Test device

Abstract

To provide a test device for reducing a distortion rate without calculating a transfer function.SOLUTION: A test device 100 includes: a specimen setting unit 9; actuators 3 to 7 for vibrating the same; sensors S04, S06 for detecting a response of the actuators 3 to 7 and the specimen 10; a controller 2 for controlling the response of the actuators 3 to 7; and a user interface 1 having a target waveform setting unit 1b for setting an operation pattern to present test operation by operation of an operator. The controller 2 includes control command calculation units 12a, 12b for calculating a control input for eliminating difference between a target waveform and sensors S04, S06 and an experimental data analysis unit 2c for calculating a parameter. The experimental data analysis unit 2c determines a transfer function of an ideal response according to the target waveform, and adjusts a parameter of the control command calculation units 12a, 12b so that an actual response approaches the ideal response by response data of the sensors S04, S06. The determined transfer function has filter characteristics of a cutoff frequency higher than reference frequencies of the target waveform.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a test device for vibrating and evaluating a test object. [Background technology]

[0002] Test equipment is known that evaluates the seismic performance of structures by vibrating a model that mimics the characteristics of the structure (hereinafter referred to as a test specimen). These test equipment must reproduce test patterns that have a fast response in displacement and acceleration, such as earthquake waveforms, so hydraulic actuators with excellent response and driving force are often used.

[0003] It is known that the response of an actual vibration test system contains harmonic components in addition to the fundamental frequency of the excitation command value due to nonlinear elements contained in the hydraulic actuator and the test specimen. The more harmonic components there are, the lower the accuracy of reproducing the excitation command value becomes. The effect of such harmonic components is evaluated by the "distortion rate" in Equation 1.

[0004]

number

[0005] E1 is the effective value of the fundamental waveform, and Ei (i=2, ..., n) is the effective value of the harmonic (i times the frequency of the fundamental waveform). Reducing the distortion rate is used as an important index for increasing the accuracy of reproducing the vibration command value.

[0006] The actuators used in vibration testing equipment are equipped with sensors that acquire the actuator's position, speed, and acceleration. These signals are used to perform feedback control to reproduce the desired test waveform.

[0007] In general, feedback control is often based on PID control.

[0008] However, parameter adjustment for PID control requires know-how, and if the parameter adjustment is insufficient, sufficient control performance cannot be achieved.

[0009] Furthermore, since PID control is a linear control, even if the parameters are sufficiently adjusted, it is not necessarily possible to sufficiently suppress the effects of nonlinear elements that cause nonlinear distortion.

[0010] In response to such problems, Patent Document 1 proposes a test device having a control configuration that suppresses the effects of nonlinear distortion by adding a distortion correction element section to normal PID control. [Prior art documents] [Patent documents]

[0011] [Patent Document 1] JP 2001-242055 A Summary of the Invention [Problem to be solved by the invention]

[0012] The test equipment in Patent Document 1 is configured to calculate the transfer function and inverse transfer function of the test equipment, and to use a filter having these transfer functions to apply the filter to the difference between the vibration command value and the actual test equipment response to extract the influence of nonlinear elements, and to add this to the input to PID control, thereby effectively suppressing the influence of nonlinear elements.

[0013] In Patent Document 1, the transfer function and inverse transfer function of a test device are required to design a compensator. To accurately calculate the target transfer function, the input signal and the output signal need to include a variety of frequencies.

[0014] For this reason, random excitation using a random signal containing a variety of frequency components or impulse excitation using an impulse signal is generally performed.

[0015] Random signals are suitable for experiments to identify transfer functions because they generate vibration command values ​​by adding up various frequency components. However, because random signals contain significant signal changes, there is a risk of damaging the test specimen in experiments to identify transfer functions. For this reason, it is not desirable to perform random vibration when adjusting the control of the test equipment.

[0016] Since the transfer function is the Laplace transform of the impass response, which is the response obtained when impulse excitation is performed, it is theoretically possible to calculate the transfer function by performing impulse excitation.

[0017] However, even if impulse vibration is actually applied to the test equipment, it is known that if the input signal level is kept low from the standpoint of protecting the test specimen, the response of the test equipment will be slow, and a sufficient response will not be obtained, making it impossible to accurately calculate the transfer function.

[0018] As described above, in order to utilize Patent Document 1, it is necessary to derive the transfer function, but it is often impossible to obtain an accurate transfer function due to the operation of the test equipment. Furthermore, with the method of Patent Document 1, if an accurate transfer function cannot be calculated, it cannot be expected to obtain a sufficient compensation effect.

[0019] The present invention has been devised to solve the above problems, and its object is to provide a test apparatus having a control adjustment function that effectively reduces the distortion rate without calculating the transfer function of the test apparatus. [Means for solving the problem]

[0020] In order to achieve the above object, the present invention is configured as follows.

[0021] The test apparatus comprises a specimen mounting section on which a specimen to be evaluated is placed, an actuator for vibrating the specimen mounting section, a sensor for detecting responses of the actuator and the specimen, a controller for controlling the response of the actuator, and a user interface having a target waveform setting section for setting an operating pattern of the test apparatus, accepting operations from an operator and presenting test operations, wherein the controller comprises a control command calculation section for calculating a control input so as to eliminate a difference between the target waveform and the response waveform of the sensor, and an experimental data analysis section for calculating parameters of the control command calculation section based on experimental data, wherein the experimental data analysis section determines a transfer function of an ideal response of a closed loop system according to the target waveform and adjusts the parameters of the control command calculation section based on response data acquired by the sensor so that an actual response approaches the ideal response, and the determined transfer function has low-pass filter characteristics with a cutoff frequency higher than a reference frequency of the target waveform so as to suppress the effects of harmonics. Effect of the Invention

[0022] It is possible to provide a test device having a control adjustment function that effectively reduces the distortion rate without calculating the transfer function of the test device.

