JPH1164190A - Material testing machine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently set the optimum control gain of a control system in the examination of a sample having an unknown rigidity by estimating the spring constant of a sample from the detected load and displacement of the sample and setting the control gain according to it. SOLUTION: The displacement caused in a sample is monitored while gradually adding a load to the sample, and when a minute displacement, for example, a displacement of the degree exceeding the minimum resolution of a displacement gauge is detected, the load added to the sample at that time is held as it is. After the displacement is stabilized in this state, the charge and displacement caused in the sample in the state where the displacement is stabilized are detected. The spring constant of the sample is estimated from the relation between the detected load and displacement, and the optimum control gain to a servo control system is automatically set according to the estimated spring constant. By such an automatic setting processing of control gain, the optimum control gain according to the properties of a sample can be efficiently and easily set even when the rigidity of the sample is unknown.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention measures a characteristic (property) of a sample while performing feedback control of a load applied to the sample via a hydraulic servo system in accordance with a change in load, displacement or strain generated on the sample. More particularly, the present invention relates to an electrohydraulic servo type material testing machine having a function of automatically optimizing and setting a control gain for the hydraulic servo system.

[0002]

2. Related Art Recently, a closed-loop electrohydraulic servo-type material testing machine has been attracting attention because it has a wide application range for various samples and can perform highly accurate control. FIG. 1 shows a schematic configuration of an electrohydraulic servo type material testing machine of this type.
The test piece S, which is a variety of samples, is held between the actuator 3 and the hydraulic cylinder 3, and pressure oil (oil pressure) from a hydraulic source 4 is supplied to the actuator 3 via the servo valve 5 to operate the actuator 3. Thus, the test piece S is configured to apply a load.

The load generated on the test piece S by applying a load is detected by the load cell 1,
The displacement generated in the test piece S by the application of the load is detected by the displacement gauge 6, and the strain is detected by the strain gauge 7 attached to the test piece S. The control unit 8 constituted by a microcomputer or the like inputs the load, displacement, and strain detected as described above, for example, such that the deviation between the load detected by the load cell 1 and the target load becomes zero (0). Then, the operation of the servo valve 5 is feedback-controlled via the servo amplifier 9. With such a feedback control system, the hydraulically driven actuator 3 is servo-controlled to adjust the load (load) applied to the test piece S.

The electro-hydraulic servo control system is expressed as a closed loop block diagram constituting a feedback control system as shown in FIG. That is, this control system calculates a deviation Δ between a control target value for a load (for example, load or displacement) and a change (control amount of load or displacement) generated in the test piece S detected via the detection amplifier 8a. Tableware 8b
And a controller 8c for controlling the operation of the servo amplifier 9 in accordance with the deviation Δ. The actuator 8 is hydraulically driven by controlling the operation of the servo valve 5 so as to set the deviation Δ to zero (0). This can be expressed as a control loop for adjusting the load (load or displacement) applied to the test piece S.

[0005]

However, in the electrohydraulic servo type material testing machine configured as described above,
It is important to optimize the control gain in the control loop according to the specification (rigidity) of the test piece S. Incidentally, if the control gain is insufficient, control responsiveness is deteriorated. If the control gain is excessive, problems such as occurrence of a hunting phenomenon occur. Such improper setting of the control gain
In addition to deteriorating the test efficiency, there is a possibility that measurement of the test data itself may become impossible, which may lead to undesired breakage of the test piece S.

Conventionally, the stiffness of the test piece S is estimated based on the experience of the operator (tester), the control gain is temporarily set, and trial and error is performed under preliminary load / displacement control. The test is performed after correcting the control gain. For this reason, setting the control gain itself takes time, and there is a problem that the test efficiency is very poor. Further, if the estimated rigidity of the test piece S is incorrect, for example, an excessive load is applied to the test piece, and as a result, the properties of the test piece itself may be changed.

The present invention has been made in view of such circumstances, and a purpose thereof is to test a sample of unknown rigidity without applying an excessive load to the sample.
An object of the present invention is to provide a material testing machine capable of simplifying and efficiently setting a control gain for the servo control system and efficiently setting the test efficiency and improving the test efficiency.

