DEVELOPING AN INVERSE ANALYSIS FRAMEWORK FOR QUANTIFYING BLAST LOADS BY INSPECTING THE CORRESPONDING STRUCTURAL DAMAGES

Abstract

The growth in the number of terrorist attacks and accidental explosive events imposed

additional demand on post-blast investigation tools. Numerous research addressed the chemical analysis of explosive residuals. Yet, limited work in the literature adhered to quantifying the blast load depending on the existing structural damage. The structural damage assessment is a useful data collection procedure for the post-blast analysis. This research proposes an Inverse approach that can estimate the explosive charge weight based on analyzing post-blast structural damage. The newly proposed approach is based on iterative finite element (FE) analysis-using Abaqus commercial package- and involves digital image processing techniques built-in Matlab software. Furthermore, to effectively identify the size of the blast, the genetic algorithm (GA) is used for optimization. At the end of each iteration, the damage contour plot is exported from Abaqus odb file through a python script. Afterward, the binary version of both the actual damage image and the FE damage plot is investigated and compared with the aim to minimize the difference between the real state and finite element solution. The image analysis is performed by virtue of edge detection operators available in Matlab. Image comparison is done depending on the Complex Wavelet- Structural Similarity Index (CW-SSIM). The concrete behavior under blast loading is represented using the Johnson- Holmquist-2 (Jh-2) damage constitutive material model. Likewise, the blast load effect is described through the Conventional Weapons Effects Blast Loading (CONWEP) model. Finally, the approach is validated using two toy examples starting with a forward FE solution as being the real damage state. The framework was able to predict the detonation mass with an accuracy between 97 to 99%.