ABSTRACT
This paper proposes an analytical and dynamic electro-thermal model of a three-phase inverter based on wide band gap semiconductors (SiC) for application in electric vehicles (EV). The optimal design approach of the passive filters is presented for LCL filters at DC side and LC filters at AC side of the inverter. The analytical modelling is performed, which includes mathematical representation of the inverter’s power losses based on the specifications of the datasheet of the power modules. The dynamic electro-thermal simulation model of the inverter with integrated power loss model is developed. The designed state-space system thermal model is incorporated and can be used as a virtual sensor, for the conversion of power losses to temperature. With this inverter model, it’s possible to simulate designs for achieving high efficiency in a specific operating range of temperature, with high specific power and reduced cost and size. The proposed system is simulated in MATLAB/Simulink software.
INTRODUCTION
In the current era, battery-powered electric vehicle technology is very popular and gaining importance. Researchers are putting their full effort to produce an efficient and reliable powertrain system. From the design point of view, it is necessary to consider the following factors in power electronics inverters: distortion-free operation and electromagnetic compatibility, high power factor and high reliability. Semiconductor devices are a major part of power electronic converters. Presently, wide band gap (WBG) semiconductors have superior material properties and they are more efficient in high power EV applications as compared to traditional silicon (Si) technology. Currently, policy makers and other stakeholders are further increasing efforts to accelerate the market share of electric vehicles. To help achieve this goal, power electronics researchers are seeking to advance EV power electronics systems by increasing their power density, and reduction of system cost, size and volume amongst others. The DC/AC inverter is an important part of the EV, as it drives the electric motor of the car by controlling the battery power . The losses and thermal modelling of a converter is an important aspect for the development of converters with performant features such as high reliability and efficiency, cost-effectiveness and high power density. In literature, many electro thermal modelling strategies are reported. The modelling approach in software like LTspice or Saber, but the major issue regarding this approach is large computational time. The switching energies are estimated with the help of turn-on, turn-off, and rise and fall times for both Ids and Vds (MOSFET (drain-source) current and voltage), by dividing the electrical waveform period of the inverter’s half bridges into very small subintervals, and then calculating the transition time for switching losses. The electrothermal behaviour of the MOSFETs is determined using datasheet charts. A model, using Gaussian process regression (GPR) technique, is created. The kernel function is applied for analyzing the rise time and fall time of the switches, and capacitance of MOSFET varies according to the voltage-charge ratio, so it has a direct impact on the switching loss. S lookup table technique is applied in simulation to estimate the switching and on-state losses of IGBT transistors. This methodology provides a compact, automatic and accurate model of the semiconductors, for power electronic inverter analysis using standard computational time of CPUs. A thermal simulation of IGBT modules is designed to predict the responses of the system which are essential for the optimization of the system, i.e. reduce the cost of the design, increase the reliability, time to power off when overheating risks or faults occur and also to predict the maximum switching frequency, which depends on junction temperature. Power losses and thermal optimization are generally controlled by modification of electrical parameters, and power losses are estimated using the electrical loading on devices and then transformed to the thermal model for the generation of junction temperature of semiconductor devices. The reliability prediction of a power converter is necessary for pre-development, and it depends on the junction temperature, therefore, power loss and thermal dynamic simulation analysis is compulsory for the developer. Finite element method (FEM) approach is used with the combination of fast computational time simulation, and accuracy of the compact thermal model is verified by the comparison of FEM analysis and actual measurements. Passive filters are reported as essential for obtaining high-efficiency responses from the converter, and system size, weight and cost are highly impacted, hence optimal design is necessary with the variation of switching frequency. A state-space averaging method is derived, and it is an effective method for stability analysis and controller design.

Block Diagram