[0023] This makes it possible to execute control to reduce the distortion rate without performing vibration experiments to identify the transfer function. Furthermore, it is possible to reduce the man-hours required for adjusting the control parameters. [Brief description of the drawings]

[0024] [Figure 1] 1 is a schematic view of a hydraulically driven vibration testing device to which the present invention is applied; [Diagram 2] FIG. 1 is a functional block diagram according to a first embodiment of the present invention. [Figure 3A] FIG. 4 is a diagram showing a target waveform according to the present invention. [Figure 3B] FIG. 13 is a diagram showing an example of a target waveform according to the present invention and its relationship with an actuator response. [Figure 3C]FIG. 1 is a diagram showing an example of the relationship between a target waveform and a specimen response according to the present invention. [Figure 4A] FIG. 2 is a block diagram of a controller according to the present invention. [Figure 4B] FIG. 2 is a block diagram of a controller according to the present invention. [Figure 4C] FIG. 2 is a block diagram of a controller according to the present invention. [Diagram 5] FIG. 11 is a functional block diagram of a modified example of the first embodiment of the present invention. [Figure 6A] FIG. 2 is a block diagram of a controller according to the present invention. [Figure 6B] FIG. 2 is a block diagram of a controller according to the present invention. [Figure 7A] 1A and 1B are diagrams illustrating design guidelines for an ideal response for suppressing the effects of harmonics. [Figure 7B] 1A and 1B are diagrams for explaining design guidelines for an ideal response for suppressing the effects of harmonics. [Figure 8] FIG. 13 is a diagram illustrating an example of the influence of harmonics. [Figure 9] FIG. 11 is a diagram illustrating an example of a harmonic suppression effect achieved by a notch filter. [Figure 10] FIG. 1 is a diagram showing a design example of an ideal response used in the present invention. [Figure 11] FIG. 2 is a diagram showing a processing procedure (flowchart) according to the first embodiment of the present invention. [Figure 12] FIG. 11 is a functional block diagram according to a second embodiment of the present invention. [Figure 13] FIG. 11 is a block diagram of a controller according to a second embodiment of the present invention. [Figure 14] FIG. 11 is a functional block diagram according to a third embodiment of the present invention. [Figure 15] FIG. 13 is a functional block diagram of a modified example of the third embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0026] In this embodiment, for the sake of simplicity, a one-axis hydraulically driven vibration testing apparatus will be described as an example, but the present invention is not limited to a one-axis hydraulically driven vibration testing apparatus. For example, the present invention can be applied to an electric linear actuator that is driven by a hydraulic piston instead of a hydraulic piston. Furthermore, the scope of application of the present invention is not limited to vibration testing apparatuses. For example, it should be noted that the present invention can be applied to a centrifugal load testing apparatus that generates centrifugal acceleration by rotational motion. EXAMPLES

[0027] Example 1 FIG. 1 is a diagram illustrating a hydraulically driven testing device 100 according to a first embodiment of the present invention.

[0028] 1, a user interface 1 is used by a user or an operator of a test apparatus 100 to set various test patterns. The operator sets a desired excitation waveform via the user interface 1.

[0029] Since the user interface 1 is a device that generates a target signal, it is not limited to a dedicated terminal, and may be composed of multiple devices such as a combination of an operation terminal and a signal generator. The operation terminal may be a normal personal computer (PC) or a tablet PC.

[0030] The dimension of the excitation waveform set in the user interface 1 may be any of displacement, velocity, and acceleration. The operator designs the excitation waveform r=A sin(ωt) by setting the amplitude A and angular frequency ω of the waveform. t means time.

[0031] A terminal (signal generator or PC) that generates a target signal for testing corresponds to the target waveform setting unit in the present invention.

[0032] The controller 2 executes various control calculations for operating the hydraulic vibration testing device 100 in accordance with the target waveform set by the user interface 1. The controller 2 acquires the target excitation waveform set by the user interface 1 and values ​​of various sensors S01 to S06, which will be described later, and calculates the operation amount of the servo amplifier 3. The detailed configuration of the controller 2 will be described later.

[0033] The servo amplifier 3 converts the command (=voltage) output from the controller 2 into a current value for driving the servo valve 4.

[0034] The servo valve 4 adjusts the pressure oil flowing from the hydraulic power source 5 to the hydraulic cylinder 6 by opening and closing the valve in accordance with the current value received from the servo valve 3. A temperature sensor S01 is provided between the servo valve 4 and the hydraulic power source 5 to detect the temperature of the pressure oil.

[0035] The pressure oil supplied to the hydraulic cylinder 6 via the servo valve 4 drives the hydraulic piston 7. At this time, the driving direction of the hydraulic piston 7 is changed depending on whether the pressure oil is supplied from the left or right port of the servo valve 4 in FIG. 1. The pressure oil discharged from the servo valve 4 can be detected by the flow sensors S02a and S02b. Since the path through which the pressure oil flows changes depending on the driving direction of the hydraulic piston 7, the flow sensor S02a and the flow sensor S02b to be used may be changed as appropriate depending on the driving direction.

