Experimental data interpretation using genetic algorithm for global health assessment of reinforced concrete slabs subjected to cracking

Abstract

Bridge deterioration must be monitored to ensure continued safety. Strategic repair can offer asset managers the benefit of saving money and time. However, knowing prime location and appropriate repair can be challenging when relying on visual techniques alone. This work quantitatively measures global damage from progressive cracking in a reinforced concrete slab. A novel combined framework of frequency methods, damage metrics, and genetic algorithm techniques is employed for maximum effectiveness of health evaluation. Through experimental modal analysis, tri-axial accelerometer data is interpreted via genetic algorithm using 24 damage indices based upon mode shapes. Along with easily interpretable visualization, four decision criteria validate the optimized results. Both general location and relative severity are estimated: when one artificial crack was induced by cutting, a relative strength change of 58% was detected. With four to eight cuts, change over 90% was detected.