JP5420621B2 - Non-Gaussian vibration control device - Google Patents

Description

  The present invention relates to a vibration control apparatus that controls a vibration testing machine so that a target non-Gaussian vibration is applied to a specimen.

  In order to simulate the effects of vibrations during transportation and operation, vibration tests that give desired vibrations to specimens have been conducted. In the vibration test, it is a vibration control device that controls the vibration testing machine so as to obtain a desired vibration.

  If vibration that is actually applied is recorded and this vibration can be applied to the specimen, an accurate vibration test can be performed. However, in order to record and reproduce the actual vibration waveform itself, an enormous recording capacity is required, so that it is not commonly used.

  On the other hand, the test which gives the vibration by a sine wave is also done. In this case, control is easy because it only outputs a sine wave, but there is a problem that the deviation from the vibration that is actually applied is too large.

  Therefore, a random vibration test is performed in which frequency characteristics (power spectral density) of vibrations actually applied are calculated, and vibrations having the target power spectral density are applied to the specimen.

  FIG. 20 shows a vibration control device for a conventional random vibration test disclosed in Patent Document 1. Since the vibration tester 2 itself has frequency characteristics, even if vibration having a target spectrum is given, the vibration is not given to the specimen 4. Therefore, feedback control is performed by the vibration control device so that the spectrum (response spectrum) of the vibration waveform of the specimen 4 is equal to the target spectrum.

  The specimen 4 fixed to the vibration testing machine 2 is vibrated by the vibration testing machine 2. The vibration of the specimen 4 is detected by the acceleration sensor 6 and converted into a response signal which is a digital signal by the A / D converter 8. The Fourier transform means 10 performs a Fourier transform on the response signal and calculates a response spectrum.

  The control spectrum calculation means 12 compares the target spectrum with the response spectrum and calculates the control spectrum so that they are equal. The drive signal generation means 14 gives a random phase to each frequency component of the control spectrum and performs inverse Fourier transform to generate a drive signal.

  The D / A converter 16 converts the generated drive signal into an analog signal and supplies it to the vibration testing machine 2.

  As described above, it is possible to control to give the specimen 4 vibration having the target spectrum.

JP-A-8-068718

  However, in the vibration control apparatus of Patent Document 1, although vibration having a target spectrum can be given to the specimen 4, the probability density distribution of the vibration is a Gaussian distribution (normal distribution) as shown by the solid line in FIG. ) However, even though actual vibrations often have a non-Gaussian distribution, the conventional vibration control device cannot impart vibrations having non-Gaussian characteristics to the specimen.

  An object of the present invention is to solve the above-described problem and to provide a vibration control device capable of controlling not only a target spectrum but also a vibration having a target non-Gaussian characteristic to a specimen. And

(1) (3) (9) The vibration control device according to the present invention is a vibration tester so that the specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic. A vibration control device for controlling a drive signal applied to the sensor, wherein a response signal from a sensor for detecting vibration of the specimen is converted into a frequency function to obtain a response spectrum, a response spectrum and a target spectrum. A control spectrum calculating means for calculating a control spectrum such that both are equal, a non-Gaussian random signal generating means for converting the control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal; Non-Gaussian characteristics with the inverse characteristics of the transfer characteristics of the system including the vibration tester and the specimen as control characteristics so that the non-Gaussian random signal and response signal are equal. Drive signal generating means for transforming a random signal to obtain a drive signal, and control characteristic updating means for calculating the transfer characteristic based on the drive signal and a response signal from the sensor and updating the control characteristic. I have.

  Feedback control is performed so that vibration having a target spectrum is given to the specimen, and further, vibration equal to the generated non-Gaussian random signal is given to the specimen. Vibration having Gaussian characteristics can be given.

(2) (4) (10) The vibration control device according to the present invention is a vibration tester so that the specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic. A vibration control device for controlling a drive signal applied to the sensor, wherein a response signal from a sensor for detecting vibration of the specimen is converted into a frequency function to obtain a response spectrum, a response spectrum and a target spectrum. A control spectrum calculating means for calculating a control spectrum such that both are equal, a non-Gaussian random signal generating means for converting the control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal; The non-Gaussian characteristic calculating means for calculating the response non-Gaussian characteristic of the response signal from the sensor is compared with the response non-Gaussian characteristic and the target non-Gaussian characteristic. As a control characteristic, the inverse characteristic of the transfer characteristic of the system including the vibration tester and the test specimen is set so that the non-Gaussian random signal and the response signal are equal to each other. Drive signal generation means for transforming a non-Gaussian random signal to obtain a drive signal, and a control characteristic update for calculating the transfer characteristic and updating the control characteristic based on the drive signal and a response signal from the sensor Means.

  Therefore, since the control characteristics are updated in real time, the same vibration as the non-Gaussian random signal can be given to the specimen more accurately.

(5) In the vibration control device according to the present invention, the non-Gaussian random signal generating means converts the control spectrum into a time function and obtains a Gaussian random signal, and the calculated Gaussian property. Conversion means for converting a random signal based on the conversion characteristics and converting it into a non-Gaussian random signal is provided.

  Therefore, a desired non-Gaussian random signal can be easily obtained by controlling the conversion characteristics.