[0008]

SUMMARY OF THE INVENTION In order to achieve the above-mentioned object, the present invention provides feedback control of a load applied to a sample via a hydraulic servo system, and measures the properties of the sample from the load, displacement or strain generated on the sample. The present invention relates to a closed-loop electrohydraulic servo-type material testing machine, and particularly to monitor a change occurring in a sample while gradually applying a load to the sample before starting the test of the sample, for example, monitoring a load, displacement or strain. (1) when the displacement of the sample is detected,
A second freeze of the load applied to the sample at that point
Means for detecting the load applied to the sample and the displacement of the sample after the load is frozen and then waiting for the displacement to stabilize, and the detected load and displacement And a fifth means for estimating the spring constant of the sample from the above (fourth means) and initial setting the control gain for the hydraulic servo system according to the estimated spring constant.

More specifically, the control gain for the hydraulic servo system is gradually increased from its minimum value over time by gradually increasing the load applied to the sample with the passage of time. (Load and displacement)
The load applied to the sample is frozen when a displacement exceeding the minimum resolution of the displacement meter is detected. Then, while waiting for the stabilization of the displacement occurring in the load in a state where the load is frozen, after estimating the spring constant of the sample according to the load and the displacement in a stable state of the sample, for example, as described in claim 3, A control gain to be set for the hydraulic servo system is obtained in accordance with a table showing a relationship between a spring constant and a control gain that has been obtained in advance, and the optimization setting of the control gain is automatically performed.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to the drawings, a material testing machine according to an embodiment of the present invention will be described below. The function for automatically setting the control gain for the control will be described. The material tester according to this embodiment is basically configured as shown in FIG. 1 described above. One of the functions of the control unit 8 is that a sample (test piece S) is framed. Prior to mounting (chucking) between the actuator 2 and the actuator 3 and driving the actuator 3 to start the fatigue test, the control gain for the servo control system is set to the property (rigidity) of the sample (test piece S). It is characterized in that it has a function of automatically setting the optimization according to it.

The automatic setting function of the control gain is a function of monitoring a change occurring in the chucked sample S while gradually applying a load to the sample S, for example, a load, a displacement or a strain. A function of monitoring the displacement occurring in the sample while gradually applying a load to S (first means), and a minute displacement of the sample S, for example, a displacement exceeding the minimum resolution of the displacement meter 6 was detected. At this time, the function of holding (freezing) the load applied to the sample S as it is at that time (second means), and in this state, after waiting for the displacement to stabilize, in the state where the displacement of the sample is stabilized. A function of detecting the load and displacement generated in the sample (third means), a function of estimating the spring constant KL of the sample S from the relationship between the detected load and displacement (fourth means), and estimation Did Consisting a function to automatically set (fifth means) the control gain for the servo control system according constant KL.

That is, the automatic setting process of the control gain is as follows.
As shown in FIG. 3 as a temporal process, the operation of the actuator 3 is controlled while gradually increasing the control gain for the control system from its minimum value, and the load (load) applied to the sample S is gradually increased. It starts with increasing. The application of this load is performed while monitoring the displacement of the sample S detected by the displacement meter 6 until a minute displacement (judgment point) slightly exceeding the minimum resolution of the displacement meter 6 is detected ( Processing period T1). At this time, the maximum load applied to the sample S is limited as a guide to 30 to 50% of the target maximum load value obtained from the estimated characteristics of the sample S, and a large load is applied to the sample S before the test starts. It is preferable not to add them.

If the above-mentioned minute displacement of the sample S accompanying the application of the load is detected, the operating state of the actuator 3 is frozen, and the load value (load) applied to the sample S at that time is kept as it is. I do. That is, the load state on the sample S is frozen as it is. The load state is maintained for a predetermined time until the displacement of the sample S is stabilized (processing time T2). At this time, when the rigidity of the sample S is high and the spring constant KL is large, the displacement is stabilized in a relatively short time as shown by the characteristic A in FIG. However, when the rigidity of the sample S is low and its spring constant KL is small, the displacement stabilizes after a certain amount of displacement fluctuation occurs as shown by the characteristic B in FIG. The state holding waiting time T2 is set in consideration of the time required for stabilizing the displacement due to such a difference in the spring constant.

