A control method for stationary non-Gaussian random vibration test

Abstract

The present invention discloses a control method for a stationary non-Gaussian random vibration test, which comprises the following steps: S1, defining target power spectrum density, slope and kurtosis; S2, designing a filter and calculating relevant parameters; S3, transforming a Gaussian random signal into a non-Gaussian random signal, and converting the non-Gaussian random signal into a reference signal according to the filter; S4, calculating a time-domain driving signal according to the reference signal, and sending the time-domain driving signal to a vibration test system for the vibration test; S5, collecting a control signal of the vibration test, and estimating the power spectrum density, slope and kurtosis of the control signal; S6, calculating an error matrix between a control spectrum and a target spectrum, and correcting the driving signal according to the error matrix.

Description

Technical Field

[0001] The present invention relates to the technical field of random vibration testing, in particular to a control method for a steady non-Gaussian random vibration testing. Background Art

[0002] Random vibration environmental testing is an effective means of assessing the structural integrity, functionality, and reliability of equipment in the laboratory. Currently, random vibration testing primarily simulates a stationary Gaussian random vibration environment, where the controlled vibration signal follows a Gaussian distribution. With the rapid advancement of computer technology and the increasing demand for precision in environmental testing, random vibration testing technology is gradually evolving toward non-Gaussian and non-stationary vibrations. Many vibrations measured in the field are strongly non-Gaussian, such as the vibrations experienced by missiles when missile launchers maneuver on uneven land surfaces, or the vibrations experienced by onboard weapons and electronic equipment on warships during sea cruises in inclement weather. To improve the accuracy and reliability of random vibration testing in the laboratory and to simulate the vibration environment to which equipment will be subjected in actual service as realistically as possible, the development of control technology for stationary non-Gaussian random vibration testing is of paramount importance and significance. Summary of the Invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention implements a method for simultaneously controlling power spectral density, slope and kurtosis in a stationary non-Gaussian random vibration test.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] The present invention is a control method for a stationary non-Gaussian random vibration test, the control method comprising the following contents:

[0006] S1, defines the target power spectrum density, slope and kurtosis;

[0007] S2, design the filter and calculate the relevant parameters;

[0008] S3, transforming the Gaussian random signal into a non-Gaussian random signal, and converting the non-Gaussian random signal into a reference signal according to the filter;

[0009] S4, calculating a time domain driving signal according to the reference signal, and sending the time domain driving signal to a vibration test system for a vibration test;

[0010] S5, collecting a control signal of a vibration test, and estimating a power spectrum density, a slope, and a kurtosis of the control signal;

[0011] S6, calculating an error matrix between the control spectrum and the target spectrum, and correcting the driving signal according to the error matrix.

[0012] Furthermore, the S1 is:

[0013] Set the target power spectral density S t , slope W t and kurtosis K t , then the control objective is expressed as:

[0014] S y =S t ,W y =W t ,K y =K t

[0015] Among them, S y 、W y and K y They represent the control spectrum, control slope and control kurtosis respectively.

[0016] Furthermore, the S2 is:

[0017] S21, using the target spectral density to design a linear phase finite length unit impulse response filter and calculate the parameters; using the frequency sampling method, the designed filter is expressed as

[0018]

[0019] Where P is the filter length; parameter M satisfies P=2M+1; L0 and L k Calculated from the target power spectral density;

[0020] S22, calculate intermediate parameters Expressed as

[0021]

[0022] Furthermore, the S3 is:

[0023] The Gaussian random signal is transformed into a non-Gaussian random signal s(t) using the nonlinear transformation method. The nonlinear transformation function is expressed as

[0024] g(x)=(a+b)x+(a 2 -b 2 )x 2 +(a 3 +b 3 )x 3

[0025] Parameters a and b are obtained using nonlinear optimization methods; the reference slope value W used for nonlinear transformation is nt and kurtosis value K nt The calculation is as follows

[0026]

[0027] in, as well as The parameters are calculated using the formula in step 3.

