JPH11230850A - Testing device and method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute a vibration test even when a nonlinear element is included at a numerical modeling part by generating a control signal for a driving means with a boundary point response with a sample as a target value. SOLUTION: A known outer force is calculated in a predetermined step (i), a counterforce from a sample part 201 that is measured by a counterforce sensor 118 is collected, and a counterforce ff1 being applied to the numerical modeling part is calculated. Then, a counterforce ff2 for generating a nonlinear element within the numerical modeling part is calculated based on a response displacement pf2 of the nonlinear element within the numerical modeling part. According to the result, a vibration response vector (x) at the numerical modeling part in a step i+1 is calculated. According to the result, a vibration response pf1 of the boundary point with the sample part 201 at time i+1 is calculated, and a command value 107 to an actuator controller 106 for vibrating the sample part 201 is calculated with the response pf1 as a target value. Also, the vibration response pf2 of the nonlinear element within the numeric modeling part in a step x+1 is calculated and outputted from the vector x calculation result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a test apparatus and method for performing a vibration test on a structure, and more particularly to a vibration test for only a part of a structure and a numerical analysis of a vibration response of another part in real time. The present invention relates to a test apparatus and a method for performing the test in combination.

[0002]

2. Description of the Related Art Conventionally, when a vibration test is performed on a structure, a test using a vibration table is often used. Further, when the target structure becomes large and exceeds the loading limit of the shaking table, a shaking table test using a reduced model or a shaking table test using a partial model has been performed. However, in a structure such as a nonlinear structure in which the similarity rule is hardly established or a structure having a large coupling of vibration with the surrounding structure, a shaking table test using the reduced model or a shaking table test using the partial model is used. Could be inaccurate.

In order to solve this problem, a vibration test using an actuator of only a part of a structure (hereinafter, referred to as a specimen) and a vibration response numerical analysis of a numerical model of another part are performed in real time. A method of a vibration test of a structure (hereinafter, referred to as a real-time hybrid vibration test) is described in JP-A-5-10846.

[0004]

In the real-time hybrid vibration test, in order to perform a vibration response numerical analysis, it is necessary to numerically model a target portion (hereinafter referred to as a numerical modeling section) and solve a vibration response equation. is there. Here, a nonlinear stiffness element that exhibits nonlinear stiffness with respect to displacement and a nonlinear damping element that exhibits nonlinear decay with respect to velocity (hereinafter collectively referred to as nonlinear elements) are included in the numerical modeling unit. The vibration response equation when included (hereinafter referred to as non-linear vibration response equation) is the vibration response equation when only the linear element is not included in the numerical modeling unit (hereinafter referred to as linear vibration response equation). ), It may be necessary to repeat the convergence operation in some cases, and the time required to obtain the solution greatly increases. Therefore, it has been difficult to set a numerical modeling unit including such a non-linear element in a real-time hybrid vibration test that requires high-speed vibration response analysis of the numerical modeling unit in real time.

[0005] It is an object of the present invention to provide a test apparatus and method capable of executing a vibration test in the real-time hybrid vibration test even when the numerical modeling unit includes a nonlinear element.

[0006]

According to the present invention, there is provided a driving means for driving a real object or a model which is a part of an evaluation object as a test object, and a reaction force generated from the test object in the driving means. A reaction force measuring means for measuring as a first reaction force, and a portion other than the test piece of the object to be evaluated is divided into a first portion and a second portion, and each is defined as a first and a second numerical model. A first calculating unit for calculating a reaction force applied to the first digitized model from the first reaction force measured by the reaction force measuring unit; and a response to the response of the second digitized model. A second calculation means for calculating a reaction force generated by the second numerical model as a second reaction force; and a first and second reaction force calculated by the first and second calculation means. Calculating a response of the first numerical model from the applied external force; Third calculating means for calculating the response of the boundary point with the specimen from the calculation result of the means, and the response of the boundary point calculated by the means as a target value. Fifth calculating means for generating a control signal to the driving means and outputting the control signal to the driving means, and calculating a response of the second digitized model from a response calculated by the third calculating means The sixth calculating means for outputting the result to the second calculating means, and the processing by the first, second, third, fourth, fifth and sixth calculating means are cyclically performed. A test device comprising: a control unit for performing control so as to be executed.

[0007] The present invention also provides a second numerical model which divides a portion other than the specimen of the object to be evaluated into the first portion and the second portion and represents the second portion. Disclosed is a test apparatus including a numerical model division designation unit for generating a constant to be used.

Further, according to the present invention, the first to sixth calculation means are constituted by a plurality of processors, and
While the processor constituting the calculating means is operating, the second
A test apparatus characterized in that processors constituting other calculation means are configured to operate in parallel.

The present invention also discloses a test apparatus, wherein the second part is a part including an element in which a relationship between a response and a reaction force to the response is non-linear.