(6) In the vibration control device according to the present invention, the response spectrum generation means converts the response signal from the sensor that detects the vibration of the specimen into a Gaussian response signal based on the conversion characteristics of the conversion means. It is characterized by comprising Gauss-Gauss conversion means and Gaussian response spectrum generation means for converting the Gaussian response signal into a frequency function and obtaining a Gaussian response spectrum.

  Therefore, the response spectrum calculation means can calculate the response spectrum based on the Gaussian response signal, and can perform feedback control with high accuracy.

(7) In the vibration control device according to the present invention, when the non-Gaussian random signal generating means converts the time function based on the control spectrum into the time function, the probability density distribution of the phase difference sequentially given to each component of the control spectrum is It is characterized by controlling the non-Gaussian characteristic of the generated non-Gaussian random signal by controlling it as a conversion characteristic.

  Therefore, since the non-Gaussian random signal is generated directly from the control spectrum, there is little possibility of generating unnecessary spectral components, and the accuracy of feedback control by the control spectrum calculation means is improved.

(8) In the vibration control device according to the present invention, the non-Gaussian random signal generation means generates a non-Gaussian random signal generated by controlling at least an average or variance of the probability density distribution of the phase difference as the conversion characteristic. It is characterized by controlling the non-Gaussian characteristics of the signal.

  Therefore, a desired non-Gaussian random signal can be generated.

(9) In the vibration control device according to the present invention, the vibration tester has a plurality of vibration generators that apply vibrations from a plurality of axial directions to the specimen, and the response spectrum calculation means, A control spectrum calculating means and the non-Gaussian random signal generating means are provided for each of a plurality of axes, and the control characteristic updating means includes a response signal for each axis and a drive signal for each axis. Based on the control characteristics related to the combination of the response signal for each axis and the drive signal for each axis, the drive signal generating means, the non-Gaussian random signal for each axis and the A drive signal for each axis is generated based on the control characteristics related to the combination.

  Therefore, vibration control in the multi-axis direction can be performed.

(10) The vibration control device according to the present invention is provided in the vibration tester so that the specimen placed on the vibration tester vibrates with a waveform according to a target spectrum and a target arbitrary probability density distribution. This is a vibration control device that controls the drive signal to be applied. The response signal from the sensor that detects the vibration of the specimen is converted into a frequency function, and the response spectrum calculation means for obtaining the response spectrum is compared with the response spectrum and the target spectrum A control spectrum calculating means for calculating a control spectrum such that both are equal, a Gaussian random signal generating means for converting the control spectrum into a time function and obtaining a Gaussian random signal having a probability density distribution by a Gaussian distribution, Based on the ZMNL function, a Gaussian random signal having a probability density distribution by the Gaussian distribution is set as the target. A non-Gaussian random signal generating means for converting into a non-Gaussian random signal having a desired probability density distribution, and a system including a vibration tester and a specimen so that the non-Gaussian random signal and the response signal are equal. Drive signal generating means for obtaining a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic as a control characteristic is provided.

  Therefore, by giving the probability density distribution of the target vibration, the vibration matching this can be given to the specimen.

  In the embodiment, “response spectrum calculation means” corresponds to step S12.

  In the embodiment, “control spectrum calculation means” corresponds to step S13.

  In the embodiment, “non-Gaussian random signal generating means” corresponds to step S2.

  In the embodiment, “Gaussian random signal generating means” corresponds to steps S21 to S25.

  In the embodiment, “conversion means” corresponds to step S26.

  In the embodiment, the “waveform control unit” corresponds to step S3.

  In the embodiment, the “drive signal generation unit” corresponds to steps S31 to S33.

  In the embodiment, the “control characteristic updating unit” corresponds to steps S35 and S36.

  “Non-Gaussian characteristic calculation means” corresponds to steps S251 and S252 in the embodiment.

  In the embodiment, the “conversion characteristic calculation unit” corresponds to step S253.

  The “program” is a concept that includes not only a program that can be directly executed by the CPU, but also a source format program, a compressed program, an encrypted program, and the like.

  In the present invention, “control to be equal” means that control is performed in a direction in which both are equal. For example, feedback control is included in this concept.

It is a functional block diagram of a vibration control device by a first embodiment. It is a hardware configuration of a vibration control apparatus. 4 is a flowchart of a control program 44. It is a flowchart of a control spectrum generation process. It is a flowchart of the production | generation process of a non-Gaussian random signal. It is a flowchart of a waveform control process. It is a figure which shows a target spectrum and a target non-Gaussian characteristic. It is a figure which shows the produced | generated several frames Gaussian random signal. It is a figure for demonstrating the process which superimposes while multiplying a Gaussian random signal by a window function and shifting. It is an example of a ZMNL function. It is a figure for demonstrating the process for obtaining a drive signal based on a non-Gaussian random signal. It is a functional block diagram of the vibration control apparatus by 2nd embodiment. 4 is a flowchart of a control program 44. It is a flowchart of the correction process of conversion characteristics. It is a figure which shows the relationship between a standard deviation and kurtosis. It is a figure which shows the example at the time of providing the reverse conversion means 19. FIG. It is a figure for demonstrating the random signal generation by other embodiment. It is a flowchart for demonstrating the process which determines a ZMNL function. It is a figure which shows the example in the case of performing multi-axis control. It is a figure which shows the conventional vibration control apparatus. It is a graph which shows the probability density distribution of a response signal.

