JPS6134438A - Testing method of structure - Google Patents

Abstract

PURPOSE:To provide a strength testing method of a structure, by which the total behavior of the structure can be grasped highly precisely and correctly, by using the local structural models having the actual sizes of various kinds of structures, and utilizing the advantages of simulated experiment. CONSTITUTION:A structure 1 is divided into a part 1a, which can be theoretically analyzed, and a part 1b, in which prediction of elastic and plastic deformation and broken parts is difficult and theoretical analysis is also hard. With respect to a part 1a, the behavior of an internal force R against an external force P2 is obtained by using a well known method. Meanwhile, with respect to the part 1b, an experiment using the model of a local structural part is conducted, and the behavior R against an external force P1 is obtained. One result is fed back to the other. In consideration of the mass of the part 1b, the effect of the inertial force of the part 1b is evaluated. the theoretical analysis and the experiment are repeated until the continous condition of the boundary part of both parts is satisfied. Thus the total behavior is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a structure testing method capable of measuring the strength of various structures against external forces with high precision.

[Conventional technology]

A pseudodynamic experiment method has been proposed as a new structural experiment method for experimentally investigating the response of various structures to external forces such as earthquakes, and confirming the strength of the structures based on the experimental results.

This experimental method applies the load and
The restoring force characteristics, which are indicators of deformation, are directly determined through a force experiment on a model specimen of the entire structure, and the restoring force characteristics are fed back to the computer. This is a method in which the nonlinear response of the structure, which changes from moment to moment, is traced online while determining the response of the structure at the next time using information on the restoring force. Therefore, in this experimental method, an actual vibration phenomenon that changes continuously with respect to time is assumed to be a minute time Δt, and that Δ is used as a response for each time to reproduce the actual vibration phenomenon. □In addition, during actual loading, the time axis of the short-time vibration phenomenon is magnified several hundred to several thousand times compared to real time, and the force is applied quasi-statically to the model specimen. Become.

FIGS. 3 and 4 show an overview of such conventional maneuverability experiments. This experiment is to find the stress in the structure 1 when an external force P is acting on the structure 1 as shown in Fig. 4'. The experiment is carried out by interconnecting an experimental system 4 including a full-scale full-scale model specimen 3 of the above-mentioned structure online.

This experiment is conducted, for example, in the following manner.

■ Now, in the computer system 2, the response displacement of the model specimen 3 at the time (n steps).

(Xn) is the response displacement (Xn-1), response acceleration (Xn-1), and minute □ of the previous time (n-1 steps).
time interval (ΔT), and one more previous time (n-
(1) This response displacement (Xn) is transmitted to the controller 5 of the experimental system 4 and given as a command displacement (Xn) to the model specimen 5. It will be done.

(2) The actuator 6 applies this command displacement (Xn) to the model specimen 3, and the displacement (xn) actually generated in the model specimen 3 is measured by the displacement meter 7. While performing this measurement, the error between the commanded displacement (Xn) and the actually measured displacement (xn) is kept within an allowable range.

(2) At this time, the restoring force (Fn) generated in the model specimen 3 is measured by the load cell 8, and the measurement result is fed back to the computer system 2.

■ In the computer system 2, the mass of the model specimen 3 (
m), the restoring force (Fn) obtained as above, and the time history of the external force (P).
From the external force (Pn) of n steps), the response acceleration (Xn) at this step is calculated from the equation of motion.

′[■ After that, the response displacement (X n+
1) Perform the calculation. This calculation consists of the response acceleration (X'n) found in the 0 term above, the response displacement (X'n) found in the 0 term above,
), and the n-1 step response displacement (Xn-1), which is known at this point, is used.

By repeating such a procedure, the nonlinear response of the structure (model specimen 3) at each time is sequentially progressed, and its strength is measured. , [Problems to be solved by the invention] However, such conventional mobile experiments have had the following problems.

In other words, in this case, the overall response of the structure is determined by model specimen 3.
It is undeniable that the manufacturing cost of the model specimen 3 would be extremely high.

