Optimization of ultrasonic tomography in concrete using non-linear paths through bio-inspired algorithms

Abstract

This paper aims to challenge the conventional approach of ultrasonic computerized tomography in concrete structures by using bio-inspired algorithms (genetic algorithms and ant colony optimization) to determine non-linear wave paths for generating tomographic images. Numerical UPV tests were conducted in two phases followed by an experimental validation. Initially, all algorithms were used to determine the shortest paths of ultrasonic waves, considering a known velocity map in a continuous medium. The time of flight (TOF) and wave trajectories were analyzed by adjusting the parameters of the bio-inspired algorithms. In the subsequent phase, the algorithms were evaluated for their efficiency in generating tomographic images of concrete cross sections through an iterative path update process. An experimental specimen was constructed, and ultrasonic tests were conducted to validate the proposed approach. The results demonstrated that all evaluated methods could determine paths that were faster than a straight line. Moreover, they exhibited the capability to generate improved images with enhanced precision regarding particle size and location. This conclusion highlights the potential of using bio-inspired algorithms as a promising alternative for optimizing ultrasonic wave paths and enhancing the accuracy of tomographic imaging.