Numerical study on LDLT decomposition-based damage locating vector method for truss structures

Abstract

Real-time structural health monitoring is very important for truss structures especially those having large-spans.In recent years, many methods have been proposed for damage monitoring of truss structures. However, damagesensitivity of these methods is still required to be improved. In this work an efficient damage localizationtechnique for truss structures is proposed, which is based on the LDLT decomposition of the flexibilitydifference matrix and the Damage Locating Vectors (DLV) method. Compared with the present Stochastic DLV(SDLV) method, the proposed method is modified in two ways. First of all, the way of calculating the damagelocating vectors is modified by using LDLT decomposition instead of Singular Value Decomposition. Secondly,in order to compute the flexibility, the mass matrix which is obtained from the finite element model is used tomass-normalize mode shapes identified from ambient excitations. As a result, the proposed LDLT-DLV methodhas a higher sensitivity to damage for different types of truss members. The effectiveness of the proposed LDLTDLVmethod is validated with the numerical example of a laboratory-scale Bailey truss bridge.