[0028] Furthermore, the S4 is:

[0029] S41, the non-Gaussian random signal s(t) generated in step S3 is passed through the designed filter, that is, a non-Gaussian random signal that meets the set requirements is obtained, which is expressed as

[0030]

[0031] Where h is the filter; t is the time variable;

[0032] S42, perform RMS correction on the signal to obtain the reference signal u(t), which is expressed as

[0033]

[0034] Among them, r t is the target RMS value, which can be calculated from the target power spectral density. It's a signal The root mean square value of

[0035] S43, calculate the driving signal spectrum, expressed as

[0036] X(ω)=Z(ω)U(ω)

[0037] Among them, the transfer function is set before the experiment, Z(ω) is the inverse of the transfer function, and U(ω) is the Fourier transform spectrum of the reference signal u(t);

[0038] S44, performing an inverse Fourier transform on the driving signal spectrum to obtain a time domain driving signal, which is output by the data transmission acquisition system to the vibration test system.

[0039] Furthermore, the S5 is:

[0040] The acceleration sensor collects the vibration control signal and transmits it to the computer system through the input module of the data acquisition and transmission system; the power spectrum density, slope and kurtosis of the collected control signal are estimated.

[0041] Furthermore, the S6 is:

[0042] S61, calculate the error matrix between the control spectrum and the target spectrum, the error matrix between the control slope and the target slope, and the error matrix between the control kurtosis and the target kurtosis, which are recorded as:

[0043]

[0044] S62, judging the error matrix threshold, if the test control target is not reached, then go to step 12, otherwise exit the correction algorithm.

[0045] S63, spectrum correction algorithm, slope correction algorithm and kurtosis correction algorithm are defined as

[0046]

[0047] W (k+1) =α W E W +W (k)

[0048]

[0049] The superscript (k) indicates the number of iterations. When k = 0, that is, when the initial drive is generated, S (0) =S t , W (0) =W t and K (0) =K t , α S , α W and α K The ratio is the spectrum correction convergence factor, the slope correction convergence factor and the kurtosis correction convergence factor;

[0050] S64, substitute the iteratively corrected S, W and K back to step 2 to calculate the new time domain driving signal.

[0051] Beneficial effects

[0052] The stationary non-Gaussian random vibration test algorithm provided by the present invention can independently control the power spectrum density, slope and kurtosis simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flow chart of the control method for a stationary non-Gaussian random vibration test;

[0054] Figure 2 A schematic diagram of an experiment set up for an embodiment of the present invention;

[0055] Figure 3 This is a diagram showing the power spectrum density control effect of a steady non-Gaussian random vibration test according to an embodiment of the present invention;

[0056] Figure 4 This is a diagram showing the effect of slope control in a steady non-Gaussian random vibration test according to an embodiment of the present invention;

[0057] Figure 5 This is a diagram showing the kurtosis control effect of a stationary non-Gaussian random vibration test according to an embodiment of the present invention;

[0058] Figure 6 This is a time domain control signal segment of a stationary non-Gaussian random vibration test according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to embodiments and drawings. The contents mentioned in the embodiments are not intended to limit the present invention.

[0060] Reference Figure 1 As shown, the power spectrum density, slope and kurtosis decoupling control method of the present invention for the stationary non-Gaussian random vibration environment test is divided into the following steps:

[0061] 1. Set the target power spectrum density S t , slope W t and kurtosis K t , then the control objective is expressed as:

[0062] S y =S t ,W y =W t ,K y =K t

[0063] Among them, S y 、W y and K y They represent the control spectrum, control slope and control kurtosis respectively.

[0064] 2. Use the target spectral density to design a linear phase finite length unit impulse response filter and calculate the parameters. Using the frequency sampling method, the designed filter can be expressed as

[0065]

[0066] Where P is the filter length; parameter M satisfies P=2M+1; L0 and L k Calculated from the target power spectral density;

[0067] 3. Calculate intermediate parameters Expressed as

[0068]

[0069] 4. Use the nonlinear transformation method to transform the Gaussian random signal into a non-Gaussian random signal s(t). The nonlinear transformation function is expressed as

[0070] g(x)=(a+b)x+(a 2 -b 2 )x 2 +(a3 +b 3 )x 3

[0071] The parameters a and b are obtained by conventional nonlinear optimization methods and calculated based on common means; the reference slope value W used for nonlinear transformation is nt and kurtosis value K nt The calculation is as follows

[0072]

[0073] in, as well as The parameters are calculated using the formula in step 3.

