FOC of Induction Motor for Elevator Application

Abstract

The demand for high-performance elevator systems in modern infrastructure requires advanced motor control strategies that ensure precise speed regulation, smooth operation, and enhanced ride comfort. Traditional control methods such as scalar control (V/f) and direct torque control (DTC) often fail to meet the stringent performance requirements of elevator applications, particularly in terms of torque ripple reduction and dynamic response. The presented study carried out an exhaustive investigation of Field-Oriented Control (FOC) enhanced with Adaptive Neuro-Fuzzy Inference System (ANFIS) controllers for induction motor drives in elevator applications. The study developed and compared three FOC control variants: conventional PI-controlled FOC, un-optimized ANFIS-based FOC, and optimized ANFIS controllers using Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). A 1.1 kW three-phase induction motor was mathematically modeled and simulated in MATLAB/Simulink environment, with comprehensive data collection of 500,000 samples over a five-second simulation period. The ANFIS controllers were designed to replace conventional PI controllers in four critical control loops: speed control, flux control, d-axis current control, and q-axis current control. Performance evaluation was conducted using multiple metrics including Mean Square Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R²). Dynamic performance characteristics such as rise time, settling time, overshoot, and steady-state error were analyzed for both speed and torque responses. The results demonstrate that PSO and GA-optimized ANFIS controllers achieved superior performance in speed control, with PSO and GA variants reducing speed control error by up to 80% compared to un-optimized ANFIS, achieving zero steady-state error and the fastest dynamic response with rise times of 0.06 seconds and settling times of 0.15 seconds. However, the study revealed a critical finding that single-objective optimization significantly improved speed control performance at the expense of other control variables, highlighting the need for multi-objective strategies. While optimized ANFIS controllers excelled in speed regulation, they exhibited performance degradation in flux and current control, with some parameters showing error increases of up to 82%. Statistical analysis confirmed significant differences between optimized and un-optimized variants (p < 0.05), while PSO and GA-optimized controllers showed no statistically significant difference in performance.