Further, according to the present invention, the object to be evaluated is divided into a real model part or a real model part thereof and a remaining part, and the remaining part is numerically modeled. A display device for displaying a numerical model in a test apparatus that obtains an overall vibration response of an evaluation object by performing a vibration test and a vibration response calculation of the numerically modeled part in real time, Pointing means for specifying a non-linear element having a non-linear relationship between the response of the element and the reaction force generated by the element on the numerical model displayed by the display means, Position information creating means for creating position information indicating the position of the nonlinear element specified by the means on the digitized model; and a response information and a reaction force of the nonlinear element. It discloses a test apparatus characterized by having a calculating means for performing a vibration response calculation of the numerical modeling portion using the position information created by the model and the position information creating means for indicating the engagement.

Further, according to the present invention, a real object or a model forming a part of an object to be evaluated is used as a specimen, and a portion other than the above-mentioned specimen of the object to be evaluated is divided into a first portion and a second portion, and each is numerically divided. The first calculation is performed by modeling the first and second digitized models to measure the first reaction force generated when the specimen is driven and calculating the reaction force applied to the first digitized model from the measured first reaction force. Executing the reaction force generated by the second numerical model with respect to the response of the second numerical model
Executing a second calculation for calculating the reaction force, and calculating a response of the first numerical model from the first and second reaction forces calculated by the first and second calculations and the applied external force; A third operation to calculate the response of the boundary point with the specimen from the third operation result, and calculate the response of the boundary point calculated by the fourth operation. A fifth operation is performed in which the specimen is driven as a target value, a response of the second numerical model is calculated from the response calculated by the third operation, and the response is used in the second operation. And performing a test on the evaluation object by controlling the first, second, third, fourth, and fifth operations to be performed cyclically. A method is disclosed.

Further, according to the present invention, the object to be evaluated is divided into a real model part or a real model part thereof and a remaining part, and the remaining part is converted into a numerical model. By performing the vibration test and the vibration response calculation of the numerically modeled portion in real time in combination with each other, an element included in the numerical model in a test method for obtaining the overall vibration response of the evaluation object. ,
A non-linear element having a non-linear relationship between the response of the element and the reaction force generated by the element is specified by a pointing device on a numerical model displayed on the display means, and the specified numerical model of each specified nonlinear element is displayed on the numerical model. The position information indicating the positional relationship with the other elements is created from the digitized model, and the model indicating the relationship between the response of the nonlinear element and the reaction force and the position information are used for the numerically modeled part. A test method characterized by performing a vibration response calculation is disclosed.

[0013]

Embodiments of the present invention will be described below in detail. Note that, in the following description, symbols enclosed by ｛｝ represent vectors, and symbols enclosed by <> represent a matrix. For example, the vector a is expressed as {a}, and the matrix B is expressed as <B>.

FIG. 2 shows an example of an object to be evaluated in the real-time hybrid vibration test, in which a mass m1, a spring element k1,
The damping element c1 forms one degree of freedom system, and the mass m2, the spring element k2, and the damping element c2 form another degree of freedom system. Here, the spring element k1 is used as the specimen 201, a vibration experiment is performed by an actuator, and the numerical modeling unit 202 including the remaining elements is simulated by a computer. now,
<M>, <C>, and <K> are the mass matrix, damping matrix, and stiffness matrix of the numerical modeling unit, respectively, and {x} is the basis of the numerical model (the stage on which the sample is placed).
, {P} is an external force vector due to an earthquake or the like, and {ff} is a vector of a force (hereinafter referred to as a reaction force) generated at a boundary point between the numerical model and the real model. In the case of an evaluation target consisting only of linear elements, which has been targeted by the prior art, each of the matrices <M>, <C>,
<K> is a constant, and the equation of motion of the numerical modeling unit is

(Equation 1) And multiply. Here, the reaction force {ff} is measured by a vibration experiment, and this is used as an external force to solve (Equation 1) to calculate a vibration response {x} after a short time Δt. The actuator is driven by setting {pf} as a target value, and this displacement is added to the specimen 201.
The reaction force {ff} to this deformation is measured. By repeating the above, the vibration response of the entire structure can be evaluated.

As a solution of the above (Equation 1), <M>,
Since <C> and <K> are constants, they can be solved using, for example, the central difference method. The result is

(Equation 2)

(Equation 3) Becomes Where {xb} is the displacement of each node at the current time, {xb
bb} is the displacement of each node before the time Δt, <FI> is the reaction force direction matrix, {ff} is the reaction force, <AI> is the acceleration input direction matrix, {a} is the input acceleration vector,
{X} is the displacement of each node after the time Δt, {pf} is the excitation displacement vector to be given to the specimen, and <XO> is the displacement output direction matrix. That is, <FI> {ff}
Indicates the force acting on the component of {x} by the reaction force {ff},
<AI> {a} indicates a force acting on each component of {x} by the input acceleration {a} (that is, an external force due to the input acceleration).
XO> {x} is {pf} caused by displacement {x}
Are shown.

