Dynamic modeling of induction motors in developing tool for automotive applications.

The current electrification processes involving hybrid and electric vehicles require increasingly accurate tools to evaluate electric powertrain performance and reliability. In this context, the virtualization of electric powertrain components needs to be performed. In this way, several analyses like energetic assessments can be performed without using physical components, saving costs and development times. Among the virtualization methods, an ever more used simulation approach in automotive industries consists of the Hardware in the Loop (HIL). The HIL is a real-time simulation method that allows testing one or more real components of an electric powertrain without the need for the other hardware components usually connected to it. For example, using HIL makes it possible to test a motor control unit without connecting to it a real electronic power converter or even an electrical machine. Hence, the main advantages of this simulation approach are evident. In this thesis, the HIL simulation approach is used to virtualize the electric machine of a hybrid or electric powertrain. The simulation hardware consists of the dSPACE rapid prototyping board and using the library “XSG Electric Library” provided by dSPACE. The work activity has been developed through the cooperation between Politecnico di Torino and POWERTECH ENGINEERING S.r.l., a consulting company specialized in simulation and development of conventional powertrains, and more recently, hybrid and electric ones, too. This thesis focuses on the validation and analysis of dynamic models for electrical machines using dSPACE. The considered components are the Induction Machine (IM) fed by a Voltage Source Inverter (VSI). Concerning the IM, it has been modeled using the following reference frames: phase values time-domain (abc), stationary axes (αβ0), and rotating rotor flux (dq) axes. The aim of the work was the comparison of the models, so the iron losses and the magnetic saturation were neglected. Also, to handle and figure out the models, their analysis and validation have been performed in the Matlab/Simulink environment. Thus, comparing the obtained responses with those provided by dSPACE. Finally, the VSI has been modelled considering its average behaviour, without considering the Pulse Width Modulation (PWM), and so making the models’ simulations faster to be performed. A real IM has been tested on a dedicated test rig in Politecnico di Torino’s laboratories to get a preliminary validation of the proposed models. The IM parameters have been obtained by performing the standard no-load and locked rotor tests.