1. First embodiment
1.1 Functional Block Diagram FIG. 1 shows a functional block diagram of the vibration control device according to the first embodiment of the present invention. The specimen 4 fixed to the vibration testing machine 2 is vibrated by the vibration testing machine 2. The vibration of the specimen 4 is detected by the acceleration sensor 6 and converted into a response signal which is a digital signal by the A / D converter 8. The response spectrum calculation means 20 performs Fourier transform on the response signal to calculate a response spectrum.

  The control spectrum calculation means 22 compares the target spectrum with the response spectrum and calculates the control spectrum so that they are equal. That is, feedback control is performed.

  The Gaussian random signal generation means 26 gives a random phase to each frequency component of the control spectrum, performs Fourier transform, and generates a Gaussian random signal. The conversion means 28 converts the Gaussian random signal based on the conversion characteristic, and converts it into a non-Gaussian random signal. In this embodiment, the Gaussian random signal generating means 26 and the converting means 28 constitute a non-Gaussian random signal generating means 24.

  The drive signal generation means 32 generates a drive signal by transforming the non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic. The drive signal is converted into an analog signal by the D / A converter 16 and given to the vibration testing machine 2.

  The control characteristic update unit 34 calculates the transfer characteristic of the system based on the drive signal and the response signal, and updates the control characteristic which is the reverse characteristic. The updated control characteristic is given to the drive signal generation means 32. In this embodiment, the waveform control means 30 is constituted by the control characteristic update means 34 and the drive signal generation means 32.

  The above processing is summarized as follows. The response spectrum calculation means 20 and the control spectrum calculation means 22 perform the food back control so that the response spectrum becomes equal to the target spectrum, and calculate the control spectrum. The non-Gaussian random signal generating means 24 generates a non-Gaussian random signal having this control spectrum. The waveform control means 30 controls the specimen 4 to vibrate according to the waveform of the non-Gaussian random signal thus generated.

  As described above, by setting the conversion characteristic of the conversion means 28 for the specimen 4, vibration having a target non-Gaussian characteristic and having a target spectrum can be given to the specimen 4. .

1.2 Hardware Configuration FIG. 2 shows a hardware configuration when the vibration control device of FIG. 1 is realized using a digital signal processor (DSP).

  Connected to the DSP 40 are a flash ROM 42, a memory 46, a D / A converter 16, and an A / D converter 8. The D / A converter 16 converts the drive signal as digital data into an analog signal and gives it to the vibration testing machine 2. The vibration of the specimen 4 vibrated by the vibration testing machine 2 is detected by the acceleration sensor 6 and output as an analog response signal. The A / D converter 8 converts the response signal as an analog signal into digital data. A control program 44 for control is recorded in the flash ROM 42.

  Note that an MPU, a touch panel display, etc. (not shown) are connected to the DSP 40 so that an operator can set a target spectrum or a target Gaussian characteristic.

1.3 Control Processing FIG. 3 shows a flowchart of the control program 44. Here, as shown in FIG. 7, the description will proceed assuming that the target spectrum, target kurtosis (kurtosis) and skewness (distortion) are set by the operator and recorded in the memory 46.

  Here, kurtosis is a scale representing the degree of sharpness of the distribution shape in the probability density distribution of the vibration waveform shown in FIG. In the Gaussian vibration waveform, the kurtosis is 3. In the non-Gaussian vibration waveform, the kurtosis is larger (see the broken line in FIG. 21). This indicates that in the non-Gaussian vibration waveform, the peak value of the vibration waveform protrudes greatly with respect to the average value of the vibration waveform. Skewness is distortion (asymmetry degree) of the vibration waveform. In the Gaussian vibration waveform, the skewness is zero. In a non-Gaussian vibration waveform, the skewness may deviate from zero. For example, as indicated by the wavy line in FIG. 21, the average value (ie, skewness) is 0.5.

  First, the DSP 40 acquires a response signal from the A / D converter 8 and generates a control spectrum (step S1). Next, a non-Gaussian random signal having the frequency characteristic of the control spectrum is generated (step S2). Subsequently, waveform control is performed based on the non-Gaussian random signal, and vibration is given to the specimen 4 (step S3). Such control is repeated, and vibration having a target spectrum and a target non-Gaussian characteristic is given to the specimen 4.

  FIG. 4 shows a detailed flowchart of control spectrum generation (step S1). The DSP 40 first obtains a response signal from the A / D converter 8 (step S11).

  Next, the DSP 40 performs Fourier transform (FFT) on the acquired response signal, and calculates a power spectrum (referred to as response spectrum) of the response signal (step S12). Since the FFT is performed on the response signal as digital data, the response spectrum becomes a line spectrum. Subsequently, the DSP 40 compares the response spectrum with a target power spectrum (referred to as a target spectrum), and corrects the control spectrum so that they match each other (step S13). This control spectrum is also a line spectrum. As will be described later, since the signal having the control spectrum thus corrected is supplied to the vibration testing machine 2, the response spectrum becomes equal (or close) to the target spectrum by feedback control.

  FIG. 5 shows a detailed flowchart of the non-Gaussian random signal generation process (step S2). First, the DSP 40 sequentially gives a random phase θ to each line of the control spectrum (each frequency component of the line spectrum) (step S21). Here, uniform random numbers from 0 to 2π are given as the random phase θ.