(Equation 1) to (Equation 3) are general formulas. When a two-degree-of-freedom system shown in FIG. 2 is subjected to a unidirectional vibration as shown in FIG. 2, the reaction force from the specimen 201 is ff1. given that,
In (Equation 2) and (Equation 3),

(Equation 4)

(Equation 5)

(Equation 6)

(Equation 7)

(Equation 8)

(Equation 9)

(Equation 10)

[Equation 11]

(Equation 12) It is good.

The solution of the numerical modeling unit according to (Equation 3) to (Equation 12) is directed to a linear element, but cannot be analyzed as it is for a nonlinear element. Therefore, a two-degree-of-freedom system which gives a one-way excitation shown in FIG. 2 is considered as a vibration response evaluation object, and a case where the spring element k2 of the numerical modeling unit 202 is a non-linear element and the other elements are linear elements. Think.
In this case, in the present invention, in (Equation 2) and (Equation 3),
The dimensions of the reaction force vector {ff} and the excitation displacement vector {pf} are extended by the number of nonlinear elements. In this case, since the number of nonlinear elements is one, the above vectors {ff}, {pf}
Are respectively extended from a 1 × 1 matrix to a 2 × 1 matrix. Accordingly, the matrix <FI> is also expanded from the 2 × 1 matrix to the 2 × 2 matrix, and the matrix <XO> is also expanded from the 1 × 2 matrix to the 2 × 2 matrix. That is, the vectors and matrices of (Equation 2) and (Equation 3) are replaced with (Equation 4) to (Equation 12).

(Equation 13)

[Equation 14]

(Equation 15)

(Equation 16)

[Equation 17]

(Equation 18)

[Equation 19]

(Equation 20)

(Equation 21) And Here, the 2-row, 1-column element pf2 of the expanded vector {pf} is a vibration response displacement generated in the nonlinear element k2, and the 2-row, 1-column element ff2 of the expanded vector {ff}
Is a reaction force generated by the nonlinear element k2 with respect to the vibration response displacement pf2. Therefore, the expanded <FI> {ff}
Indicates the force acting on each component of {x} as a result of reflecting both the reaction force from the specimen and the reaction force from the nonlinear element k2,
<XO> {x} is {p} caused by displacement {x}
Each component of f｝ is shown. Therefore, for the vibration response displacement pf2, the reaction force ff2 generated by the nonlinear element k2 is calculated according to the characteristic of the nonlinear element k2, and this is substituted into (Equation 2) to obtain {x} at the next time. Can be calculated. That is, the reaction force ff2 generated by the nonlinear element k2 is modeled in response to the vibration response displacement pf2 of the nonlinear element k2 according to the characteristics of the nonlinear element k2, and the vibration response displacement pf2 is calculated according to this model. And the nonlinear element k2
Is calculated.

FIG. 1 is a flowchart showing the procedure of the above-described test method of the present invention.
The algorithm 102 which is a feature of the present invention is installed. The control device 101 is equipped with a means 103 for managing time intervals for calculating the algorithm 102. To execute the algorithm 102, the following preparations are required. This preparatory work will be described with reference to the example of FIG. 2. An actual test in which a part of the object to be evaluated is used as a test part 201 using a real object or a model, and the test part 201 can be excited by the actuator 105 A unit 119 is prepared. Further, the numerical modeling unit 202, which is another part of the evaluation object, is divided into a first part consisting of masses m1 and m2 which are linear elements and damping elements c1 and c2, and a second part which is k2 which is a nonlinear element. Separate. Then, the first part is modeled as a first numerical model, and the second part is
At least one element (k2) is provided, and the response of each element is referred to as a second response (pf2). Each reaction force generated by each element of the second portion corresponding to the second response is When a second reaction force (ff2) is set, at least one model indicating a relationship between the second response and the second reaction force is modeled as a second numerical model. When the preparation is completed in this way, the processing of the algorithm 102 shown in the following 1) to 7) is executed.

1) The known external force in step i determined in advance is calculated (step 110). As this method, for example, a method of collecting a predetermined input acceleration {a} and calculating an external force due to the input acceleration based on the collected input acceleration {a} may be adopted.

2) The reaction force from the specimen 201 measured by the reaction sensor 118 is collected (step 11).
6) The reaction force ff1 applied to the numerical modeling unit from the specimen unit 104 is calculated (step 111).

3) Based on the response displacement pf2 of the non-linear element in the numerical model part, the reaction force ff2 generated by the non-linear element in the numerical model part is correspondingly calculated by the reaction force model (the second numerical model). (Step 11
2).

4) From the results of 1), 2) and 3) above, based on (Equation 2)) and (Equations 13) to (Equation 21) (based on the first numerical model), a numerical modeling unit in step i + 1 Calculate the vibration response {x} (step 11)
3).

