An Automated Vehicle System Architecture Exploration, Evaluation and Uncertainty-Based Design Optimization Framework

Abstract

System integrators across industry adhere to the so called V-model to manage their product development process. The V-model starts with the conceptualization of a system architecture, guided by functional and non-functional requirements and ending with the verification and validation of the designed product. In the conceptual design stage, crucial decisions impacting the performance of the final product are made. However, due to limited knowledge about the behavior of emerging technologies, unknown effects of connecting various components, and the low technology readiness levels (TRL) of components, the architectural design space for innovative design projects becomes large and uncertain. This uncertainty is traditionally managed through engineering judgement and large safety factors, which restrict design space exploration and often lead to sub-optimal solutions.

To address this, an uncertainty-based system architecture design exploration and optimization framework is developed in which system architecture design inputs and models, commonly used in model-based systems engineering, are treated as stochastic. This is integrated into a multidisciplinary multi-objective surrogate based optimization framework aimed at finding robust optimal vehicle system architectures.

The proposed framework has been successfully applied to the design of a heavy duty electric vehicle drive train. This resulted in increased system-specific knowledge of the vehicle’s performance in real-world applications by finding the sensitivity of the optimal design solution to operational- and model-uncertainties, thereby facilitating better informed design decisions, enabling trade-offs and reducing project risks.