Design and simulation of a single-phase shepwm inverter-fed induction motor drive using generalized hopfield neural network

Abstract

This paper discusses the design of a single-phase (1 − φ) inverter-fed induction motor drive which uses GHNN controlled selective harmonic elimination pulse width modulation (SHEPWM) technique. SHEPWM is a popular switching technique which reduces the lower-order harmonics by proper selection of the switching angles. The initial switching angles for modulation index (M) varying from 0.1 to 0.9 with an interval of 0.1 are derived by solving five nonlinear algebraic transcendental equations using genetic algorithm technique. In this paper, a generalized Hopfield neural network (GHNN) is used for reduction of the fifth-, seventh-, eleventh-, and thirteenth-order harmonics in the output voltage of the inverter-fed 1 − φ induction motor for any modulation index which changes continuously. A set of five ordinary differential equations (ODEs) describing the behavior of GHNN is obtained from the energy function which is formulated using the solution set of nonlinear algebraic transcendental equations. In some application, the modulation index (M) changes continuously due to various factors and the switching angles to eliminate the low-order harmonics are determined by solving the above-obtained ODEs using Runge–Kutta fourth-order method. The performance of the proposed inverter drive was simulated for both 1 − φ capacitor start and 1 − φ capacitor start and run induction motors (IM), for varying modulation index, and with the sudden change in the load torque, with and without a filter are presented. The proposed SHEPWM using GHNN for induction motor drive has been simulated using MATLAB/Simulink with satisfactory results.