Field-Oriented Sensorless Sliding Mode Current Control of Single-Stage PV-Grid Interactive Induction Motor Drive for Water Pumping

Traditional proportional-integral (PI) controllers used in field-oriented control (FOC) suffer from poor dynamic response, parameter sensitivity, and inadequate performance during solar irradiance variations. To overcome these challenges, this paper presents a unified control framework for a single-stage photovoltaic (PV) grid-interactive induction motor drive (IMD) intended for water pumping applications. Unlike conventional two-stage architectures that employ separate DC–DC converters, the proposed system integrates maximum power point tracking (MPPT) and motor control within a single three-level neutral-point-clamped (NPC) inverter operating under inherently variable dc-link voltage conditions. To achieve robust performance, a tanh-based sliding mode current controller is combined with a recurrent neural network (RNN)–based back-EMF model reference adaptive system (MRAS) for sensorless speed estimation. The proposed control strategy mitigates chattering, enhances disturbance rejection, and maintains stable operation under irradiance fluctuations and motor parameter variations. A second-order generalized integrator enables seamless grid synchronization and bidirectional power flow without requiring an intermediate dc–dc stage. Comprehensive simulation studies validate robust performance during irradiance transitions. Full-scale experimental validation confirms excellent power quality characteristics with grid current total harmonic distortion (THD) of 2.39% under grid-only operation and 1.98% under PV-only operation, while stator current THD remains below 3.97%.