A New Structural Damage Evaluation Method Based on Modified Extreme Learning Machine

Structural damage assessment is an important part of structural health monitoring. Nevertheless, there is the defect of low accuracy in the existed evaluation methods. In order to solve this problem, a new model for structural damage assessment using particle swarm optimization (PSO) and extreme learning machine (ELM) named PELM is proposed in this paper. Firstly, PSO is used to calculate the best parameters of ELM, and then the optimized ELM model is adopted to evaluate the structural damage. PELM model and other two different models are used to assess the structural damages of a pile structure as an experiment, the results show that PELM model has a higher prediction accuracy than SVM model and ELM model.