  Next, the DSP 40 performs inverse Fourier transform (inverse FFT) on a control spectrum given a random phase θ, and generates a Gaussian random signal (time axis signal) for a plurality of frames (for example, 12 frames). Generate. Here, the frame refers to a predetermined time length.

  Since a random phase is sequentially given to one control spectrum, Gaussian random signals having different waveforms can be obtained for a plurality of frames. Incidentally, the probability density distribution of the drive signal generated in this way follows a Gaussian distribution (see the solid line in FIG. 21). The DSP 40 records the generated Gaussian random signal in the memory 46 (step S22).

  Subsequently, a Gaussian random signal for a desired frame is generated by superimposing the generated Gaussian random signal by applying a window function.

  FIG. 8 shows the generated Gaussian random signals GD1, GD2, GD3... GDn for n frames. The DSP 40 multiplies each Gaussian random signal GD1, GD2, GD3... GDn by a window function (step S24). FIG. 9 shows the respective Gaussian random signals WGD1, WGD2, WGD3... WGD4 multiplied by the window function. For reference, the window function is shown as a broken line in FIG.

  Next, the DSP 40 superimposes (adds together) the Gaussian random signals WGD1, WGD2, WGD3... WGDn while shifting by 1/4 frame as shown in FIG. 9 (step S25). In this way, a continuous Gaussian random signal SGD can be obtained. Since the amplitude becomes larger than the original waveform due to the superposition, the amplitude is adjusted.

  The control spectrum is a discrete value calculated digitally and becomes a line spectrum. Therefore, the frequency characteristic of the Gaussian random signal obtained based on this control spectrum is a line spectrum. However, in this embodiment, a line spectrum is made a continuous spectrum by superimposing a plurality of different Gaussian random signals by applying a window function. In this embodiment, a Hanning window is used as the window function.

  The probability density distribution of the amplitude of the generated Gaussian random signal SGD exhibits a Gaussian distribution as shown in FIG. That is, the kurtosis (projection) is 3.0 and the skewness (distortion) is 0. On the other hand, the target kurtosis is set to 5.0 and the target skewness is set to 0.5 (see FIG. 7). That is, the maximum amplitude must be larger and the signal must be distorted to the plus side.

  Therefore, in this embodiment, a signal having a desired non-Gaussian characteristic is converted using a ZMNL function (Zero-memory nonliner sysytem function) as shown in FIG. 10 (step S26). In FIG. 10, the horizontal axis represents a given signal, and the vertical axis represents an output signal. When the Gaussian random signal x (t) is converted by the ZMNL function, a non-Gaussian random signal y (t) is obtained.

  The ZMNL function is expressed by the following mathematical formula 1.

x is a Gaussian signal, and y is a non-Gaussian signal. Details of each coefficient are as shown in Equation 2.

  By substituting the values of the target skewness and the target kurtosis into the skewness S and kurtosis K in the above equation, a ZMNL function that can generate a desired non-Gaussian random signal by conversion can be obtained.

  In the above equation, a target kurtosis smaller than 2.8 cannot be set. Therefore, a function such as Equation 3 may be used.

The coefficient Cn can be determined by solving an optimization problem that minimizes the error variable ε shown in Equation 4.

  The non-Gaussian random signal obtained as described above is recorded in the memory 46.

  FIG. 6 shows details of the waveform control process (step S3) based on the non-Gaussian random signal.

  The DSP 40 reads one frame of the non-Gaussian random signal SNGD and multiplies it by a window function (step S31). In this embodiment, a Hanning window is used as the window function.

  Next, based on the inverse transfer function of the vibration tester 2 and the specimen 4 updated in the previous process, the drive signal WD1 is generated for one frame of the non-Gaussian random signal multiplied by the window function (step S32) (see FIG. 11). Even if the above-described non-Gaussian random signal is given to the vibration testing machine 2, the specimen 4 cannot be given vibration according to the non-Gaussian random signal. This is because the vibration tester 2 and the specimen 4 have transmission characteristics. Therefore, the transfer characteristics (transfer function) of the vibration tester 2 and the specimen 4 are measured, and a non-Gaussian random signal multiplied by the inverse characteristic is given to the vibration tester 2 as a drive signal. Thereby, the specimen 4 vibrates according to a desired non-Gaussian random signal.

  The transfer function is not constant depending on the spectrum of the given signal. Therefore, the inverse characteristic calculated based on the transfer function is updated in real time. However, when the inverse characteristic is updated, the drive signals before and after the update are not continuous, which may cause unnecessary frequency components. Therefore, in order to make this continue smoothly, a window function is multiplied and overlapped as described later.

  Next, the DSP 40 superimposes the drive signal to which the window function has been applied by shifting the drive signal generated last time by 1/2 frame (step S33). For example, as shown in FIG. 11, if the currently generated drive signal is WD2, the memory 46 is overlaid (added) by ½ frame from the previously generated drive signal WD1.

  Thereby, as shown in FIG. 11, a continuous drive signal SD can be obtained.

  Next, the DSP 40 reads out one frame of the generated drive signal SD and outputs it to the D / A converter 16 (step S34). Therefore, the D / A converter 16 gives an analog drive signal SD to the vibration testing machine 2. Thereby, the vibration testing machine 2 generates vibration. The DSP 40 records the applied drive signal in the memory 46.

