Towards Interpretable Machine Learning for Automated Damage Detection Based on Ultrasonic Guided Waves.

Christopher Schnur,Jens Prager,Jannis Bulling,Yevgeniya Lugovtsova,Jochen Moll,Payman Goodarzi,Andreas Schütze,Kilian Tschöke,Tizian Schneider

doi:10.3390/s22010406

Abstract

Data-driven analysis for damage assessment has a large potential in structural health monitoring (SHM) systems, where sensors are permanently attached to the structure, enabling continuous and frequent measurements. In this contribution, we propose a machine learning (ML) approach for automated damage detection, based on an ML toolbox for industrial condition monitoring. The toolbox combines multiple complementary algorithms for feature extraction and selection and automatically chooses the best combination of methods for the dataset at hand. Here, this toolbox is applied to a guided wave-based SHM dataset for varying temperatures and damage locations, which is freely available on the Open Guided Waves platform. A classification rate of 96.2% is achieved, demonstrating reliable and automated damage detection. Moreover, the ability of the ML model to identify a damaged structure at untrained damage locations and temperatures is demonstrated.

Highlights

Machine learning (ML) techniques require a large number of measurements for adequate training and reliable decision-making
The methodology presented in this paper focuses on interpretability, meaning that the ML results must be physically interpretable to enable the use of ML in safety-relevant applications
We describe our approach to find a robust model with a high classification rate

Summary

Introduction

Machine learning (ML) techniques require a large number of measurements for adequate training and reliable decision-making. ML is well suited for structural health monitoring (SHM) applications in which one or multiple sensors are permanently attached to the structure so that structural measurements can be recorded frequently. This rich data pool can be exploited by ML techniques to train a model that can detect damages or anomalies, allowing for fully automated damage detection. Several ML methods have been developed in the last few years to solve various. Even though ML methods are already well established in vibration-based SHM [6], their use in guided wave-based SHM is currently rising [7,8,9]. Roy et al [7]

Objectives

Results

Conclusion