JPH0379658B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an improvement in a transfer function measuring method for measuring a transfer function of a structure.

The transfer function of a structure can be determined by calculation from the excitation force signal and response signal, which are simultaneously measured after applying an excitation force to the structure.

Excitation methods that apply excitation force to structures include sine wave sweep excitation method, random excitation method,
Impulse vibration method etc. are used.

These vibration methods will be described below. First of all,
In the sine wave sweep excitation method, as shown in Fig. 1, a sine wave signal S is supplied to an electrodynamic vibrator, a hydraulic vibrator,
By driving an unbalanced weight type vibration exciter 1 (hereinafter simply referred to as a vibration exciter), an excitation force is applied to the structure 2 by the vibration exciter 1. At this time, the acceleration signal 4, which is a response signal obtained from the accelerometer 3 in contact with the structure 2, and the excitation force signal 6 detected by the load cell 5 are taken into the mechanical impedance measuring device 7, where the This is a method of finding the transfer function 8 at the frequency of .

In addition, if the sine wave signal S is continuously changed,
The frequency response of the transfer function 8 is generally obtained in the form of a Bode plot.

Next, in the random vibration method, as shown in FIG. 2, the vibration exciter 1 is driven using a random wave signal RS with a sufficiently wide frequency band instead of the sine wave signal S.
Apply an excitation force to the structure 2. Similarly, the acceleration signal 4 and the excitation force signal 6 are taken in, and a Fourier transform device 9 performs processing such as Fourier transform to obtain the transfer function 8.

Furthermore, in the impulse vibration method, as shown in FIG. 3, the structure 2 is hit with a hammer or the like 10,
This is a method of determining the transfer function 8 from the acceleration signal 4 and the excitation force signal 6 at this time through processing such as Fourier transformation. This method utilizes the fact that when the structure 2 is struck with a hammer or the like 10, the impulse-like striking force has a wide frequency component.

Of the three methods described above, method 1 does not require the vibrator 1 and has the advantage of being able to measure the transfer function 8 in a short time, but it requires a large excitation force. I can't. Therefore, it cannot be used for large structures 2 such as bridges and high-rise buildings. Therefore, in the case of such a structure 2, it is customary to measure the transfer function 8 using a large-sized exciter 1 such as a rotating unbalanced weight type exciter. The time it takes to do so has to become longer.

By the way, there is a free vibration experiment as a vibration experiment method for the large structure 2, and there are many types of methods such as a seine method and a rocket method. These methods are
It uses a wire and the propulsion force of a winch or a rocket to apply an initial displacement, and when this displacement is released, the vibration characteristics are determined from the free vibration.
It is also called step vibration due to the nature of the excitation force, and the natural frequency and damping coefficient of the structure 2 are determined.

That is, the conventional step vibration method is carried out by measuring the vibration force using a load cell 11 as shown in FIG. 6, which will be described later. However, since a truncation error occurs in the measured value by this load cell, it cannot be used for calculating the transfer function, and is merely used as reference data. Therefore, in conventional vibration testing, the transfer function is not determined, but the free vibration of the structure, that is, the waveform of one of displacement, velocity, and acceleration, is observed and the natural frequency is calculated from the period. Alternatively, the natural frequency is determined by calculating the power spectrum of the free vibration waveform.

Therefore, like the above-mentioned impulse excitation method, this method does not require the vibrator 2, has the advantage of being able to measure in a short time, and can easily obtain a large excitation force; It is not possible to accurately determine the transfer function 8. The reason is as follows. First, for reference, FIG. 4 shows the original step function and its Fourier spectrum. Figure 4a shows the time waveform of the step function, and Figure 4b and c
are diagrams showing the amplitude and phase of the Fourier spectrum, respectively. Now, the actual excitation force is the fifth
As shown in Figure a, it is approximately a step function. However, even if this is directly subjected to finite Fourier transformation, a truncation error will occur, making it impossible to perform correct calculations.

This truncation error occurs when a finite length signal is Fourier transformed, and can be defined as follows. When a signal of finite length is Fourier transformed, the signal is interpreted as repeating itself infinitely. This is considered to be the same continuous signal starting again from where the signal ended. However, in general, the end and beginning parts of a signal are not smoothly continuous. This results in discontinuities in the signal and disturbances in the Fourier spectrum. This is the truncation error.

Therefore, when a truncation error occurs as described above, the result will not be as shown in FIG. 4b but as shown in FIG. 5b. Therefore, when this excitation force is actually applied to a one-degree-of-freedom system (natural frequency: approximately 20Hz), the response is measured and the transfer function 8 is determined from them. Although there is only one point (mountain), one more pseudo resonance point appears, and the phase fluctuates regardless of the resonance point, resulting in a very large error.