  Next, the DSP 40 takes in a response signal from the A / D converter 8 (step S35). Subsequently, based on the drive signal recorded in the memory 46 and the corresponding response signal, a transfer characteristic (transfer function) is calculated. Based on this transfer function, an inverse transfer function is calculated, and the inverse transfer function is updated and recorded (step S36). That is, the updated inverse transfer function is used in the generation of the drive signal from the next time.

As described above, vibration equal to the non-Gaussian random signal can be applied to the specimen 4.

2. Second embodiment
2.1 Functional Block Diagram FIG. 12 shows a functional block diagram of the vibration control device according to the second embodiment of the present invention. In the first embodiment, control is performed by the waveform control means 30 so that the waveform of the generated non-Gaussian random signal matches the waveform of the response signal. As a result, vibration having a target spectrum and a target non-Gaussian characteristic can be given to the specimen 4.

  In the second embodiment, the conversion characteristic is feedback-controlled so that the non-Gaussian characteristic (response non-Gaussian characteristic) of the response signal matches the target non-Gaussian characteristic. Thereby, the target non-Gaussian characteristic is realized with higher accuracy.

  FIG. 12 shows a functional block diagram of the vibration control device according to the second embodiment. In addition to the vibration control apparatus according to the first embodiment, a non-Gaussian characteristic calculating unit 31 and a conversion characteristic calculating unit 33 are provided.

  The non-Gaussian characteristic calculation means 31 calculates a non-Gaussian characteristic (for example, kurtosis K, skewness S) of the response signal. The conversion characteristic calculation means 33 compares the response non-Gaussian characteristic and the target non-Gaussian characteristic, and calculates the conversion characteristic so that both are equal. The conversion means 28 converts the Gaussian random signal into a non-Gaussian random signal based on this conversion characteristic.

2.2 Hardware Configuration The hardware configuration of the vibration control apparatus according to the second embodiment is the same as that shown in FIG.

2.3 Control Processing FIG. 13 shows a flowchart of the control program 44. First, the DSP 40 acquires a response signal from the A / D converter 8 and generates a control spectrum (step S1). Next, a non-Gaussian random signal having the frequency characteristic of the control spectrum is generated (step S2). Next, the non-Gaussian characteristic of the response signal is calculated, and the conversion characteristic when generating the non-Gaussian random signal is corrected (step S25).

Subsequently, waveform control is performed based on the non-Gaussian random signal, and vibration is given to the specimen 4 (step S3). Such control is repeated, and vibration having a target spectrum and a target non-Gaussian characteristic is given to the specimen 4.

  Steps S1, S2, and S3 are the same as the control in the first embodiment. Details of the modification of the conversion characteristics (step S25) are shown in FIG.

  The DSP 40 acquires a response signal for one frame from the A / D converter 8 (step S251). Subsequently, non-Gaussian characteristics (for example, skewness S and kurtosis K) of the response signal are calculated (step S252). The skewness S and the kurtosis K can be calculated by Equation 5.

Also, Z3 and Z4 can be calculated by Equation 6.

The average value X and the standard deviation σ appearing in Equation 6 can be calculated by Equations 7 and 8.

  Next, the DSP 40 compares whether or not the calculated response skewness Sr and response kurtosis Kr are equal to the target skewness So and the target kurtosis Ko (see FIG. 7), and performs feedback control by changing the conversion characteristics so that both approach each other. This is performed (step S253). Specifically, when the response skewness Sr (Cultosis Kr) is small, control is performed so that S in Equations 1 and 2 is increased. Conversely, when the response skewness Sr (Cultosis Kr) is large, control is performed so that S in Equations 1 and 2 is reduced.

As described above, the non-Gaussian characteristic can be controlled with higher accuracy by controlling the conversion characteristic in real time.

3. Other embodiments
(1) In the above embodiment, a non-Gaussian random signal is obtained from a Gaussian random signal using a ZMNL function. When conversion by the ZMNL function is performed, the frequency characteristic changes. Of course, since this change appears in the response spectrum, the control spectrum calculation means 22 performs feedback control including this change to realize the target spectrum.

  However, since the conversion by the ZMNL function is a non-linear conversion, an error that cannot be captured by feedback control occurs. For example, the target spectrum and the response spectrum may not match in the high frequency component.

  Therefore, as another embodiment, after generating a Gaussian random signal, the Gaussian random signal may not be converted into a non-Gaussian random signal, but a non-Gaussian random signal may be generated directly from the control spectrum. .

  In the above embodiment, the drive signal is generated by sequentially giving a random phase θ to each line of the control spectrum. However, a non-Gaussian random signal may be directly generated by sequentially randomizing the phase difference Δθ given to each line of the control spectrum. In this case, the kurtosis K of the generated non-Gaussian random signal can be controlled by adjusting the standard deviation of the probability density distribution (here, the normal distribution is used) of the phase difference Δθ.

  For reference, FIG. 15 shows the relationship between standard deviation and kurtosis. For example, when it is desired to obtain a non-Gaussian random signal with a kurtosis of 4.5, it is understood that the standard deviation of the probability density distribution of the phase difference Δθ should be set to 200.

  Further, the skewness S of the generated non-Gaussian random signal can be controlled by adjusting the initial phase of the probability density distribution of the phase difference Δθ by φ0.

