Document Type
Thesis
Date of Award
5-31-1990
Degree Name
Master of Science in Mechanical Engineering - (M.S.)
Department
Mechanical and Industrial Engineering
First Advisor
Youngjin Park
Second Advisor
E. S. Geskin
Third Advisor
Harry Herman
Abstract
Based on the model reduction technique called SEREP (System Equivalent Reduction Expansion Process), systematic ways to select the master degrees of freedom (the degrees of freedom kept in the reduced model) are developed and their usefulness is demonstrated by the examples. The following two criteria are ' developed for two specific objectives.
When we need to identify the response of the full model from the response of the reduced model (usually obtained from the experiments,) we hope to reduce the estimation error between the estimated response and the actual one. We also hope to keep the points which will have large response in the reduced model because larger response is easier to measure. A scalar value, named singular value group (SG), obtained from the singular value decomposition of the reduced modal matrix is defined and used as a criterion to check the estimation error. Another scalar value called participation factor (PF) is also defined and used to find the points which have a large response.
The concepts of keeping large response based on participation factor (PF) and reducing estimation error based on singular value group (SG) can help save time during the reduction of a large model because we can use SV and PF to predict the dynamic properties of reduced model before we reduce it. Two analytical models, a cantilever beam with a concentrated mass at free end and a frame structure, obtained from Finite Element Method are studied to demonstrate the selection technique based on these two criteria.
Recommended Citation
Qi, Yixing, "Selection of master degrees of freedom in model reduction techniques" (1990). Theses. 2907.
https://digitalcommons.njit.edu/theses/2907
