Extract Natural Frequencies Research Articles

Selecting sampling interval of transient response for the improved prony method

A sampling interval selection scheme for use with the improved Prony method is proposed. Since natural frequency extraction is a numerically ill-conditioned problem, a small change in the sampling interval results in a large change in the extracted natural frequencies. From the perturbation of the eigenvectors of the matrix, the accuracy criterion of the extracted natural frequencies is derived. The proposed scheme is validated using natural frequency extraction from the synthetic response, the method of moments (MM) response, and the measured response.

Image sequences and wavelets for vibration analysis: Part 1: Edge detection and extraction of natural frequencies

The paper presents the application of the wavelet edge detection algorithm for structural vibration analysis based on image sequences. A number of different edge detection operators are briefly discussed and illustrated using a simple image example. The wavelet-based edge detection algorithm is then used to identify natural frequencies of a simple cantilever beam. The paper discusses a number of important issues that need to be confronted when using image sequences for vibration analysis. These include correspondence of image features from image to image and image calibration. The paper compares the natural frequencies obtained using such an approach and also another non-contact method and finds them in good agreement.

Natural Frequency Extraction From Multiple Frequency Response Sample Sets Using the Fitting Error

We propose a method for selecting the most accurate natural frequency set from all of the natural frequency sets extracted from multiple frequency response sample sets. The accuracy of the natural frequency set is estimated from the proposed fitting error. The method is applied to synthetic data, the Method of Moments (MM) data and measured data, and variations of the error of the natural frequencies and the fitting error show an agreement.

Improvement of natural frequency extraction in frequency domain

Natural frequency extraction in the frequency domain using measured frequency responses is an ill-conditioned problem. A method is proposed to improve the accuracy of the previously extracted natural frequencies. It is shown from results that the method can improve the accuracy of the natural frequencies extracted from synthetic and measured frequency response samples.

Selection of Sampling Interval for the Gpof Method

A sampling interval selection scheme for the generalized pencil-of-function (GPOF) method is presented. The sensitivity which quantifies the natural frequency perturbation is defined, and the sampling interval which minimizes the sensitivity is selected as an optimum sampling interval. To show the validity of the proposed scheme, the results for the natural frequency extraction from synthetic data and measured data in a compact range are presented.

Asymptotic formulae for correcting finite element predicted natural frequencies of gravity and embankment dams

Extraction of natural frequencies of a gravity dam or an embankment dam plays an important role in the seismic design of the dam because the seismic response of a dam is dependent largely on the dynamic characteristics of the dam. Owing to the lack of exact solutions and the geometry of a dam, numerical methods such as finite element methods have been often used to extract the natural frequencies of the dam. Since the finite element method is an approximate one, the resulting finite element solution to the natural frequency of a dam cannot be safely used unless its accuracy is evaluated within the acceptable range for the seismic design of the dam. To solve this problem, some asymptotic formulae for correcting the finite element predicted natural frequencies of a gravity dam and an embankment dam have been developed in this paper. Since the present asymptotic formulae are derived from the fact that the finite element solution tends to the exact one if the finite element size used approaches zero, they provide a corrected solution of higher accuracy for the natural frequency of a dam so that the accuracy of a finite element solution can be evaluated against this corrected solution. After the correctness and usefulness of the present formulae are assessed, two practical examples have been given to show how the asymptotic formulae can be used to correct and evaluate the discretization error for the finite element predicted natural frequencies of gravity dams and embankment dams.

Bridge Assessment Using Forced-Vibration Testing

Full-scale forced-vibration tests conducted before and after structural repairs on a multispan reinforced-concrete highway bridge are described in the paper. The tests were conducted to investigate any correlation that may exist between the repair works and changes in the dynamic characteristics of the bridge. The approach adopted is suitable for assessing the structural condition of a bridge using vibration data. A purpose-built hydraulic vibrator was used to artificially excite the bridge, and the dynamic response was measured by accelerometers placed on the bridge deck. A single-degree-of-freedom model was fitted to the response functions in order to extract natural frequencies, modal damping ratios, and mode shapes. It was found that the repair works caused a slight reduction in the natural frequencies but there was no definite trend in the changes to the modal damping ratios. Comparison of the mode shapes before and after repairs using modal analysis procedures was found to give an indication of the ...

Extraction of natural frequencies of radar target by L2 space optimisation

A new method (L2M) for extracting the natural frequencies of a radar target is presented. The method is based on L2 space optimisation. The superiority of this method to the tradional L2 space optimisation method (l2M) in the presence of random noise is analysed and demonstrated by simulation computations.

Natural resonance extraction from multiple data sets using physical constraints

This paper presents a scheme to extract natural frequencies from multiple data sets using an E pulse technique. Physical constraints are placed on the amplitudes of the E pulse basis functions to prevent the extracted damping coefficients from being positive, as often happens with low signal‐to‐noise ratio (snr) data. The constrained technique is compared to the usual E pulse extraction technique to show the improvement in the estimated natural frequencies.