A simple LMS-based approach to the structural health monitoring benchmark problem

Abstract

A structure's health or level of damage can be monitored by identifying changes in structural or modal parameters. However, the fundamental modal frequencies can sometimes be less sensitive to (localized) damage in large civil structures, although there are developing algorithms that seek to reduce this difficulty. This research directly identifies changes in structural stiffness due to modeling error or damage using a structural health monitoring method based on adaptive least mean square (LMS) filtering theory. The focus is on computational simplicity to enable real-time implementation. Several adaptive LMS filtering based approaches are used to analyze the data from the IASC–ASCE Structural Health Monitoring Task Group Benchmark problem. Results are compared with those from the task group and other published results. The proposed methods are shown to be very effective, accurately identifying damage to within 1%, with convergence times of 0.4–13.0 s for the twelve different 4 and 12 degree of freedom benchmark problems. The resulting modal parameters match to within 1% those from the benchmark problem definition. Finally, the methods developed require 1.4–14.0 Mcycles of computation and therefore could easily be implemented in real time. Copyright © 2004 John Wiley & Sons, Ltd.