Please join us for a UW Data Science Seminar featuring UW Herbold Fellow and Mechanical Engineering graduate student Chinmay Ratnaparkhe on Wednesday, May 20th from 4:30 to 5:20 p.m. PT. The seminar will be held in IEB G109.

“Seeing Through Composites: Automating Defect Detection from Ultrasonic Sensor Data”

Abstract: Modern commercial aircraft are built with over 50% composite materials by weight. Manufacturing these layered structures can trap hidden wrinkle defects, out-of-plane fiber waviness that reduces structural strength by up to 73%. Industry inspects these parts using ultrasonic scanners at multiple frequencies: low-frequency systems penetrate deep into thick structures for broad defect detection, while high-frequency systems provide the image clarity needed to characterize wrinkle geometry. But wrinkle characterization today remains largely manual. Trained experts interpret complex waveforms scan by scan, a process that is slow, subjective, and increasingly bottlenecked by a shortage of qualified inspectors.

This talk presents a data-driven approach to automating wrinkle characterization from ultrasound scans, built around Dynamic Mode Decomposition (DMD) and its variants as the primary feature extraction framework. The method leverages spectral decomposition to separate sequential sensor data into physically meaningful patterns, enabling interpretable defect measurements with reduced reliance on manual analysis. The talk walks through the pipeline, from data collection and signal preprocessing to feature extraction and defect characterization, and discusses preliminary results on real composite samples. It concludes with a look at alternative approaches, open challenges around scaling and automation, and broader connections to sequential sensor data problems across disciplines.

Speaker Bio: Chinmay Ratnaparkhe is a Herbold Fellow and M.S. student in Mechanical Engineering: Data Science at the University of Washington, advised by Prof. Krithika Manohar and Prof. Xu Chen. His research at the Boeing Advanced Research Center (BARC) applies spectral decomposition techniques to ultrasonic inspection data for automated defect characterization in aerospace composites. He completed his B.Tech in Mechanical Engineering from the Indian Institute of Technology (IIT) Bhilai