JP2002214068A - Vibration testing device for structure, digital computer used therefor, and method for vibration test - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily perform vibration tests with high accuracy even when the natural period of a numerical analysis model is short. SOLUTION: During the course of the vibration tests in which vibration tests by means of an exciter 4 and numerical calculations by means of a computer 5 are simultaneously advanced through the medium of a load applied by means of the exciter 4 and a command value outputted from the computer 5 to the exciter 4, vibration response analysis is performed in a mode space. In addition, a partial mode is calculated as a first-order lag system from which an inertial term is removed. Consequently, the time increment of the vibration calculations can be made larger and the execution of the vibration tests in real time becomes easier, because the time increment is decided depending upon the shortest natural period of a natural mode in which the time ticking is calculated as a second-order lag system.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vibration test apparatus for a structure which combines a vibration test of only a part of the structure with a numerical analysis of vibration response, a digital computer used for the apparatus, and a vibration test method. The present invention relates to an apparatus and a method for testing a vibration of a structure which are suitable for performing the method with high accuracy. Further, the present invention relates to a structure vibration test apparatus and a method suitable for a structure having a short natural period.

[0002]

2. Description of the Related Art A vibration test performed by combining a vibration test with a vibrator for only a part of a structure and a vibration response analysis is performed as follows.
It is called "test machine-computer online test" or the like. For example, Architectural Institute of Japan Transactions No. 229 (Showa 50) 77
An example of the system is shown from page 83 to page 83. There is also a technique described in JP-A-61-34438.

In the numerical analysis combined with the above-mentioned vibration test, the vibration response calculation of a numerical model is directly integrated by using a matrix. JP-A-2-82132
And Japanese Patent Application Laid-Open No. 5-332876 describe a technique for performing a vibration test by performing this numerical analysis in a mode space. Further, Japanese Patent Application Laid-Open No. Hei 5-10846 describes a technique in which the time axis of the calculation and the test are made to coincide with each other in order to conduct an experiment in consideration of the dynamic characteristics of the structure to be tested, that is, a technique of performing the test in real time.

[0004]

In the numerical calculation used in the prior art, the range in which the calculation can be performed stably is determined by the shortest natural period of the structure to be calculated and the calculation time interval. Therefore, when the natural period to be evaluated is very short, it is necessary to reduce the calculation time interval to a stability limit determined by the calculation method. Therefore, in the case of the quasi-dynamic test, the number of steps required to evaluate the response to the seismic excitation for a certain period of time increases, and the experiment time may become extremely long. In addition, since the amount of change in the vibration displacement of the vibration machine in one step is small, the vibration accuracy is deteriorated.

In order to solve the above problem, there is a method of calculating the eigenmode of the entire structure assuming the rigidity of the structure to be tested, performing numerical calculation in the mode space, and removing the one having a short natural period. Conceivable. However, this method has a problem that if the assumed spring constant is different from the actual spring constant, an error occurs in the obtained vibration characteristics. In addition, if a vibration test exciter having dynamic characteristics including response delay is generally included in the real-time test, this will cause an error,
It was possible that the vibration response would diverge. Further, in the prior art, since a vibration test using a vibrator is used, only the reaction force related to the relative deformation of both ends of the test object structure is considered. However, in the case of a real-time test, the inertia force related to acceleration may not be negligible in the measured reaction force.
This is different from the inertial force obtained by the fundamental excitation, and may cause a test error.

SUMMARY OF THE INVENTION It is an object of the present invention to easily and accurately perform a vibration test of only a part of a structure in combination with a vibration response numerical analysis, especially when the numerical analysis model has a short natural period. An object of the present invention is to provide a vibration test apparatus for a structure capable of performing a vibration test or suitable for performing the main vibration test with high accuracy, a digital computer used for the apparatus, and a vibration test method.

[0007]

According to the present invention, vibration evaluation of a structure having a short natural period is performed at relatively large time intervals.
In order to achieve high accuracy, it is assumed that the spring constant of the vibration test object is assumed, the eigenvalue analysis of the entire structure is performed using the assumed spring constant, and the numerical calculation is performed in the mode space. This can be achieved by substituting a first-order lag system or a proportional system and performing a test by calculating a second-order lag system as a second-order lag system, and by using a test apparatus capable of performing the above test.

Further, canceling the negative attenuation caused by the shaker response delay is to predict the vibration response ahead of the shaker response delay of the second-order lag system and to generate the shaker command signal. It is achieved by using it and making it an implementable device.

In order to predict the vibration response in a short time, the predicted value is obtained by multiplying a past response value at predetermined time intervals by a coefficient determined by the number of response values, and the above can be performed. It is achieved by using a simple device.

Implementing the prediction with high accuracy involves measuring the displacement of the vibrator, comparing the calculated value with the vibration response to be vibrated, and correcting the predetermined time interval in the prediction so that the two coincide with each other. Is achieved by

In order to remove the error caused by the inertial force included in the reaction force, the mass of the structure to be tested for vibration is evaluated in advance, and the generated inertial force is obtained by measuring or calculating the excitation acceleration. This can be achieved by removing the measured reaction force and using it in the vibration response calculation, and by making the device executable. It can also be achieved by mounting a vibration exciter and a structure to be subjected to a vibration test on a vibration table and performing a vibration test.

In order to easily carry out a vibration experiment, it is necessary to compare a natural frequency input to a computer with a judgment condition input in advance and determine whether to treat it as a second-order lag system or a first-order lag system. This is achieved by providing a function of making a judgment and passing it to the vibration response calculation processing means.

Further, to easily carry out a vibration experiment,
Before the actual excitation test, a pre-excitation with a small amplitude was performed to measure the spring constant of the vibration test target structure, and the eigenvalue analysis of the entire vibration response evaluation structure was performed using this. This is achieved by implementing them in an efficient manner, and by making them operable devices.

That is, the invention of the apparatus and the method for testing the vibration of a structure according to the present invention is characterized by one of the following. (1): One or a plurality of vibrators connected to a structure to be excited (hereinafter, simply referred to as a structure) to vibrate the structure, a control device of the shaker, and the structure from the shaker Measuring means of a load applied to an object, measuring means of a displacement applied from the shaker to the structure, a digital computer, a signal transmitting means from the digital computer to the shaker control device, and a load measuring means A signal transmission unit to the computer, and a signal transmission unit from the displacement measurement unit to the computer, the digital computer includes a time management unit, an input unit of an output of the load measurement unit, an input unit of an output of the displacement measurement unit, From the measurement time of the load measurement value or signal using the external force value storage means and / or the external force signal input means, the external force value or signal, the load measurement value or signal by the load measurement means, and the measurement value by the displacement measurement means. A means for calculating a command signal value to the shaker after a predetermined period of time, a means for generating a command signal for the shaker, an output means thereof, and an input / output means for external data. Means for calculating a command signal value to the vibrator, input calculating means, one or more secondary delay systems and one or more primary delay system or proportional system response calculating means, Signal value calculating means.

(2): In (1), the computer performs (a)
Input of load value, displacement value and external force value, (b) calculation of a shaker command signal after the fixed time from the load measurement time of the shaker,
(C) generating and outputting a command signal so that the shaker command signal after the predetermined time is input to the shaker control device, and managing the step by time management means to execute the step at the fixed time. , A predetermined number of times or continuously until a stop signal is input.

(3) In (2), the input calculation means obtains a difference between a load measurement value or a signal and a product of a predetermined coefficient and a displacement measurement value or a signal, and the difference is determined for each system. The sum of the product obtained by multiplying the coefficient and the value obtained by multiplying the external force value or signal at the current step by a coefficient separately determined for each system is obtained as an input to each system.

(4) In (3), the response calculation means calculates a response to the input calculated for each system by the input calculation means after the predetermined time. (5): In (4), the signal value calculation means may include a coefficient previously determined for each system in the response calculation value of each system of the secondary delay system among the response calculation values of each system of the response calculation means. (Hereinafter referred to as the sum of the response calculation values of the second-order lag system), and the value of the sum is predicted after a predetermined time, and the first-order lag or the proportional system is calculated. To calculate the sum (hereinafter referred to as the sum of the response calculation values of the first-order lag system) obtained by multiplying the response calculation value of each system by a coefficient predetermined for each system, and further calculate the sum with the predicted value. Is the command signal value to the vibrator after the predetermined time.

(6): In (5), the computer comprises means for storing the sum of the past calculated values of the response of the second-order lag, and the signal value calculating means stores the sum of the calculated values of the response of the second-order lag for a predetermined time. The later prediction is the sum of the response calculation values of the current step and the sum of the past response calculation values at predetermined time intervals for the predetermined number of different predictions among the values stored in the response calculation value storage means. Are multiplied by a predetermined coefficient and added to each other as a predicted value.

(7) In (5), the computer has a predicted time correcting means, and the predicted time correcting means calculates the sum of the measured value of the displacement and the calculated value of the response of the secondary delay system and the response of the primary delay system. The calculated signal is compared with a value obtained by adding the sum, and the time used for prediction is corrected by the excitation signal calculating means.

(8): In (7), the predicted time correcting means includes a calculated value of the sum of the response calculated values of the secondary delay system, a displacement measured value, and a first-order delayed system which is used for prediction for a predetermined time. The difference between the calculated values of the sums of the calculated response values is compared to calculate a correction value.

(9) In any one of the constitutions (1) to (8), the apparatus further comprises: a vibrating machine vibration acceleration measuring device; and a data transmission means from the vibration acceleration measuring means to the vibration acceleration input means. The computer includes an excitation acceleration input means, and uses a difference between an actual load measurement value and a product of a predetermined coefficient and an acceleration measurement value as a load measurement value.

(10) In any one of (1) to (8), the computer has a vibration acceleration calculating means, and the acceleration calculating means calculates the vibration acceleration from the vibration displacement measurement value. Yes, the difference between the actual measured load value and the product of the predetermined coefficient and the calculated acceleration value is used as the measured load value.

