Fault-Tolerant Technique Against Current Sensors for Model Predictive Control of Induction Motor Drives
Rajarshi Chakraborty, Sukanta Das
Abstract
This paper illustrates a Fault-Tolerant Control (FTC) algorithm for Induction Motor (IM) drives operating under Model Predictive Control (MPC) regime. In the proposed work, any type of faults against current sensors/ current sensing devices (CSDs) are detected by a two-section mathematical fault detector and isolator (MFDI). One section uses a third-difference operator (TDO) for fault detection at higher speeds, and the other section incorporates a single discrete derivative operator (SDDO) for detecting faults at lower speeds. Fault isolation is executed by a soft single-pole double throw (SPDT) switch. A current synthesizer (CS), for estimating machine stator currents, is formulated in stationary two-axis reference frame employing the fundamental modelling equations of an IM. This eliminates the use of slip speed and synchronous speed for current estimation. The action of the MFDI is independent of the primary control algorithm making the fault detection routine resilient and independent of parameter variations. The effectiveness of the proposed FTC scheme against current sensor failures is validated through extensive hardware experimentation.