基于动态响应灵敏度同步反演结构物理参数与输入的算法

提出一种基于动态响应灵敏度概念,仅利用结构少数测点上的动态响应同步反演结构物理参数和输入的方法。首先,将结构的输入力进行正交多项式分解,推导了结构动态响应对输入力的正交系数和结构物理参数的灵敏度计算公式;然后,构造结构参数与输入同步反演的识别方程,并采用阻尼最小二乘法求解方程,同时识别结构物理参数变化与输入力时程;最后,分别利用一个五层平面框架结构和一个九榀三维刚架结构验证了方法的有效性。研究结果表明,该方法可以克服测量噪声和初始模型误差的影响,同步精确地识别结构的输入力和局部损伤引起的结构物理参数变化。     

有限测点下结构复合反演的混合法

回顾了近年来输入信息未知条件下结构识别的复合反演算法，并对输出信息测试不完备条件下的复合反演问题做了深入的研究。提出了时频域混合的复合反演算法。该算法首先在时域内利用子结构上的测试信息反演未知输入，然后在频域内依据反演得到的未知输入和有限测点上的响应信息构造目标函数来识别结构参数。该算法适用于具有多个未知输入的情况，与仅适用于单个未知输入条件下的复合反演算法相比，具有更广泛的应用价值。最后用一数值算例说明了本文算法的可行性。     

环境激励下桥梁结构模态参数识别方法的研究

研究了在环境激励下对桥梁结构进行模态参数识别的一种时域联合算法:自然环境激励技术(NEXT)和特征系统实现算法(ERA)的结合(简称NExT/ERA法).通过对某一预应力混凝土连续梁桥结构的仿真,其中考虑了利用加速度信号、外界白噪声因素的影响,验证了方法的有效性,并指出NExT/ERA法适合于环境振动下桥桑结构的监测、诊断.     

随机激励下随机结构动力可靠性灵敏度分析

对于随机激励下随机结构动力可靠性的灵敏度分析问题,在加权非线性响应面法的基础之上建立了随机结构动力可靠性灵敏度分析方法.所提方法从随机结构无条件动力可靠度的表达式出发,首先将随机结构的动力可靠性分析问题转化成传统的静力可靠性分析问题,然后采用基于加权非线性响应面法的Monte-Carlo可靠性灵敏度分析方法求解动力可靠性灵敏度值.算例表明该方法的计算结果是合理的,并且由于加权非线性法具有较高的效率和精度,因而所提方法具有一定的工程意义.     

结构动力参数识别方法研究

本文通过将脉动风对结构作用的特点转化为分析中的辅助计算条件，给出了一种在无需风作用过程信息的前提下，直接从结构风振动中应识别结构动力参数的方法，数值算例表明了该方法良好的实用性能。     

非稳态环境激励下线性结构的模态参数辨识

假定任意随机激励信号由白噪声与非白噪声信号组成，由此导出线性结构响应之间的相关函数由两部分组成，一部分与脉冲响应具有相同的数学形式，另一部分为其它形式，利用模态分解法的基本原理，把相关函数分解为各个模态函数的叠加与余项之和，这样，第一部分信号已经分解为不同的模态函数，第二部分中的周期信号也变成了模态函数，这就把非稳态环境激励下多自由度线性结构系统的模态参数辨识问题转化为类似于已知各个单自由度系统的脉冲响应进行模态参数辨识问题，理论和模拟实验表明，本文成功地利用模态分解法进行非稳态环境激励下多自由度线性结构系统的模态参数辨识，其主要优点是，无论是白噪声激励，稳态随机激励还是非稳态随机激励，仅根据结构的响应不仅能辨识线性结构的模态参数，而且能有效地识别出环境激励中的周期成分。     

环境激励下线性结构模态参数辨识方法的最新进展

本文根据国内外最新文献,对目前环境激励线性结构模态参数的辨识方法如频域分解法、ARMA模型法、NExT法、随机子空间法、时频分析方法和基于相关函数分解法等进行了评述,讨论了各方法在理论上和实际应用中存在的问题,对环境激励的模态参数识别方法的发展方向和趋势进行了探讨.     