5) From the result of the above 4), (Equation 13)
Based on (Equation 18), (Equation 19) and (Equation 20), the vibration response pf1 at the boundary point with the specimen at time i + 1 is calculated and output (Step 114), and this pf1 is set as a target value. Calculates and outputs a command value 107 to the actuator controller 106 that excites (step 117).

6) From the result of 4), (Equation 13)
Based on (Equation 18), (Equation 19), and (Equation 20), calculation and output of the vibration response pf2 of the nonlinear element in the numerical model unit in step i + 1 are performed (step 115).

6) If the test is to be terminated, the calculation process is terminated; otherwise, the process returns to 1).

FIGS. 3 and 4 show examples of the nonlinear element model used in step 112 of FIG. In these figures, corresponding to the model in FIG. 2, the response on the horizontal axis is the vibration response displacement pf2, and the reaction force on the vertical axis is the reaction force ff2 applied to the nonlinear element.
These nonlinear element models are prepared in a computer in the form of mathematical expressions or look-up tables. FIG. 3 shows a hysteresis characteristic which is often seen, and FIG. 4 shows a hysteresis characteristic obtained by a polygonal line approximation in which the degree of approximation generally decreases but the calculation speed is high.

According to the test method of the present invention described above, the calculation of the non-linear element characteristic is involved only in step 112 of FIG. 1, and as described later, this step and other steps are performed in parallel. Easy to do. Therefore, compared to solving a single nonlinear equation including a nonlinear element as in the related art, the calculation is speeded up, so that the test time interval can be reduced and the real-time hybrid vibration test can be easily realized. .

Next, the case of the evaluation object in FIG. 2 has been described above as an example. However, even when the vibration direction with respect to the specimen is not one direction but multiple directions, (Equation 2) and (Equation 3) , (Equation 1
Since 3) to (Equation 21) are described using generalized coordinates, it can be easily handled by increasing the dimensions of matrices and vectors. At this time, a plurality of actuators may be connected to the specimen using a link mechanism. In this case, in processing steps 111 and 116, a numerical model is formed based on the reaction force from the specimen measured by the reaction force sensor. A calculation for obtaining the reaction force ff1 applied to the portion is performed, and in step 117,
From the excitation displacement to be applied to the specimen, a displacement target value of each actuator is calculated in consideration of the link mechanism, and the command value to the servo controller of the actuator that excites the specimen is calculated and output using the target value as the target value. By doing so, it is possible to respond. Further, processing for compensating for the delay of the actuator may be added to step 117. At this time,
The delay of the response realization value of the actuator from the displacement target value of the actuator can be suppressed to a predetermined value or less. Further, the algorithm 102 may be executed in real time.

Further, even when the number of nonlinear elements included in the numerical modeling unit becomes plural, the vector {ff},
This can be dealt with by extending the dimensions of {pf} and the matrices <FI> and <XO>. That is, the present invention is established even when there are a plurality of vibration responses of a plurality of non-linear elements and a plurality of reaction forces generated corresponding thereto. Similarly, the present invention is established even when there are a plurality of reaction force models.

Further, in FIG. 1, a model showing a relationship between a vibration response at a boundary with the specimen and a reaction force generated by the specimen when the vibration response is given to the specimen is defined as a specimen model. And a reaction force generated in response to this response from the specimen unit using the response input from the control device 101 based on the specimen model, and output the calculated reaction force to the control device 101. In step 117, a test simulator is used in place of the actual test unit 119, and in step 117, the vibration response pf1 of the boundary point with the test unit output from step 114 is output to the test simulator. Step 11 calculates the reaction force output from the sample simulator.
The present invention can be configured and realized without using an actual specimen, an actuator, an actuator controller, and a reaction force sensor by using the collected reaction force of 6.

Further, in the above embodiment, an example in which the vibration response is a displacement has been described.
And any one of acceleration and acceleration or a combination thereof. In addition, the example in which the reaction force modeling of the nonlinear element is performed in the form of the reaction force generated for the displacement response has been described, but in the modeling of the reaction force for the response of the nonlinear element, the response includes displacement, velocity, and acceleration. One of
Alternatively, a combination of these may be used.

Next, various configurations of the computer for executing the algorithm shown in FIG. 1 will be described below. First, a single computer may be used for this purpose, but in order to realize a sufficient processing speed, it is preferable that the processing is shared among a plurality of processors. FIG. 5 shows algorithm 1
02 is 3 for processor A, processor B, and processor C
An example of a method of processing using one processor will be described. The number of each processing in the figure is the number of each processing step in the algorithm 102 shown in FIG. The dotted lines in the figure indicate the synchronization processing between the processes in charge of each processor. Steps 110 and 111 and 116 and 112 can be processed in parallel, respectively, and the same applies to steps 114 and 115. Therefore, the processing time of the algorithm 102 can be reduced by the parallel processing shown in FIG. Therefore, according to the embodiment, it is possible to shorten a calculation time interval (hereinafter, referred to as a calculation time step), and it is possible to improve a test accuracy.

