Application of dynamic multi-swarm particle swarm optimization and sparse decomposition in ultrasonic thickness measurement of thin coating
Here,based on sparse decomposition matching tracking algorithm,ultrasonic testing signals of prefabricated steel structure protective coating were represented in an overcomplete Gabor time-frequency library,and the time-domain information of coating was further extracted to obtain the thickness information of coating.Aiming at problems of high complexity and large computation amount of the matching tracking algorithm,the dynamic multi-group particle swarm algorithm with characteristics of fast convergence and strong optimization ability was used to optimize the matching tracking algorithm.Based on chaos strategy,inertia weights were generated,and learning factors and inertia weights were linked together through trigonometric relations.In position update,time factors and influence factors of chaos disturbance strategy were added to balance local and global optimization capabilities of the algorithm.Simulation and experiments showed that the improved algorithm's testing accuracy is more largely increased,it can satisfy practical applications and greatly enhance the efficiency of sparse decomposition operations;compared with metallographic detection results,the relative error of fireproof coating testing is-4.65％,and the relative error of anti-corrosion coating testing is 1.33％.
万方数据
维普