基于有限多个测点信息的结构破损诊断研究

测试振型不完整是阻碍结构破损诊断技术发展的重要原因，本文以结构各自由度所含破损信息的大小为条件，根据灵敏度分析方法提出了结构测点数目和位置的优化方法，确保了有限多个测点能获取足够的结构破损信息，从而避免了测试振型扩充的困难，同时基于有限多个传感器测点信息提出了结构破损定位的诊断方法，算例研究表明该方法结构破损定位效果好，算法简单，应用方便。     

混沌激励下振动系统的非线性参数识别

本文对基于时-频相结合的非线性振动系统的参数识别问题进行了研究。首先建立含非线性参数单自由度振动系统的力学模型,将已知非线性系统产生的混沌响应作为该系统的激励, 假定其响应有若干不稳定的周期轨道组成,从混沌响应的状态空间中提取出近似周期轨道, 采用谐波平衡法识别出系统的参数,然后对识别出的参数进行误差分析。最后通过数值模拟,验证了混沌信号作为激励源对非线性系统进行参数识别的可行性。     

环境激励下结构模态参数识别的改进ITD法

结合随机子空间法（SSI）提出了环境激励下结构模态参数识别的改进ITD法。随机子空间法的识别精度高,其中数据的协方差计算可以保留原始数据中的所有信息,同时去除了噪声,得到的Toeplitz矩阵中的数据可以作为ITD法的输入数据,这样ITD法不再需要采用随机减量法或者自然激励技术（NExT法）进行前处理,从而避免了这两种前处理方法的不准确性带来的误差。通过环境振动下四层钢框架模型试验的位移响应,对提出的改进ITD法进行了验证。与ITD法相比,改进的ITD法明显提高了对频率和阻尼比等结构模态参数的识别精度,表明改进ITD法可应用于结构的模态参数识别;与SSI法相比,改进ITD法精度没有降低,同时缩短了计算时间,这将为该方法应用到结构的实时监测提供了可能。     

环境激励下基于信号降噪的模态参数识别研究

提出一种环境激励下基于测试信号降噪的模态参数识别方法。该方法首先对测试的随机响应数据采用自然激励技术(NEx T)获得互相关函数,进而基于结构矩阵低秩逼近(SLRA)方法得到降噪后的信号,最后通过复指数(Prony)方法识别结构的模态参数。数值算例和模型实验结果表明,该方法对测量信号有很好的降噪作用,识别精度高。     

基于不变环境振动的结构损伤识别方法

建立了受随机外力激励作用的损伤识别理论模型和计算方法 ,提出一种在线结构损伤识别方法。讨论一个15层框架结构 ,数值仿真结果表明 ,该方法具有较高的辨别率和计算精度。它能够使用在桥梁和建筑的健康监测中。     

多点输入下大跨输电单塔和塔线体系抗震性能分析

为了分析多点输入下大跨越输电单塔和塔线体系结构响应的差异,采用时程分析法系统地分析了多点输入下其地震响应,从节点最大位移、结构最大扭转角、杆件最大轴力和基底剪力等方面找出了2种结构体系地震响应的差异.通过分析可知,单塔和塔线体系结构响应存在较大的差异,且其差异与其所处的位置有关,并且与视波速的变化有关.因此对于大跨输电塔结构,应该考虑塔和线的耦合作用,在进行抗震分析建模时应采用塔线体系模型,否则得出的结论将偏于保守或偏于危险.     

Simultaneous identification of structural parameters and input time history from output-only measurements

A new time-domain method is suggested in this paper for simultaneous identification of the structural parameters and the time history of the input excitation using output-only measurements. The proposed method is based on an iterative identification procedure consisting of the least-squares identification technique and a modification process between each iterative step. The modification process is introduced to convert the spatial information of the external excitation into mathematical conditions. First, the unknown force vector is conjectured through the equation of motion using the initial guess of the structural parameters and the measured structural responses. The estimated input force vector is then modified to force it comply with the spatial distribution of the external excitations. The modified input force vector is further used to provide new estimation of structural parameters. Repeat the aforementioned procedure until the structural parameters satisfy the preset convergence criterion. Numerical examples as shear building and truss bridge model are employed to evaluate the feasibility of the proposed method. In the numerical examples, typical scenario of complete and noise-free as well as incomplete and noise-contaminated output measurements are considered. The results demonstrate that the proposed method can accurately identify both the structural parameters and the input time history for the cases that the structural responses are not polluted or slightly contaminated by measurement noise.The writers are grateful for the financial support from the National Natural Science Foundation of China (NSFC) through its Project of Young Scientists Fund (No. 50308020) awarded to the first author.     

