Impact Localization and Severity Estimation on Composite Structure Using Fiber Bragg Grating Sensors by Least Square Support Vector Regression

Abstract

Low-velocity impact (LVI) events in carbon fiber reinforced plastic (CFRP) structures are a major concern as they result in barely visible impact damage which can reduce the strength and stiffness of impacted structure. Detection of such impact events in terms of location and severity will be a significant milestone for aerospace structural health monitoring applications. This can result in well-targeted inspection programs thereby reducing the time and cost of periodic inspections. One of the easy and efficient ways of getting the impact response of a structure is through strain measurement. LVI event monitoring using fiber Bragg grating sensors has evolved as an attractive choice in recent years along with various soft computing algorithms and advanced signal processing techniques. Machine learning techniques such as artificial neural networks and support vector machine are widely used to localize impact events. Getting information regarding the ensuing damage due to LVI event from the strain response is not straight forward as it could involve estimation of intermediate parameters like energy/force of impact, which are further related with probable damage size. This paper demonstrates least square support vector regression-based algorithm to localize impact event in terms of $X$ - and $Y$ -coordinates and its energy on a CFRP plate-like structure and its comparison with other algorithms cited in the literature.