基于粒子滤波的结构参数和荷载同步识别方法

在输入荷载未知的情况下,为解决结构参数和荷载的同步识别问题,将速度、位移、结构参数和输入荷载同时作为状态值构建状态空间方程,选用非线性性能较好的粒子滤波辨识结构参数与荷载.针对利用加速度进行荷载识别时出现的漂移问题,采用梯形积分和零相位高通滤波,通过测试得到加速度计算速度、位移数据,并运用计算值更新粒子滤波求得的速度、位移状态值,从而消除漂移现象.数值算例表明,文中方法能同时识别结构参数和荷载值,与真实值的误差较小,可有效解决识别过程中荷载识别值漂移的问题.

Simultaneous identification of structural parameter and external force with particle filter

To accomplish identification of parameters and force under unknown external force, the velocity, displacement, structural parameters and external force are simultaneously augmented into to the state vector for model formulation, and particle filter with good performance in nonlinear system is then used to identify both the parameters and force. There will be drift of the identified force during force identification due to the lack of velocity and displacement measurement. This paper uses trapezoidal integration and zero-phase high-pass filtering to calculate the velocity and displacement with acceleration. The calculated values are used to adjust the velocity and displacement obtained by particle filter at a specific interval to eliminate the drift. It is demonstrated in simulation that this method can provide relatively accurate estimation results and ameliorate the drift phenomenon during force identification.