(2) In the above embodiment, kurtosis and skewness are set as the target non-Gaussian characteristics. However, other characteristics may be used as the control characteristics. Also, the shape of the probability density distribution itself as shown in FIG. 21 may be the target non-Gaussian characteristic. Thereby, it becomes possible to reproduce the vibration according to the probability density distribution obtained from the measured data. In this case, if the probability density distribution is determined, the corresponding ZMNL function can be easily obtained.

  FIG. 18 describes the processing of the ZMNL function determination processing program in the case where the shape of the probability density distribution is given as the target non-Gaussian characteristic and the non-Gaussian vibration according to this probability density distribution is given.

In step S181, initial values are given to the coefficients (c0, c1, c2, c3) of the ZMNL function. Here, it is the ZMNL function is a function expressed by ^{y = c0 + c1x + c2x 2} + c3x 4. Next, y is obtained by substituting an appropriate vector variable x into the ZMNL function (step S182).

  Subsequently, a probability density distribution is obtained (step S183). That is, the probability density distribution of the non-Gaussian signal obtained by converting the Gaussian signal with the ZMNL function having the coefficient determined in step S181 is calculated.

  Next, the obtained probability density distribution is compared with the target probability density distribution, and the error ε is calculated (step S184). Here, the error ε is calculated as a sum of absolute values of errors in the respective amplitudes.

  Next, the coefficient is changed and step S181 and the subsequent steps are executed. It is determined whether the error ε at this time is smaller or larger than the previous time.

  Further, the coefficient is changed and step S181 and the subsequent steps are repeatedly executed to find a point where the error ε converges to a predetermined value or less. The ZMNL function to be determined is determined by the coefficient at this time.

(3) In the above embodiment, a continuous Gaussian random signal SGD is obtained by superimposing a plurality of frames of Gaussian random signals GD1 to GD4 by applying a window function, and this is converted into a non-Gaussian random signal SNGD. Furthermore, a drive signal SD is obtained by applying a window function to one frame of the non-Gaussian random signal SNGD and superimposing it with control characteristics.

  However, the processing may be performed as follows. A window function is applied to the Gaussian random signals GD1 to GD4 of a plurality of frames, and the Gaussian random signals of the plurality of frames are converted into non-Gaussian random signals to obtain non-Gaussian random signals NGD1 to NGD4. The drive signal SD may be obtained by superimposing the non-Gaussian random signals NGD1 to NGD4 with control characteristics. In this case, the window function may be applied immediately before superposition.

(4) In the above embodiment, a transfer function is used as the transfer characteristic, but an impulse response may be used.

(5) In the above embodiment, the waveform control means 30 controls the waveform of the non-Gaussian random signal and the response signal to be equal by updating the transfer characteristics. However, feedback control may be performed so that the waveforms of the non-Gaussian random signal and the response signal are equal.

(6) In the above embodiment, the response spectrum calculation unit 20 calculates the spectrum of the response signal having non-Gaussian characteristics, and the control spectrum calculation unit 22 calculates the control spectrum by comparing the response spectrum with the target spectrum. doing.

  Here, for calculating the response spectrum, a response signal having Gaussian property is preferable to a response signal having non-Gaussian property. Therefore, as shown in FIG. 16, a response signal having non-Gaussian characteristics is converted into a response signal having Gaussian characteristics using the inverse characteristic of the conversion characteristics used in the converting means 28, and the response signal having this Gaussian characteristics. The spectrum may be calculated.

  FIG. 16 is constructed based on the second embodiment, but can also be applied to the first embodiment.

(7) In the above embodiment, the control characteristics used in the drive signal generation means 32 are corrected during operation. However, by adding vibration due to white noise to the vibration tester 2 and obtaining the response signal, the transfer characteristics between the vibration tester 2 and the specimen 4 are measured in advance, and the control characteristics are calculated and corrected. You may make it use without. In this case, the control characteristic updating means 34 is not necessary, and the process of superimposing by multiplying by the window function is also unnecessary.

(8) In the above embodiment, when a non-Gaussian random signal of a plurality of frames is generated from one control spectrum, a different phase is given for each frame, that is, the intensity of each line of the control spectrum is set. When P1, P2, P3, P4... Pn, φ1, φ2, φ3, φ4... Φm are given as follows, and the first frame P1 ( φ1), P2 (φ2), P3 (φ3). . . Pn (φn)
Second frame P1 (φn + 1), P2 (φn + 2), P3 (φn + 3). . . Pn (φn + n)
3rd frame P1 (φ2n + 1), P2 (φ2n + 2), P3 (φ2n + 3). . . Pn (φ2n + n)
...
mth frame P1 (φn (m-1) +1), P2 (φn (m-1) +2), P3 (φn (m-1) +3). . . Pn (φn (m-1) + n)
As described above, by giving different phases for each frame, the frequency characteristics of the waveform superimposed with the window function become a continuous spectrum instead of a line spectrum.

However, it is also possible to obtain a signal having a line spectrum by giving the same phase to a plurality of frames, generating a non-Gaussian random signal of a plurality of frames, and superposing them by applying a window function. That is,
First frame P1 (φ1), P2 (φ2), P3 (φ3). . . Pn (φn)
Second frame P1 (φ1), P2 (φ2), P3 (φ3). . . Pn (φn)
3rd frame P1 (φ1), P2 (φ2), P3 (φ3). . . Pn (φn)
...
mth frame P1 (φ1), P2 (φ2), P3 (φ3). . . Pn (φn)
In other words, when the same non-Gaussian random signal is generated for a plurality of frames and superimposed on the window function, a signal having a line spectrum can be obtained.

