针对多模态振动信号的在线监测和跟踪,提出基于随机子空间(SSI)和粒子滤波(PF)算法的仿真振动信号在线监测和跟踪方法.通过SSI算法提取得到振动系统的模态主频和阻尼比,根据振动系统模型模态主频和阻尼比的计算公式,得到系统的状态矩阵和输出矩阵.将计算所得状态矩阵和输出矩阵代入状态方程,利用PF算法进行信号的在线监测和跟踪,实现信号的降噪处理和预测分析.对于大型机械、桥梁等建筑物,对其进行在线监测保障其正常营运对社会经济发展具有深远影响.利用SSI算法提取系统的模态参数,进一步构建振动系统的状态矩阵和输出矩阵,并利用PF算法进行信号滤波抑噪和预测,在此基础上可以对结构状态实施在线监测及预警控制,实际大桥斜拉索振动信号测试也表明该算法可以提供稳定可靠的信号跟踪与预测技术.%Aiming at on-line monitoring and tracking of multi-mode vibration signals, a simulation method for online monitoring and tracking of vibration signals based on stochastic subspace identification (SSI) and particle filter (PF) techniques was proposed. Using SSI technique, the modal frequencies and damping ratios of a vibration system were extracted. Then, with the formulas of these modal frequencies and damping ratios, the state matrix and output one of the system could be obtained. Substituting these two matrices into the state equations of the vibration system, its vibration signals' filtering, denoising and predicting were performed with PF technique. Afterward, the on-line monitoring and early warning control for the state of the system could be conducted. The inclined cables' vibration signal measurement of a bridge indicated that the proposed method can provide a reliable and stable method for signal tracking and predicting.
展开▼
