Cornell University
METHODS COMPARISON ON FLOW MODEL CONSTRUCTION AND PARAMETER ESTIMATION
Abstract
dc:description.abstractKnowing the equation of an unknown dynamical system is essential when trying to apply optimal control. Sometimes researchers do not have a comprehensive knowledge to a nonlinear system. The unknown part might be the function representing the relation between states (e.g. transfer function), or key parameters of a dynamical system (e.g. proportional constant of spring in a linear spring system). Various methods have been developed to identify the dynamics of an unknown system. In this thesis, multiple approaches include Neural Network polynomial Extraction (NN-poly), Sparse Identification of nonlinear Dynamics (SINDy) and Non-Uniform Discrete Fourier Transform (NUDFT) are compared over their ability to find the expression of unknown systems or to estimate key parameters of a dynamical system. Multiple tasks with different purposes are created to test the performances of these methods.
Degree
thesis:*- Name thesis:degree_name
- M.S., Mechanical Engineering
- Level thesis:degree_level
- Master of Science
- Discipline thesis:degree_discipline
- Mechanical Engineering
- Grantor
- Cornell University
- Year dc:date.issued
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Jing, Dongheng
- Committee member dc:contributor.committeemember
-
- Petersen, Kirstin Hagelskjaer
Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Other
-
ProQuest Submission ID: 11058
ProQuest Publication ID: 28089177 - OAI identifier oai:identifier
- oai:ecommons.cornell.edu:1813/103120