Structural Damage Detection and Classification Based on Clone Selection Algorithm of Particle Swarm Mutation

Inspired by a study on the ability of biological immune systems to identify antigens, we investigated adamage detection and classification problem in the health monitoring of large-scale structures and proposed a clone-selection algorithm of particle swarm mutation based on the autonomous, adaptive, and evolutional artificial immune theory and method. The algorithm sampled data from a structure model of an antigen that stimulates antibody sets. To improve the quality of the memory cells and achieve damage detection and classification of the measured data, the antibodies go through a learning and evolving process that included cloning, mutation, and selection. In particular, particle swarm mutation was introduced to solve the problem of binary encoding complexity and the uncertainty of the mutation direction in the clone selection algorithm. The experimental results obtained for the proposed algorithm using the Benchmark structure model indicated that the algorithm could effectively identify failure modes and improve the success rate of structure damage classification.