[0036] The hydraulic piston 7 applies a force to a table (specimen mounting portion) 9 via a coupling 8, thereby vibrating the table 9. The hydraulic piston 7 is equipped with a speed sensor S03 and a displacement sensor S04. It is not necessary to provide both the speed sensor S03 and the displacement sensor S04. For example, when only the displacement sensor S04 is provided, the difference between the detected values ​​may be used as an approximation of the speed. Similarly, when only the speed sensor S03 is provided, the integral value may be used as the amount of displacement. Also, an acceleration sensor may be provided instead of or in addition to the speed sensor S03 and the displacement sensor S04. These sensors for detecting the displacement of the hydraulic piston 7 correspond to the internal sensors of the present invention.

[0037] The hydraulic cylinder 6 is provided with pressure sensors S05a and S05b for detecting the front and rear pressures of the hydraulic piston 7.

[0038] As described above, the hydraulic actuator including the components from the servo amplifier 3 to the hydraulic piston 7 corresponds to the actuator of the present invention.

[0039] A structure to be tested (specimen 10) is attached to the table 9, and various evaluations are performed by exciting the specimen 10 according to the target waveform set by the user interface 1. The table 9 corresponds to the specimen installation section in this invention.

[0040] There are two possible operating methods for the test device 100: one in which the table 9 is vibrated with a target waveform, and one in which the specimen 10 is vibrated with a target waveform. The hydraulic actuator and table 9 are connected with a high-rigidity coupling 8, and therefore behave in the same way. Similarly, if the specimen 10 has extremely high rigidity, it is expected that the vibration of the table 9 and the vibration of the specimen 10 will match. However, if the specimen 10 is flexible, it is desirable to have a configuration in which the vibration of the specimen 10 is controlled rather than that of the table 9.

[0041] Therefore, when configuring control such that the response of the specimen 10 follows a target waveform, it is necessary to prepare a non-contact sensor S06 in order to measure the behavior (displacement, speed, acceleration) of the specimen 10. As the non-contact sensor S06, for example, a laser displacement meter can be used.

[0042] Output signals from the sensors S01 to S05 attached to the test device 100 and an output signal from a sensor S06 that measures the behavior of the specimen 10 are all input to the controller 2.

[0043] The features of the first embodiment of the present invention will be described in detail below through some specific examples.

[0044] For the sake of simplicity, the term "response" will be used without specifying the dimension of the excitation waveform. It goes without saying that the scope of the present invention includes any of the following "response": displacement, velocity, and acceleration.

[0045] Specific components of the user interface 1 and the controller 2 in the first embodiment of the present invention will be described with reference to Fig. 2. Note that in Fig. 2, the components included in Fig. 1 are omitted in order to make it easier to understand the contents of the invention.

[0046] <User Interface 1> In FIG. 2, the user interface 1 is a GUI ( G rhetorical U ser I The system consists of a GUI 1a and a target waveform setting unit 1b. The GUI 1a corresponds to an operation terminal (such as a PC or tablet) where the operator inputs various information, and the target waveform setting unit 1b corresponds to a signal generator that generates test patterns. The GUI 1a is equipped with a monitor, which allows the operator to visually check the experiment results and the adjustment results of the controller, as described later.

[0047] <Controller 2> The functions of the controller 2 include a feedback control section 2a, a feedforward control section 2b, an experimental data analysis section 2c, and a response prediction section 2b.

[0048] The feedback control unit 2a calculates a feedback control input based on the output result of an internal sensor (corresponding to the displacement sensor S04) so ​​that the behavior of the actuator (corresponding to the servo valve 3 to the hydraulic piston 7) is stabilized. PID control or the like is used for the feedback control.

[0049] The feedforward control unit 2b calculates a feedforward control input based on the target waveform set by the target waveform setting unit 1b. The calculation of the feedforward control input will be described in detail later.

[0050] The actuator is applied with a command value obtained by adding the feedback control input and the feedforward control input. In other words, the feedback control 2a and the feedforward control 2b are added together in the adder 11 to form the control command calculation unit 12a (feedback control 2a, feedforward control 2b, adder 11) in the first embodiment of the present invention.

[0051] The experimental data analysis unit 2c adjusts the parameters of the above-mentioned feedforward control unit 2b and the parameters of the response prediction unit 2d, which will be described later, according to the actuator response (displacement sensor detection value) acquired by the internal sensor S04 and the specimen response (non-contact sensor detection value) acquired by the non-contact sensor S06, which is an external sensor. The experimental data analysis unit 2c adjusts the parameters of the transfer function of the feedback control unit 2a and the parameters of the transfer function of the feedforward control unit 2b.

[0052] The details of the calculations of the experimental data analysis unit 2c will be described later. It is not necessary for both the actuator response and the specimen response to be input to the experimental data analysis unit 2c. If the user of the test device 100 wishes to adjust the response of the test device 100, the actuator response is input, and if the user wishes to adjust the specimen response, the specimen response is input.

[0053] Although details will be described later, the experimental data analysis unit 2c has two functions: an ideal response design unit that analyzes the experimental data and sets the ideal response to be achieved, and an automatic parameter adjustment unit that adjusts the parameters of the feedback control 2a and feedforward control 2b so as to realize the ideal response.

[0054] The response prediction unit 2d calculates a response prediction value of the actuator or the specimen 10 according to the actuator response acquired by the displacement sensor S04, which is an internal sensor, or the specimen response acquired by the non-contact sensor S06, which is an external sensor, and the parameters calculated by the experimental data analysis unit 2c. As with the experimental data analysis unit 2c, the response prediction unit 2d only needs to receive an input of a signal corresponding to the adjustment target desired by the user of the testing device 100.

