Mechanical Engineering MS Thesis Defense by Mr. Bolong Shi
When: Monday,
November 7, 2016
2:30 PM
-
4:30 PM
Where: Textiles Building 101E
Description: Mechanical Engineering MS THESIS DEFENSE by Mr. Bolong Shi
November 22, 2016
2:30 p.m. 4:30 p.m.
Textile Building, Room 101E
TOPIC:
Statistical GD&T Tolerance Verification by Predictive
Confidence Interval
ABSTRACT:
Flatness, surface profile and straightness are three of the most fundamental GD&T (Geometri Dimensioning and Tolerancing) tolerances for evaluating dimensional and geometrical characteristics on critical features of a part. The wildly accepted tolerance verification methods are Least Square Method (LSM) and Minimum Zone Method (MZM). However, the limitations of these methods hindered their application in some important aspects, for example, they are not able to ensure the complete conformance of GD&T tolerances even when a feature passed their verification. This is caused by the paradox of the finite and limited measurement samples and the complete conformance verification of a continuous feature in a tolerance zone. In this work, a Statistical GD&T Tolerance Verification Method (SVM), based on cosine discrete transform (DCT) regression and statistical predictive interval estimation, has been proposed and developed to verify the GD&T tolerances. The core idea of the new method is to predict the variation range of all measured/unmeasured locations with predefined confidence level, i.e. resolving the paradox with statistical inference method. Simulation experiments are designed and the algorithms are developed for validation. The results showed its efficiency and effectiveness. This method can also be extended for verification of other GD&T tolerances with minor modifications.
ADVISOR:
Dr. Wenzhen Huang, Mechanical Engineering Dept.
(whuang@umassd.edu, 508-910-6568)
COMMITTEE MEMBERS:
Dr. Vijaya B. Chalivendra, Mechanical Engineering Dept.
Dr. Jun Li, Mechanical Engineering Dept.
Open to the public.
All MNE students are encouraged to attend.
November 22, 2016
2:30 p.m. 4:30 p.m.
Textile Building, Room 101E
TOPIC:
Statistical GD&T Tolerance Verification by Predictive
Confidence Interval
ABSTRACT:
Flatness, surface profile and straightness are three of the most fundamental GD&T (Geometri Dimensioning and Tolerancing) tolerances for evaluating dimensional and geometrical characteristics on critical features of a part. The wildly accepted tolerance verification methods are Least Square Method (LSM) and Minimum Zone Method (MZM). However, the limitations of these methods hindered their application in some important aspects, for example, they are not able to ensure the complete conformance of GD&T tolerances even when a feature passed their verification. This is caused by the paradox of the finite and limited measurement samples and the complete conformance verification of a continuous feature in a tolerance zone. In this work, a Statistical GD&T Tolerance Verification Method (SVM), based on cosine discrete transform (DCT) regression and statistical predictive interval estimation, has been proposed and developed to verify the GD&T tolerances. The core idea of the new method is to predict the variation range of all measured/unmeasured locations with predefined confidence level, i.e. resolving the paradox with statistical inference method. Simulation experiments are designed and the algorithms are developed for validation. The results showed its efficiency and effectiveness. This method can also be extended for verification of other GD&T tolerances with minor modifications.
ADVISOR:
Dr. Wenzhen Huang, Mechanical Engineering Dept.
(whuang@umassd.edu, 508-910-6568)
COMMITTEE MEMBERS:
Dr. Vijaya B. Chalivendra, Mechanical Engineering Dept.
Dr. Jun Li, Mechanical Engineering Dept.
Open to the public.
All MNE students are encouraged to attend.
Topical Areas: Faculty, General Public, Staff and Administrators, Students, Students, Graduate, Students, Undergraduate, University Community, College of Engineering, Mechanical Engineering, Lectures and Seminars