FIG. 6 shows an example of a method of processing the algorithm 102 by using two processors, the processor A and the processor B. The number of each processing in the figure is the number of each processing step in the algorithm 102 shown in FIG. The dotted lines in the figure indicate the synchronization processing between the processes in charge of each processor. This processing method is effective when the step 112 dealing with nonlinearity takes a long time, and as shown in FIG.
Processing time can be shortened.

FIG. 7 shows an example of a method of processing the algorithm 102 by using four processors A, B, C and D. The number of each process in the figure is the number of each process in the algorithm 102 shown in FIG. The dotted lines in the figure indicate the synchronization processing between the processes in charge of each processor. Here, FIG. 7 shows an example in which the processing step 112 has a higher load. As an example of this, for example, when the number of nonlinear elements becomes two, the processing step 112 is performed by processing each of the two nonlinear elements 112 (a) and 112 (b).
And each of these processes can be processed in parallel, so that the processors C and D respectively perform the processes. Further, since the processing step 113 is a matrix operation, it can be processed in parallel, for example, by dividing every several rows. The processing divided in this way is the processing steps 113 (a), 113 (b), 113
(C) and 113 (d), and each of these processes is processed by each of the processor A, the processor B, the processor C, and the processor D.

FIG. 8 shows an example of a method of processing the algorithm 102 by using two processors, the processor A and the processor B. The number of each processing in the figure is the number of each processing step in the algorithm 102 shown in FIG.
The dotted lines in the figure indicate the synchronization processing between the processes in charge of each processor. In this example, algorithm 10
The processing steps other than the processing step 112 in 2 are processed by the processor A, and only the processing step 112 is processed by the processor B. In this example, since the results of the processing steps 110, 111, and 112 are used in the processing step 113, the processor A and the processor B can operate in parallel until the start of the processing step 113. This example corresponds to the case where steps 114 and 115 in the case of FIG. 6 are processed by the same processor.

In FIGS. 5 to 8 described above, several parallel processing methods have been described, but all of them schematically show the processing relationship between the processors. In practice, it is necessary to accurately determine the method of synchronization, the relationship of data between the processing steps, and the like. This will be described in detail below using the processing method of FIG. FIG. 11 shows the processing method of FIG. 8 in more detail, and is suitable when the processing step 112 is under a heavy load. This is, for example, the case where the number of nonlinear elements is plural, or the case where a complicated nonlinear element model is adopted as a model showing the relationship between the response pf2 and the reaction force ff2.

In FIG. 11, at times ti, ti + 1..., The two processors are synchronized and the algorithm 10 in FIG.
This is the time when 2 has made one round. The interval ΔT is constant, and each time slot is represented by the code SLi−1, SLi + 1.
... are indicated. Now, the vibration response displacement pf of the nonlinear element calculated in step 115 of time slot SLi-1
Since 2 is the displacement at time ti, this is called pf2 (i)
Is shown. Then, the response pf2 (i) is input, and in step 112 of the time slot SLi, the response p
The generated reaction force ff2 (i) corresponding to f2 (i) is calculated by the processor B. This reaction force ff2 (i) should be originally used in the processing step 113 in the time slot SLi. However, if the processing load of the processing step 112 is heavy, this calculation cannot be performed in time.
(I) is the processing step 11 of the time slot SLi + 1.
3, and a shift of time ΔT occurs. Although this deviation is a factor of an error in the test, on the other hand, the calculation time interval ΔT can be reduced by the parallel processing, and the error can be reduced as the time interval ΔT decreases. The parallel processing has an effect that the test accuracy can be improved comprehensively.

In the description of FIG. 11, the processing of step 110 by processor A and the processing of step 11 by processor B
2 is the start time of each time slot ti, ti + 1 ...
At the same time, the start times of these processes do not necessarily have to coincide. FIG. 12 shows an example of a processing method in which the start times do not match each other, and the phases of the time slots SLAi, SLAi + 1... Of the processor A and the time slots SLBi, SLBi + 1. It is. This phase shift ΔS may be a negative value, and the processing in step 112 in the time slot SLBi is different from the processing in step 1 in the time slot SLAi-1.
15 which is started after the completion of the processing and the time slot SL
It suffices that the setting is made so as to end before the start time of step 113 of Ai + 1.

In FIG. 12, two processors A,
Although the phase shift ΔS of the time slot B is fixed, a timing synchronization system is required between the two processors. But again, this need not be constant. That is, when the response pf2 is calculated in step 115 of the processor A like a data flow computer, a flag is set on the shared memory of the two processors, the result is stored, and the processor B is notified that the preparation is completed. Processor B
Always checks this shared memory, and when it finds that the flag is set, starts the processing of step 112.
The result can be notified to the processor A in a similar manner, so that the processing can be performed with synchronization between the two processors. FIGS. 11 and 12 show details when the processing method of FIG. 8 is used with two processors.
It goes without saying that the same applies to the case of FIG.

