Search

Power module heat dissipation with spatial inhomogeneity and induced nonuniform temperature distribution presents a challenging concern for system reliability due to thermo-mechanical stresses. These reliability challenges are especially important for nonplanar designs using 3-D packaging principles. Here, we develop a heat spreader thermal switch capable of actively reducing temperature gradients between silicon carbide (SiC) devices in a three-level T-type power converter, which can change depending on electronic operating conditions. The heat spreader thermal switch consists of a stainless-steel (SS) heat spreader and a copper sliding switch embedded within the spreader. Heat transfer from the SiC devices and within the spreader can be controlled by moving the slider between positions within the spreader. To understand heat transfer mechanisms and design the heat spreader thermal switch, we conducted 3-D finite element method simulations to calculate the minimum attainable temperature difference between SiC devices. We used the finite element simulation to quantify the reduction in junction temperature swing during dynamic operation and coupled the results to the Coffin–Manson reliability model to quantify lifetime to failure. We integrated the heat spreader thermal switch with one phase of a three-phase T-type converter and demonstrated isothermalization at different working conditions. At 2.4-kW converter power, each hot SiC device dissipated 4.7 W of heat, resulting in a device case temperature of 43 °C, with each cold SiC device dissipating 1.7 W at 38 °C. The device-to-device temperature difference was decreased from 5 °C to 0 °C by moving the switch a distance of 20 mm. Finite volume method (FVM) simulations of the conjugate heat transfer problem validate the experimental results and support analysis of key performance parameters. This work demonstrates successful isothermalization of a power converter with an active heat spreader thermal switch and develops simulation and design guidelines for successful electro-thermal codesign and implementation of thermal switches for other electronics applications for which isothermalization and enhanced device reliability is a key issue. [ABSTRACT FROM AUTHOR]

As a critical energy-conversion system component, power semiconductor modules and their layout optimization has been identified as a crucial step in achieving the maximum performance and density for wide bandgap technologies (i.e., GaN and SiC). New packaging technologies are also introduced to produce reliable and efficient multichip power module (MCPM) designs to push the current limits. The complexity of the MCPM layout is surpassing the capability of a manual, iterative design process to produce an optimum design with agile development requirements. An electronic design automation tool called PowerSynth has been introduced with on-going research toward enhanced capabilities to speed up the optimized MCPM layout design process. As a part of this continuing research, in PowerSynth v1.9, a constraint-aware layout engine has been developed, which enables integrating heterogeneous components, handling complex geometry, exploring a larger solution space, improved success rate, and providing options for multiobjective optimization algorithms. The layout engine is generic, scalable, and efficient in performing electro-thermal optimizations on both 2-D and 2.5-D power modules. To validate these enhanced design capabilities, a 2.5-D full-bridge power module layout is designed, optimized, fabricated, and tested with measurement results matching closely with model prediction. This result closes the loop in the power electronics design process with an experimentally validated module design automation flow. [ABSTRACT FROM AUTHOR]

This paper presents a practical strategy for utilizing the submodule (SM) redundancy of a modular multilevel converter (MMC) for its fault tolerance. This strategy provides a systematic framework for balancing the tradeoff between two conventional methods for using the active redundancy and, thus, achieves operational flexibility. One of the existing methods improves SM reliability owing to less voltage stress on the SM components by employing all of the SMs to form the ac or dc voltages (voltage-sharing mode). The other avoids transients by keeping the average SM voltage constant at the cost of slightly increased stress on the SM components (fixed-level mode), which, however, can be controlled to provide the grid-adaptive operation by reserving the energy of the SMs not in service. We, thus, develop a new redundancy management scheme by integrating these two methods and exploiting their technical benefits to meet the PQ requirements and MMC control performance. This research provides a theoretical basis and a technical guide to determining the number of SMs, which can further increase the voltage steps as per the MMC and grid conditions. This paper also connects the remaining PQ capability of the MMC at a particular operating point with the SM redundancy concept by defining a potential redundancy, especially useful when the physical redundancy is exhausted. The theoretical findings and efficacy of the proposed strategy are validated through PSCAD/EMTDC time-domain simulations followed by experiments using a nine-level single-phase MMC system. [ABSTRACT FROM AUTHOR]