System Level Modelling, Evaluation, and Trade-Off/Optimization of Solid-State & Hybrid DC Circuit Breakers for an EV Eco-System Using AI/ML in an MBSE Framework

Abstract

<div class="section abstract"><div class="htmlview paragraph">With the increasing demand for efficient & clean transport solutions, applications such as road transport vehicles, aerospace and marine are seeing a rise in electrification at a significant rate. Irrespective of industries, the main source of power that enables electrification in mobility applications like electric vehicles (EV), electric ships and electrical vertical take-off & landing (e-VTOL) is primarily a battery making it fundamentally a DC system. Fast charging solutions for EVs & e-VTOLs are also found to be DC in nature because of several advantages like ease of integration, higher efficiency, etc. Likewise, the key drivers of the electric grid are resulting in an energy transition towards renewable sources, that are also essentially DC in nature. Overall, these different business trends with their drivers appear to be converging towards DC power systems, making it pertinent. However, DC circuit protection poses serious challenges compared to AC due to the absence of natural zero-crossing points and the high rise rate of fault current demanding the need for fast-acting protective devices. Circuit breakers as protective devices are advantageous because of their resettable feature. This has prompted engineers to look beyond their conventional purview and explore developments in mechanical technologies of fast-acting switches & power electronics technologies to develop circuit protection solutions for DC systems. The result has been the development of solid-state circuit breakers (SSCB) and hybrid circuit breakers (HCB), making their design and analysis crucial.</div><div class="htmlview paragraph">Thus, this paper focuses primarily on the model-based analysis of SSCB & HCB used for DC circuit protection in mobility applications. System-level models of the breakers are developed and integrated with a DC power distribution system. Key critical to quality (CTQs) of the design like response time, power loss, temperature rise, rate of rise of fault current, current commutation, and drop in DC bus voltage are predicted. Also, sensitivity to design parameters is analyzed. As there are a variety of architectures for these breakers, a Model-Based System Engineering (MBSE) approach is adopted to have an ideal framework for tracking and evaluating different architectures. Further, a Machine Learning (ML) based Reduced-Order Model (ROM) is developed, embedded in the MBSE framework for quick model-based sensitivity and optimization studies for faster product-to-market without compromising on accuracy.</div></div>