(11): In any one of the constitutions (1) to (10), there is provided a structure to be vibrated and a vibration table on which a vibrator is mounted. (12): In any one of (1) to (11), the computer omits the second derivative term when the input natural frequency of the second-order lag system is larger than a predetermined value. It has the function of replacing it as a first-order lag system.

(13) In any one of the constitutions (1) to (12), the computer has a function of measuring the spring constant of the structure by vibrating the vibrator minutely, and inputting the spring constant in advance. The function of performing the eigenvalue analysis using the obtained numerical value and the mode in which the eigenfrequency is smaller than the given value in the analysis result are set as a second-order lag system, and the mode is set to be smaller than the given value in the analysis result. The mode having a large natural frequency has a function of setting as a first-order lag system.

(14) One or more vibrators connected to the structure to vibrate the structure, a control device of the vibrator, and a means for measuring a load applied to the structure from the vibrator. ,
The apparatus includes a vibration displacement measuring unit of the vibrator and a digital computer, and the digital computer includes an input unit of an output of the load measuring unit, an input unit of an output of the displacement measuring unit, a storage unit of the external force value, and / or Means for inputting an external force signal, means for calculating a command value to the vibrator using the external force value or signal and the load measurement value or signal by the load measuring means, output means for the command signal of the vibrator, Means for measuring the displacement of the machine and the command value, means for determining whether or not the difference between the displacement of the vibrator and the displacement command value is within a predetermined allowable value, and means for input / output of data with the outside. And a signal transmitting means from the digital computer to the exciter control device, a signal transmitting means from the load measuring means to the computer, and a signal transmitting means from the displacement measuring means to the computer. Means for calculating the command signal value of Means out, one or a plurality of 2
It has a second delay system, one or more first-order delay systems or proportional system response calculation means, and signal value calculation means.

(15): In (14), the computer:
(a) Input of load value and external force value, (b) Calculation of shaker command signal, (c) Generate and output command signal, (d) Displacement measurement, (e) Convergence judgment, (f) as necessary Steps (c) to (e) are repeated a predetermined number of times or continuously until a stop signal is input.

(16) In (15), the input calculation means obtains a difference between the measured load value or signal and a product of a predetermined coefficient and a command signal value of the previous step, and determines a difference for each system in advance. The sum of the value obtained by multiplying the obtained coefficient and the value obtained by multiplying the external force value at the current step by a coefficient separately determined for each system is obtained as an input to each system.

(17) In (16), the response calculating means calculates a response after a predetermined period of time to the input calculated for each system by the input calculating means. (18): In (17), the signal value calculation means calculates the sum of the response calculation value of each system of the response calculation means multiplied by a coefficient predetermined for each system, and the excitation in the next step is performed. Command signal value to the machine.

(19) The whole structure to be evaluated is converted into a numerical model and mounted on a digital computer, and a part of the structure to be evaluated is used as a real model, and is shaken by one or a plurality of vibrators connected to the real model. Shaking, the load applied to the real model by the vibrator is measured by the load measuring means, the displacement applied to the real model by the vibrator is measured by the displacement measuring means, and (a) the load value, the displacement value and the external force value Input,
(b) The displacement of the boundary point between the real model and other parts after a certain time from the load measurement time of the shaker is calculated as a shaker command signal, and (c) The shaker command signal after a certain time The step of generating and outputting a command signal so as to be input to the shaker control device is time-managed in the digital computer and executed at the above-mentioned fixed time, and is continuously performed until a predetermined number of times or a stop signal is input At the time of repetition, the numerical model is a linear numerical model of the excitation target structure, the natural frequency by eigenvalue analysis, the natural mode obtained by calculating the natural mode, at least one mode is a single-degree-of-freedom vibration system, That is, a second-order lag system is used, and at least one of the remaining eigenmodes is omitted from the inertia term and input to the digital computer as a first-order lag system or a proportional system. Calculation stage And flop, a response step of calculating one or a plurality of secondary delay system and one or a plurality of first-order system or a proportional system, carried out by the signal value calculation step.

(20): In (19), the digital computer includes time management means, load measurement result input means,
Based on the numerical model, an external force value or an external force signal, a load measured value or a signal equivalent to the load measured value, and a displacement measured value are input based on the displacement measurement result input means, the external force value storage means and / or the external force signal input means. Means for calculating the displacement of the boundary point between the real model and the other part after a predetermined time from the measurement time of the load measurement value using the load measurement value, and calculating a command signal value to the vibrator; And an input / output means for data input / output with the outside of the computer.

(21): In (20), in the input calculation step, the difference between the load measurement value and the product of the spring constant and the displacement measurement value of the numerical model of the real model part is determined, and the mode vector of each mode is determined. The sum of the product obtained by multiplying the difference between the elements corresponding to both ends of the internal real model and the product obtained by multiplying the external force value at the current step by the stimulation coefficient of the external force for each mode determined in advance is obtained as the input of each mode. . (22): In (21), the response calculation step calculates a response to the input calculated for each mode in the input calculation step after the predetermined time.

(23): In (22), in the signal value calculation step, among the response calculation values of each mode in the response calculation step, the mode calculation vector of each mode is added to the response calculation value of each mode obtained as a secondary delay system. Is calculated by multiplying the difference between elements corresponding to both ends of the internal real model (hereinafter referred to as a second-order lag system response value), and a predetermined shaker response delay of the sum is calculated. A value obtained by estimating the preceding value by the time and further multiplying the response calculation value of each mode obtained as a first-order delay or a proportional system by the difference between elements corresponding to both ends of the internal real model of the mode vector of each mode (hereinafter, referred to as (Referred to as a first-order lag-system response value), and the sum of the second-order lag-system response value and a predicted value is used as a command signal value to the vibrator after the predetermined time.

(24): In (23), in the signal value calculating step, the prediction of the second-order delay system response value after a predetermined time is performed by using a predetermined number of past vibrator response delay times for each of the vibrator response delay times. The delay system response value and the second-order delay system response value of the current step are each multiplied by a predetermined coefficient and then added to each other.

(25) In (23) or (24), the computer compares the measured displacement value with the sum of the second-order lag-based response value and the first-order lag-based response value, and makes a prediction when calculating the excitation signal. Correct the set value of the shaker response delay time used for.

(26): In (25), the vibration exciter response delay time is corrected by the second-order delay system response value and the displacement measurement value and the vibration exciter response delay time set value before the set value used for prediction. A correction value is calculated by comparing the difference with the next delay response value.

(27): In any one of (19) to (26), the excitation acceleration is measured or calculated, and the measured load is determined in advance as an equivalent mass of the actual measured load and the actual model. The difference between the coefficient and the product of the measured acceleration value or the calculated value is used. (28): In any one of (19) to (27),
The shaker and the structure to be tested are mounted on a vibrating table, and are excited by the acceleration applied to the structure simultaneously with the excitation by the shaker.

(29): In any one of (19) to (28), the computer may have the input natural frequency of the second-order lag system larger than a predetermined value before performing the vibration experiment. In this case, the method includes a step of replacing the second-order differential term as a first-order lag system omitting the second-order differential term.

(30) In any one of the constitutions (19) to (29), the computer may include: a step of minutely vibrating the vibrator to measure a spring constant of the structure to be vibrated; Performing a natural value analysis using a numerical value input in advance, and setting a mode in which the natural frequency is smaller than a predetermined value in the analysis result as a second-order lag system;
A mode in which the natural frequency is larger than a given value in the analysis results is set as a first-order lag system, and then a vibration experiment is performed under given conditions.

(31): The whole structure to be evaluated is numerically modeled and mounted on a digital computer, and a part of the structure to be evaluated is used as a real model, and one or more vibrators connected to the structure The load applied to the real model by the shaker is measured, and (a) input of the load value, displacement value and external force value, (b) calculation of the shaker command signal, and (c) the command signal The steps of generation output, (d) displacement measurement, (e) convergence judgment, (f) repeating (c) to (e) as necessary, are repeated a predetermined number of times, or until a stop signal is input. At the time of repetition, the numerical model is a linear numerical model of the excitation target structure, and at least one of the eigenmodes obtained by calculating the eigenfrequency and the eigenmode by eigenvalue analysis is a one-degree-of-freedom vibration system. That is, a second-order delay system is used, and at least one of the remaining eigenmodes is used. In this mode, the inertia term is omitted and the signal is input as a first-order lag system or a proportional system. The calculation of the command signal value to the vibrator is performed by an input calculation step and the one or more second-order lag systems. This is performed by one or a plurality of first-order lag-based or proportional-based response calculation steps and a signal value calculation step.

(32): In (31), the digital computer uses the external force value or the external force signal based on the numerical model based on the load measurement result inputting means, the external force value storing means and / or the external force signal inputting means. Calculate the displacement of the boundary point between the real model and other parts after a predetermined time from the measurement time of the load measurement value using the load measurement value or a signal equivalent to the load measurement value, and send a command to the vibrator. Means for calculating a signal value, means for outputting a command signal of the shaker, a displacement measured value of the shaker and the command value, and a difference between the displacement of the shaker and the displacement command value is predetermined. It is provided with means for determining whether or not convergence is within an allowable value, and means for inputting / outputting data with the outside of the computer.

(33): In (31) or (32), the input calculation step calculates the difference between the measured load value and the product of the spring constant of the numerical model of the actual model and the command signal value of the previous step. The sum of the product of the mode vector of each mode multiplied by the difference between the elements corresponding to both ends of the inner real model and the product of the external force value at the current step multiplied by the stimulation coefficient of the external force value for each mode previously determined is calculated. , Input of each mode. (34): In (33), in the calculation step, a response to the input calculated for each mode in the input calculation step after a predetermined time has been calculated.

