EPE Association: EPE 2022 - LS7d: Health Monitoring of Power Converters

EPE 2022 - LS7d: Health Monitoring of Power Converters
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 2022 ECCE Europe - Conference > EPE 2022 - Topic 04: Electrical Machines and Drive Systems > EPE 2022 - LS7d: Health Monitoring of Power Converters

   [return to parent folder]

   Climatically Induced Insulation Degradation in Power Semiconductor Modules of Wind Turbines
By Timo LICHTENSTEIN [View]
[Download]

Abstract: Power converters are among the most frequently failing subsystems in wind turbines. Humidity-induced degradation plays a key role, especially in locations with geographically high absolute humidity. In a laboratory experiment, a failure of a power module induced by humidity and condensation could be replicated under extreme climatic conditions.

   Cognitive Insights into Metaheuristic Digital Twin based Health Monitoring of DC-DC Converters
By Abdul Basit MIRZA [View]
[Download]

Abstract: This paper extends the concept of Digital Twin (DT) to component level health monitoring in DC-DC converters through metaheuristic optimization methods. The efficacy of the proposed DT-based health monitoring is demonstrated on a higher order two-phase interleaved boost converter with coupled inductor with more parameters to estimate. The performance of Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) is compared on HIL and hardware prototype. According to the results, GA outperforms PSO with better accuracy sup. 95 \% .

   Real-Time Thermal Characterization of Power Semiconductors using a PSO-based Digital Twin Approach
By Johannes KUPRAT [View]
[Download]

Abstract: Thermal impedance is essential for assessing the state-of-health of power semiconductors and to use thermal observers. This work proposes a Particle-Swarm-Optimization-based Digital Twin approach to extract the thermal impedance for online monitoring. A proof of concept of the approach is achieved in a real-time simulation with a digital reference model by showing the convergence to the given parameter set. Further, the convergence of the algorithm to a fixed parameter set is validated in the laboratory.

EPE 2022 - LS7d: Health Monitoring of Power Converters
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 2022 ECCE Europe - Conference > EPE 2022 - Topic 04: Electrical Machines and Drive Systems > EPE 2022 - LS7d: Health Monitoring of Power Converters
	[return to parent folder]

	Climatically Induced Insulation Degradation in Power Semiconductor Modules of Wind Turbines By Timo LICHTENSTEIN	[View] [Download]
Abstract: Power converters are among the most frequently failing subsystems in wind turbines. Humidity-induced degradation plays a key role, especially in locations with geographically high absolute humidity. In a laboratory experiment, a failure of a power module induced by humidity and condensation could be replicated under extreme climatic conditions.

	Cognitive Insights into Metaheuristic Digital Twin based Health Monitoring of DC-DC Converters By Abdul Basit MIRZA	[View] [Download]
Abstract: This paper extends the concept of Digital Twin (DT) to component level health monitoring in DC-DC converters through metaheuristic optimization methods. The efficacy of the proposed DT-based health monitoring is demonstrated on a higher order two-phase interleaved boost converter with coupled inductor with more parameters to estimate. The performance of Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) is compared on HIL and hardware prototype. According to the results, GA outperforms PSO with better accuracy sup. 95 \% .

	Real-Time Thermal Characterization of Power Semiconductors using a PSO-based Digital Twin Approach By Johannes KUPRAT	[View] [Download]
Abstract: Thermal impedance is essential for assessing the state-of-health of power semiconductors and to use thermal observers. This work proposes a Particle-Swarm-Optimization-based Digital Twin approach to extract the thermal impedance for online monitoring. A proof of concept of the approach is achieved in a real-time simulation with a digital reference model by showing the convergence to the given parameter set. Further, the convergence of the algorithm to a fixed parameter set is validated in the laboratory.