An induction machine-based drive system and supervisory control for application to smart water-pumped storage and networks

Abstract

The authors present an effective induction machine (IM) drive system and intelligent supervisory control for application to smart water-pumped storage and networks. The main IM drive unit comprises reduced voltage starter, regenerative clutch/variable speed mode fluid coupling and reversible pump to allow bidirectional power flow. Under the wide umbrella of supervisory control, the following objectives are achieved: (1) Recurrent neural network (RNN) and RNN-aided Kalman filter are proposed for wired and wireless sensor fault-tolerant control of the plant based on Wald's likelihood ratio test. (2) Cheap fuzzy logic-based availability assessment scheme is presented and deployed on cloud to arrange for preventive maintenance of the starter. (3) Robust energy saving mode decision-making algorithm is developed for best energetic cost reduction. (4) Remote synchronisation method of multi-motor starting with a mechanically switched capacitor is introduced to minimise voltage dip. Further, an auxiliary inverter fed pump unit is adopted to improve system availability and revenue. The validity and superiority of the proposed drive system and control are confirmed through theoretical analysis, computer simulations, cost study, efficiency evaluation and application example to a single-pump water network during motor and generator operational modes using preliminary test results.