FIG. 13 shows another configuration example for executing the processing method by the processor B in FIG. 8, and is particularly suitable for the processing method shown in FIG. In FIG. 13, a processor A 1301 and a processor B 1302
/ A converter 1303, signal line 1307, A / D converter 1
304, D / A converter 1306, signal line 1308, A /
It is connected via a D converter 1305, and the transmission of information between the two processors is performed by temporarily converting the information into an analog signal. Computers (processors) used for vibration test devices and the like often have an analog boat as an interface with actuators and sensors, and therefore, may have built-in A / D and D / A converters in FIG. Many. Therefore, the configuration in FIG. 13 has an advantage that such a computer can be configured only by connecting the computers with signal lines.

FIG. 14 shows another example of a two-processor configuration, in which two processors 1401 and 1402 are connected via communication devices 1403 and 1404. The feature of this configuration is effective, for example, when the two processors are separate systems, particularly when they are separated from each other.

Next, when executing the algorithm shown in FIG. 1, it is necessary to set a constant according to the evaluation object, particularly when specifying a nonlinear element in the numerical modeling unit.
There are a matrix <FI> and a matrix <X0> as constants relating to this, and these are determined by the structure to be evaluated. Then, compared to the case of the linear system, it is necessary to increase the number of columns of the matrix <FI> and the number of rows of the matrix <X0> by the number of nonlinear elements being numerically modeled. For this purpose, a tool that displays an input screen for each matrix on a display and inputs each matrix from this screen to the numerical model may be prepared. Alternatively, a file containing the information of each matrix may be separately prepared and read. When the evaluation target does not change, a nonlinear element may be designated in the numerical modeling unit in the program that executes the algorithm 102.

The following relationship may be provided between matrix <FI> and matrix <X0>.

(Equation 22)

(Equation 23) At this time, the matrix <FI> or the matrix <
XO> may be input using a display or the like as described above, and the other may be calculated as a transposed matrix as shown in (Equation 22) or (Equation 23).
As the matrices <FI> and <XO> determined as described above, when the evaluation target is shown in FIG.
It goes without saying that this is given by (Equation 18). An example of the <FI> input screen at this time is shown in FIG.
FIGS. 16 and 17 show examples of input screens for both I> and <XO>.

Further, for example, as shown in FIG. 15, a case where a three-degree-of-freedom system composed of a numerical modeling unit 302 including two nonlinear elements (spring elements k2 and k3) and a specimen unit 301 is to be tested. Is that information in the second and third columns of the matrix <FI> indicates the position of the nonlinear element,

(Equation 24) And the matrix <XO> is given by, for example, this transposed matrix. Further, (Equation 24) can be further generalized. Assuming that element kh in FIG. 15 is a specimen portion, element ki is a first nonlinear element, and element kj is a second nonlinear element, a matrix < The nonzero element of FI> is

## EQU25 ## FI <h-1, 1> =-1

## EQU26 ## FI <h, 1> = 1

## EQU27 ## FI <i-1,2> =-1

## EQU28 ## FI <i, 2> = 1

## EQU29 ## FI <j-1,3> =-1

## EQU30 ## FI <j, 3> = 1, and the other elements are zero. (Equation 24) corresponds to the case where h = 1, i = 2, and j = 3 in these equations. Therefore, the model shown in FIG. 15 is displayed on the screen, and the user designates the non-linear element or the specimen by using a mouse or the like, so that
The matrix <FI> can be set from (Equation 30). Also, as shown in FIG. 18, the matrix <FI
> May be input.

FIG. 20 shows another embodiment of the present invention. In this embodiment, an experiment preparation process 2004 is added to the algorithm 102 shown in FIG. The experiment preparation process 2004 includes a parameter input process A2001, a non-linear element designation process 2002, and a parameter input process B2003. In the parameter input processing A2001, parameters necessary for an experiment, for example, setting of a model of the first numerical modeling unit (linear modeling unit) and the like are performed. In the parameter input processing B2003, parameters necessary for the experiment, for example, setting of input speed,
The second model (non-linear characteristic) is set. In the non-linear element designation process 2002, the matrix <F
I> and matrix <XO>
Specify the nonlinear element.

FIG. 20 shows the parameter input processing A of FIG.
FIG. 1 shows a detailed example of a non-linear element designation process 2001 and a matrix <FI><XO> in FIG.
Although input may be made while looking at the screen as shown in FIG. 8, here, a method for automatically creating these is shown. This figure 2
0 will be described as an example of the evaluation target shown in FIG. In this case, the number of test pieces (F_hyb) is one.

