE), Mines, Industry and Geology School, Niamey, Niger Boureima Seibou Laboratory for Research in Engineering and Environmental Technical Sciences (STIE), Mines, Industry and Geology School, Niamey, Niger. Journal Article 2025 12 12 Photovoltaic water pumping systems (PVWPSs) are increasingly deployed in remote and off-grid regions where grid extension is technically infeasible or economically prohibitive. However, their performance is significantly affected by fluctuating solar irradiance, temperature variations, and the nonlinear dynamics of the electric motor and mechanical load. To address these challenges, this paper proposes an integrated control framework combining reinforcement learning (RL)–based strategy for the PV optimal operating point tracking and ADRC scheme enhanced by dynamic surface control (DSC) and RL-based induction motor (IM) speed regulation. The RL-based MPPT estimates the MPP, while the ADRC–DSC structure with RL compensation ensures smooth speed tracking while effectively rejecting load torque disturbances. The proposed control strategy is evaluated through comprehensive simulations and benchmarked against conventional P&O/IM-based PID and sliding mode control (SMC) approaches. Quantitative results demonstrate that the proposed method achieves up to 97.7% and 89% reduction in power oscillations compared to P&O/PID and SMC, respectively. Moreover, the normalized speed mean square error (NSMSE) is reduced by 13.9% and 2.1% compared with P&O/PID and SMC, respectively. Additional improvements in torque estimation accuracy, current quality, flux ripple mitigation, and water pumping efficiency confirm the robustness and effectiveness of the proposed control framework. https://jser.ut.ac.ir/article_106004_d85d4769c280ef242c4648b268a471e4.pdf