Files
Abstract
Modern power and energy systems rely heavily on power electronics converters. The most significant component of these converters is the power semiconductor switching devices MOSFETs/IGBTs. Two critical parameters that provide real-time information about the stress and aging of these devices are the junction temperature (Tj) and On-state resistance (RDSON) of the MOSFETs. By actively controlling and modeling these parameters, it is possible to increase the converter’s useful lifetime and predict failure. Additionally, if real-time stress information is available, it is possible to balance stress among multiple converters by balancing converter operating conditions. Power semiconductor devices such as MOSFETs/IGBTs degrade through different degradation mechanisms in long-term applications. Failure of MOSFETs/IGBTs is one of the primary causes of power electronics failure. The RDSON and Tj of a MOSFET are key health-indicating parameters. This dissertation presents sensing circuit designs that enable real-time monitoring of the oRDSON and Tj of the MOSFETs/(IGBTs). The sensing circuits are integrated into a three-phase 1000 VDC to480 VAC DC-AC inverter for a complete online in-situ health monitoring system of the MOSFETs. A new on-state drain-source voltage (VDSON) sensing circuit has been used to monitor the RDSON of the high-side MOSFETs. This sensor references the drain of the high-side transistors for their VDSON measurement and allows VDSON measurement of multiple high-side transistors with respect to the same ground reference. The high-side VDSON measurement circuits combined with low-side VDSON measurement circuits have been used for a complete RDSON monitoring of all the MOSFETs in the inverter. The drain current (IDS) is captured from measurements using an off-the-shelf current sensor located at the output filter inductor. Accounting for propagation delays in the measurement circuitry, both the VDSON and IDS are sampled and converted into digital data multiple times in a switching cycle, filtered, and stored in a digital signal processor (DSP). The DSP, originally used for the inverter control, then processes the sensor data captured over one grid cycle and calculates the average RDSON of the MOSFETs of the inverter. Validations of all the sensing circuits, using theoretical analysis and hardware experiments along with the software implementation for data processing and handling are presented in the dissertation for this real-time, in-situ RDSON measurement. Furthermore, a method of mapping the MOSFET’s RDSON to the Tj for real-time accurate Tj estimation of the MOSFETs.