[0074] 5. The non-Gaussian random signal s(t) generated in step 4 is passed through the designed filter to obtain a non-Gaussian random signal that meets the set requirements, which is expressed as

[0075]

[0076] Where h is the filter designed in step 2 and t is the time variable.

[0077] 6. Perform RMS correction on the signal to obtain the reference signal u(t), which is expressed as

[0078]

[0079] Among them, r t is the target RMS value, which can be calculated from the target power spectral density. It's a signal The root mean square value of .

[0080] 7. Calculate the driving signal spectrum, expressed as

[0081] X(ω)=Z(ω)U(ω)

[0082] The transfer function used is the one set before the experiment, Z(ω) is the inverse of the transfer function, and U(ω) is the Fourier transform spectrum of the reference signal u(t) in step 6.

[0083] 8. Perform inverse Fourier transform on the driving signal spectrum to obtain the time domain driving signal, which is expressed as

[0084]

[0085] in, The generated driving signal is output from the data acquisition system to the vibration test system.

[0086] 9. Use the accelerometer to collect the vibration control signal and transmit it to the computer system through the input module of the data acquisition and transmission system. Estimate the power spectrum density, slope, and kurtosis of the collected control signal.

[0087] 10. Calculate the error matrix between the control spectrum and the target spectrum, the error matrix between the control slope and the target slope, and the error matrix between the control kurtosis and the target kurtosis, which are recorded as:

[0088] E W =W t -W y ,

[0089] 11. Determine the error matrix threshold. If the test control target is not reached, go to step 12; otherwise, exit the correction algorithm.

[0090] 12. The spectrum correction algorithm, slope correction algorithm and kurtosis correction algorithm are defined as

[0091]

[0092] W (k+1) =α W E W +W (k)

[0093]

[0094] The superscript (k) indicates the number of iterations. When k = 0, that is, when the initial drive is generated, S (0) =S t , W (0) =W t and K (0) =K t α S , α W and α K The ratios are spectrum correction convergence factor, slope correction convergence factor and kurtosis correction convergence factor.

[0095] 13. Substitute the iteratively corrected S, W, and K back into step 2 to calculate the new time domain driving signal.

[0096] Figure 2 Shown is a schematic diagram of a non-Gaussian random vibration test according to an embodiment of the present invention.

[0097] In this example, the data acquisition and transmission system uses equipment from National Instruments. The controller module, output module, and input module are integrated into the PXIe-1071 chassis. The controller module connects to the computer and is responsible for sending and acquiring transmission instructions and data. The output module is a signal source module responsible for sending signals to the vibration table test system. The input module is an acquisition module responsible for receiving signals from the accelerometer.

[0098] In this example, a unidirectional vibration table system is used. The test piece is fixed to the vibration table with a fixture, and the acceleration vibration response signal of the test piece base is picked up by a PCB acceleration sensor.

[0099] The test mainly consists of the following steps:

[0100] 1. Set the target power spectrum density, slope and kurtosis and related test parameters for the random vibration test. In this example, the target power spectrum density is set as follows: Figure 3 The dashed line shows the slope and kurtosis, respectively, set to 0.2 and 6. The test frequency range is 20 Hz to 2000 Hz, and the full-scale test duration is 0.5 hours.

[0101] 2. Start the non-Gaussian random vibration control test. Generate the initial time-domain drive signal, transmit it to the power amplifier system through the output module, and finally load it onto the vibration table.

[0102] 3. Use the acceleration sensor to pick up the vibration control signal, collect it through the input module and transmit it to the computer system for analysis through the controller module, calculate the power spectrum density, slope and kurtosis of the control signal, and use the closed-loop control method to update and send the drive signal.

[0103] 4. Determine the error threshold of the control signal power spectrum. If the test control condition is not met, enter the iterative correction phase; otherwise, exit the correction algorithm.

[0104] 5. Determine whether the vibration test duration has been met. If so, end the test; otherwise, continue to execute the closed-loop vibration control program.

