Structural Damage Detection in Steel Frame Using ANFIS

Abstract

Damage to various structural elements of a structure will cause the structure to deteriorate and ultimately cause its complete failure. It is important to detect such damages in the initial stage. There are many methods that have proved to be effective in detecting the damages using natural frequencies and mode shapes. Artificial Intelligence technique is an efficient method for detecting structural damages. In this study, an attempt was made to find the best ANFIS model to detect the single and multiple damages caused due to the reduction in modulus of elasticity within the frame. Various membership functions are used to predict the accurate ANFIS model. The Gaussian membership function has proved to be the most accurate ANFIS model which has generated almost similar to the same natural frequencies with coefficient of determination (R²) close to 1. This allows ANFIS to be used effectively to predict damage in the framed structure.