The present invention was made in view of the above circumstances, and uses a step excitation method, which is a simple excitation method, and approximates an impulse response by passing the signal obtained by this excitation method through a filter. The purpose of the present invention is to provide a method for measuring the transfer function of a structure by which a correct transfer function can be determined.

Before explaining the embodiments of the method of the present invention, the reason why the transfer function could not be accurately determined by the conventional step excitation method will be explained below. In the conventional step excitation method, a step-like excitation force signal is sampled and converted into a digital quantity, and this digital signal is subjected to Fourier transformation. This is because a truncation error occurs due to certain reasons. That is, the magnitude of the signal is zero at time zero in FIG. 5a, but the magnitude of the signal becomes 1 at the end of the data length (time 2 in FIG. 5a). Therefore, when the finite Fourier transform is performed, the data becomes discontinuous and a truncation error occurs.

Therefore, the method of the present invention removes the truncation error and measures the transfer function by a simple step excitation method, and FIG. 6 is a diagram showing one embodiment thereof. In this step vibration method, a tensile force F is applied to the structure 12 via the load cell 11 to give an initial displacement to the structure 12, and then the load cell 11 and the structure 1 are
By separating the structure 12 from the structure 12, a step-like displacement is applied to the structure 12, and the acceleration and excitation force at this time are transferred to the accelerometer 13 and the load cell 1 that are in contact with the structure 12.
Detect using 1. Then, the acceleration signal 14 detected by the accelerometer 13 and the excitation force signal 15 detected by the load cell 11 are passed through high pass filters 16 and 17 having appropriate cutoff frequency characteristics, respectively, to cut out low frequency signal components. Thereafter, if the outputs of both filters 16 and 17 are subjected to processing such as Fourier transform using the Fourier transform device 18, the transfer function 19 can be accurately determined.

The reason for this is as follows. That is, in conventional step vibration, the signal starts at 0, but the end is not 0, as shown in FIG. 4a. Therefore, repeating this signal creates discontinuities that cause truncation errors. However, if high-pass filters 16 and 17 are provided as shown in Figure 6, the excitation force will be converted as shown in Figure 7a, and both the beginning and end of the signal will be 0, so this can be repeated and arranged. Even if there is no discontinuity at the connection point, the transfer function can be determined accurately.

In other words, if the signal in FIG. 5a is passed through the high pass filters 16 and 17, it will not be a step-like signal as shown in FIG. Equal signals can be obtained, and even when such signals are subjected to finite Fourier transformation, no truncation error occurs, and correct results are obtained as shown in FIG. 7b. Therefore, if the transfer function of the structure 12 is measured with the configuration shown in FIG. 6, the result shown in FIG. is obtained. In other words, Fig. 7d shows the measurement results using the impulse excitation method, which has been sufficiently reliable in the past;
In Fig. 7(c), only one resonance point appears, and at that resonance point, the phase changes only once from 180° to -180°, and almost the same measurement results as in Fig. 7(d) can be obtained.

As detailed above, according to the present invention, the signal obtained by the step excitation method is passed through a high-pass filter and then subjected to processing such as Fourier transformation, thereby achieving accurate transmission that is not affected by truncation errors. It is possible to provide a method for measuring a transfer function of a structure, which can obtain a function, and can also measure a transfer function of a large structure using a relatively simple step excitation method.

[Brief explanation of drawings]

Figures 1 to 3 are block diagrams explaining different conventional excitation methods, Figures 4 a to c are diagrams explaining the original step function and its Fourier spectrum, and Figures 5 a to c are diagrams explaining conventional excitation methods. The step function, Fourier spectrum, and transfer function characteristic diagram obtained by actually applying the excitation method. Figure 6 is a block diagram explaining one embodiment of the method of the present invention. Figures 7 a to c are diagrams of the method according to the present invention. The characteristic diagram thus obtained, FIG. 7d, is a transfer function characteristic diagram obtained by the impulse excitation method. 11...Load cell, 12...Structure, 13...
... accelerometer, 16, 17 ... high pass filter, 18 ... Fourier transform device, 19 ... transfer function.

Claims (1)

[Claims] 1 A step-like displacement is applied to a structure via a load cell, an acceleration signal obtained from the structure is detected by an accelerometer attached to the structure, and an excitation force signal acting on the structure is detected by the load cell. A structure characterized in that the acceleration signal and the excitation force signal are passed through a high-pass filter, and the transfer function is obtained by Fourier-transforming both signals output from the high-pass filters. A method for measuring the transfer function of objects.