JPS6217765B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a vibration control device for a vibration table, and in particular, a vibration table characterized by faithfully reproducing a target time-series waveform and obtaining an unlimited vibration duration. Regarding control method.

Figure 1 shows the system configuration of the shaking table control device. By setting and inputting a target vibration waveform to be applied to the vibration test body 10 using the target waveform input device 1, the target vibration waveform is provided to the computer 2. calculator 2
Now, the Fourier transform device 3 is used to create a waveform to be applied to the vibration control system most suitable for the vibration control system, and this waveform is applied to the servo control device 6 via the high-speed analog input/output device 4. This analog waveform signal is converted from an electric signal to a hydraulic signal by the servo valve 7, and the hydraulic signal causes the vibrator 8 to vibrate, vibrating the table 9 connected to the vibrator 8, and moving the test specimen 10. It vibrates with a vibration waveform that targets the The servo control device 6 performs analog calculations using signals such as the spool position signal of the servo valve 7, the displacement signal of the vibrator 8, the differential pressure signal, and the acceleration signal of the table 9, as well as the signal from the computer 2, to control the servo valve 7. By applying a signal to the computer 2, the table 9 is vibrated in accordance with the command signal from the computer 2. The computer 2 also has a servo control device 6.
The actual vibration signals of the servo valve, vibrator, and table are input via the high-speed analog input device 5.
It has a function of ascertaining the control characteristics of the vibration control system and correcting the command waveform to the servo control device 6 based on the characteristics so that a target vibration waveform can be obtained. This operation will be referred to as input compensation. The present invention relates to this input compensation. Below, regarding input compensation, see Figures 2 and 3.
The principle will be explained using diagrams.

As shown in Figure 2, the transfer characteristic of a certain system is G
Suppose that it has a frequency characteristic of (ω). If the Fourier spectrum of F _O is input to this system, and the waveform of the Fourier spectrum of F _A is obtained as the output, the following equation holds true.

F _A = G (ω) × F _O (1) Here, for the system of G (ω), in order to make the input waveform and output waveform equal, that is, F _A = F _O ( If formula 1) is transformed to F _A =G(ω)×{1/G(ω)×F _O } (2), F _A and F _O become equal.

As shown in Figure 3, by multiplying F _O by G(ω) ^-1 , making this F _C , and inputting this F _C into the system (G(ω)), the target waveform F _O and the output Waveform F _A
can be made equal.

Here, the characteristic equation of the control system from the high-speed analog output device 4 to the table 9 shown in FIG.
(ω), calculate F _C =G(ω) ^-1 ×F _O inside the computer 2 using the input compensation method described above, and output this F _C to the high-speed analog output device 4. By doing so, the waveform to be reproduced in the table can be set to the target waveform F _O .

Next, according to Figure 4, the input compensation by the computer,
The calculation procedure will be explained. First, the target vibration waveform f
Extract _O 's time series data. If the sampling pitch of this time-series waveform data is ΔT [seconds] and the number of pieces of data is N, then the waveform connection time T _O is T _O =NXΔT [seconds] (3). This data is then subjected to Fourier transformation to convert it into frequency domain data _FO . Here, the aforementioned F O is multiplied by the reverse transfer characteristic G(ω) ^-1 of G(ω), which was previously measured as the frequency characteristic of the control system, and this is set _{as F C .} This F _C is
This is frequency domain data after input compensation. Continue to F
When _C is inversely Fourier transformed, it becomes time series data _fC , which is time series data after input compensation.By applying this time series waveform to a high-speed analog output device, it can be displayed on the table of the shaking table. The target waveform f _O can be faithfully reproduced.
However, it should be noted here that the excitation duration resulting from the above processing is limited to T _O . In other words, it is impossible to excite more than T _O . Here, T _is
As shown in equation (3), it is determined by N and ΔT.
ΔT is the sampling output interval, and according to the sampling theorem, ΔT=1/2·····(4): Maximum frequency It is determined by the maximum frequency to be output. Therefore, the excitation duration depends on the number N of data. However, this N is the data error of the Fourier transform, and is determined by the processing capacity of the Fourier transform device 3 shown in FIG. 1, and due to the configuration of the device,
There is a limit. Usually it is 2000-4000 points. Therefore, as shown in Fig. 5, if we call this section a frame as the time T _O that can be Fourier transformed, the first section frame 1 and the second section frame 2 can be used as they are. , if the connection is possible, input compensation can be added without any time limit by repeating the above process. However, theoretically, it has been proven that frame 1 and frame 2 cannot be connected. In other words, when performing input compensation in the frequency domain, there is a limit to the excitation duration due to the processing power of the Fourier transform device. There is a limit to this.

The present invention has been made with attention to this point, and enables continuous vibration excitation for a long period of time without subjecting the target waveform to an accurate target waveform and the vibration duration being subject to the above-mentioned limitations. . That is, an object of the present invention is to compensate the time series waveform for each interval in the frequency domain in order to improve the accuracy of the waveform, and to
To provide a vibration table control method capable of accurately maintaining a target waveform and sustaining vibration without time limit by connecting waveforms of each section with high precision without deviation.