(35): In (34), in the signal value calculating step, the response calculation value of each mode of the response calculating means is multiplied by a difference between elements corresponding to both ends of the mode model of each mode at both ends of the real model. Is calculated as a command signal value to the vibrator.

(36): Using the vibration test apparatus according to any one of (1) to (13), a part of the structure to be evaluated is subjected to vibration as a real model, and the entire structure is a linear numerical model. The eigenvalue analysis is performed in advance, and at least one of the eigenmodes obtained thereby is used as a second-order delay system in the response calculation means, and at least one of the remaining eigenmodes is used as a first-order delay system or a linear system. The spring constant of the excitation target structure part in the modeling is a coefficient multiplied by the command signal value of the previous step in the input calculation means,
The stimulus coefficient for each mode of the external force calculated based on the eigenvalue analysis is a coefficient determined for each system multiplied by the external force value at the current step in the input calculation means,
The eigenmode vector of each mode calculated by the eigenvalue analysis is used as a coefficient determined for each system by the input calculation means using the difference between elements corresponding to both ends of the inner real model, and the eigenmode of each mode calculated by the eigenvalue analysis The difference between elements corresponding to both ends of the internal model of the vector is set as a coefficient multiplied by the response calculation value of the second-order delay system or the first-order delay system or the proportional system in the signal value calculation means, and the response delay time of the vibrator is set to 2 It is the time used for predicting the sum of the response calculation values of the secondary delay system.

(37): Using the vibration test apparatus according to any one of (14) to (18), a part of the structure to be evaluated is subjected to vibration as a real model, and the whole structure is a linear numerical model. The eigenvalue analysis is performed in advance, and at least one of the eigenmodes obtained thereby is used as a second-order delay system in the response calculation means, and at least one of the remaining eigenmodes is used as a first-order delay system or a linear system. The input constants are used as the coefficients to be multiplied by the command signal value of the previous step in the input calculation means, and the stimulus coefficients for each mode of the external force calculated based on the eigenvalue analysis are input calculation means. In the above, the coefficients determined for each system to be multiplied by the external force value at the current step are used, and the internal real model of the eigenmode vector of each mode calculated by the eigenvalue analysis And a coefficient defined for each system in the input-calculating means the difference between the two ends to the corresponding element of the Le,
The difference between the elements corresponding to both ends of the eigenmode vector of each mode calculated by the eigenvalue analysis and the internal model is multiplied by the response calculation value of the second-order lag system or the first-order lag system or the proportional system by the signal value calculation means. It is a coefficient.

By the way, the digital computer used in the vibration test apparatus and test method for the present structure has a plurality of processors, means for exchanging data between the plurality of processors, and a relation among the plurality of processors. Means for waiting for processing, wherein a plurality of modes are divided into a plurality of groups, and arithmetic processing of each group is processed by the plurality of processors.

Further, the digital computer has a plurality of processors, a unit for exchanging data between the plurality of processors, and a unit for waiting for a related process among the plurality of processors. The processor may be divided into a plurality of processor groups, a plurality of modes may be divided into a plurality of groups, and arithmetic processing of each group may be assigned to each of the plurality of processor groups for processing. Furthermore, the processing of each mode group, which is performed by each processor group, may be divided into a plurality of processing stages, and the processing of each stage may be performed by each processor in each processor group.

Assuming the spring constant of the vibration test object, performing eigenvalue analysis of the entire structure using the spring constant, and performing numerical calculation in mode space, the short-period eigenmode is a first-order lag system or a proportional system. And the other is calculated as a second-order lag system to perform the test, and by using a test device capable of performing the above test, the numerical calculation time step replaced by the first-order lag system or the proportional system is large. Since the divergence of the vibration response does not occur, the numerical calculation time step may be determined for the mode having the shortest natural period among the modes handled as the second-order lag system. In addition, since the deformation caused by the short-period eigenmode is evaluated as the response of the first-order lag system or the proportional system, the value used for the eigenvalue analysis of the spring constant of the substructure to be tested and the actual value are Even if different, the error is reduced, so that the vibration evaluation of the structure having the short natural period can be performed at a relatively large time interval and the accuracy can be made high.

By predicting the vibration response ahead of the exciter response delay of the second order delay system and using it for generating the exciter command signal, and by using a device capable of performing the above, the predicted time is obtained. However, since the previous command signal is delayed by the shaker, the resulting displacement is realized by the shaker at the time to be realized, and the negative damping caused by the shaker response delay Can be offset.

The prediction value is obtained by multiplying the past response value at predetermined time intervals by a coefficient determined by the number of response values to obtain a prediction value. Since it is possible, the prediction of the vibration response can be performed in a short time.

The displacement of the vibrator is measured by measuring the displacement of the vibrator and comparing it with the calculated value of the vibration response to be vibrated so as to correct the predetermined time interval in the prediction so that the two agree with each other. Therefore, the prediction can be performed with high accuracy.

The mass of the structure to be subjected to the vibration test is evaluated in advance, and the generated inertia force is obtained by measuring or calculating the excitation acceleration, and is removed from the measured reaction force to be used for the vibration response calculation. In addition, by using an apparatus capable of performing the above, the inertial force included in the measured reaction force can be removed, and the vibration response calculation can be performed using only the load related to the relative deformation of both ends of the vibration test target structure. Therefore, it is possible to remove an error caused by the inertial force included in the reaction force.

Also, by mounting a vibration exciter and a structure to be subjected to a vibration test on a vibration table and conducting a vibration test, the inertia force generated by the basic vibration can be correctly included in the reaction force. The error caused by the inertial force included in the force can be eliminated.

The natural period input to the computer is
Alternatively, an amount related thereto (for example, natural frequency) is compared with a judgment condition input in advance, and it is determined whether to treat as a second-order delay system or a first-order delay system. By providing the function of delivering, it is not necessary for the tester to determine which eigenmode is to be treated as the first-order lag system, so that the vibration experiment can be easily performed.

Further, before the actual vibration test, a preliminary vibration with a small amplitude is performed to measure the spring constant of the structure to be subjected to the vibration test, and the eigenvalue analysis of the entire vibration response evaluation structure is performed by using this. It is not necessary to evaluate the spring constant of the substructure to be vibrated in advance and to perform eigenvalue analysis using the same in advance by automatically performing the above-mentioned operations, and by using a device capable of performing these operations. Experiment easily
And can be implemented accurately.

[0055]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. For example, consider a structure 1 as shown in FIG. As shown in FIG. 3, the partial structure 2 is vibrated by a vibrator 4, and the other parts 3 are numerically modeled and mounted on a computer 5. The vibrator 4 is provided with a load measuring device 6, and measures a reaction force applied from the partial structure 2 to the vibrator 4. The measured reaction force is input to the computer 5. The vibrator 4 is controlled by the control device 7, and a command signal is calculated in the computer 5 and input to the control device 7.

Exciting force such as seismic force applied to the structure is sequentially input to the computer 5 or stored in a memory (not shown) in the computer. A step of calculating an external force vector from the obtained reaction force, a step of calculating a vibration response of the numerical model with respect to the external force vector after a predetermined time, and a step of calculating a command value of the vibrator based on the vibration response calculation result. As described above, the vibration response in a state where the partial structure 2 is coupled to the other portion 3 (the structure 1) can be evaluated by the vibration test of only the partial structure.

The vibration response calculation performed by the computer 5 is as follows. The part 3 other than the substructure 2 to be tested is numerically modeled as a mass matrix M, a damping matrix C, and a stiffness matrix K. Then, the equation of motion for the relative displacement vector x with respect to the foundation is written as follows using the external force vector q. In addition, * represents differentiation with respect to time.

[0058]

(Equation 1)

Here, the external force vector is composed of two components, an exciting force q1 such as an earthquake force and a measured reaction force q2. That is,

[0060]

(Equation 2)

Using this, the vibration response after a short time Δt is calculated. When the vibration response calculation result is realized by the vibrator 7, when only a static reaction force such as the spring constant of the test object structure 2 is handled, the calculation result after Δt is calculated after a time larger than Δt. It can be realized. Hereinafter, this test method is referred to as a “quasi-dynamic” test.

Further, in order to measure a dynamic component such as a damping force included in the reaction force q2 by a test and reflect the result in a vibration response analysis, it is necessary to perform the measurement in real time. This test method is hereinafter referred to as a "real time" test.

Various algorithms can be used as a numerical calculation method. For example, in the known central difference method, the time ti + 1 (ti + 1 = ti + Δt) is obtained from the known information at the time ti as shown in the following equation. ) Can be calculated.

[0064]

(Equation 3)

The above equation is calculated by a matrix operation, but it is also possible to implement this in the mode space.

In the numerical calculation used in the above technique, the range in which the calculation can be performed stably is determined by the shortest natural period Tmin of the structure to be calculated and the calculation time step Δt. For example, in the central difference method

[0067]

(Equation 4)

Must be satisfied. Therefore, when the natural period to be evaluated is very short, it is necessary to reduce the calculation time step Δt according to (Equation 4) or the stability limit determined by the calculation method.

[0069]

Therefore, in the case of the quasi-dynamic test, the number of steps required to evaluate the response to the earthquake excitation for a certain period of time increases, and if the time required for each time step is constant, the experimental time Is very long. Further, since the amount of change in the vibration displacement of the vibrator 7 in one step is small, there is a problem that the vibration accuracy is deteriorated. Further, when performing the real-time test, it is necessary to perform the response calculation process after the calculation time interval Δt at least during Δt. However, if Δt is too small, it may not be possible to perform the calculation.

As one method for solving the above problem, the eigenmode of the entire structure 1 is calculated assuming the rigidity of the test object structure 2 and numerical calculation is performed in the mode space. Techniques for removing short ones have been proposed. However, this method has a problem that if the assumed spring constant is different from the actual spring constant, an error occurs in the obtained vibration characteristics.