First, the degree of freedom DOF for the entire evaluation object is set (step 2101). In the example of FIG. 15, DOF =
It becomes 3. Next, the input drawing is displayed on the display (step 2102). FIG. 15 may be shown as an example of the input diagram. Next, the specimen section is designated (step 2103). For example, the element k shown in FIG.
When 1 is the specimen section, k1 is selected as the specimen section by, for example, a pointing device. Thus, the degree of freedom to which the element k1 belongs is 1 in the specimen part indicating variable h (1).
Set. That is, h (1) = 1 is set. Next, the number F_hyb of the test part is set (step 2104).
In the case of this example, F_hyb = 1 is set. The above corresponds to the process 2001 in FIG.

Next, the designation of the nonlinear element and the nonlinear element F_NL
Is set (step 2105). For example, when the elements k2 and k3 shown in FIG. 15 are non-linear elements, these elements are selected as a non-linear element part by a pointing device, for example. As a result, the number of nonlinear elements F_NL = 2 is set, and the degrees of freedom to which the elements k2 and k3 belong are set to 2 and 3 in the nonlinear element part designating variables NL (1) to NL (F_NL). That is, in the case of this example, NL (1) = 2 and NL (2) = 3 are set. Next, the sum F_N of F_hyb and F_NL is calculated (step 2106). In this example, F_hyb = 1, F_NL = 2
Therefore, F_N = 3.

Next, each component FI (i, j) of the matrix <FI> is set for i = 1 to DOF and j = 1 to F_N. For this purpose, first, all components in the range of i, j are initialized to 0 (step 2107). In this case,
Since DOF = 3 and F_N = 3, i = 1 to 3, j = 1 to
The range of 3 is the effective range. Next, the matrix <FI
> Is set to -1 for the component FI (h (1) -1,1) and 1 is set to FI (h (1), 1) (step 2108). In this example, h
Since (1) = 1, FI (0,1) = − 1, FI
(1,1) = 1. However, since FI (0, 1) is outside the valid range of i, its value is not set in the matrix. Next, the other FI components are obtained by the general case formula shown in step 2109. In our example
Since F_NL = 2 and F_hyb = 1 and NL (1) = 2 and NL (2) = 3, FI (1,2) = − 1, FI (2,2) = 1,
FI (2,3) = − 1 and FI (3,3) = 1 are set. This is the value shown in (Equation 24). Next, for example, a transposed matrix of the matrix <FI> is obtained and set as a matrix <XO> (step 2110). As described above, the designation of the nonlinear element and the preparation of the parameters necessary for the analysis can be performed.

According to the embodiment shown in FIG. 19, there is an effect that the operator can easily specify the nonlinear element in the numerical modeling unit without error. In this way, even when there is a nonlinear element in the numerical modeling unit, the test for obtaining the vibration response of the entire evaluation object can be performed by combining the excitation test of the real model unit and the vibration response calculation of the numerical model in real time. It will be easier.

Although the present invention has been described with reference to an example in which the present invention is applied to a vibration test, it is needless to say that the present invention can be applied to other test methods such as a vehicle test apparatus.

[0052]

According to the present invention, in a hybrid test apparatus, even when a nonlinear element is included in the numerical modeling unit, the vibration response calculation of the numerical modeling unit can be processed at high speed. This has the effect of enabling a hybrid vibration test in time. Further, according to the present invention, the calculation time interval (hereinafter, referred to as a calculation time step) can be shortened, and the test accuracy can be improved. Further, according to the present invention, there is an effect that a nonlinear element in the numerical modeling unit can be easily specified in a real-time hybrid vibration test.

[Brief description of the drawings]

FIG. 1 is a flowchart showing the entire flow of a test method according to the present invention and the processing of a numerical modeling unit.

FIG. 2 is an example of a two-degree-of-freedom system evaluation target;

FIG. 3 is an example of a reaction force versus response characteristic of a nonlinear element.

FIG. 4 is another example of a reaction force versus response characteristic of a nonlinear element.

FIG. 5 is an example of a parallel processing method of a numerical modeling unit.

FIG. 6 is another example of the parallel processing method of the numerical modeling unit.

FIG. 7 is another example of the parallel processing method of the numerical modeling unit.

FIG. 8 is another example of the parallel processing method of the numerical modeling unit.

FIG. 9 is a configuration example of a processor that performs parallel processing.

FIG. 10 is another configuration example of a processor that performs parallel processing.

FIG. 11 is a detailed explanatory diagram of a parallel processing method.

FIG. 12 is a detailed explanatory diagram of a parallel processing method.

FIG. 13 is another configuration example of a processor that performs parallel processing.

FIG. 14 is another configuration example of a processor that performs parallel processing.

FIG. 15 is an example of a three-degree-of-freedom system evaluation target.

FIG. 16 is an example of an input screen of a matrix <FI>.

FIG. 17 is an example of an input screen for a matrix <FI><XO>.

FIG. 18 is an example of an input screen of a matrix <FI>.

FIG. 19 is a flowchart showing another embodiment of the present invention.

