Design Optimization Lab - Research Area

Research Area

Methodologies

Design Optimization
Reliability-Based Design Optimization (RBDO)
Robust Design Optimization (RDO)
Reliability Analysis
Reliability-Based Design for Market Systems (RBDMS)
Bayesian Optimization
Design for Market Systems (DMS)
Analytical Target Cascading (ATC)
Artificial Intelligence/Deep Learning
Artificial Neural Network (ANN)

Applications

Drone System Design
Solar Energy Optimization
Laser Weapons
Tank Armor Design
Shared Autonomous Electric Vehicle System Design
Battery Electric Vehicle System Design
Fuel Cell Electric Vehicle System Design
Battery Cell/Module Design

An experiment-informed optimization framework for drone-based surveillance system

This study proposes an experiment-informed optimization framework for a drone-based surveillance system that is energy-efficient and cost-effective. The system defines the surveillance gap as the time interval between consecutive drone visits to the same location and minimizes this gap to ensure continuous monitoring. Electric quadcopters are adopted for the system, and a battery degradation model is incorporated to account for long-term operation, ensuring the sustainability of the system over extended periods. Flight experiments are conducted to measure the impact of varying weights and cruise speeds on energy consumption. Two charging scenarios—wireless charging and battery swapping—are examined to evaluate their effect on operational time and energy consumption. The baseline optimization results show that the battery swapping scenario is more cost-effective than the wireless charging scenario, reducing the total cost by 20.2% ($11,837). Additionally, the drone-based surveillance system is demonstrated to be energy sustainable, as the electricity cost is negligible compared to other system costs. Parametric studies highlight the changes in system design under various operational conditions and confirm the energy independence of the system.

Future transportation systems: Comparison of electric robo-taxis (eRTs) and

electric unmanned aerial vehicles (eUAVs)

Electric robo-taxis (eRTs) and electric unmanned aerial vehicles (eUAVs) are emerging as promising solutions to urban mobility challenges. While eRT system optimization has been extensively explored, research on eUAV optimization for human transportation remains limited, with few comparative studies. Therefore, this study proposes a design framework for eRT and eUAV systems to assess their feasibility and utility, offering a comprehensive cost and performance comparison to address a critical gap in transportation research. The optimization problem minimizes total system cost under passenger travel time constraints. Results show that eRTs are more cost-effective due to lower infrastructure requirements, whereas eUAVs significantly reduce travel time, offering a compelling solution for fast urban transport. Although the eUAV system more frequently achieves early arrivals, greater variability in travel time is observed, emphasizing the importance of reliability management in aerial mobility services. The eUAV system requires a substantially smaller fleet than the eRT system, demonstrating superior transport capacity. Despite lower energy efficiency, eUAVs incur lower electricity costs under short travel time targets. Parametric studies further explore key design factors for eUAVs. This study provides insights into the design direction and optimal operation of eRT and eUAV systems, offering practical implications for policymakers and urban planners to optimize real-world implementation of both transportation systems.

Reliability-based design optimization (RBDO) framework for high-energy laser weapons (HELWs) using artificial neural network ensemble

This study proposes a reliability-based design optimization (RBDO) framework for laser penetration to address the reliability issues in high-energy laser weapons (HELWs) caused by system uncertainties. After quantifying the uncertainties associated with laser irradiation, Monte Carlo simulation (MCS) based on laser irradiation simulations can be performed to obtain the penetration time distribution for a given nominal laser power. To alleviate the computational burden inherent in MCS, this study employs an ensemble of artificial neural networks (ANNs) that combines multiple ANN models to enhance predictive performance. For various target penetration times, the proposed RBDO framework immediately derives the optimum nominal laser power required to satisfy the given target penetration time with target reliability, thereby increasing the likelihood of mission success in combat. The global sensitivity analysis for system uncertainties is performed using Sobol method, and the results show that the uncertainty in absorptivity has the greatest effect on penetration time.

Shared Autonomous Electric Vehicle (SAEV) System Design

Shared autonomous electric vehicles (SAEVs), which combine electric vehicles (EVs), autonomous driving technology, and car-sharing service, are expected to become essential parts of transportation systems. SAEV system design includes optimization of charging infrastructure design, vehicle design, and vehicle operation, and can result in optimal SAEV system designs that improve customer satisfaction while reducing total cost. Battery electric vehicle (BEV) and fuel cell electric vehicle (FCEV) can be considered for SAEV, and an idle vehicle relocation strategy based on deep learning is proposed for efficient and smart relocation. Additionally, a comprehensive framework for SAEV system design and optimization in accordance with the dynamic nature of battery degradation under varying load conditions is proposed.

Multi-Scale Design Optimization of Electric Vehicles

An efficient multi-scale design optimization framework for electric vehicles (EVs), including four levels from the battery cell level to the marketing level, is proposed. In the engineering domain, three levels of fundamental physics-based models are developed—an electrochemical cell, thermal module, and EV dynamics—to describe the performance of EVs. In the marketing domain, a utility model is used to maximize profit while considering customer preferences and costs. In addition, the vehicle model at the top level of the engineering domain is linked with the marketing model. Thus, all four model levels are hierarchically connected and can pass information. The multi-scale design optimization problem is solved by applying an analytical target cascading (ATC) method specialized for multi-level optimization. Through the ATC framework, optimization problems at each level can be independently solved while satisfying the targets from upper levels. Results show that each level successfully communicates with the others and carries out optimization to maximize profit while satisfying various constraints.

Reliability-Based Design Optimization (RBDO) framework for Battery Module Structure Design

A generalized optimization framework to increase the energy density of the battery module while satisfying the mechanical and thermal safety requirements is proposed. A reliability-based design optimization (RBDO) framework for module structures that considers system uncertainties to mitigate the possibility of safety failures and obtain a reliable design is also developed. The mechanical behavior of the module shows how swelling, breathing, material types, and properties affect the final module length and stress. The effects of the C-rate and design parameters on the maximum temperature and temperature deviation are also investigated. All designs satisfy the target probability of failure, verifying that the RBDO framework of the module structure has been successfully established. The proposed RBDO framework provides guidance for the design and operation of battery modules.

Reliability-Based Design for Market System (RBDMS)

In real engineering field, there exist many uncertainties which can cause a failure of a system. So, reliability-based design optimization (RBDO) is about maximizing functionality of a system while securing reliability. And in order to find the optimum reliability from the view point of market, design for market systems (DMS) is integrated with RBDO. DMS is also a kind of design optimization, but more focused on maximizing profit or social welfare from the perspective of manufacturer. For a product there are many attributes, and by estimating the consumer preferences on these attributes, we can calculate the market share and finally the profit. So, how to design a product to maximize the profit of a company is the main focus on DMS. By integrating RBDO with DMS, how reliability affects both engineering and marketing model can be figured out.

Page updated

Google Sites

Report abuse