  Usually, a waveform having a continuous spectrum is preferable, but a line spectrum can be used depending on the purpose of the test.

(9) In the above embodiment, control processing is performed using a DSP. However, it may be executed using a CPU.

(10) In the above embodiment, random signals of a plurality of frames are generated from one control spectrum by giving different phases.

  However, as shown in FIG. 17, a random signal of a plurality of frames may be generated from one control spectrum by shifting the start position of the waveform.

  Specifically, one frame of random signal W1 shown in FIG. 17A is generated from one control spectrum. Next, the start position is shifted cyclically to T2. That is, when reading to the end of one frame, returning to the beginning and continuing reading. In this way, the random signal W2 shown in FIG. 17B is obtained. Similarly, the reading start position is changed at random, such as T3, T4,..., And random signals W3, W4,.

(11) In the above embodiment, the case where vibration in one direction (X-axis direction) is applied has been described. However, the present invention can also be applied to the case where vibration is applied in multiaxial directions (for example, X axis, Y axis, Z axis orthogonal to each other). In this case, the above-described control circuit may be provided for each vibration generator that generates vibration in each axial direction. FIG. 19 is a functional block diagram of a vibration control apparatus that controls the vibration testing machine 2 that applies vibrations in two axial directions of the X axis and the Z axis. A / D converters 8Z, 8X, response spectrum calculation means 20X, 20Z, control spectrum calculation means 22X, 22Z, non-Gaussian random signal generation means 24X, 24Z, non-Gaussian characteristic calculation means 31X, 31Z, conversion characteristic calculation means 33X , 33Z are provided for the X-axis and the Z-axis, respectively.

  When generating a drive signal based on a non-Gaussian random signal, it is necessary to correct crosstalk (mutual interference) between the axes. For this reason, when the drive signal generating means 32 receives the non-Gaussian random signal for X-axis and the non-Gaussian random signal for Z-axis and generates the X-axis drive signal and the Z-axis drive signal, The matrix control characteristics are used. That is, the control characteristic XX for the drive signal for the X-axis and the response signal for the X-axis, the control characteristic ZX for the drive signal for the Z-axis and the response signal for the X-axis, the drive signal for the Z-axis and the response for the Z-axis The non-Gaussian random signal generating means is converted into a drive signal by using the control characteristic ZZ for the signal, the control characteristic ZX for the Z-axis drive signal and the response signal for the X-axis.

  Here, H is a 2 × 2 transfer function matrix. G is a control characteristic and is a matrix corresponding to a 2 × 2 inverse transfer function.

  Specifically, the control characteristics XX and ZX are used to generate the X-axis drive signal, and the control characteristics ZZ and XZ are used to generate the Z-axis drive signal (see JP-A-10-105252). ).

Thus, by treating the transfer function as a matrix, non-Gaussian random control can be easily extended to a multi-axis / multi-point vibration test.

Claims (12)