Furthermore, if the actual structure is a very large-scale structure, the model specimen 3 will also be large-scale, making it virtually impossible to conduct experiments using the actual large-scale model specimen 3. Therefore, it was necessary to use a model specimen 3 that was reduced in size at a predetermined ratio, resulting in incomplete reproduction of the restoring force characteristics and a decrease in measurement accuracy. This made it difficult to accurately grasp the complex restoring force characteristics that occur in structures when external forces are applied, and it was not possible to fully utilize the advantages of mobile experiments.

In addition, in the above-mentioned maneuver experiments, the structure of the structure was relatively simple;
Moreover, if the direction in which the external force is applied is fixed as one direction, highly accurate experiments are possible; When a device has a complex structure such as a supporting steel frame, its behavior in response to external forces such as earthquakes becomes complicated, and there is a problem in that it is not possible to accurately reproduce that behavior just by measuring the displacement in the input direction of the external force. Ta. In other words, as structures became more complex, highly accurate experiments were not possible.

The present invention was created in view of these problems, and its purpose is to utilize full-scale local structural models of various structures while taking advantage of mobile experiments.
It is an object of the present invention to provide a highly practical structural strength testing method that can accurately and accurately grasp the overall behavior of the structure.

(Means for Solving the Problems) The present invention basically consists of a portion 1a of a structure 1 that can be analyzed theoretically, as shown in FIG. The analyzable portion 1a is divided into a theoretically unanalyzable portion 1b that is expected to exhibit difficult and complex nonlinear characteristics, and the analyzable portion 1a is analyzed using a known finite element method, etc., as shown in Figure 6. A theoretical analysis is performed to determine the behavior of the internal force R with respect to the external force P2, and for the portion 1b that cannot be analyzed, an experiment is performed using a model of the local structure to determine the behavior of the internal force R with respect to the external force P1. is fed back to the other, and the theoretical analysis and experiment are repeated until the continuity condition (internal force R) at the boundary between the two is satisfied, thereby determining the overall behavior of the structure. At this time, while evaluating the effect of the inertial force accompanying the response of the local structure part, taking into account the mass of the local structure part 1b, the displacement of the boundary part and the displacement of the nodes inside the local structure part are respectively calculated. The above-mentioned theoretical analysis and experiment are repeatedly executed through measurement control.

[Function of the invention and its effects]

Thus, according to the present invention, large steel structures such as chimneys, power transmission towers, long suspension bridge main towers, plant towers, fixed offshore platform salt tower structures, boiler support steel frames, various frames, and various steel frame buildings, etc. The strength against external force is
Divide into parts that can be analyzed theoretically and parts that cannot be analyzed theoretically, and perform experiments on the parts that cannot be analyzed theoretically using a local structure model to determine the boundary conditions between the experimental results and the theoretical analysis results. By repeating the above-mentioned process (experiment and analysis) to satisfy the above-mentioned conditions, the behavior of the entire structure in response to external forces can be ascertained. Therefore, it is possible to accurately measure the strength, etc. of the above-mentioned structure in response to external forces, which exhibits nonlinear behavior. It becomes possible to do so.

Furthermore, since the behavior of the structure is determined by considering the mass of the local structural part and evaluating the force associated with the response of the structure, it is possible to conduct experiments using a scaled model of the local structural part. , it becomes possible to obtain highly accurate experimental results. Therefore, it is possible to easily and economically carry out experiments on structures having complex structures and measure their strength, which has great practical effects.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to the drawings.

゛□ Fig. 1 schematically shows the experimental procedure (structure) of the structure test method according to the example. Also, Figure 2 shows the external force P (nll) of the structure used in the experiment.

ff1P) is a schematic diagram showing a structure for obtaining a response to ff1P).

As shown in Fig. 2, in this example, the entire structure of the structure is divided into a nonlinear part N where plasticization and other nonlinearity are expected to occur, and a linear part where other linear analyzes can be performed. , and a boundary portion B that forms the boundary between them, an experimental system is constructed, and a model of the local structure of the nonlinear portion N is particularly prepared, and this is used as a model specimen 11. Using this model specimen 11, a model experiment system 1 is constructed as shown in FIG.
2 and a computer system 13 that performs theoretical structural analysis of the linear portion, etc., are interconnected online to construct the entire experimental system. At this time, each part N
, L at corresponding positions in the boundary portion B are in a state in which the magnitudes of the internal forces R are equal and the directions of the forces are opposite to each other (boundary condition). Therefore, the response of the structure to the external force P is determined by the nodes ℃ of the linear part, the nodes n of the nonlinear part N, and the boundary part 8 shown in FIG.
It can be obtained as each response of node 0 of .