[0055] The response prediction unit 2d predicts the transfer function using the parameters calculated by the experimental data analysis unit 2c and the response data acquired by the sensors S04 and S06, thereby predicting the response waveform that will be obtained after the parameters are changed, and presents the prediction result on the user interface 1.

[0056] The feedforward control unit 2b, the experimental data analysis unit 2c, and the response prediction unit 2d in the controller 2 will be specifically described below with reference to several figures.

[0057] The two functional blocks, the feedforward control unit 2b and the response prediction unit 2d, are desirably not functioning during the initial excitation, and are preferably operated so that their respective functions become active only after test excitation is performed and the data on the specimen response and actuator response are stored and analyzed in the experiment data analysis unit 2c.

[0058] Assume that test excitation was performed with the target waveform 1bw (black solid line) as shown in Fig. 3A. At this time, the displacements of actuators 3 to 7 acquired by the displacement sensor S04, which is an internal sensor, are shown by the dashed line (Fig. 3B), and the specimen displacement acquired by the non-contact sensor S06, which is an external sensor, is shown by the dashed line (Fig. 3C).

[0059] The control objective of the test apparatus 100 in the first embodiment of the present invention is to match the target waveform 1bw with either the response of the actuator (the waveform of the displacement sensor S04 which is an internal sensor) or the response of the specimen 10 (the waveform of the non-contact sensor S06 which is an external sensor). It should be noted that when the stiffness of the specimen 10 is low, that is, when the specimen 10 has a response characteristic with respect to the response of table 9, the response of the actuator and the response of the specimen 10 do not match, and therefore it is not possible to match both the response of the actuator and the response of the specimen 10 with the target waveform 1b at the same time.

[0060] In general, the feedback control section 2a included in the controller 2 is used to match the response of the actuator (the response of the displacement sensor S04, which is an internal sensor) to the target waveform 1bw. In other words, the feedback control section 2a is used to match the two waveforms in FIG. 3B.

[0061] Therefore, it should be noted that this is not suitable for matching the response of the test piece 10 with the target waveform 1bw. For this reason, in the first embodiment of the present invention, a configuration is adopted in which a feedforward control 2b is used in addition to the feedback control 2a.

[0062] A specific method for designing the feedforward control section 2b will be described.

[0063] 4A is a control block diagram when the feedforward control section 2b is in an inactive state and only the feedback control section 2a is active. In FIG. 4A, C indicates the feedback control section 12a (in the figure, FB control: F eed Back) has a transfer function (having adjustable parameters), P is the transfer function of the actuators (3 to 7), and G is the transfer function of the test piece 10 to be controlled. Note that when adjusting the actuator response, it is not necessary to take the transfer function G into consideration. In such a case, G=1 should be set.

[0064] At this time, the transfer function from the target value r (corresponding to the target waveform 1bw) to the actuator response y (corresponding to the response of the displacement sensor S04, which is an internal sensor) can be given by Equation 2.

[0065]

number

[0066] Furthermore, the response z of the specimen 10 (corresponding to the response of the non-contact sensor S06, which is an external sensor) can be given from the response y of the actuator by Equation 3.

[0067]

number

[0068] From Equation 2 and Equation 3, the response from the target value r to the response z of the test piece 10 can be given by Equation 4.

[0069]

number

[0070] Note that, although it is necessary to express the above formulas 2 to 4 in a more precise manner using convolution integrals, they have been expressed in a way that prioritizes ease of understanding. Also, the same notation will be used hereafter.

[0071] FIG. 4B is a block diagram when the feedforward control unit 2b is enabled, in which F stands for feedforward control (in the figure, FF control: F eed F orward).

[0072] When feedforward control is effective, the transfer function from the target value r to the actuator displacement y can be expressed by Equation 5.

[0073]

Equation

[0074] Similar to before, the response from the target value r to the response z of the specimen 10 can be given by Equation 6.

[0075]

Equation

[0076] Here, when the response y of the actuator obtained when only feedback control is effective is denoted as y1, and the response y of the actuator obtained when feedforward control is added is denoted as y2, from Equations 2 and 5, the relational expression of Equation 7 can be obtained.

[0077]

Equation

[0078] Similarly, when the response z of the specimen 10 obtained when only feedback control is effective is denoted as z1, and the response z of the specimen 10 obtained when feedforward control is added is denoted as z2, from Equations 4 and 6, the relational expression of Equation 8 can be obtained.

[0079]

Equation

[0080] That is, if there are the responses y1 and z1 obtained when only feedback control is effective, the responses y2 and z2 after adding feedforward control can be calculated by Equations 7 and 8.

[0081] <When FF is effective> It should be noted that the above calculation is based on the premise that the experimental data is acquired in a state where the feedforward control is disabled, but the same processing can also be performed using experimental data in a state where the feedforward control is enabled.

[0082] If the response obtained when the feedforward control F1 is enabled is y1 and the response obtained when the feedforward control F2 is enabled is y2, it can be easily derived that the relationship between y1 and y2 can be expressed as in Equation 9. The same applies to the prediction of the response of the specimen 10.

[0083]

number

[0084] To adjust the feedforward control F, the feedforward control F is designed to minimize the evaluation function J of Formula 11 so that the difference between the responses ym, zm (Formulas 10a and 10b in Formula 10) for a reference model M expressing an ideal response as shown in Fig. 4C and the responses y2, z2 after adding the feedforward control is minimized. Note that N in Formula 11 is the number of data.