The vibrator 4 used for the vibration test is
Generally, it has dynamic characteristics including response delay. That is, as shown in FIG. 4, the excitation displacement is realized with a slight delay with respect to the input signal. When the vibrator 4 having such characteristics is included in the real-time test, the linear spring having the spring constant k is evaluated as having the characteristics shown in FIG. 5, and when the delay time is δt,

[0073]

(Equation 5)

Has a negative attenuation ceq as shown below.
Therefore, this causes an error. In particular, when the absolute value of ceq is larger than the damping of the structure itself, there is a problem that the vibration response diverges.

Furthermore, in the prior art, since the vibration test using the vibrator 4 is used, only the reaction force related to the relative deformation of both ends of the test object structure 2 is considered. However, in the case of a real-time test, the inertia force related to the acceleration may not be negligible in the measured reaction force, but this is different from the inertia force obtained by basic excitation, and may cause test errors. was there.

The present invention has been made in order to solve such a problem, and a vibration test apparatus and a method for performing a vibration test and a vibration response numerical analysis of only a part of a structure in combination with the numerical analysis are particularly preferred. Provided is a vibration test apparatus and method for a structure that can easily and accurately perform a vibration test even when a model has a short natural cycle or that is suitable for performing the present vibration test with high accuracy. It is an object.

First, an outline of a vibration test method for performing a vibration response calculation using the mode space will be described. For example, the structure 1 shown in FIG. It is considered that the substructure 2 between the foundation and the lowermost mass point is a target of the vibration test, and the others are numerically modeled. As the external force, consider the acceleration of the foundation. The equation of motion relating to the basic excitation of the structure 1 including the substructure 2 is as follows if the damping term is omitted for simplicity.

[0078]

(Equation 6)

Here, M * is a mass matrix, K * is a rigidity matrix, x * is a relative displacement vector with respect to the foundation, i
Is a vector of all 1's, and z is the basal acceleration.
Since these matrices and vectors include the substructure 2 to be subjected to the vibration test and are different from (Equation 1), they are distinguished by adding *.

The actual stiffness of the substructure 2 and the stiffness of the numerical model may differ due to an estimation error in numerical modeling, a change in element characteristics during excitation, and the like. Assuming that the actual spring constant of the substructure 2 is k, the spring constant assumed as a numerical model is k ′, and the assumed rigidity matrix as a numerical model is K ′, the rigidity matrix K * of (Formula 6) is as follows. It is on the street.

[0081]

(Equation 7)

Here, E is a unit element rigidity matrix of the substructure 2. For example, if a member is at a support point as shown in FIG.

[0083]

(Equation 8)

By substituting (Equation 7) into (Equation 6) and rearranging,
The following equation is obtained.

[0085]

(Equation 9)

Here,

[0087]

(Equation 10)

Therefore, if the vibration response numerical calculation is performed by adding the load f shown in (Equation 10) as an external force to the equation of motion using the assumed rigidity matrix K ′ as in (Equation 9), (Equation 6) is obtained. It turns out that it is equivalent to calculating. By the way, rewriting (Equation 10) is as follows.

[0089]

[Equation 11]

Here, Δx is a relative displacement of both ends of the substructure 2 and is a vibration displacement by the vibrator 4, and r is 1 or 1 for an element corresponding to a node where an internal force load of the substructure 2 acts. -1 (depending on the coordinate system), the other is a vector that is zero.

KΔx is the reaction force q ′ actually generated in the substructure 2 and can be measured, and k′Δx is an amount that can be calculated by measuring the displacement of the vibrator. Therefore, it is possible to perform a vibration test using (Equation 9).
If the substructure has damping characteristics or non-linearity, these characteristics are reflected in the measured reaction force, and are taken into account in the vibration response numerical calculation.

Now, (Equation 9) can be converted into a mode space using the eigenmode vector by performing eigenvalue analysis by setting the right side to 0. In the eigenvalue analysis, the following equation is obtained when n is the number of degrees of freedom.

[0093]

(Equation 12)

Here, u is a mode displacement vector, P is a mode matrix, and pi is an i-th mode vector.
In the following discussion, it is assumed that the eigenmode is standardized so that the mode mass is 1. That is, with I as the unit matrix,

[0095]

(Equation 13)

The generality is not lost by this. Using this, (Equation 9) can be rewritten as the following equation.

[0097]

[Equation 14]

Here, ωi is the ith natural circular frequency, and T
Denote transposition

[0099]

(Equation 15)

Where b is a stimulus coefficient vector for the basic excitation, g is a stimulus coefficient vector for the excitation force generated from the substructure 2,

[0101]

(Equation 16)

[0102]

[Equation 17]

Is as follows. The mode parameter obtained here is not necessarily a true value because it is obtained from the rigidity matrix K ′ of the assumed numerical model.

The above is summarized as a flow as shown in FIG. That is, an external force caused by the test object structure is calculated from the measured load value and the measured displacement value (step i), and a modal external force is calculated from the calculated output and an external force from seismic force or the like (step ii). The mode response is calculated (step iii), and then converted into the physical space (step iv). Thereby, the same calculation as the vibration test using (Equation 3) can be performed by the vibration calculation in the mode space. In calculating the vibration numerical value according to (Equation 14), the eigenmode having a short period can be removed. However, when the mode is removed, it is known that a difference between the spring constant k ′ assumed in the eigenvalue analysis and the actual spring constant k causes an error.

Hereinafter, an embodiment of the present invention will be described with reference to FIG. The substructure 2 is vibrated by the vibrator 4. The shaker 4 is provided with a means 6 for measuring the load applied to the partial structure 2 from the shaker and a means 8 for measuring the displacement applied to the partial structure 2 by the shaker 4. The value measured by the load measuring means 6 is transmitted by the load transmitting means 9 and input to the computer 5 which is a digital computer via the load input means 11.

Further, the value measured by the displacement measuring means 8 is transmitted by the displacement transmitting means 10 and is similarly input to the computer 5 through the displacement input means 12. The computer 5 is provided with one or both of the external force storage means 15 and the external force input means 16 to store or input an external force such as a seismic force applied to the vibration response evaluation target structure 1.

The computer 5 has a vibrator command signal calculating means 17
And calculates a command signal to the vibrator 4 using the load measurement value, the displacement measurement value, and the external force value.
Further, a command signal generating means 18 is mounted, generates a command signal based on the command signal calculation result, and outputs the command signal via the command signal output means 13. The command signal is input to the control device 7 of the vibrator 4 via the command signal transmitting means 14 to drive the vibrator.

Input of external force value, load measurement value, displacement measurement value,
Calculation of the shaker command signal, generation of the shaker command signal, and output of the shaker command signal are managed by the time management means 19,
For example, as shown in FIG. 7, the process is repeatedly performed at regular intervals.

The load / displacement measurement value and the shaker command signal are transmitted as, for example, a voltage signal. At this time, the signal transmission means 9, 10, and 14 are cable lines. The signal input means 11 and 12 are A / D converters, and the signal output means 1
3 is a D / A converter. However, these signals may be in other forms, and the transmission means and input / output means will differ accordingly.

The command signal calculating means 17 calculates a signal in the following procedure. However, the parameters (e.g., natural frequency, mode vector, stimulus coefficient, etc.) required for the following calculations are input in advance from data input means (not shown) attached to the computer 5 and stored in data storage means (not shown) according to the following procedure. Stored for use in.

(1) In the input calculation means 171, (Equation 1)
The load applied to the evaluation target structure 1 is calculated by 1).
That is, the difference between the load measurement value q ′ and the product of the displacement measurement value Δx and the spring constant k ′ assumed in the numerical model (hereinafter, referred to as a partial structure-induced external force) is obtained. Also, a vibration acceleration value which is an external force is input. Then, an external force term for each mode is calculated from the right side of (Equation 14). The excitation force due to the acceleration for a certain mode is the product of the element for the mode in the stimulus coefficient vector b and the acceleration value input in the previous step. For the external force caused by the substructure, the stimulus coefficient vector g for the mode is The product of the element and the external force caused by the substructure is calculated. What changed the sign of the sum of the two is the excitation force for each mode.

(2) The response calculating means 172 obtains the vibration response after Δt of each mode. At this time, the mode selected in advance is treated as a proportional system in which the secondary differential term of (Equation 14) is omitted or a first-order lag system in which an attenuation term is added. Other modes are treated as a second-order delay system and are calculated by an algorithm such as (Equation 3). Modes that are considered to have little effect on the vibration response can be removed.

(3) The signal value calculating means 173 calculates the command signal value of the vibrator (displacement Δx at both ends of the substructure 2 to be added after Δt). First, in the second-order lag system, for each of the modes for which the vibration response has been calculated, each mode vibration response is multiplied by the difference between the elements corresponding to the nodes at both ends of the substructure 2 of the mode vector, and the sum is obtained. (Equation 18).

[0114]

(Equation 18)

The same processing is performed for the mode calculated as the first-order lag system or the proportional system to obtain Δxh.
Next, a value ahead of time corresponding to the shaker delay time of Δxl is predicted and set to Δxl ′. Then, the sum of the two is defined as a shaker command signal Δx (Equation 19).

[0116]

[Equation 19]

Thus, the command signal value to the vibrator 4 is calculated. Further, the signal generating means 18 generates a signal according to the signal output means 13 and the vibrator 4 based on the calculated value. However, the output of the signal is performed by the time management means 19 according to the time table of FIG.

The reason why a short-period mode is treated as a first-order lag system or a proportional system as in this embodiment will be described. FIG.
Fig. 3 shows a schematic diagram of the frequency response of the mode. The short-period or high-frequency eigenmode has a substantially constant amplitude in the test range and can be considered as a simple spring. Also, considering the energy dissipated by this element, it can be modeled as a first-order lag system.