Force identification of dynamic systems using genetic programming

One obvious limitation of the traditional force identification techniques is that they are unable to obtain the explicit expression of the force. Moreover, some techniques need both the displacement and velocity data of all freedoms, and some need the Markov parameters from numerical calculation or experimental test before the force identification can be carried out. This paper presents a genetic programming (GP) based method for excitation force identification of dynamic systems to overcome these traditional methods' disadvantages. GP is employed as a search and optimization method to obtain the optimal, if not the best, force expression from the known dynamic response. One obvious merit of the proposed method is that it can obtain the explicit expression of the unknown force. Another advantage is that it only needs the dynamic response data at one point, i.e. displacement or velocity or acceleration of one freedom. Illustrative examples demonstrate that the GP based method is able to identify the excitation force of a single‐degree, a three‐degree dynamic systems and a frame structure, depicting its potential for force forecast problems. Copyright © 2005 John Wiley & Sons, Ltd.     

Analysis of the sensitivity of a vibration-based procedure for structural identification

We analyze the accuracy of estimation of structural component material properties (elastic modulus, Poisson’s ratio, density, etc.) using the technique for measuring the eigenfrequencies of oscillations. The accuracy of frequency estimation using the Fourier method and its influence on the final result are evaluated. Relationships between the errors in the frequency estimation and those in the identified parameters are derived using the metamodel method. Application of the method is illustrated by the example of a shell element with a stiffening rib. __________ Translated from Problemy Prochnosti, No. 4, pp. 140–147, July–August, 2006.     

Parametric identification of systems with non-Gaussian excitation using measured response spectra

The problem of estimating parameters in dynamic systems excited by stochastic processes is addressed. Attention is focused on situations where the response processes are measurable but the excitation processes are non-Gaussian, unmeasurable and known only in terms of parameterised stochastic process models. General techniques for simultaneously estimating system and excitation process parameters are developed, based on the use of both normal, second order spectra and higher order, trispectra. The method is validated through application to some simulated data, relating to an oscillator driven by two specific kinds of non-Gaussian stochastic excitation.     

An enhanced response sensitivity approach for structural damage identification: convergence and performance

This paper develops an enhanced response sensitivity approach for structural damage identification. The whole work is mainly two‐fold. Firstly, the general response sensitivity approach has been shown to perform well for small damage, but may not work for moderate or large damage. Therefore, to overcome this drawback, the trust‐region restriction is additionally enforced to enhance the general response sensitivity approach. In doing so, the Tikhonov regularization is invoked which is simple for practical manipulation and is shown equivalent to trust‐region considerations. Secondly, concrete convergence analysis is presented to guarantee the performance of the enhanced response sensitivity approach. Numerical examples are studied to verify the proposed approach. Copyright © 2016 John Wiley & Sons, Ltd.     

On-line identification, flutter testing and adaptive notching of structural mode parameters for V-22 tiltrotor aircraft

New algorithms and results are presented for flutter testing and adaptive notching of structural modes in V-22 tiltrotor aircraft based on simulated and flight-test data from Bell Helicopter Textron, Inc. (BHTI). For flutter testing and the identification of structural mode frequencies, dampings and mode shapes, time domain state space techniques based on Deterministic Stochastic Realization Algorithms (DSRA) are used to accurately identify multiple modes simultaneously from sine sweep and other multifrequency data, resulting in great savings over the conventional Prony method. Two different techniques for adaptive notching are explored in order to design an Integrated Flight Structural Control (IFSC) system. The first technique is based on on-line identification of structural mode parameters using DSRA algorithm and tuning of a notch filter. The second technique is based on decoupling rigid-body and structural modes of the aircraft by means of a Kalman filter and using rigid-body estimates in the feedback control loop. The difference between the two approaches is that on-line identification and adaptive notching in the first approach are entirely based on the knowledge of structural modes, whereas the Kalman filter design in the second approach is based on the rigid-body dynamic model only. In the first IFSC design, on-line identification is necessary for flight envelope expansion and to adjust the notch filter frequencies and suppress aero-servoelastic instabilities due to changing flight conditions such as gross weight, sling loads, and air speed. It is shown that by tuning the notch filter frequency to the identified frequency, the phase lag is reduced and the corresponding structural mode is effectively suppressed and stability is maintained. In the second IFSC design using Kalman filter design, the structural modes are again effectively suppressed. Furthermore, the rigid-body estimates are found to be fairly insensitive to both natural frequency and damping factor variations and therefore stability is maintained. The Kalman filter design might be a better choice when the rigid-body dynamics are well known because no adaptation is necessary in this case.