[0085]

number

[0086]

number

[0087] For example, when the feedforward control is defined as a transfer function having an adjustment parameter ρ by Equation 12, the adjustment parameter ρ can be calculated so as to minimize Equation 11.

[0088]

number

[0089] The above parameter adjustment function corresponds to the automatic parameter adjustment section of the present invention, and is provided as a function of the experiment data analysis section 2c.

[0090] Note that as the order m of Equation 12 increases, the range of feasible control operations increases, making it possible to realize waveforms closer to the ideal responses ym and zm; however, the time required for optimization calculations increases.

[0091] For this reason, it is desirable to set the order m to a small value and increase the order m only when the desired accuracy cannot be achieved.

[0092] Furthermore, according to formulas 7 to 9, the responses y2, z2 after the addition of the designed feedforward control can be predicted using the pre-designed feedback control C, the additionally designed feedforward control F or the feedforward controls F1, F2 before and after the change, and the actually measured responses y1, z1.

[0093] Using this prediction result, it is possible to calculate in advance whether the designed feedforward control F will achieve the desired response. It is desirable to have a mechanism that allows the operator to adjust the order m by presenting the results of this pre-calculation on the user interface 1.

[0094] As described above, an algorithm for designing the feedforward control unit 2b using the experimental data (actuator response and specimen response) is implemented in the experimental data analysis unit 2c.

[0095] <When adjusting feedback control as well> Up until now, it has been assumed that only the feedforward control is changed, but as shown in FIG. 5, it is also possible to change the feedback control section 2a by using the experimental data analysis section 2c.

[0096] When changing the feedback control, the feedback control is defined as a transfer function having an adjustment parameter θ by Equation 13, and the controller defined by the initial parameters is designated as C1 (shown in FIG. 6A).

[0097]

number

[0098] For example, if the initial feedback control is a PI control having a proportional gain Kp and an integral gain Ki, then it can be given by Equation 14.

[0099]

number

[0100] This is the state in which the adjustment parameters in Equation 13 are set as θ0=Kp, θ1=Ki, and θi=0 (i>1).

[0101] When the control input obtained under such conditions is u1 and the actuator response is y1, the pseudo reference signal can be obtained as shown in Equation 15.

[0102]

number

[0103] It is known that by using this pseudo reference signal to calculate the optimal parameter θ* that minimizes Equation 16, the response of the closed-loop system after applying the adjusted parameters (Equation 17) will match the ideal response M of Equation 10.

[0104]

number

[0105]

number

[0106] Furthermore, in adjusting the feedback control, in a situation where a disturbance d is applied as shown in FIG. 6B, it is also possible to calculate the parameter θ so as to suppress the disturbance d.

[0107] In addition, although both the feedback control 2a and the feedforward control 2b are configured to be adjustable in Fig. 5, this does not necessarily mean that both must be adjusted in order to use the system. In other words, only the feedback control 2a or only the feedforward control 2b may be adjusted, and this choice is left to the operator.

[0108] Both feedback control and feedforward control parameters can be designed using experimental data. In addition, there is no need to directly calculate the transfer functions of the controlled object (the actuator's transfer function P and the test piece's transfer function G), which has the advantage of eliminating the need for random excitation for identification.

[0109] However, the design method of the ideal response M can be realized by the present invention. In the present invention, a design method of the ideal response M capable of efficiently reducing the distortion rate is provided. The function of designing this ideal response M corresponds to the ideal response design unit of the present invention, and is executed in the experimental data analysis unit 2c.

[0110] <Specific design method for ideal response> As shown in Equation 1, the distortion factor is defined as the ratio of the effective value Ei (i ≧ 2) of the harmonic to the effective value E1 of the fundamental waveform. Therefore, the distortion factor can be suppressed by performing two operations: (1) increasing the effective value E1 of the fundamental waveform, which corresponds to the denominator, and (2) decreasing the effective value Ei of the harmonic, which corresponds to the numerator.

[0111] In the following explanation, the excitation waveform is assumed to be a sine wave r = Asin(ωt) with amplitude A and angular frequency ω. Note that angular frequency (rad / s) and frequency (Hz, 1 / s) have different units, but in the following explanation, for simplicity, they will not be distinguished and will simply be referred to as frequency.

[0112] To achieve the above item (1), the effective value at the fundamental frequency, i.e., ω, should be as large as possible. However, it is undesirable for the response amplitude to be greater than the amplitude A of the target excitation waveform. Therefore, it is desirable for the amplitude A of the ideal response waveform at frequency ω, i.e., the gain of the response value, to be 0 dB (1x).

[0113] Harmonics refer to frequency components that are integer multiples of the fundamental frequency ω. In other words, when an integer i (i ≧ 2) is used, harmonics affect the frequency i × ω. Therefore, it can be said that the effect of item (2) can be suppressed if the gain at frequencies 2ω or higher is 0 dB or less.

[0114] Therefore, as shown in Fig. 7A, if a low-pass filter with a cutoff frequency ωcut set between ω and 2ω is designed for ideal response M, the distortion rate can be effectively reduced. Such a low-pass filter can be designed to have the characteristics shown in Equation 18 in a first-order lag system using piecewise linear approximation.

[0115]

number

[0116] The analysis of effective values ​​uses the fast Fourier transform (FFT). F AST F Ourier T When using FFT, the components that are even multiples of the fundamental frequency become 0. In this case, the effective value E2 of the frequency 2ω does not affect the distortion rate, so a low-pass filter with a cutoff frequency ωcut set to less than 3ω (less than 3 times) as shown in Figure 7B can be used as the ideal response M.

