A Multi-Ridge Recognition Method Using Modified MOPSO Algorithm and its Application on Modal Parameter Identification

The time-frequency image ridges can carry much information of system characteristic parameters. This paper presents the first modified multi-objective particle swarm optimization (MOPSO) algorithm designed to address this problem. The proposed method applies the modified velocity updating and modified position updating formulas to each subswarm such that better particle is generated. Its concrete implementation steps are given. Besides, to improve the calculation efficiency, this paper presents an energy threshold reject-ridge method. Finally, an example applied to the modal parameters identification of designed two-degree-of-freedom (2DOF) system to explore the effectiveness of the proposed method. The experimental results demonstrate that the method can achieve fast recognition speed, high accuracy, less iterations and small noise. This paper has proved that PSO algorithm is a promising method to solve ridge recognition problems regarding, in particular, according to the results herein obtained.