FPGA Hardware Co-Simulation Implementation of Real-Time Doubly Fed Induction Machine Parameters Estimation Using Genetic Algorithms

Abstract

This contribution deals with a detail of an approach based on genetic algorithms (GAs) exclusively adapted for the implementation of on-line doubly fed induction machine parameter estimation on FPGA. Thus, knowing just the stator/rotor voltage ratio initially, the proposed solution offers a significantly reduced computational complexity without any simplifying assumptions. The resulting identification problem is linear in a similar way to the separated DC machine case. The considered identification problem algorithm is performed using the Matlab/Simulink software; this simulation model is used then to implement the algorithm using the Xilinx Tool Box. Finally, the System Generator (SysGen) tool in the Xilinx block set is used to produce the corresponding bit file, this last will be downloaded on the FPGA via hardware co-simulation. The proposed procedure illustrates that the obtained results are in perfect concordance with respect to the simulated ones with reasonable usage of the considered FPGA’s available resources.