If this mode is removed, there is no displacement to be caused by this mode, and the displacement of the excitation point is reduced. That is, the same result as when the spring constant is overestimated occurs, and causes an error such as an increase in the natural frequency. If a proportional system or a first-order lag system is used, the time step Δt
Since the stability does not depend on, the time interval may be determined by the minimum period in the mode treated as the second-order delay system. Therefore, the time interval can be made large, and FIG.
Can be easily realized.

Also, as in the present embodiment, the response depending on the mode treated as the second-order delay system is predicted by the vibrator delay time and used as a command signal for the vibrator. As shown in FIG. 9, since the response is delayed by the vibrator for the predicted time, the response to be realized at the time to be realized can be realized.

The reason why the prediction is not performed for the mode treated as the first-order lag system or the proportional system is that the error is generally large in the response prediction with respect to the vibration whose cycle is shorter than the prediction time. In the second-order lag system, the response is almost zero in the frequency range higher than the natural frequency, so that this error is offset. However, in the proportional system or the first-order lag system, the prediction error appears as it is. Get bigger,
In some cases, it cannot be performed stably.

According to the present embodiment, since the time interval can be set large, realization of a real-time test becomes easy. Further, since the response delay of the vibrator is corrected, a stable and highly accurate vibration test can be performed.

Another embodiment will be described with reference to FIG. FIG. 10 is obtained by adding a response calculation value storage unit 20 to the embodiment of FIG. In the present embodiment, the prediction of the response Δxl relating to the second-order delay system is performed by calculating the response calculation value Δx at the time preceding the current time by the vibrator response delay time δt as shown in FIG.
lk is multiplied by a coefficient determined in the following Table 1, and the sum is calculated. However, the number m of data used for prediction can be arbitrarily selected.

[0124]

[Table 1]

That is, the following equation is obtained.

[0126]

(Equation 20)

According to the present embodiment, since the prediction of the response can be performed by a simple product-sum operation, the calculation time is short, and therefore, the processing shown in FIG. 7 can be easily performed. However, since the response calculation value is calculated for each calculation time interval Δt, the value Δxlk for each shaker response delay time δt is not always obtained. In that case, the response calculation value at the time closest to the interval of δt is used as Δxlk, or Δxlk is calculated by interpolating the response calculation value, and the calculation of Expression 20 is performed.

Another embodiment will be described with reference to FIG. This embodiment is provided with a predicted time correcting means 27 in addition to the embodiment of FIG. In the embodiment of FIG. 1, the predicted time of the response value is input to the computer in advance, but if the set value differs from the actual value or changes during the test, the predicted time of the predicted signal is changed. It will cause errors and, consequently, errors in the vibration test.

Thus, the accuracy of the experiment is improved by comparing the value to be excited with the actual displacement and correcting the predicted time. The correction of the predicted time is performed, for example, as follows.
Calculated relative displacement Δx = Δxl in substructure 2
The following integration is performed over one cycle with respect to + Δxh and the measured value Δx ′ of the displacement of the shaker 4.

[0130]

(Equation 21)

The following equation is used for the integral value I, the oscillation period T, the amplitude A, and the error (ie, the difference between the actual delay time and the delay time used for prediction) δt ′ between the shaker response delay time. Holds.

[0132]

(Equation 22)

This calculation is performed sequentially, and the estimated time is changed as appropriate. According to the present embodiment, the prediction time is always set to the actual value, so that the prediction accuracy, that is, the test accuracy can be increased. In addition, in the correction of the prediction time of the present embodiment, the value Δ represented by the following equation is used instead of Δx and Δx ′ instead of Δx.
x "can also be used.

[0134]

(Equation 23)

According to this, it is necessary to compare only the portion that has been subjected to prediction, so that there is an effect that the accuracy is further improved.

Another embodiment will be described with reference to FIG.
In this embodiment, in addition to the embodiment of FIG. 1, an acceleration measuring means 21 is provided at an excitation point,
The acceleration measurement value input means 22 transmitted to the computer 5
In addition, the calculation processing is performed using q ″ obtained by the following equation instead of the load measurement value q ′.

[0137]

[0138]

(Equation 24)

The measured load value includes the inertial force due to the mass of the substructure 2, but in this test, the inertia force is determined by the relative acceleration at both ends of the substructure. However, actually, an inertia force proportional to the absolute acceleration must be considered, which causes an error. Therefore,
In this embodiment, the inertial force is removed from the measured load value q ′, and only the load related to the relative deformation is used for the vibration response calculation. In this embodiment, the numerical model needs to be modeled including the mass of the substructure. In this embodiment, since the inertial force can be removed, a highly accurate vibration test can be performed.

Still another embodiment will be described with reference to FIG.
In this embodiment, an acceleration calculation unit 23 is provided instead of the acceleration measurement unit 21 and the acceleration input unit 22 in the embodiment of FIG. The acceleration calculation means 23 calculates the excitation acceleration by differentiating the measured displacement value. Using this value, the same processing as in the above embodiment is performed. According to this embodiment, the same effect as that of the above embodiment can be obtained, and there is an effect that a measuring unit or the like is not required.

Another embodiment will be described with reference to FIG.
In this embodiment, the partial structure 2 and the vibrator of the embodiment of FIG. 1 are mounted on a shaking table 28, and a portion where the shaking table 28 is supported is vibrated when a vibration test is performed. It is driven by a power acceleration. According to this embodiment, since the measured reaction force includes the inertial force corresponding to the absolute acceleration, a highly accurate vibration test can be performed. In this embodiment, it is necessary to make a numerical model excluding the mass of the substructure. It is also possible to install an acceleration measuring means above the vibration, and to sequentially input and test the external force signal using the input means 16.

Another embodiment will be described with reference to FIG. In the present embodiment, in addition to the embodiment according to FIG. 1, a means 23 for automatically determining whether the eigenmode is treated as a second-order delay system, a first-order delay system, or a proportional system as preprocessing is added to the computer 5. It was done. The determination method is, for example, the magnitude of the natural frequency of each mode with respect to the reference value, and the reference value is determined in advance from the time interval of the vibration calculation and is input in advance. According to the present embodiment, the user can easily use the test apparatus.

Another embodiment will be described with reference to FIG. In the present embodiment, in addition to the embodiment according to FIG. 1, a means 24 for executing a vibration test of a small amplitude as a preprocessing on the computer 5 to measure the spring constant of the substructure, and the structure using the spring constant 1
The means 25 for performing the eigenvalue analysis is added. According to the present embodiment, the user can easily use the test apparatus, and the spring constant becomes accurate, so that the vibration test can be performed accurately.

Another embodiment will be described with reference to FIG. This embodiment is different from the embodiment of FIG. 1 in that the displacement convergence determination means 26 is provided, and the response calculation value is not predicted.
In this embodiment, the time management is not performed as in the time table of FIG. Also in this embodiment, since the time interval of the calculation can be made large, there is an effect that the number of test steps can be reduced and the test time can be shortened.

In the above description of the embodiment, the number of the test object structure 2 and the number of the vibrators 4 are one. However, even if one test object partial structure is vibrated by a plurality of the vibrators. No problem. Further, there may be a plurality of partial structures to be tested. A vibration test can be performed by a combination of the above description.

In the above embodiments, the part connected to the foundation as a partial structure has been described as an object of the vibration test. However, the present invention is not limited to this, and any part of the structure may be used.

Although the description has been made assuming that the calculation processing is performed by one computer, the data may be exchanged and shared by a plurality of computers as long as the above contents can be implemented. The computer is not limited to a computer having one CPU, and for example, a parallel computer having a plurality of CPUs can be used. For example, one C
By causing the PU to perform one mode of response calculation, it is possible to reduce the calculation time required for one step.

Further, in the above-described embodiment, the computer does not include a data input / output device, a display device, and the like, but has the same components as those of a normal computer. In addition, a system is used in which a plurality of computers are used, and one computer is used for processing such as data input and eigenvalue analysis, from which data is transferred to another computer and a vibration response calculation of a vibration test is performed. It is also possible.

Further, in the above description of the embodiment, a uniaxial vibrator was used, but in practice, it is necessary to adopt a vibrating method corresponding to the degree of freedom of the boundary of the numerical model. For example, it is conceivable to use a six-axis vibrator as shown in FIG. 49a-f are 6 actuators, 5
Reference numerals 0a to 0f denote bearings for mounting the respective actuators on the support base or the stage 51, and reference numeral 51 denotes a stage on which a structure is mounted.

In the above embodiment, the description has been given focusing on the vibration test portion. However, according to the present invention, it is also possible to calculate the vibration response other than the excitation point. Therefore, these calculation results can be stored in the computer or in a storage device attached to the computer, and can be processed after the test is completed. It is also possible to sequentially output the data to the outside of the computer during the test.

A vibration test method according to one embodiment of the present invention will be described with reference to FIGS. First, in step 31, a load value, a displacement value, and an external force value are input to a computer,
Calculate the input value for calculating the displacement, and then go to step 32
Calculates the displacement of the boundary point between the real model and the other part after a fixed time from the load measurement time of the vibrator, and calculates this displacement as a vibrator command signal in step 33. To enter. In this way, the real model is vibrated by the vibrator (step 34), and the measurement of the load (step 35) and the measurement of the displacement (step 36) (these two steps may be performed in reverse order or simultaneously). Then, the end is determined as appropriate (step 37). If not, the process returns to step 31.

Next, the contents of step 31 (input calculation step portion) will be described in detail with reference to FIG. First, the product of the measured displacement value and the spring constant is calculated (step 101), and the difference between the result and the measured load value is calculated (step 102). For example, when the number of modes using n and j is a counter, j = 0 is set (step 103), j = j + 1 (step 104) is obtained, and in step 105, both ends of the real model of the j-order mode vector are determined. The product of the difference between the elements to be performed and the result of step 102 is calculated. Next, the product of the external force value and the j-order mode stimulation coefficient is calculated (step 10).
6), the sum of the calculated value of step 105 and the calculated value of step 106 is calculated, and the result is input to the j-th mode (step 107). J <
Repeat as long as n (step 108). Note that the determination for exiting this loop may be different.