However, this experiment is performed as follows.

First, in the computer system 13, the response displacement <aX> of the many nodes of the linear part, the response displacement (nX) of the node n of the nonlinear part N, and the internal node (14 in FIG. 1) of the model specimen 11 are calculated. ) is calculated from the equation of motion. In addition, the boundary part B of the model specimen 11
Since the commanded displacement given to node b (15 in Figure 1) cannot be uniquely determined by theoretical analysis, it is determined by the internal force at the boundary of the linear portion determined by the computer system 13 and by the model experiment system 12. □ is determined after repeating experiments until a state is reached where the internal force at the boundary of the nonlinear portion N converges to satisfy the boundary condition. Then, when the internal force of the node 15 of the boundary portion B converges, the node 1
The actual stress displacement and internal force of 5, and the restoring force at the node 14 of the nonlinear portion N are measured by the displacement meter 16 and the load cell 1.
7, and the measurement results are provided to the computer system 13.

Obtaining such information, the computer system 13 calculates each of the above-described response displacements in the next step from the external force Pn in step n, the response displacement, internal force, and restoring force. Such a process is repeated at each time (each step) to obtain the response characteristics of the structure. That is, ■ First, in the computer system 13, the time (n
The response displacement (ρ
Xn) is the response displacement (nXn-1) of the previous time (n-1 step), the response acceleration (nXn-1), the minute time interval (ΔT), and one more previous time (n-2 steps). The response displacement (nXn) of the node n of the nonlinear portion N is calculated from the response displacement (j2Xn-2) of the previous time (n-1 step), and the response displacement (nXn-1>
, response acceleration < nXn-1), the minute time interval (ΔT), and the response displacement (nXn-2) at one more previous time (n-2 steps). This calculation can be uniquely performed, for example, by integrating the equation of motion using the known central difference method.

■ Regarding the n-step response displacement (b
Since fn) is unknown, the response displacement (bXn) of the boundary portion B is obtained by repeating the experiment using the model test system 12 described above to obtain a state in which the internal force is converged, and then performing an approximate calculation. Find the internal force (bfn). In other words, the internal force (bfn-1) at the n-1 step is converted into the internal force (bfn-1) calculated by the linear approximation at the n-step
)'.

■ Next, this linear approximation calculated internal force (bfn)' and the stiffness of the linear section (K ffi
) and the response displacement (βXn) of the node β, a first-order approximate value (bXn) of the response displacement of the node S in the boundary portion B is calculated.

■ The response displacement of each node obtained in this way (nX
n), (bXn) in the model experiment system 12
controllers 18 and 19, and gives it as a command displacement to the model specimen 11.

■ In response to the command displacement from the controllers 18 and 19, the actuator 20 moves to the node n (first
The commanded displacement (nXn) is applied to the node b (15 in FIG. 1) of the boundary portion B, and the actuator 21 applies the commanded displacement (bXn) to the node b (15 in FIG. 1) of the boundary portion B. As a result, the displacements (nxn) and (bxn) actually generated in the model specimen 11 are measured by the displacement meter 16, respectively. The experiment is repeated while performing this measurement, and the error between the commanded displacement (nXn) and the actually measured displacement (nXn), and the difference between the commanded displacement (bXn) and the actually measured displacement (bxn) Both errors are within the allowable range.

■ At this time, the restoring force (nfn) generated at the node n of the nonlinear portion N of the model specimen 11 and the internal force (bfn) at the node gb of the boundary portion B are measured by the load cell 17. The measurement results are fed back to the computer system 13.

■ In the computer system 13, since the internal force (bfn) fed back in this way is an improved first-order approximation value of the internal force assumed earlier, this is obtained as its second-order approximation value. .

Then, the convergence of the internal force is determined by comparing this second-order approximation value with the first-order approximation value, and by repeating the above process, the value at the time when the internal force has converged is determined at the node of the boundary part in n steps. The restoring force at the node n of the nonlinear portion N measured at this time is extracted as a value obtained by a model experiment.