The present invention is based on the characteristics of Fourier transform. In FIG. 5, it is assumed that an interval frame 1 with a time-series waveform is taken, the number of data points during this period is N, and the data sampling pitch is ΔT. When Fourier transform is performed on the N data of frame 1, the characteristic of Fourier transform is that the time series waveforms exactly the same as those of frame 1 are repeatedly arranged before and after frame 1. Get results. Therefore, even with a repetitive waveform such as a sine wave, the Fourier transform result will vary depending on the selection of the section. In order to reduce this influence, when a time-series waveform of a certain section is taken, the time-series wave is corrected at both ends as shown in FIG. 6, and then Fourier transformed. Using the frequency domain data obtained by this Fourier transform method, input compensation is performed by the method described in Section 3-2, and inverse Fourier transform is performed. After performing this operation on frame 1 as shown in FIG.
When selecting the section of frame 2, calculate frame 2 so that it has an overlapping section with frame 1,
Since the overlapped section of frame 1 and frame 2 should have the same waveform, within this section,
Search for a connectable point in frame 2 that follows the data in frame 1. In this way, by sequentially repeating the connection of frames 2 and 3, it is possible to connect them without limit.

A method of checking connection points is shown in FIG. That is, suppose that frame i-1 and frame i are connected at point p. This connection checking method will be explained in detail with reference to FIG. At the connection point between frames i-1 and i,
An evaluation function F indicating the connection status is introduced as follows from three elements: the midpoint value at the connection point P, the slope, and the error area in a certain section after P.

F=W _A × ΔA + W _R × ΔR + W _S × ΔS where ΔA: amplitude value deviation of connection point P ΔR: slope deviation of connection point P ΔS: frames i and i-1 in a certain check range after connection point P Area deviation of the difference in amplitude values W _A ; Weighting coefficient of ΔA W _R ; Weighting coefficient of ΔR W _S ; Weighting coefficient of ΔS Using this evaluation function value F, FΔF(ΔF;
Points that satisfy the evaluation function tolerance value are defined as waveform connectable points.

Using the above evaluation function, frame i is sequentially shifted within the connection check range to find a point where the waveform can be connected, and after this point, by arranging the waveform data of frame i-1, frames i and i -1 data is connected.

By connecting waveform data without limit as described above, it is possible to excite a target vibration waveform accurately and without any time limit.

The system configuration of the present invention is generally the same as the configuration shown in FIG. In order to accurately reproduce the target waveform f _O set by the target waveform input device 1 in the table 9 in FIG.
_C is created and applied to the analog output device 4. The present invention relates to the creation of this input compensation waveform f _C , and detailed embodiments thereof will be explained with reference to FIGS. 9 to 12.

In FIG. 9, data in section f _Oi of the time series waveform is first taken out, and both ends are multiplied by a correction pattern for Fourier transformation. This data is Fourier transformed to obtain F _Oi , which is then multiplied by the compensation characteristic G ^-1 of the system. As a result, F _Ci is inversely Fourier transformed to obtain a compensated time series waveform of f _Ci . Next, find the connection point of the waveform in the second half of f _Ci-1 and the first half of f _Ci , and
Arrange the waveform data of f _Ci after the connection point of _Ci-1 . By repeating this operation, input compensation waveforms can be connected without restrictions. In order to perform this process, the data structure inside the computer 2 is
Shown in Figure 0. That is, in this process, a target time series waveform f _O is given as an input in block 15, and a time series waveform f _C is output after input compensation. As data during the calculation, blocks 30, 40, 50, and 60 have F _Oi , F _Ci , and f _Ci, and block 20 gives the reverse transfer characteristic G ^-1 of the system.

FIG. 11 shows a detailed flow of the waveform matching point search. First, the value of the evaluation function F at the search starting point is calculated. Next, it is determined whether F is within the allowable value ΔF, and if the condition is not satisfied, the search start address is updated and the process is repeated again. The point that satisfies the conditions is set as the waveform connection point P, and then the data arrangement process shown in FIG. 12 is performed.
That is, the data starting from the address P in the f _Ci table FCIT is transferred to the corresponding address starting from the corresponding address in the f _C table FCT. By connecting the data of frame i-1 and the data of frame i in this way and repeating this process, waveforms can be connected continuously without any restrictions.

As described above, according to this embodiment, by continuously generating input compensation waveforms that take into account the frequency characteristics of the control system, the target waveform can be accurately and
It became possible to vibrate the vibration table without any time limit, and it was possible to provide an advanced vibration control device.

In the above embodiment, the evaluation function was formed using the three elements of amplitude, inclination, and area, but it is also possible to create a certain evaluation function using two of these elements.
Further, elements other than the three elements may be added.

According to the present invention, it is possible to continuously create input waveform compensation in the frequency domain in consideration of the response characteristics of the control system, and since the compensation is in the frequency domain, it is possible to continuously create input waveform compensation in the frequency domain, taking into account the response characteristics of the control system. Regardless of the waveform type, it is possible to accurately reproduce a time-series target waveform on the vibration table without any time limit for any waveform, realizing high-performance vibration control of the vibration table. can.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the system configuration of the present invention, FIGS. 2 to 5 are diagrams for explaining the conventional technology,
6 to 8 are diagrams for explaining the principle of the present invention, and FIGS. 9 to 12 are diagrams for explaining embodiments of the present invention. 1... Target waveform pulse input device, 2... Calculator, 1
0...Test specimen.

Claims (1)

[Claims] 1. For a vibration table device having a vibration table, an exciter, a servo valve, and a servo control device, the transfer characteristics of the vibration table control system are measured in advance, and frequency control by Fourier transform is performed using this characteristic equation. In a vibration control method that performs input waveform compensation to a servo control device, input waveform compensation is performed for each fixed interval of the target time series waveform, and in order to connect these waveforms, weighting coefficients are added to elements that indicate waveform characteristics. By evaluating the degree of connection of waveforms using the evaluated function, connecting the waveforms at the optimal points, and continuously performing waveform compensation, we demonstrated that it is possible to excite the target waveform accurately without any time limit. Characteristic shaking table control method. 2. The shaking table control method according to claim 1, wherein the above-mentioned elements are amplitude, inclination, and area.