[0117] 7A and 7B, the gain over the entire high frequency band can be reduced. However, the effective response of the test apparatus 100 may have large values ​​at odd harmonics, as shown in FIG.

[0118] Such large values ​​cannot be suppressed by a low-pass filter such as those shown in Figures 7A and 7B. In order to suppress the effects of such harmonics, it is desirable to use a notch filter to design an ideal response by suppressing the gain at a specific frequency. In other words, it is desirable for the transfer function to have a low-pass filter characteristic with a cutoff frequency higher than the reference frequency of the target waveform, and one or more notch filter characteristics with a band that is an integer multiple of the reference frequency.

[0119] The transfer function Ni of the notch filter can be given by Equation 19.

[0120]

number

[0121] In Equation 19, η is a parameter that adjusts the notch width, and d' is a parameter that adjusts the notch depth (gain). The smaller η is, the smaller the notch width becomes, and the smaller d' is, the greater the effect of suppressing the gain. These parameters can be designed by analyzing the waveform data obtained by applying vibration in a pre-test for control adjustment.

[0122] If a notch filter is applied to i×ω, it is expected that the effects of harmonics will be suppressed. The notch filter designed using Equation 19 has notch filter characteristics that suppress only a specific gain, as shown in Figure 9.

[0123] The Bode diagram of the transfer function obtained when using the product of each transfer function can be expressed by superposing each element. In other words, the Bode diagram obtained when the ideal response M is designed using Equation 20 is as shown in Figure 10.

[0124]

number

[0125] Note that Equation 20a is a simplified notation that takes advantage of the fact that the effective values ​​of even orders become 0 when FFT is used. As shown in Equation 20b, an ideal response may be designed to suppress the effects of all harmonics. Note that in Equation 20, Nn is a value that represents the number of notch filters.

[0126] It is believed that the distortion rate can be effectively reduced by utilizing the ideal response M described above and then designing the feedback control C and feedforward control F.

[0127] <Response prediction unit 2d> As mentioned above, if the actuator response y and the specimen response z are acquired while only feedback control is enabled, the response after adding feedforward control can be calculated using Equations 7 to 9. Furthermore, by analyzing this predicted value, it becomes possible to predict the effective value of each frequency after the control change.

[0128] Therefore, by utilizing this mechanism, it is possible to present a prediction of improvement in the distortion rate by showing a response prediction on the user interface before adding feedforward control.

[0129] <Runtime flow> A method for driving the test apparatus 100 realized by the above-described first embodiment of the present invention will be described with reference to the flowchart of FIG.

[0130] First, in step S1, an operator performs various settings (initial settings) in preparation for starting up the test apparatus 100. This corresponds to the operator operating the user interface 1 to set a test pattern, and by performing this operation, the test pattern is set in the target waveform setting unit 1b. The frequency ω of the target waveform set here becomes the fundamental frequency.

[0131] Furthermore, in the controller 2, at the timing when the operator sets a test pattern in the target waveform setting unit 1b, all of the coefficients ρi (i=0 to m) in the numerator part of the transfer function F (Formula 11) of the feedforward control unit 2b are set to 0. By setting in this way, the feedforward control unit 2b always outputs 0, becoming an ineffective state.

[0132] As mentioned above, the present invention can use data for which feedforward control is effective without any problem. However, when using such an operation, it is necessary to note that the operator must manually design the initial feedforward control.

[0133] 5, when the feedback control C (Formula 13) is also adjusted, step S1 also includes a process in which the operator designs a PID control such as Formula 14. When all the processes in step S1 are completed, the process proceeds to step S2.

[0134] In step S2, a vibration test is performed using a test pattern in the target waveform setting unit 1b. From the start of vibration, data output from the displacement sensor S04, which is an internal sensor, and the non-contact sensor S06, which is an external sensor, are started to be logged, and when the vibration pattern ends, the logging of data ends. This operation corresponds to step S3. When adjusting the feedback control, the control input u output from the controller 2 is also logged at the same time. The logging is performed in an internal memory (not shown) in the experiment data analysis unit 2c.

[0135] In step S4, FFT is performed using the data acquired during vibration. By performing FFT, it is possible to calculate the effective value as shown in Figure 8. This function is executed by the experimental data analysis unit 2c.

[0136] In step S5, a low-pass filter L (Formula 18) is determined based on the fundamental frequency ω set in step S1, and a notch filter Ni (Formula 19) is determined so as to correspond to the effective value calculated in step S4, and the ideal response M (Formula 20) is determined using these. This function is also executed by the experimental data analysis unit 2c.

[0137] In step S6, parameters for feedback control and feedforward control are adjusted using the ideal response determined in step 5. This function is also executed by the experimental data analysis unit 2c.

[0138] Execution of step S7 is optional, but if feedback control is adjusted in step S6, the prediction formula using the response prediction (Equations 7 and 8) will no longer hold. For this reason, if feedback control is adjusted (YES in step S7), it is recommended to return to step S2 to perform test excitation again. If feedback control is not adjusted in step S7 (NO) or the operator does not request the presentation of the prediction results, the process proceeds to step S8.

[0139] In step S8, response prediction (Equations 7 and 8) is performed using the experimental data and the feedforward control after the control change, and the result is presented on the user interface 1. The information presented on the user interface 1 is not limited to the graph display of the predicted waveforms y2 and z2 (Equation 9), but may also include numerical information such as the calculation result of the distortion factor using these.

[0140] Note that since response prediction is possible when only the feedforward control is changed, if the feedback control is adjusted and no additional vibration experiment is performed (if the feedback control is adjusted but NO is selected in step 7), a message indicating that response prediction is difficult will be displayed on the user interface 1. However, if response prediction is not possible, analysis results using ideal responses ym and zm that can be calculated using Equation 10 may be displayed as reference values.

[0141] In step S9, the operator checks the predicted results or the reference values ​​presented on the user interface 1 to confirm whether the performance has reached a level sufficient for testing. If the required performance has been reached (YES), the process proceeds to step SC10. On the other hand, if the required performance has not been reached (NO), the process returns to step S5.

[0142] When the process returns to step S5, the low-pass filter L (Formula 18) and the notch filter Ni (Formula 19) are redesigned. Furthermore, the dimension n of the feedback control and the dimension m of the feedforward control to be adjusted in step S6 are changed, and the control adjustment is redone.

[0143] In step S10, the actual vibration test is carried out using feedback control and feedforward control after the control parameters have been adjusted.

[0144] The entire process of control adjustment is completed up to step S10. After that, the process for retesting will be carried out.

[0145] In step S11, it is determined whether or not to perform a retest. If it is a one-time test, retesting is not necessary (NO), and the process ends. If retesting is to be performed (YES) in step S11, the process proceeds to step S12.

[0146] In step S12, a decision is made as to whether or not to perform a retest. Even if a retest is performed, if there is no change in the specimen 10 (NO), the adjusted controller can be used, and therefore the process proceeds to step S10, where the actual vibration test can be performed immediately.

[0147] On the other hand, if there is a change in the specimen 10 in step S12 (YES), the process returns to step S1 and the normal work procedure is repeated.

[0148] If the response adjustment is to be performed on the actuator response, it may be possible to reuse the control parameters that have been adjusted in advance, even if the specimen 10 changes. However, if the specimen 10 changes, there is a high possibility that the specimen reaction force (corresponding to the disturbance d in Fig. 6B) will also change. Therefore, it is desirable to display on the user interface 1 that such an operation is not recommended.

[0149] As described above, according to the first embodiment of the present invention, it is possible to provide the test apparatus 100 having a control adjustment function for effectively reducing the distortion rate without calculating the transfer function of the test apparatus 100.

[0150] In addition, control for reducing the distortion rate can be performed without performing vibration experiments for identifying the transfer function, and the number of steps required for adjusting the control parameters can be reduced.

[0151] Example 2 Next, a second embodiment of the present invention will be described.

[0152] In the first embodiment, the feedforward control 2b is assumed to exist in parallel with the feedback control 2a. However, the feedforward control 2b is not limited to this implementation form.

[0153] In the second embodiment, a configuration will be described in which a feedforward control 2a and a feedback control 2b are arranged in series as shown in Fig. 12. A control command calculation unit 12b in the second embodiment of the present invention includes a feedback control 2a and a feedforward control 2b.

[0154] As in FIG. 5, FIG. 12 is configured so that the control parameters of both feedback control 2a and feedforward control 2b can be adjusted by the experimental data analysis unit 2c. However, as in FIG. 2, it is also possible to adjust only the feedforward control 2b, or only the feedback control 2a.

[0155] Since most of the configuration of the second embodiment is similar to that of the first embodiment, only the differences from the first embodiment will be described.

[0156] In the second embodiment, a block diagram including the feedforward control can be shown as in FIG.

[0157] In other words, the transfer function from the target waveform r to the actuator response y when the feedforward control is effective can be given by Equation 21.

[0158]

number

[0159] If the response waveform obtained when only feedback control is effective (FIG. 4A, Equation 5) is y1, then the response when feedforward control is added can be given by Equation 22 using a similar discussion to that in Example 1.

[0160]

number

[0161] In this way, if the response prediction can be performed, the subsequent processing will be the same as in the first embodiment.

[0162] The processing procedure of the second embodiment can also be explained using the flowchart shown in Fig. 11. Only the differences from the first embodiment will be explained.

[0163] In the first embodiment, in order to disable the feedforward control in step S1, all of the coefficients ρi (i=0 to m) in the numerator of the transfer function F (Formula 12) of the feedforward control are set to 0. If such a setting is made in the second embodiment, F=0 will result, and the corrected target waveform r'=Fr will also become 0. Since the purpose of step S1 is to disable the feedforward control, it is necessary to use a value that satisfies r'=r, that is, F=1, as the initial value.

[0164] Therefore, in the case of the second embodiment, all the adjustment parameters ρi (i=0 to 2m) should be set to 1. The rest is the same as in the first embodiment, except that the method of initial setting of the feedforward control is changed.

[0165] In the second embodiment, the same effects as those in the first embodiment can be obtained.

[0166] Example 3 Next, a third embodiment of the present invention will be described with reference to FIG.

[0167] In the second embodiment, it is assumed that the feedforward control is implemented in the controller. In the case of an existing test device 100, it may be assumed that the device is driven only by feedback control and that it is difficult to rewrite the program.

[0168] In the third embodiment, a test apparatus 100 is provided that is suitable for use at a site where it is difficult to update the program of the controller 2, such as an existing test apparatus 100.

[0169] In the third embodiment, an additional controller 20 is provided in addition to a normal controller 2. The additional controller 20 has functions of a feedforward control 20b, an experimental data analysis unit 20c, and a response prediction unit 20d. The difference between the second embodiment and the third embodiment is whether the feedforward control unit 20b, the experimental data analysis unit 20c, and the response prediction unit 20d are implemented in the controller 2 or in the additional controller 20.

