输入信息不完备下预报-校正类识别法探讨

Prediction-correction identification method with incomplete input information

输入信息测试不完备是结构动力测试中常遇到的问题,本文在回顾这方面文献的基础上,重点对预报-校正迭代算法进行了研究,针对该算法提出的第一类迭代过程,分别从部分输入未知和全部输入未知两方面进行讨论,从理论分析和数值算例说明了该迭代算法存在的不足,证明了学者们长期以来对该算法提出的疑问。

In this paper, structural system identification with incomplete input information is studied. On the basis of study of the existed methods for multiple inversion, prediction-correction iterative approaches for input and system identification are further studied. Because this class of methods were not proved theoretically, they have been doubted all the time, this paper put emphases on one iterative approach for multiple inversion, and the other method for multiple inversion with unknown ground motion is studied in another paper. From theory analysis and numerical examples, it can be seen： this method has the same applicable condition as Compensation Method when the iterative produce is convergent. And a new two-step method for multi-inversion is proposed.