Overview
This research presents a decomposition-based framework for the joint optimization and spatial packaging of interconnected systems with physical interactions (SPI2). The framework specifically addresses three-dimensional component placement challenges, aiming to improve solution reliability and computational tractability for complex engineering applications. A primary application detailed is within the automotive sector, focusing on both initial design generation and system-level coordination for integrated components.
Research Context
The problem of spatial packaging of interconnected systems with physical interactions (SPI2) involves optimizing the placement of components that are physically connected and interact within a three-dimensional space. For automotive applications, SPI2 must facilitate the creation of initial designs, including component alignment, and support robust system-level coordination. Existing SPI2 approaches faced challenges in achieving high solution reliability and manageable computational costs, necessitating improvements in numerical robustness, convergence rates, and computational efficiency.
Approach
The proposed methodology enhances existing SPI2 approaches by focusing on several key areas. It improves the convergence rate and solution quality through advancements in numerical robustness, particularly in gradient-based optimization. Concurrently, it reduces computational load. The framework extends SPI2 capabilities by incorporating alignment functionalities, enabling the explicit representation of port-to-port alignments between various components. Furthermore, the methodology expands the applicability of SPI2 by treating component placement locations as design variables. This allows for the integration of penalty-based coordination mechanisms to ensure design feasibility. Such an approach enables seamless integration within broader system-level optimization frameworks.
Validation of the approach was performed using a multi-objective optimization framework. Specifically, the Nondominated Sorting Genetic Algorithm II (NSGA-II) was employed. This validation was executed within a combined powertrain optimization and battery chassis integration problem, chosen as a representative automotive use case. This allowed for the assessment of SPI2's effectiveness within a system-level design context.
Findings
The research demonstrates a twofold application of the proposed SPI2 framework in an automotive use case:
- Initial Design Generation: The framework effectively functions as a tool for creating initial designs for interconnected systems, including component alignment.
- System-Level Design Coordination: When integrated as part of a system-level design coordinator, the SPI2 framework achieved superior performance compared to a discretized exhaustive search. This improved performance was observed while simultaneously requiring lower computational cost.
The application to a combined powertrain optimization and battery chassis integration problem specifically indicated the effectiveness of SPI2 within a system-level design context. The methodology improved convergence rate and solution quality by enhancing numerical robustness in gradient-based optimization, concurrently reducing computational load.
Why This Matters
The demonstrated framework provides a technical approach for optimizing complex system architectures in domains such as automotive engineering. Its ability to generate initial designs and coordinate system-level interactions efficiently could support development processes characterized by interconnected physical components. The findings suggest a method for achieving design feasibility with reduced computational demand compared to existing exhaustive search methods.