[0105] The power spectrum density, slope and kurtosis control effects of the stationary non-Gaussian random vibration test in this example are as follows: Figure 3 、 Figure 4 and Figure 5 As shown. Figure 3 It can be seen from the graph that the controlled spectrum density is within the ±3dB alarm limit of the target spectrum, and the control effect is satisfactory. Figure 4 and Figure 5 It can be seen that the slope and kurtosis of the control signal are stably controlled near the target value after two corrections. Figure 6 A fragment of a controlled stationary non-Gaussian random vibration signal is shown.

[0106] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Those skilled in the art may make various modifications and adjustments within the technical scope disclosed in the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. A control method for stationary non-Gaussian random vibration testing, characterized in that: The control method includes the following contents: S1, defines the target power spectrum density, slope and kurtosis; S2, design the filter and calculate the relevant parameters; S2 is: S21, using the target spectral density to design a linear phase finite length unit impulse response filter and calculate the parameters; using the frequency sampling method, the designed filter is expressed as Where P is the filter length; parameter M satisfies P=2M+1; L0 and L k Calculated from the target power spectral density; S22, calculate intermediate parameters Expressed as S3, transforming the Gaussian random signal into a non-Gaussian random signal, and converting the non-Gaussian random signal into a reference signal according to the filter; The S3 is: The Gaussian random signal is transformed into a non-Gaussian random signal s(t) by using the nonlinear transformation method. The nonlinear transformation function is expressed as g(x)=(a+b)x+(a 2 -b 2 )x 2 +(a 3 +b 3 )x 3 Parameters a and b are obtained using nonlinear optimization methods; the reference slope value W used for nonlinear transformation is nt and kurtosis value K nt The calculation is as follows in, as well as The parameters are calculated by the formula in step S22; S4, calculating a time domain driving signal according to the reference signal, and sending the time domain driving signal to a vibration test system for a vibration test; S5, collecting a control signal of a vibration test, and estimating a power spectrum density, a slope, and a kurtosis of the control signal; S6, calculating an error matrix between the control spectrum and the target spectrum, and correcting the driving signal according to the error matrix; The S6 is: S61, calculate the error matrix between the control spectrum and the target spectrum, the error matrix between the control slope and the target slope, and the error matrix between the control kurtosis and the target kurtosis, which are recorded as: HAVE BEEN W =W t -W y , S62, judging the error matrix threshold, if the test control target is not reached, proceeding to step S63, otherwise exiting the correction algorithm; S63, spectrum correction algorithm, slope correction algorithm and kurtosis correction algorithm are defined as W (k+1) =a W E W +W (k) The superscript (k) indicates the number of iterations. When k = 0, that is, when the initial drive is generated, S (0) =S t , W (0) =W t and K (0) =K t , α S , α W and α K The ratio is the spectrum correction convergence factor, the slope correction convergence factor and the kurtosis correction convergence factor; S64, the iteratively corrected S, W and K are substituted back into step S4 to calculate a new time domain driving signal.

2. A control method for stationary non-Gaussian random vibration test according to claim 1, characterized in that: Said S1 is: Set the target power spectral density S t , slope W t and kurtosis K t , then the control objective is expressed as: S y =S t ,W y =W t ,K y =K t Among them, S y 、W y and K y They represent the control spectrum, control slope and control kurtosis respectively.

3. A control method for stationary non-Gaussian random vibration testing according to claim 1, characterized in that: The S4 is: S41, the non-Gaussian random signal s(t) generated in step S3 is passed through the designed filter, that is, a non-Gaussian random signal that meets the set requirements is obtained, which is expressed as Where h is the filter; t is the time variable; S42, perform RMS correction on the signal to obtain the reference signal u(t), which is expressed as Among them, r t is the target RMS value, calculated from the target power spectral density, It's a signal The root mean square value of S43, calculate the driving signal spectrum, expressed as X(ω)=Z(ω)U(ω) Among them, the transfer function is set before the experiment, Z(ω) is the inverse of the transfer function, and U(ω) is the Fourier transform spectrum of the reference signal u(t); S44, performing an inverse Fourier transform on the driving signal spectrum to obtain a time domain driving signal, which is output by the data transmission acquisition system to the vibration test system.

4. A control method for stationary non-Gaussian random vibration testing according to claim 1, characterized in that: The S5 is: The acceleration sensor collects the vibration control signal and transmits it to the computer system through the input module of the data acquisition and transmission system; the power spectrum density, slope and kurtosis of the collected control signal are estimated.

Images