Neural network based robust optimal energy control of pulse width modulation-inverter fed motor driving pump

Abstract

This paper revisits loss model control (LMC) of the 3-phase induction motor (IM) and presents a robust LMC algorithm for medium-sized pump drives. Compared with other power loss reduction algorithms for IM, the presented one has the advantages of fast and smooth flux adaptation, high accuracy, and versatile implementation. An improved loss-model for IM drive has been developed. The model considers the surplus power loss caused by inverter voltage harmonics and magnetic saturation using closed-form equations. Further, the resistance-temperature change is considered by a first-order thermal model. To determine the optimal flux level that achieves maximum drive efficiency, an artificial neural network (ANN) controller is synthesised and trained offline. The voltage and speed control loops are connecting via the stator frequency to avoid the possibility of excessive magnetization. Beside, the obtained thermal information enhances motor protection and control. These together have the potential of making the proposed algorithm robust and reliable. The system reliability is investigated and assessed in terms of energy saving using ramp start/stop. Theoretical analysis, computer simulations, and experimental studies are performed on 5.5 kW variable speed water pump using the proposed control. The test results are provided and discussed to validate the effectiveness.