A Probabilistic Model-based Crack Detection Method for Partially Obstructed Structural Element

This research project aims in developing a practical probabilistic model-base damage detection method for identifying cracks on structural members, such as plates, shells, and tubes, utilizing dynamic measurement. Many research works have been carried out on the detection of cracks on beams and columns. However, the research work in crack detection on 2-dimensional plate type structural components is very limited. Some researchers have a wrong impression that the detection of cracks on a plate is very similar to that on a beam. This impression is completely wrong, as the crack depth is not constant throughout the crack. It is a difficult task to capture the effect of this varying crack depth on the dynamic characteristics of the cracked plate. The main objective of most non-model base method is to detect the crack location. It is very difficult if not impossible for a non-model base method to quantify the crack extent by calculating the crack depth and width. In order to solve this problem, a model-base approach is adopted in this research project. Owing to the problem of measurement noise, the results of crack detection are uncertain. In this project, the probability theory will be employed to explicitly address this uncertainty problem. Hence, the project not only calculates the crack locations and severities but also the corresponding confidence level. This information is very valuable for engineers to make judgments on the remedial works. One of the requirements of the crack detection method in this research project is that it must be applicable even when the cracks are obstructed and measurement at or near the cracks are not available. Under such situation, almost all existing non-model base methods in the literature are not applicable. The research project will contain both theoretical development and experimental verification, which include the set-up of equipment, computer modeling and testing on plate type structures.