Classification of ultrasonic echo signals to detect embedded defects in carbon fibre reinforced plastic laminates

Poudel, Anish; Kanneganti, Raghuveer; Li, Shanglei; Gupta, Lalit; Chu, Tsuchin Philip

doi:10.1504/ijmmp.2015.072917

Cited by 3 publications

(1 citation statement)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In another work, Poudel et al [ 19 ] use ultrasound to identify foreign objects, impact damage, and porosity. An artificial neural network is used to analyze a-scans using a nearest mean technique.…”

Section: Introductionmentioning

confidence: 99%

Improvement in the Quantification of Foreign Object Defects in Carbon Fiber Laminates Using Immersion Pulse-Echo Ultrasound

2021

View full text Add to dashboard Cite

This research presents a new technique using pulse echo ultrasound for sizing foreign objects within carbon fiber laminates. Carbon fiber laminates are becoming increasingly popular in a wide variety of industries for their desirable properties. It is not uncommon for manufacturing defects to occur within a carbon fiber laminates, causing waste, either in the discarding of failed parts or the overdesign of the initial part to account for these anticipated and undetected errors. One such manufacturing defect is the occurrence of a foreign object within the laminate. This defect will lead to a localized weakness within the laminate including, but not limited to, stress risers, delamination, and catastrophic failure. This paper presents a method to analyze high-resolution c-scan full waveform captured data to automatically capture the geometry of the foreign object with minimal user inputs without a-priori knowledge of the shape of the defect. This paper analyzes twelve samples, each a twelve-lamina carbon fiber laminate. Foreign objects are made from polytetrafluoroethylene (PTFE) measuring 0.05 mm (0.002 in.) thick and ranging in diameter from 12.7 mm (0.5 in) to 1.588 mm (0.0625 in), are placed within the laminates during fabrication at varying depths. The samples are analyzed with a custom high-resolution c-scan system and smoothing, and edge detection methods are applied to the collected c-scan data. Results are presented on the sizing of the foreign objects with an average error of 6% of the true area, and an average absolute difference in the estimation of the diameter of 0.1 mm (0.004 in), an improvement over recently presented ultrasonic methods by a factor of three.

show abstract

Section: Introductionmentioning

confidence: 99%