Experimental evaluation of online parameter identification method based on cluster technique for EV and HEV

In electric and hybrid electric vehicles, the torque is calculated based on a machine model. Thus, precise knowledge of the machine's electrical parameters is required for optimal field oriented control and torque accuracy. This work presents the experimental application of a novel online parameter identification method. It is based on clustering technique and uses the information of stator currents, voltages, and electrical frequency. Using this method, all electrical parameters can be simultaneously identified with low computational costs. The algorithm is tested on the test bench with a 75-kW permanent magnet synchronous machine for electric and hybrid electric vehicles made by Bosch. Experimental results confirm the efficiency of the proposed methodology.