Analytical sensitivity analysis methodology for the co-design of thermal management systems

Wagenmaker, Minda Joy

Permalink

https://hdl.handle.net/2142/113089 Copy

Description

Title

Analytical sensitivity analysis methodology for the co-design of thermal management systems

Author(s)

Wagenmaker, Minda Joy

Issue Date

2021-07-22

Director of Research (if dissertation) or Advisor (if thesis)

Alleyne, Andrew G

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Sensitivity Analysis
• Graph-Based Modeling

Abstract

This thesis presents an analytical sensitivity analysis methodology to be used in a co-design approach to simultaneously optimize plant parameter variables and controller gains. Sensitivity analysis methods are commonly used in system design. The sensitivity of the system output to design variables can be calculated either analytically or numerically. These sensitivities inform the engineer which variables should be the focus of the majority of time and effort to yield the optimal result or, conversely, which variables are inconsequential and can be removed from the optimization design space. Sensitivity analysis methods are also used in control system design. The controller design is influenced by the analysis of how system parameters affect the controlled states. This thesis presents a plant and controller co-design approach where the system parameters which affect the plant output and the ones which affect the controlled states the most are identified. These parameters are optimized using a brute-force numerical method at the same time as the controller gains. This improved plant and controller pair is designed to lower the control input without sacrificing performance. By designing both the plant and the controller simultaneously, better results can be achieved than by only optimizing either the plant or controller. The methodology explored in this thesis is widely applicable to many other systems. However, it has been specifically designed to work with a graph-based modelling framework. This project was carried out with the intention of improving thermal management systems for electrified vehicles. A fluid loop with additional heat loads was used as the example framework to showcase the sensitivity analysis methods. This thesis demonstrates how optimizing the most influential parameters along with the controller gains, lowers the necessary control input, or pump energy, while improving tracking of a reference signal.

Graduation Semester

2021-08

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/113089

Copyright and License Information

Copyright 2021 Minda Wagenmaker

Owning Collections