The contents of step 32 (displacement calculation step) following step 31 will be described in detail with reference to FIG. Note that n
1 is the number of modes to be calculated as a secondary delay. j = 0
(Step 111), j = j + 1 (step 11)
After obtaining 2), the response of the second-order lag system is calculated using the result of step 107 as an input (step 113). The product of the result of step 113 and the difference between the elements corresponding to both ends of the real model of the j-order mode vector is calculated (step 11).
4), j <n1 in steps 112 to 114
Is repeated (step 115). If j <n1 is not satisfied, the sum of the modes from the 1st to the n1th order is calculated (step 116), and the predicted value of the delay time ahead of the result of step 116 is calculated (step 117), and j = n1
(Step 118), j = j + 1 (step 11)
After obtaining 9), the response of the first-order lag system is calculated using the result of step 107 as an input (step 120). Next, in step 121, the product of the difference between the elements corresponding to both ends of the real model of the j-order mode vector and the result of step 120 is calculated.
Is repeated (step 122). If j <n, the sum of the n-th mode is calculated from (n1 + 1) of the result of step 121 (step 123), and then the sum of the result of step 117 and the result of step 121 is calculated (step 124). ).

FIG. 23 and FIG. 11 show an example of step 117 in step 32. m is a score used for prediction (see Expression 20), ak is a coefficient set in advance (see Table 1), and k is a counter. First, the result of step 116 is stored (step 201), and k = 0.
(Step 202), k = k + 1 (step 20)
After obtaining 3), in step 204, step 2
Then, Δxlk is obtained from 01, and the product of this and ak is calculated. Steps 203 and 204 are repeated as long as k <m (step 205). If k <m, step 204
Is calculated as a predicted value (step 20).
6). Thus, in the signal value calculation step, the prediction after a predetermined time of the sum of the response calculation values of the second-order lag system is performed by calculating the sum of the response calculation values of the past and the response calculation value of the current step at predetermined time intervals. The sum is multiplied by a predetermined coefficient and added together.

FIG. 24 is a flow chart of FIG.
An alternative to the third example is shown. That is, at step 210, the displacement measurement value and the displacement calculation value, that is, the prediction time is corrected by comparing the sum of step 116 and step 123,
Next, a predicted value is calculated (step 211). That is, in this example, the displacement applied to the real model from the shaker is measured by the vibration displacement measuring means of the shaker, and the measured value is input to the computer, and the computer calculates the displacement measured value and the shaker command signal calculation. The calculated value in the step (the sum of step 116 and step 123) is compared, and the set value of the shaker response delay time used for prediction by the shake signal calculating means is corrected.

FIG. 25 shows FIG.
3 shows still another example. That is, the difference between the measured displacement value and the value of the result of step 121 before the shaker response delay time is calculated (step 220), and the prediction time is corrected by comparing the result of step 220 with the result of step 116 (step 221). ). Next, a predicted value is calculated (step 211). In other words, in this example, the response of the shaker response delay time is corrected by multiplying the response calculation value of each system of the second-order delay system by the difference between the elements corresponding to both ends of the internal real model of the mode vector of each mode. The sum of the calculated value and the difference between the element corresponding to both ends of the mode model of each mode of each mode of the primary delay or the proportional system before the shaker response delay time set value used for the displacement measurement value and the prediction are calculated. The correction value is calculated by comparing the difference between the product and the sum.

Next, an alternative to the contents of step 102 in step 31 will be described with reference to FIG. That is, the excitation acceleration is measured or calculated in step 231, the product of the acceleration measurement or the calculated value and the coefficient determined as the equivalent mass of the real model is calculated in the next step 232, and the actual load is calculated in step 233. Measurements and Step 2
A difference between the result of step S <b> 32 is calculated, and a difference from step S <b> 101 is calculated using the result of step S <b> 233 as a load measurement value.

FIG. 27 shows an example of step 34 (see FIG. 20). That is, the shaker and the structure to be tested are mounted on the shaking table, and simultaneously with the vibration by the shaker (step 241), the vibration is applied by the acceleration applied to the structure, and the vibration is caused by the shaking table (step 242).

Next, a vibration test method according to another embodiment of the present invention will be described with reference to FIG. The difference from the embodiment of FIG. 20 is that the computer omits the second derivative term when the input natural frequency of the second-order lag system is larger than a predetermined value before conducting the vibration experiment. That is, a step 38 for substituting the first-order lag system. That is, step 38 is a process of allocating the secondary delay system and the primary delay system based on the natural frequency.

Another embodiment will be described with reference to FIGS. 29 and 30. First, the input is calculated in the same manner as in the embodiment of FIG. 20 (step 29), and then the displacement of the boundary point is calculated (step 42). Next, a shaker command signal is generated (step 43), and the real model is shaken by the shaker (step 44). Unlike the example of FIG. 20, in the next step, the displacement is measured (step 45), and it is determined whether or not convergence has occurred (step 46).
And if it converges, measure the load (Step 4)
7). Thereafter, an end determination is made (step 48), and the process returns to the input calculation step (step 41) unless the end is determined.

FIG. 30 is a detailed diagram of step 42 (calculation of the displacement of the boundary point) in FIG. 29. First, j = 0 (step 111), and then j = j + 1 (step 11)
After obtaining 2), the result of step 107 is input and 2
The response of the next-delay system is calculated (step 113). At step 114, the product of the difference between the elements corresponding to both ends of the real model of the j-order mode vector and the result of step 113 is calculated. Up to j
Repeat as long as <n1 (step 115), j <n1
If not, the sum of the result of step 114 for the 1-n 1st mode is calculated (step 116). Next, after obtaining j = n1 (step 118) and j = j + 1 (step 119), the response of the first-order lag system is calculated using the result of step 107 as an input (step 120). The product of the difference between the elements corresponding to both ends of the real model of the mode vector and the result of step 120 is calculated, and the above steps 119 to 121 are determined by j <
This is repeated as long as n is satisfied (step 122). j <n
, The result of step 121 is (n1 + 1) to n
The sum for the next mode is calculated (step 123),
Finally, the sum of the result of step 116 and the result of step 123 is calculated (step 124).

FIG. 31 shows still another embodiment of the present invention. This embodiment shows a method of performing the processes 31 and 32 in FIG. 20 in parallel. Processing 31 and processing 32 in FIG. 31 correspond to processing 31 and processing 32 in FIG. 20, respectively. The processes 301 and 302 in FIG. 31 correspond to the processes 101 and 102 in FIG.

Since the processing 105 in FIG. 21 can be independently processed for each mode j, FIG. 31 shows the processing 105 to 305b, 305c to 305d for each mode.
In parallel. That is, the processes of modes 1 to n are executed in parallel. Similarly, processing 106 and 107 in FIG.
Can be processed independently for each mode j.
Then, the processing is performed for each mode 306 to 306
b, 306c to 306d; processes 307 to 307b, 30
Parallel processing is performed in 7c to 307d.

Further, since the processes 113 and 114 in FIG. 22 can be independently processed for each mode j, the processes 313 to 313b;
Parallel processing is performed at 314b. The process 316 in FIG.
The process 317 in the second process 116 corresponds to the process 117.

Since the processes 120 and 121 in FIG. 22 can also be independently processed for each mode j, FIG. 31 shows processes 320 to 320b and 321 to 32 for each mode.
1b performs parallel processing. The process 323 in FIG. 31 corresponds to the process 123 in FIG. 22, and the process 324 corresponds to the process 124.
In FIG. 31, processes 313 to 313b and 314 to 31
4b, 316, 317 and processing 320 to 320b, 321
321b and 323 can be processed in parallel, and as shown in FIG. 31, these are processed in parallel.

One embodiment of a method for performing the above processing by a digital computer will be described below. The computer has a plurality of processors, means for exchanging data between the plurality of processors, and means for waiting for related processing among the plurality of processors, that is, inter-processor synchronization means.

As a method of exchanging data between the plurality of processors, for example, there are a method of communicating between the processors, a method of providing a shared memory accessible by each processor, and the like. As one method of configuring the shared memory, for example, there is a method using a shared resource access control device described in Japanese Patent Application Laid-Open No. 3-257649. According to this method, each processor can access the shared memory at high speed and efficiently.

As a method of waiting for related processes among the plurality of processors, for example, communication between the processors collects inter-processor synchronization information in a specific processor, and the specific processor synchronizes with another processor. A method of processing, a method of providing an inter-processor synchronization flag on the shared memory and performing a synchronization process using the flag, a method of using a hardware logic-based synchronization mechanism that directly executes a synchronization process in accordance with an instruction from the processor, and the like. is there. As an example of a hardware logic-based synchronization mechanism, there is a synchronization processing device between processors described in JP-A-5-2568. According to this synchronization method,
Synchronous processing between processors can be performed quickly and efficiently.

A shared memory system using the device described in JP-A-3-257649 and JP-A-5-256
The use of the synchronous processing device between processors described in Japanese Patent Publication No. 8 can reduce overhead caused by parallel processing,
The size of tasks processed by the processor can be reduced while maintaining high parallel processing efficiency. As a result, the parallelism possessed by the processing target can be sufficiently brought out, and the processing time can be reduced.

FIG. 32 shows an example of a digital computer for performing the above processing. The computer 348 is a processor (1) 341
~ Processor (n1) 342, Processor (n1 + 1) 3
43-processor (n) 344, for a total of n processors. A local memory is provided in each processor. In order to exchange data between a plurality of processors, each of the processors 341 to 342 and 343 to 3
44 has a shared memory 346 accessible via the bus 347. As an inter-processor synchronization means for waiting for related processing among a plurality of processors, an inter-processor synchronization flag is provided on a shared memory, and means for performing a synchronization process using the flag is provided. Here, the external input / output device 345 is configured such that
7 is a device for inputting and outputting data to and from the outside through

FIG. 34 shows another embodiment of the digital computer for performing the above-described processing. The digital computer 375 includes a total of n processors, i.e., a processor (1) 363 to a processor (n1) 364 and a processor (n1 + 1) 365 to a processor (n) 366. In each processor, a local memory and a local shared memory LCM3
71 to 374 are provided. The local shared memory is a shared memory between processors that accesses only the local shared memory in the processor i when reading from the processor i and writes write data to the local shared memory in all processors when writing. As means for configuring the local shared memory, for example, there is a means disclosed in JP-A-3-257649. By using the local shared memory described in this publication, contention for access to the shared memory can be reduced, access overhead to the shared memory can be reduced, parallel processing efficiency can be increased, and the parallel processing time can be reduced.

As an inter-processor synchronization means for waiting for related processing among a plurality of processors, an inter-processor synchronization mechanism 370 implemented as hardware logic is provided. An example of the inter-processor synchronization mechanism is disclosed in Japanese Patent Laid-Open No. 5-2 / 1993.
There is a synchronous processing device between processors described in Japanese Patent Application Publication No. 568. By using this synchronous processing device, the access overhead for the inter-processor synchronous processing can be reduced, the parallel processing efficiency can be increased, and the parallel processing time can be reduced. 368 and 369 indicate buses. The external input / output device 345 is a device in which each processor performs data input / output with the outside via the bus 347.

In the computer of this embodiment, each processor i may have a plurality of CPU units. Then, a shared memory accessible from a plurality of CPUs may be provided. When each processor i has two CPU units, a dual port RAM (DPR) accessible from each of the two CPUs may be provided. Incidentally, this kind of related technology is disclosed in Japanese Patent Application Laid-Open No. 61-1361.
No. 57 has a detailed description.

An example in which the processing shown in FIG. 31 is performed by a digital computer will be described with reference to FIG. 32 and subsequent figures. The processing 325 and 329 in FIG. 31 are performed by the processor (1) in FIG.
In FIG. 321, the processes 326 and 330 in FIG.
The processing (327, 331) in FIG. 31 is performed by the processor (n1) 342 in FIG.
In (+1) 343, the processes 328 and 332 in FIG. 31 are processed by the processor (n) 344 in FIG. Similarly, processing for the (n1-1) next mode is performed by the processor (2), the processor (3),.
..., processed by the processor (n1-1), (n1 + 2),
(N1 + 3),... (N-1) Processing for the next mode is performed by the processor (n1 + 2), the processor (n1 + 3),
..., Processed by the processor (n-1). Here, data to be transferred between the processors is stored in the shared memory 346.

In FIG. 31, the result of the process 302 is stored in the shared memory 346,
b, when the result of the process 302 is referred to in 305c to 305d, the result of the process 302 on the shared memory is referred to. Also, the results of the processes 314 to 314b in FIG. 31 are stored on the shared memory 346, and the results on the shared memory 346 are referred to in the process 316. Process 32 in FIG. 31
The results of 1 to 321b are stored in the shared memory 346, respectively, and the process 323 refers to the results in the shared memory 346. The results of the processes 317 to 323 in FIG. 31 are respectively stored in the shared memory 346, and the process 324 refers to the results in the shared memory 346. When synchronizing access to data on the shared memory between the processors, a synchronization flag is provided on the shared memory 346 and synchronization is performed using software.

As another method for synchronizing access to data in the shared memory between processors, there is a method using a synchronization processing device between processors described in Japanese Patent Application Laid-Open No. H5-2568. According to this method, there is an effect that synchronization processing can be performed at a higher speed than in the method of synchronizing using the above-described software. Further, as one method of configuring the shared memory 346, for example, Japanese Patent Application Laid-Open No. 3-25764.
There is a method using the shared resource access control device described in Japanese Patent Application Laid-Open No. 9-209. According to this method, the shared memory 346 can be accessed efficiently. That is, even if the computer shown in FIG. 34 is used, the processing shown in FIG. 31 can be performed in the same manner as in the above-described embodiment. Further, as a method of exchanging data between processors, communication between processors may be performed without using the shared memory 346. Here, input / output of data with the outside is performed using the external input / output device 345.

According to this embodiment, the time required for the processes 31 and 32 can be reduced. This is particularly effective when the degree of freedom n is large. In this embodiment, the processing for each next mode j is performed by the processor j.
The processing in the second and third order modes may be performed by the processor 1, the processing in the fourth, fifth and sixth order modes may be performed by the processor 2, and so on.

Another method for performing the processing shown in FIG. 31 using a digital computer will be described. A plurality of processors are divided into a plurality of processor groups, and a plurality of modes are respectively divided into a plurality of groups. Then, the arithmetic processing of each group is assigned to each of the plurality of processor groups. As an example of this, there is a method in which the processing of each group in the mode processed by each processor group is divided into a plurality of processing stages, and the processing in each stage is performed by each processor in each processor group. One embodiment is shown in FIG.
This will be described with reference to FIG.

FIG. 33 shows an embodiment of a digital computer. The computer has 2n processors of processor (1) a351, processor (1) b352, processor (n) a357, and processor (n) b358. Except for this point, the configuration is the same as that of the computer shown in FIG.

In FIG. 31, the processing groups are determined as follows. That is, the processing group 1 includes the processing 325 and the processing 329 for the primary mode, the processing group n1 including the processing 326 and the processing 330 for the n1st mode, the processing 327 and the processing 3 for the (n1 + 1) th mode.
31, a processing group (n1 + 1), a processing group n corresponding to the processing 328 and the processing 332 corresponding to the n-th mode, and similarly, a processing group for the 2, 3,..., (N1-1) next mode To the processing group (2),
Processing group (3),..., Processing group (n1-1)
, (N1 + 2), (n1 + 3),..., (N-1) The processing group for the next mode is the processing group (n1 +
2), processing group (n1 + 3),..., Processing group (n-1).

Each processing group defined in this way is divided into the following stages. That is, the processing group (1) is processed in the processing 325 and the processing 329, and the processing group (n1) is processed in the processing 32
6, processing 330, and processing group (n1 + 1) as processing 32
7 and processing 331, and processing group (n) are divided into processing 328 and processing 332 stages. 2,3, ..., (n
1-1) Processing group for next mode, (n1 + 2),
The processing group for the (n1 + 3),..., (N-1) next mode is also divided in the same manner.

In FIG. 33, the processor group is defined as follows. That is, a processor group i (i = i = i) including the processor (i) a and the processor (i) b
1, 2,..., N).

Then, the processing of the stage 325 is performed by the processor (1) a: 351 of the processor group 1, and the processing of the stage 329 is performed by the processor (1) b: 352. Similarly, the processors (n
1) The processing of the stage 326 is performed at a: 353, the processing of the stage 330 is performed at the processor (n1) b: 354, and the processor (n1) of the processor group (n1 + 1) is performed.
+1) a: The processing of the stage 327 is performed at 355, the processor (n1 + 1) b: The processing of the stage 331 is performed at 356, and the processor (n) a of the processor group n is:
The processing of the stage 328 is performed at 357, and the processing of the stage 332 is performed at the processor (n) b: 358. The processing of the other stages is performed similarly. Here, the processor ia
Data transfer between the processor and the processor ib is performed using the shared memory 360. Other data exchange between processors is performed in the same manner as in the above-described embodiment. With such a configuration, there is an effect that the processing shown in FIG. 31 can be performed at high speed.

In the present embodiment, taking the processing of the n-order mode as an example, the stage includes a stage of a process 328 including processes 305d, 306d and 307d shown in FIG. 31, and processes 320b and 321b. Although the stage is divided into the stage of the process 332, the process 328 and the process 3
The plurality of processes constituting the process 32 may be divided into other combinations of stages. For example, the processing 305d and the processing 306d may be set as one stage, and the processing 307d, the processing 320b, and the processing 321b may be set as another stage. Here, from the viewpoint of improving processing efficiency, it is preferable to configure the stages so that the processing time of each stage is substantially the same, and it is preferable to provide several processors in the processor group.

The processing shown in FIG. 31 can be performed using the computer shown in FIG. In this case, there is an effect that processing can be performed at a higher speed than in the case of using the computer shown in FIG. In particular, each processor shown in FIG. 34 has two CPU units and a DPR which can be accessed from both CPUs.
When the processing in each mode is divided into two stages and the processing is performed by each of the CPUs, data transfer between the CPUs can be performed by using the DPR.
Can be performed.

Although the present invention has been described with reference to the typical embodiments, the present invention is not limited to the above embodiments.
Various forms can be adopted without departing from the gist of the present invention.

[0187]

According to the present invention, since the divergence of the vibration response caused by the short-period eigenmode can be avoided, the time step of the vibration response calculation can be made large, and the real-time test can be easily performed. Further, in the semi-dynamic test, there is an effect that the test time can be reduced. Further, since various factors of errors in the real-time test can be removed, a highly accurate test can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of a vibration test apparatus according to the present invention.

FIG. 2 is a schematic diagram illustrating an example of a vibration response evaluation target structure.

FIG. 3 is a schematic view of a conventional vibration test method.

FIG. 4 is a characteristic diagram of dynamic characteristics of a vibrator.

FIG. 5 is a characteristic diagram of negative attenuation caused by a shaker response delay in a real-time test.

FIG. 6 is a flowchart of a vibration test using a mode space.

FIG. 7 is a time table diagram for performing a vibration test.

FIG. 8 is a schematic diagram of a mode frequency response.

FIG. 9 is an explanatory diagram of an effect of response delay prediction.

FIG. 10 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

11 is an explanatory diagram of a response prediction method according to the embodiment of FIG.

FIG. 12 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

FIG. 13 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

FIG. 14 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

FIG. 15 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

FIG. 16 is a flowchart of one embodiment according to the vibration test method of the present invention.