[0170] With such a change in configuration, the feedforward control section 20 b in the third embodiment is treated as a target waveform correction section that is used to correct the target value provided to the controller 2 .

[0171] In the third embodiment, when the controller 2 cannot be completely rewritten, the mechanism for adjusting the parameters of the feedback control section 2a is omitted.

[0172] If the control parameters are adjustable even if the control program cannot be rewritten, the parameters of the feedback control section 2a may be provided from the additional controller 20 as shown in FIG.

[0173] In the third embodiment, the feedback control unit 2a corresponds to the control command calculation unit.

[0174] In the third embodiment, the functions implemented in one controller 2 in the second embodiment are simply assigned to an additional controller 20, and therefore a detailed description thereof will be omitted.

[0175] In the third embodiment, the same effect as in the first embodiment can be obtained, and even in a site where it is difficult to update the program of the controller 2, such as an existing test apparatus 100, a test apparatus having a control adjustment function for effectively reducing the distortion rate without calculating the transfer function of the test apparatus 100 can be provided. [Explanation of symbols]

[0176] 1···User interface, 1a···GUI, 1b···Target waveform setting section, 2···Controller, 2a···Feedback control section, 2b, 20b···Feedforward control section, 2c, 20c···Experiment data analysis section, 2d, 20d···Response prediction section, 3···Servo amplifier, 4···Servo valve, 5···Hydraulic source, 6···Hydraulic cylinder, 7···Hydraulic piston, 8 ···Coupling, 9···Table (specimen installation section), 10···Specimen, 11···Addition section, 12a, 12b···Control command calculation section, 20···Additional controller, 100···Test device, S01···Temperature sensor, S02a, S02b···Flow rate sensor, S03···Speed ​​sensor, S04···Displacement sensor, S05a, S05b···Pressure sensor, S06···Non-contact sensor

Claims

1. a specimen installation section in which a specimen to be evaluated is placed; An actuator for vibrating the specimen installation section; a sensor for detecting a response of the actuator and the specimen; a controller for controlling a response of the actuator; a user interface having a target waveform setting unit for setting an operation pattern of the test apparatus, accepting an operation from an operator and presenting a test operation; Equipped with the controller includes a control command calculation unit that calculates a control input so as to eliminate a difference between a target waveform and a response waveform of the sensor, and an experimental data analysis unit that calculates parameters of the control command calculation unit based on experimental data; the experimental data analysis unit determines a transfer function of an ideal response of a closed loop system in accordance with the target waveform, and adjusts parameters of the control command calculation unit based on response data acquired by the sensor so that an actual response approaches the ideal response; A test apparatus characterized in that the determined transfer function has low-pass filter characteristics with a cutoff frequency higher than a reference frequency of the target waveform so as to suppress the effects of harmonics.

2. 2. The test device according to claim 1, a test apparatus characterized in that the determined transfer function has the low-pass filter characteristic having a cutoff frequency higher than a reference frequency of the target waveform, and one or more notch filter characteristics having a band that is an integer multiple of the reference frequency.

3. 2. The test device according to claim 1, A test apparatus characterized in that the determined transfer function has a low-pass filter characteristic having a cutoff frequency equal to or higher than a reference frequency of the target waveform and less than three times the reference frequency, and one or more notch filter characteristics having a band that is an odd multiple of the reference frequency.

4. 2. The test device according to claim 1, the control command calculation unit includes a feedback control unit having a transfer function with an adjustable parameter, and a feedforward control unit that is installed in parallel with the feedback control unit and has a transfer function with an adjustable parameter; The test apparatus according to claim 1, wherein the experimental data analysis unit adjusts a parameter of the transfer function of the feedback control unit and a parameter of the transfer function of the feedforward control unit.

5. 2. The test device according to claim 1, the control command calculation unit includes a feedback control unit having a transfer function with an adjustable parameter, and a feedforward control unit connected in series with the feedback control unit and having a transfer function with an adjustable parameter; A test apparatus, wherein the experimental data analysis unit adjusts parameters of the transfer function of the feedback control unit and parameters of the transfer function of the feedforward control unit.

6. a specimen installation section in which a specimen to be evaluated is placed; An actuator for vibrating the specimen installation section; a sensor for detecting a response of the actuator and the specimen; a target waveform setting unit that sets an operation pattern of the test device; a controller for controlling a response of the actuator; an additional controller that modifies the target waveform set by the target waveform setting unit and commands the modified target value to the controller; Equipped with The controller has a control command calculation unit that calculates a control input so as to eliminate a difference between a target waveform and a response waveform of the sensor, the additional controller has a target waveform correction unit that corrects a target waveform so as to eliminate a difference between the target waveform set by the target waveform setting unit and a response waveform of the sensor, and an experimental data analysis unit that calculates parameters of the control command calculation unit based on experimental data; The test apparatus is characterized in that the experimental data analysis unit determines a transfer function of the ideal response of a closed loop system according to the target waveform, and adjusts the parameters of the target waveform correction unit so that the actual response approaches the ideal response based on the response data acquired by the sensor.

7. 2. The test device according to claim 1, a user interface that accepts an operator's operation and presents a test operation; the controller comprises a response prediction unit that predicts a response of a testing device based on the parameters changed by the experiment data analysis unit; The response prediction unit predicts a transfer function using the parameters calculated by the experimental data analysis unit and the response data acquired by the sensor, thereby predicting a response waveform that will be obtained after changing the parameters, and presents the predicted result on the user interface.

Images