A novel model for predicting nonlinear response of advanced V/F drive induction motor in maximizing motor efficiency

Abstract

The squirrel-cage induction motors (SCIM) are the largest electrical energy consumption in the world. The efficiency of SCIM is very poor if traditional constant voltage control is used and their efficiency is further lowered as they operate other than rate condition. Therefore, the efficiency improvement of SCIM which consequently improve the power factor can play an important role in energy conservation, especially in the low-load operating periods.In this paper, VI_RBFNN, a new intelligent control strategy using Radial Based Function Neural Network (RBFNN) based on voltage/current (V/I) for maximum efficiency was proposed. Through this strategy, the SCIM performance can be improved both in terms of efficiency and power factor, regardless of the load torque and speed. The proposed VI_RBFNN utilizes the advanced voltage/frequency (V/F) controller to control SCIM operation. The optimum V/I expression of the motor is used in the optimization process, which is derived, based on standard per phase equivalent circuit and its value is obtained by performing the standard induction motor parameters determination tests. The VI_RBFNN which uses RBF Neural Network with six inputs and three outputs to operate as maximum efficiency control variables estimators of V/I controller are designed. The input data used to train the RBF networks are motor phase voltage (V1), current (I1), maximum voltage (Vm), synchronous speed (f), the modulation index (m), and the voltage (Vboost); while the output data are maximum voltage (Vm*), synchronous speed (f*), modulation index (m*) and boost voltage (vboost*). The neural network were trained to learn their inverse dynamics and then configured as RBFNN controller to the motor drive, based on the set of nonlinear input-output responses. The proposed approach is simple in structure and has the straightforward goal of maximizing the SCIM efficiency for a given load torque. The results show that the function of VI_RBFNN control scheme is fast and properly driving the SCIM at optimum V/I, hence the entire drive system is working at maximum efficiency. Key words: Radial based function neural network (RBFNN), i induction motor, efficiency, v/i, V/F drive.