Digital Twin & AI-based Control

SiC MOSFETs have emerged as a promising solution for high-power applications. However, their long-term reliability can be compromised when subjected to temperature fluctuations, which are common in many real-world operating environments. Rapid temperature cycling can induce degradation in these devices both at the semiconductor level (e.g. deep traps, interfacial stress, oxide failure, etc.) and the package level (e.g.  cracks, delamination, bond-wire lift-off, etc.). This result into non-ideal behaviour in the electrical (e.g. R_ON, V_TH, I_GSS, R_DSS, etc.) and thermal characteristics (e.g. R_TH) of the device. As a result, it is important to tackle the temperature-induced degradation and thus increase the lifetime of the SiC MOSFETs.

This work primarily focuses on implementing an efficient thermal management system based on a predictive temperature controller and smoothing algorithm to reduce the magnitude of temperature fluctuations and thus extend the SiC MOSFET module lifespan to be used in a 3-phase inverter (B6 bridge topology).

Tasks and Goals:

• Familiarization and state-of-the-art literature research for
  • different temperature sensitive electrical parameters (TSEPs) for SiC MOSFETs.
  • temperature control systems
• Determination of the setpoint MOSFET junction temperature (T_J,SP) from a given predicted temperature fluctuation profile, based on different target variables, such as range, service life, robustness and energy efficiency.
• MATLAB/Simulink-based implementation of the temperature controller, which translates the T_J,SP value into the appropriate SiC MOSFET TSEPs and other inverter parameters (e.g. dead-time, switching frequency, etc.) and thus regulate the actual junction temperature of the MOSFET (T_J,actual) .
• Experimental verification and evaluation of the temperature and smoothing control.
• Written thesis and presentation.

Expected Qualifications:

• Experience of MATLAB and Simulink.
• Knowledge of B6-bridge inverter topology.

Optional (Preferable) Qualifications:

• Attended the Robust Power Semiconductor Systems I & II courses offered by our institute.
• Experience with dSPACE equipments.

Start: Immediately

Contact: Swapnil Sunil Roge

Background:

• Various sensors installed on an electric vehicle record different road condition parameters as shown in Fig. 1. The AI/ML model predicts the velocity and acceleration of the car (t_o + N) seconds into the future, where t_o represents the current time and N represents the time-span of prediction. This data is commonly known as the future mission profile.

Tasks:

• •  Development of Car Model:
    • Development of a model which takes the predicted future mission profile as the input and then evaluates in real time the mechanical power loss P_mech(t₀ + N) of an electric car.
    • Software to be used: MATLAB and Simulink.

• •  Development of Electric Drive Model:
    •  Development of a model which simulates the drive train of an electric vehicle. This model should translate the mechanical power loss obtained in the previous step to the 3-phase load current I_UVW(t₀+N) consumed by a 3-phase electric motor within the car in real time.
    • Software to be used: MATLAB and Simulink.

Skills:

• • Knowledge of MATLAB and Simulink.
  • Knowledge of vehicle dynamics to evaluate P_mech(t₀ + N).
  • Knowledge of 3-phase electric motors to evaluate I_UVW(t₀ + N).

HiWi Contract Information: 

• • Working Hours: 40 hours per month.
  • Initial contract duration will be of 1 month, which can be extended further. 

Contact:

Swapnil Sunil Roge