FIG. 20 is a flowchart illustrating a detailed example of a preparation process in FIG. 19;

[Explanation of symbols]

 Reference Signs List 101 Control device 105 Actuator 106 Actuator controller 112 Step of calculating reaction force generated by nonlinear element 113 Step of obtaining vibration response 115 Step of obtaining vibration response of nonlinear element section 201 Specimen section

Continued on the front page (72) Inventor Masahiko Inoue 502, Kandate-cho, Tsuchiura-shi, Ibaraki Pref. Machinery Research Laboratory, Hitachi, Ltd. (72) Inventor Takao Konno 603, Kachi-cho, Tsuchiura-shi, Ibaraki Pref.

Claims (7)

[Claims] 1. A driving means for driving a real object or a model, which forms a part of an evaluation object, as a test body, and measuring a reaction force generated from the test body in the driving means as a first reaction force. The reaction force measuring means, when the portion other than the specimen of the object to be evaluated is divided into a first part and a second part, and the first and second numerical models are respectively used. 1st measured by
A first calculating means for calculating a reaction force applied to the first numerical model from the reaction force; and a second calculating means for calculating a reaction force generated by the second numerical model with respect to a response of the second numerical model. Second to calculate as reaction force
And a third calculation for calculating the response of the first numerical model from the first and second reaction forces calculated by the first and second calculation means and the applied external force. Means, a fourth calculating means for calculating a response of a boundary point with the specimen from a calculation result of the means, and controlling the driving means with the response of the boundary point calculated by the means as a target value. A fifth calculating means for generating a signal and outputting the signal to the driving means; and a second calculating means for calculating the second signal from the response calculated by the third calculating means.
Sixth calculating means for calculating a response of the digitized model and outputting the response to the second calculating means; and the first, second, third, fourth, fifth and sixth calculations Control means for controlling the processing by the means to be executed cyclically. 2. A part of the object to be evaluated other than the test piece,
Dividing into the first part and the second part and the second part
2. The test apparatus according to claim 1, further comprising a digitized model division designating unit for generating a constant used in the second digitized model representing the portion of. 3. The method according to claim 1, wherein the first to sixth calculation means are constituted by a plurality of processors, and wherein the processor constituting the second calculation means is operating.
2. The test apparatus according to claim 1, wherein the processors constituting the second calculation means operate in parallel. 4. The test apparatus according to claim 1, wherein the second portion includes a component having a non-linear relationship between a response and a reaction force to the response. 5. An object to be evaluated is divided into a real model part or a real model part thereof and a remaining part, and the remaining part is numerically modeled. Display means for displaying a numerical model in a test apparatus that obtains the entire vibration response of the evaluation object by combining the vibration response calculation of the numerically modeled part with the vibration response calculation in real time, which is included in the numerical model Pointing means for designating, on the numerical model displayed by the display means, a non-linear element having a non-linear relationship between a response of the element and a reaction force generated by the element; and designation by the means. Position information creating means for creating position information indicating the position of the selected nonlinear element on the numerical model; and a model indicating a relationship between a response of the nonlinear element and a reaction force. Test apparatus characterized by having a calculating means for performing a vibration response calculation of the numerical modeling portion using the position information created by Le and the position information creating means. 6. A test object is a real object or a model that forms a part of an evaluation object, and a part other than the test object of the evaluation object is divided into a first part and a second part, and each of the parts is numerically modeled. Performing a first calculation of measuring a first reaction force generated when the specimen is driven and calculating a reaction force to be applied to the first numerical model from the first and second digitized models; A second operation for calculating a reaction force generated by the second numerical model as a second reaction force with respect to a response of the second numerical model is executed, and the first operation calculated by the first and second operations is performed. And the second
Performing a third operation for calculating a response of the first digitized model from the reaction force and the applied external force, and calculating a response at a boundary point with the specimen from the third operation result. And driving the specimen using the response of the boundary point calculated by the fourth operation as a target value, and calculating the response of the second numerical model from the response calculated by the third operation. A fifth operation for calculating and using this in the second operation is performed, and the first, second, third, fourth, and fifth operations are performed cyclically. A test of the object to be evaluated by controlling the test method. 7. An object to be evaluated is divided into a real model part or a real model part thereof and a remaining part, and the remaining part is numerically modeled. In a test method in which the vibration response calculation of the numerically modeled portion and the vibration response calculation of the portion to be calculated are performed in real time to obtain the overall vibration response of the evaluation object, an element included in the numerical model, A non-linear element having a non-linear relationship between the response and the reaction force generated by the element is specified by a pointing device on the numerical model displayed on the display means, and each of the specified non-linear elements on the numerical model is Position information indicating the positional relationship with the element is created from the numerical model, and the numerical value is calculated using the model indicating the relationship between the response of the nonlinear element and the reaction force and the position information. Test method being characterized in that to perform the vibration response calculation Dell reduction portion.