A vibration control device that controls a drive signal applied to a vibration tester so that a specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic,
A response spectrum calculating means for converting a response signal from a sensor for detecting vibration of the specimen into a frequency function and obtaining a response spectrum;
A control spectrum calculating means for comparing the response spectrum with the target spectrum and calculating a control spectrum such that both are equal;
A non-Gaussian random signal generating means for converting a control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal;
A drive signal that obtains a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic so that the non-Gaussian random signal and the response signal are equal. Generating means;
Control characteristic updating means for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and updating the control characteristic;
A vibration control device. A vibration control device that controls a drive signal applied to a vibration tester so that a specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic,
A response spectrum calculating means for converting a response signal from a sensor for detecting vibration of the specimen into a frequency function and obtaining a response spectrum;
A control spectrum calculating means for comparing the response spectrum with the target spectrum and calculating a control spectrum such that both are equal;
A non-Gaussian random signal generating means for converting a control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal;
Non-Gaussian characteristic calculating means for calculating a response non-Gaussian characteristic of a response signal from the sensor;
A conversion characteristic calculating means for comparing the response non-Gaussian characteristic with the target non-Gaussian characteristic and calculating a conversion characteristic that makes both equal;
A drive signal that obtains a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic so that the non-Gaussian random signal and the response signal are equal. Generating means;
Control characteristic updating means for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and updating the control characteristic;
A vibration control device. Control for realizing by a computer a vibration control device that controls the drive signal applied to the vibration tester so that the specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic. A program for controlling the computer,
A response spectrum calculating means for converting a response signal from a sensor for detecting vibration of the specimen into a frequency function and obtaining a response spectrum;
A control spectrum calculating means for comparing the response spectrum with the target spectrum and calculating a control spectrum such that both are equal;
A non-Gaussian random signal generating means for converting a control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal;
A drive signal that obtains a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic so that the non-Gaussian random signal and the response signal are equal. Generating means;
A control program for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and functioning as control characteristic update means for updating the control characteristic. Control for realizing by a computer a vibration control device that controls the drive signal applied to the vibration tester so that the specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic. A program for controlling the computer,
A response spectrum calculating means for converting a response signal from a sensor for detecting vibration of the specimen into a frequency function and obtaining a response spectrum;
A control spectrum calculating means for comparing the response spectrum with the target spectrum and calculating a control spectrum such that both are equal;
A non-Gaussian random signal generating means for converting a control spectrum into a time function based on the conversion characteristics and obtaining a non-Gaussian random signal;
Non-Gaussian characteristic calculating means for calculating a response non-Gaussian characteristic of a response signal from the sensor;
A conversion characteristic calculating means for comparing the response non-Gaussian characteristic with the target non-Gaussian characteristic and calculating a conversion characteristic that makes both equal;
A drive signal that obtains a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic so that the non-Gaussian random signal and the response signal are equal. Generating means;
A control program for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and functioning as control characteristic update means for updating the control characteristic. In the vibration control apparatus or control program according to any one of claims 1 to 4,
The non-Gaussian random signal generating means includes
A Gaussian random signal generating means for converting a control spectrum into a time function and obtaining a Gaussian random signal;
Conversion means for converting the calculated Gaussian random signal based on the conversion characteristics and converting it into a non-Gaussian random signal;
A vibration control apparatus or a control program comprising: In the vibration control apparatus or control program according to claim 5,
The response spectrum generating means includes
Non-Gauss-Gauss conversion means for converting a response signal from a sensor for detecting vibration of the specimen into a Gaussian response signal based on the conversion characteristics of the conversion means;
A Gaussian response spectrum generating means for converting the Gaussian response signal into a frequency function to obtain a Gaussian response spectrum;
A vibration control apparatus or a control program comprising: In the vibration control apparatus or control program according to any one of claims 1 to 4,
The non-Gaussian random signal generating means is generated by controlling the probability density distribution of the phase difference sequentially given to each component of the control spectrum as the conversion characteristic when converting to a time function based on the control spectrum. A vibration control apparatus or control program for controlling a non-Gaussian characteristic of a non-Gaussian random signal. The vibration control apparatus or control program according to claim 7,
The non-Gaussian random signal generation means controls a non-Gaussian characteristic of a generated non-Gaussian random signal by controlling at least an average or variance of the probability density distribution of the phase difference as the conversion characteristic. A vibration control device or control program. A vibration control device or a control program according to any one of claims 1 to 8,
The vibration testing machine has a plurality of vibration generators for applying vibrations from a plurality of axial directions to the specimen.
The response spectrum calculating means, the control spectrum calculating means, and the non-Gaussian random signal generating means are provided for each of a plurality of axes;
The control characteristic updating means is configured to control characteristics related to a combination of the response signal for each axis and the drive signal for each axis based on the response signal for each axis and the drive signal for each axis. And
Said drive signal generating means, the non-Gaussian random signal based on the control characteristic according to the combination, the vibration control apparatus or a control program and generates a drive signal for each axis for each axis. A vibration control method for controlling a drive signal applied to a vibration tester so that a specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic,
Convert the response signal from the sensor that detects the vibration of the specimen into a frequency function, obtain the response spectrum,
Compare the response spectrum with the target spectrum, calculate the control spectrum so that both are equal,
Based on the conversion characteristics, the control spectrum is converted into a time function to obtain a non-Gaussian random signal,
In order to make the non-Gaussian random signal and the response signal equal, the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen is used as a control characteristic, and the non-Gaussian random signal is transformed to obtain a drive signal.
A vibration control method for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and updating the control characteristic. A vibration control method for controlling a drive signal applied to a vibration tester so that a specimen placed on the vibration tester vibrates with a waveform having a target spectrum and a target non-Gaussian characteristic,
Convert the response signal from the sensor that detects the vibration of the specimen into a frequency function, obtain the response spectrum,
Compare the response spectrum with the target spectrum, calculate the control spectrum so that both are equal,
Based on the conversion characteristics, the control spectrum is converted into a time function to obtain a non-Gaussian random signal,
Calculating a response non-Gaussian characteristic of the response signal from the sensor;
Compare the response non-Gaussian characteristics with the target non-Gaussian characteristics, calculate the conversion characteristics so that they are equal,
In order to make the non-Gaussian random signal and the response signal equal, the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen is used as a control characteristic, and the non-Gaussian random signal is transformed to obtain a drive signal.
A vibration control method for calculating the transfer characteristic based on a drive signal and a response signal from the sensor and updating the control characteristic. A vibration control device for controlling a drive signal applied to a vibration tester so that a specimen placed on the vibration tester vibrates with a waveform according to a target spectrum and a desired probability density distribution as a target. ,
A response spectrum calculating means for converting a response signal from a sensor for detecting vibration of the specimen into a frequency function and obtaining a response spectrum;
A control spectrum calculating means for comparing the response spectrum with the target spectrum and calculating a control spectrum such that both are equal;
A Gaussian random signal generating means for converting a control spectrum into a time function and obtaining a Gaussian random signal having a probability density distribution by a Gaussian distribution;
Non-Gaussian random signal generating means for converting a Gaussian random signal having a probability density distribution by the Gaussian distribution into a non-Gaussian random signal having an arbitrary target probability density distribution based on a ZMNL function;
Drive that obtains a drive signal by transforming a non-Gaussian random signal using the inverse characteristic of the transfer characteristic of the system including the vibration tester and the specimen as a control characteristic so that the non-Gaussian random signal and the response signal are equal. Signal generating means;
A vibration control device.

Images