Stiffness assessment of bridge decks repaired by steel plates using an advanced system identification method

Abstract

This work examines the identification of stiffness reduction in damaged reinforced concrete bridges under moving loads, and carries out the stiffness assessment after repairing using steel plates. In particular, the change of stiffness in each element before and after repairing, based on an advanced system identification (SI) method, is described and discussed by using a modified bivariate Gaussian distribution function. The proposed method in the study is more feasible than the conventional element-based method from computation efficiency point of view. The validity of the technique is numerically verified using a set of dynamic data obtained from a simulation of the actual bridge modeled with a three-dimensional solid element. The numerical examples show that the proposed technique is a feasible and practical method which can inspect the complex distribution of deteriorated stiffness although there is a difference between actual bridge and numerical model as well as uncertain noise occurred in the measured data.

Damage Identification in Bridges by Processing Dynamic Responses to Moving Loads: Features and Evaluation

“…Alternatively, the residual was defined as the relative difference between the measured and computed displacement responses [127]: γ=iY−iUfalse(normalqfalse), and the objective function or fitness function was expressed as [127]:minimizefalse(normalFfalse), normalF=true∑i=1N[γ]2=true∑i=1N[iY−iUfalse(normalqfalse)]2, where 1Y, …,NY denote the measured displacement data; iU: Rs→boldnormalR(i=1, …, N) is the function that satisfies normalU: Ω⊂Rs→normalΓ⊂RN; normalUfalse(normalqfalse)=[1U2…”

“…Lee et al [127] used μGA and the FEM to analyze measured dynamic displacement responses for identifying the arbitrary stiffness distribution in damaged reinforced concrete bridges subjected to moving loads. The dynamic displacement responses with noise effects were measured from a 3D solid finite element model subjected to a moving load with various velocities instead of real reinforced concrete slab bridges.…”

“…From the comparisons, acceleration-response-based detection was more practical than displacement- or velocity- response-based detection because acceleration data could be obtained directly from normal accelerometers. Unlike the procedure in reference [127], a hybrid GA was used instead of the μGA to optimize the residual of Equation (32). The hybrid GA was formulated by the incorporation of a GA with a conventional gradient-based optimization technique in light of the advantage of the former in enhancing the searching capability and that of the latter in improving convergence performance.…”