Location

Brunswick, ME

Start Date

1-1-1997 12:00 AM

Description

Pulsed eddy current and swept-frequency techniques have both been previously applied to thickness measurement and conductivity profiling of layered metallic samples. Methods of inverting measured signals for thickness measurements and conductivity profiling, in both cases, typically involve iterating with a forward-model until the measured signal value is reached. These inversion methods are extremely time consuming due to the iterative approach. More recently, fast, feature-based methods for conductivity and thickness estimation have been reported. However, these methods involve manually constructing a comprehensive look-up table.

Book Title

Review of Progress in Quantitative Nondestructive Evaluation

Volume

16A

Chapter

Chapter 3: Signal Processing and Image Analysis

Section

Signal Processing

Pages

781-788

DOI

10.1007/978-1-4615-5947-4_102

Language

en

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Neural Network Inversion for Thickness Measurements and Conductivity Profiling

Brunswick, ME

Pulsed eddy current and swept-frequency techniques have both been previously applied to thickness measurement and conductivity profiling of layered metallic samples. Methods of inverting measured signals for thickness measurements and conductivity profiling, in both cases, typically involve iterating with a forward-model until the measured signal value is reached. These inversion methods are extremely time consuming due to the iterative approach. More recently, fast, feature-based methods for conductivity and thickness estimation have been reported. However, these methods involve manually constructing a comprehensive look-up table.