[CCoE Notice] Thesis Announcement: Nikolas Austin Reuter, "Percussion-Based Single and Multi-Bolt Looseness Detection and Composite Plate Delamination Detection Using Machine Learning"

[Greenwell, Stephen J]

[Thesis Defense Announcement at the Cullen College of Engineering]
Percussion-Based Single and Multi-Bolt Looseness Detection and Composite Plate Delamination Detection Using Machine Learning

Nikolas Austin Reuter
April 29, 2025, 10 a.m. to 11 a.m. (CST)
Location: Teams Link<https://urldefense.com/v3/__https://teams.microsoft.com/l/meetup-join/19*3ameeting_Yzg5ZjU2NDEtZmFhOS00ZTYyLWE4ZjMtOTAyOThhZmM5OWI5*40thread.v2/0?context=*7b*22Tid*22*3a*22170bbabd-a2f0-4c90-ad4b-0e8f0f0c4259*22*2c*22Oid*22*3a*22d7dc0bd2-599d-4bde-bbc2-ff72c9fb1000*22*7d__;JSUlJSUlJSUlJSUlJSUl!!LkSTlj0I!Fj3eIkhhlIxqw2EGRD9cfAymgen69Qf5feDKyASctaxJvl4a8Ei1wO4Z2zP9pqZBTdft60IRqp3v1OKuyVupu1lCwg0$ >
Meeting ID: 234 784 269 480 0
Passcode: xE9g27os

Committee Chair:
Gangbing Song, Ph.D.
Committee Members:
Weihang Zhu, Ph.D. | Zheng Chen, Ph.D. | Bradley Davis, Ph.D.
Abstract
Bolted Flanges are commonly used in the energy industry. Bolt looseness is a common cause of leaks. Pipeline leakage monitoring is essential for ensuring safety and operational efficiency. This study uses a percussion-based method to apply machine learning in the detection of bolt looseness in flanged pipeline connections. A weld neck flange with eight bolts is examined under two test cases. In the first case, the objective is to identify a single loose bolt among eight, where seven bolts are torqued to 210 ft-lbs, and one remains loose. The second case investigates the ability to determine the number of loose bolts from four possible conditions: 0, 2, 4, or 8 loose bolts. A hammer is used to strike between two bolts, and the resulting audio is recorded using an iPhone and applied to machine learning algorithms. Various machine learning techniques are employed, including the shallow learning method Support Vector Machine (SVM), the deep learning method Recurrent Neural Network (RNN), and the clustering method Spectral Clustering.
With the use of machine learning becoming more common, so does the need to educate students in applying machine learning to real-world applications. Composite plates made of carbon fiber and aluminum with a sandwiched layer of closed-cell foam and epoxy are used to simulate carbon fiber delamination for the purpose of educating students in machine learning topics. A grid system is implemented, where both healthy and delaminated sections are manufactured to evaluate machine learning models for identifying delaminated regions. This study demonstrates the potential of machine learning for structural health monitoring and predictive maintenance in industrial applications.
[Engineered For What's Next]



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20250428/e438ca2d/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 28468 bytes
Desc: image001.png
Url : http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20250428/e438ca2d/attachment-0002.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 5699 bytes
Desc: image002.png
Url : http://Bug.EGR.UH.EDU/pipermail/engi-dist/attachments/20250428/e438ca2d/attachment-0003.png