实时子结构实验OS法的累积误差分析.pdfVIP

第十一届全国实验力学学术会议 大连 2005 实时子结构实验 OS 法的累积误差分析? 许国山，吴斌 （哈尔滨工业大学土木工程学院，哈尔滨 150090 ） 摘要: 在实时子结构实验中，逐步积分算法采用递推的形式，而使实验过程中的反馈误差呈 现累积特性。由拟动力实验中无条件稳定的 OS法经速度假定发展而来的实时子结构实验 OS 法也使位移、速度、恢复力的反馈误差在实验进行过程中被累积起来。这些误差就像附加外 力一样作用在系统上，甚至使结果发散。根据实时子结构实验 OS 法执行过程可知，实时子 结构实验误差的主要来源有由于控制系统的自身限制而导致的位移、速度的控制误差和由于 量测仪器的精度而导致的恢复力的测量误差，另外还有 A/D 或 D/A 的转换误差以及电子设 备的噪声的误差等。这篇文章主要是检验了实时子结构 OS 法累积误差的增长，求出了线性 单自由度系统的位移、速度、恢复力误差放大系数。从结果来看，误差放大系数与具体的结 构参数有关，不同的结构参数会得到不同的结果，同时可以得出，系统误差对于弹性试件的 影响比对阻尼试件更大。 关键词: 实时;子结构;OS法;误差累积;拟动力 OS Method for Real-time Substructure Testing Error Propagation Analysis Xu Guoshan ，Wu Bin （School of Civil Engineering, Harbin Institute of Technology, Harbin 150090） Abstract: In real-time substructure testing, experimental feedback errors are accumulated in the step-by-step integration procedure. Displacement and velocity and acceleration errors in operator-splitting method for real-time substructure testing (OSM-RST) developed from OS method for pseudo-dynamics testing (OSM-PDT) by velocity modification are cumulated along with the test carry on. These error acting on experimental substructure as an extra force, even make this arithmetic radiation. According to the implementation of OSM-RST, we can know that the major sources of errors are inherent limitations or persistently biased inaccuracies in experimental apparatus, such as inadequate response of electro-hydraulic actuators, and the analogue-to-digital (or digital-to-analogue) conversion of electrical signals and the noise in electronic equipment. In this paper, the growth of cumulative experimental errors is examined, and the amplifications of force and displacement and acceleration for linear single-degree-system are dete