Adaptive Parameter Measurement in Sensorless Speed Control of Induction Motor for Vehicle Loads

A self-constructing intelligence training-based adaptive neuro-fuzzy inference system (ANFIS) is employed to measure parameters of adaptive sensorless indirect field orientation control induction motor drive (IFOC-IMD). To intensify induction motor drive (IMD) functional performance indices, the fuzzy-neural hybridization control strategy of evolving-Takagi-Sugeno–Kang (e-TSK) is employed to measure instantaneous parameters. To avoid slow convergence, local minima, and provide a derivative-free solution, a nature-inspired metaheuristic biogeography-based optimization (BBO) is integrated into eTSK-ANFIS training. It plays an important role in designing the optimal rules and appropriate selection of membership functions (MFs). The statistical performance measurement metrics are adopted to evaluate model performance during training and testing. An objective function mean square error (mse) of a predictive model is reported as 0.004618 in training and 0.004885 in the testing phase. This will ensure the ANFIS estimator drastically enhances the measurement of parameters and consistently mimics reference speed. The BBO algorithm is also implemented to obtain the optimized fractional-order proportional integral (FOPI) gains of speed and current controllers for the generation of reference signals with minimized error. It provides the lower value of maximum peak overshot and settling time. The Performance results confirm that the proposed measurement-based adaptive control strategy ensures satisfactory robustness with accurate measurement and stable operation of the system under dynamics.