■ Thereafter, the computer system 13 calculates this step (n steps), which is determined from the mass (m2) and stiffness (Kffi) of the linear portion, and the time history of the external force (P).
The response acceleration (I 2

Similarly, regarding the nonlinear part N, the nonlinear part N
The stress acceleration (nXn) at the node n of the nonlinear portion N in n steps is calculated from the mass (Mn), the external force (nPn), and the restoring force (nfn) for the nonlinear portion N.

■ Then, from the response acceleration of each node above, the next time (
0+1 step) response displacement 1! Xn+1)
, , (nX n+1) are calculated, respectively. Response displacement (fiX n+1
) is calculated using the response acceleration (βXn) obtained in the 0 term above,
The response displacement (j2Xn) obtained in the above section (■), and the response displacement (nXn
-1). Similarly, node n of nonlinear part N
The calculation of the response displacement (nX n+i') for This is done using response displacements (nxn-1).

As described above, the response behavior of the structure to external force is analyzed by repeatedly executing the steps from item (1) to item 0 and examining the response of the entire structure at each time (each step).

Note that, unlike the above-described conventional preliminary experiment, here, control is performed not only for displacement in one direction but also for multiple directions and rotations. Therefore, as the actuators 20 and 21, those capable of applying rotation while pushing the nodes of the model specimen 11 are used. Further, as the displacement meter 16, one that can similarly measure rotation is used.

Thus, according to the experimental method according to this book, since the inertial force associated with the response is evaluated while considering the mass of the local structural part, not only the node in the boundary part B but also the node n inside the nonlinear part N is evaluated. It also becomes possible to perform displacement control. Therefore, the inventors of the present invention first applied for patent application No. 58-1212.
This enables even more accurate measurements than the experimental method proposed in No. 23.

Furthermore, even in cases where the mass of a plastic part cannot be ignored, experiments are conducted using the local structural model after taking into account the effect of inertial force associated with the response of that part.
Even if a structure includes structural parts that cannot be analyzed theoretically, its overall behavior can be measured economically and with high precision.

In other words, by effectively using a local structural model, it is possible to measure the restoring force of large structures exhibiting complex behavior with high precision by making full use of preliminary experimental methods. In addition, the response of elastic parts can be determined by highly reliable theoretical calculations such as the well-known finite element method, so for example, response calculations that take into account not only the earthquake input direction but also the perpendicular direction and rotational effects are possible. It is possible. Therefore, even for structures with complex structures that cannot be grasped by conventional preliminary experiments, tremendous effects such as the ability to calculate highly accurate nonlinear responses over the entire structure can be achieved.

Note that the present invention is not limited to the embodiments described above. For example, there may be a plurality of nonlinear parts in the structure. In this case, experiments may be performed using a plurality of local structure models. Further, the number of nodes in the local model of the nonlinear structure may also be set according to the structure. Furthermore, as the calculation method used in the theoretical analysis, various known algorithms may be used as appropriate. In short, the present invention can be implemented with various modifications without departing from the gist thereof.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a schematic configuration of an experimental model showing an embodiment of the present invention, FIG. 2 is a diagram showing a division model of a structure,
FIG. 3 is a diagram showing the configuration of a conventional experimental model, FIG. 4 is a diagram showing an example of a structure, and FIGS. 5 and 6 are diagrams showing the basic concept of the experimental method according to the present invention. 11... Model specimen, 12... Model experiment system,
13... Computer system, 14.15... Node, 1
6...Displacement meter, 17...Load cell, 18.19.
...Controller, 20.21...Actuator.

Claims (1)

[Claims] Divide the structure into theoretically analyzable parts and theoretically analyzable parts, perform theoretical analysis on the analyzable parts, and conduct experiments using models of local structural parts for the analyzable parts. , when determining the behavior of the structure by feeding back the results of one to the other and repeating the theoretical analysis and experiment until the continuity condition at the boundary between the two is satisfied, the mass of the local structural part is taken into consideration. The theoretical analysis and experiment are repeatedly performed by measuring and controlling the displacement of the boundary portion and the displacement of the nodes inside the local structure portion, while evaluating the effect of inertial force accompanying the response of the local structure portion. A structure testing method characterized by: