Abstract

The nonlinearities present in engineering structures are generally local in nature. Hence for system identification, it is required first to identify the spatial location and then the nonlinearities are characterized. Once the nonlinearity is characterized, the system identification involves identifying coefficient of nonlinear terms. In the present work, an improved version of traditional conditioned time and frequency domain reverse path method for nonlinear parameter estimation of multiple-degree-of-freedom structures is presented. Both numerical and experimental studies have been carried out to investigate the effectiveness of the proposed improved time and frequency domain parameter estimation techniques over the traditional ones. The comparisons between the time and frequency domain reverse path approaches are also carried out to highlight the strengths and weaknesses of each of these approaches. Numerical and experimental investigations reveal that both the time and frequency domain approaches proposed in this paper overcomes the major limitations of their traditional counterparts and are capable of estimating the nonlinear coefficients of multiple nonlinear attachments either grounded or attached between the masses.

摘要

工程结构中存在的非线性通常本质上是局部的。因此对于系统辨识，首先需要辨识空间位置，然后表征非线性。一旦表征了非线性，系统识别就涉及识别非线性项的系数。在目前的工作中，提出了一种用于多自由度结构非线性参数估计的传统条件时域和频域反向路径方法的改进版本。已经进行了数值和实验研究，以研究所提出的改进时域和频域参数估计技术相对于传统技术的有效性。还对时域和频域反向路径方法进行了比较，以突出每种方法的优缺点。数值和实验研究表明，本文提出的时域和频域方法都克服了传统方法的主要局限性，并且能够估计接地或附着在质量块之间的多个非线性附件的非线性系数。