FIG. 17 is a flowchart of another embodiment according to the vibration test method of the present invention.

FIG. 18 is a block diagram showing another embodiment of the vibration test apparatus of the present invention.

FIG. 19 is a conceptual diagram of a six-degree-of-freedom shaker.

FIG. 20 is a flowchart of another embodiment according to the vibration test method of the present invention.

FIG. 21 is a detailed flowchart of a part of FIG. 20;

FIG. 22 is a detailed flowchart of a part of FIG. 20;

FIG. 23 is a detailed flowchart of a part of FIG. 22;

FIG. 24 is a detailed flowchart of a part of FIG. 22;

FIG. 25 is a detailed flowchart of a part of FIG. 22;

FIG. 26 is a detailed flowchart of a part of FIG. 21;

FIG. 27 is a detailed flowchart of a part of FIG. 20;

FIG. 28 is a flowchart of another embodiment according to the vibration test method of the present invention.

FIG. 29 is a flowchart of another embodiment according to the vibration test method of the present invention.

FIG. 30 is a detailed flowchart of a part of FIG. 29;

FIG. 31 is a flowchart showing one embodiment of the present invention.

FIG. 32 is a diagram showing one embodiment of a digital computer used in the present invention.

FIG. 33 is a diagram showing another embodiment of the digital computer used in the present invention.

FIG. 34 is a diagram showing another embodiment of the digital computer used in the present invention.

[Explanation of symbols]

1 ... Vibration response evaluation target structure, 2 ... Vibration test target partial structure, 3 ... Vibration test target partial structure other than partial structure,
DESCRIPTION OF SYMBOLS 4 ... Exciter, 5 ... Computer, 6 ... Load measuring means, 7 ... Exciter control device, 8 ... Displacement measuring means, 9 ... Transmitting means of load measurement value, 10 ... Transmitting means of displacement measuring means, 11 ... load measurement value input means, 12 ... displacement measurement value input means, 13 ...
Means for outputting a vibrator command signal, 14 means for transmitting a vibrator command signal, 15 means for storing an external force signal, 16 means for inputting an external force signal, 17 means for calculating a vibrator command signal, 171 ...
Input calculation means, 172: response value calculation means, 173: signal value calculation means, 18: signal generation means, 19: time management means, 20: response calculation value storage means, 21: acceleration measurement means, 211: acceleration measurement value transmission Means, 22: Acceleration measurement input means, 23: Mode automatic determination means, 24: Spring constant measurement means, 25: Eigenvalue analysis means, 26: Displacement convergence determination means, 27: Predicted time correction means, 28: Shaking table.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Masaki Kurihara 502 Kandate-cho, Tsuchiura-shi, Ibaraki Pref.Mechanical Research Laboratory, Inc. Within the Tsuchiura Plant (72) Inventor Kazuhiro Umekita 502 Machinery, Kandachicho, Tsuchiura-shi, Ibaraki Pref.

Claims (6)

[Claims] The present invention relates to a method for calculating a numerical model of an entire structure to be evaluated and mounting the numerical model on a digital computer. A part of the structure to be evaluated is used as a real model. A load and a displacement applied to the real model by a shaker are measured, and (1) a load value, a displacement value, and an external force value are input; (2) a real model after a fixed time from the load measurement time of the shaker and others. (3) generating and outputting a command signal so that the shaker command signal after the predetermined time is input to the shaker control device; In the vibration test method for a structure, the step of performing time management in the digital computer and performing the predetermined time interval and continuously repeating until a predetermined number of times or a stop signal is input, the numerical model includes: Excitation target structure line Of the eigenmodes obtained by estimating the eigenfrequency and eigenmode by eigenvalue analysis by numerical modeling, at least one mode is a second-order lag system, and at least one of the remaining eigenmodes omits the inertial term A command signal value to the vibrator is calculated by an input calculation step, a response calculation step of a secondary delay system and a primary delay system or a proportional system, and a signal value is calculated. A vibration test method, which is performed by a calculation step, wherein the plurality of modes are divided into a plurality of groups, and a process of each group is processed by a plurality of processors included in the digital computer. 2. The whole structure to be evaluated is numerically modeled and mounted on a digital computer, and a part of the structure to be evaluated is used as a real model, which is vibrated by a vibrator connected to the real model. A load and a displacement applied to the real model by a shaker are measured, and (1) a load value, a displacement value, and an external force value are input; (2) a real model after a fixed time from the load measurement time of the shaker and others. (3) generating and outputting a command signal so that the shaker command signal after the predetermined time is input to the shaker control device; In the vibration test method for a structure, the step of performing time management in the digital computer and performing the predetermined time interval and continuously repeating until a predetermined number of times or a stop signal is input, the numerical model includes: Excitation target structure line Of the eigenmodes obtained by estimating the eigenfrequency and eigenmode by eigenvalue analysis by numerical modeling, at least one mode is a second-order lag system, and at least one of the remaining eigenmodes omits the inertial term A command signal value to the vibrator is calculated by an input calculation step, a response calculation step of a secondary delay system and a primary delay system or a proportional system, and a signal value is calculated. A plurality of processors are divided into a plurality of processor groups, a plurality of modes are divided into a plurality of groups, and arithmetic processing of each group is performed by the plurality of processor groups. A vibration test method characterized in that the vibration test is performed by assigning to a vibration test. 3. The processing of each group in a mode processed by each processor group is divided into a plurality of processing stages,
3. The vibration test method according to claim 2, wherein the processing for each stage is performed in each processor in each of the processor groups. 4. The whole structure to be evaluated is numerically modeled and mounted on a digital computer, and a part of the structure to be evaluated is used as a real model, which is vibrated by a vibrator connected to the real model. A load and a displacement applied to the real model by a shaker are measured, and (1) a load value, a displacement value, and an external force value are input; (2) a real model after a fixed time from the load measurement time of the shaker and others. (3) generating and outputting a command signal so that the shaker command signal after the predetermined time is input to the shaker control device; In the vibration test method for a structure, the step of performing time management in the digital computer and performing the predetermined time interval and continuously repeating until a predetermined number of times or a stop signal is input, the numerical model includes: Excitation target structure line Of the eigenmodes obtained by estimating the eigenfrequency and eigenmode by eigenvalue analysis by numerical modeling, at least one mode is a second-order lag system, and at least one of the remaining eigenmodes omits the inertial term A command signal value to the vibrator is calculated by an input calculation step, a response calculation step of a secondary delay system and a primary delay system or a proportional system, and a signal value is calculated. A plurality of processors of the digital computer are divided into a plurality of processor groups, and the plurality of modes of the input calculation step are divided into N (N is a positive integer of 2 or more) groups. The arithmetic processing of each group is processed by a plurality of processors of the digital computer, and a plurality of modes of the second-order lag-based response calculation step are calculated. Are divided into the N groups, the arithmetic processing of each group is processed by the plurality of processors, and the plurality of modes of the response calculation step of the primary delay system or the proportional system are divided into the N groups. A vibration test method in which arithmetic processing of each group is processed by the plurality of processors. 5. A vibrator connected to a structure to be vibrated and vibrating the structure, a control device for the vibrator, and a means for measuring a load applied to the structure from the vibrator. Measuring means for measuring the displacement applied to the structure from the shaker, and a digital computer, the digital computer, time management means,
Input means for the output of the load measuring means, input means for the output of the displacement measuring means, storage means for external force values and / or input means for external force signals, and load measurement by the external force values or signals and the load measuring means Means for calculating a command signal value to the vibrator after a predetermined period of time from the measurement time of the load measurement value or the signal using the value or the signal, and a means for outputting a command signal of the vibrator. , An external data input / output unit, a signal transmission unit from the digital computer to the shaker control device, a signal transmission unit from the load measurement unit to the computer, and the displacement measurement unit In a vibration test apparatus for a structure, comprising: a signal transmission unit for transmitting a signal to the computer; a unit for calculating a command signal value to the vibrator; an input calculation unit; a second-order delay system and a first-order delay system; system Includes a response calculation means, and a signal value calculating means, the digital computer, a plurality of processors,
Means for exchanging data between the plurality of processors and means for waiting for related processing among the plurality of processors, wherein different ones of the plurality of processors are N (N is a positive integer of 2 or more) N) groups different in a plurality of modes in the input calculation step grouped into N groups, and different groups in a plurality of modes in the response calculation step of the second-order delay system grouped into N groups; A vibration test apparatus for performing arithmetic processing on each of the groups of a plurality of modes in the first-order lag-based or proportional-system response calculation step which are grouped into a plurality of groups. 6. A vibrator connected to a structure to be vibrated and vibrating the structure, a control device for the vibrator, and means for measuring a load applied to the structure from the vibrator. Measuring means for measuring the displacement applied to the structure from the shaker, and a digital computer, the digital computer, time management means,
Input means for the output of the load measuring means, input means for the output of the displacement measuring means, storage means for external force values and / or input means for external force signals, and load measurement by the external force values or signals and the load measuring means Means for calculating a command signal value to the vibrator after a predetermined period of time from the measurement time of the load measurement value or the signal using the value or the signal, and a means for outputting a command signal of the vibrator. , An external data input / output unit, a signal transmission unit from the digital computer to the shaker control device, a signal transmission unit from the load measurement unit to the computer, and the displacement measurement unit In a vibration test apparatus for a structure, comprising: a signal transmission unit for transmitting a signal to the computer; a unit for calculating a command signal value to the vibrator; an input calculation unit; a second-order delay system and a first-order delay system; system A plurality of groups, each of which includes a response calculation unit and a signal value calculation unit, and is grouped into N groups for performing arithmetic processing on a plurality of modes grouped into N (N is a positive integer of 2 or more) groups Processor, means for exchanging data between the plurality of processors, and means for waiting for related processing among the plurality of processors, wherein processors included in different groups perform processing of modes of different groups. A digital computer used in a vibration test device, characterized in that: