Virtual reconstruction of crack with genetic algorithm and B-spline curve in bridge remote monitoring

Abstract

Smart film that was proposed in 2010 may monitor cracks of concrete bridges remotely and automatically. In the method, virtual reconstruction of crack plays an important role in providing intuitive foundations for bridge maintenance. Original crack data implies crucial information, while previous reconstruction is considered simply. In this paper, we redesign the method basing on B-spline curve optimized by using Genetic Algorithm (GA). Three steps are adopted to optimize the parameter vector and the number of control points of B-spline: (1) Build initial population model according to General Parameter Model (GPM); (2) Construct • tness function on the basis of sum square error; (3) Discuss the probability of crossover and mutation of GA. Compared with the methods of direct point connections and least-squares fit of B-spline without optimizing, results show that the novel method can achieve satisfactory fitting curves of crack. It is helpful to analysis the state stress of concrete structure, and to predict the cracking direction. It is also proved to be feasible and to have underlying using value for bridge remote monitoring.