基于遗传算法与随机减量技术的气动阻尼识别方法研究

The random decrement technique (RDT) has been widely utilized in the modal parameter identification of engineering structures due to its advantages of rapid computation and low cost. Aiming at the shortcomings of subjective selection of the truncation amplitude and sample duration in the RDT, bringing large identification error, a genetic algorithm(GA) was adopted to revise the RDT and an improved RDT (GA - RDT) method was proposed. Then the GA - RDT method was applied in the aerodynamic damping ratio identification of an aeroelastic model of a supertall building. Firstly, the free decay curve obtained by the RDT was fitted and then the error was defined. The effects of the cut-off amplitude and sample duration on the optimization target were analyzed and the genetic algorithm was utilized to search the optimal solution of these two parameters. Secondly, based on the tip acceleration time history acquired by the aeroelastic model wind tunnel tests，the aerodynamic damping ratio was identified by combinedly using the GA - RDT and Hilbert - Huang transformation method. Finally, the aerodynamic damping ratio identified by the natural excitation technique (NExT) was regarded as the benchmark, and the identification accuracy based on the GA - RDT method and traditional RDT method were discussed. The results show that, when compared with the aerodynamic damping ratio identified by the NExT method，the corresponding results identified by the GA - RDT method at different wind speeds have an average error of less than 0. 14% in the X, Y and torsional directions respectively, and the identification accuracy is much better than those identified by the traditional RDT method. Therefore, the feasibility of the aerodynamic damping identification method based on the GA - RDT method is verified.