Thereafter, the load value applied to the sample S and the displacement amount occurring in the sample S are detected over a plurality of samples (processing time T3). Detection of this load and displacement
For example, 100 to 200
It is repeated over about a sample. Thereafter, the average of the detected load value and the average of the displacement amount are obtained, and the spring constant KL of the sample S is estimated by the calculation of [average load value ÷ average displacement amount] (processing time T4). Then, in accordance with the spring constant KL estimated in this manner, for example, according to a table showing the relationship between the spring constant and the optimum control gain in the servo control system which is obtained in advance, the optimum control gain to be set in the servo control system is obtained. , Initialize this.

The relationship between the spring constant and the optimum control gain is determined by, for example, optimizing various samples S having known spring constants KL under the optimization control of a servo control system in this material testing machine in advance. By obtaining control gains and organizing these data, the control gains can be obtained in advance as table data unique to the servo control system. Further, each time the spring constant KL and the optimum control gain at that time are newly obtained with the execution of the fatigue test, the table data is updated by learning the actual data.

Here, a spring constant KL indicating the rigidity of the sample S
The relationship between and the optimum control gain for the servo control system of the material testing machine for testing the sample S will be described.
The control gain of the feedback control system for stably operating the servo control system that hydraulically drives the actuator 3 that applies a load to the sample S differs depending on the rigidity (spring constant) of the sample S, and a value specific to the control system. Take. That is, as shown in FIG. 2, the servo control system of the material testing machine includes a transmission system shown by a block diagram that forms a closed loop including the servo amplifier 9, the servo valve 5, the actuator 3, and the sample S (load). Is configured as Therefore, in order to realize a stable operation of the servo control system, a transmission system (transfer function) unique to the material testing machine such as the servo amplifier 9 and a load transmission system (transfer function) depending on the properties of the sample S are considered. Then, for example, it is necessary to set an optimum control gain in the controller 8c.

By the way, in a displacement control system which servo-controls the operation of the actuator 3 paying attention to the displacement of the sample S, optimal control gains (proportional gain KP and integral gain KI) corresponding to the spring constant KL of the sample S are as follows. For example, it is as shown in FIG. An optimum control gain corresponding to the spring constant KL of the sample S in a load control system that servo-controls the operation of the actuator 3 paying attention to the load applied to the sample S is, for example, as shown in FIG. .

Therefore, in the present invention, as a prerequisite, a plurality of samples S whose spring constant KL is known in advance within the range of the spring constant KL covering the test range in the material testing machine.
By optimizing the control system by tuning the proportional gain KP and the integral gain KI in the displacement control system and the load control system, respectively, the optimum control gains KP and KI for various spring constants KL can be respectively adjusted. Desired.
Then, the relationship between the optimum control gains KP and KI with respect to these various spring constants KL is represented in advance as data having a table structure as shown in FIG. Based on such table data, for example, the optimum control gains KP and KI as shown in FIGS. 4A and 4B, which are unique to the material testing machine, are determined by the spring constant K of the sample S as described later.
L is given according to L.

On the other hand, in order to estimate the spring constant KL of the sample S, the sample S
The load is gradually applied to the sample S until a minute displacement is detected. The control of the load applied to the sample S utilizes the relationship between the spring constant KL and the proportional gain KP at the time of step response in a load control system as shown in FIG. 5 (a). As shown, this is realized by gradually increasing the proportional gain KP over time. Specifically, for example, a proportional gain KP (proportional gain value) at the maximum stiffness prepared in the control system under a target load pattern assuming a ramp hold waveform in which the maximum load is limited as shown in FIG. Is minimum), the load control is performed while gradually increasing the gain value. At this time, the integral gain KI is set to, for example, zero (0).

The relationship between the spring constant KL and the proportional gain KP at the time of step response in the displacement control system is shown in FIG. 6 (a).
For example, as shown in FIG. 6B, a displacement may be applied to the sample S while gradually increasing the proportional gain KP with time. Even when a displacement is applied to the sample S under the control of such a displacement control system, the load state is naturally frozen when a minute displacement of the sample S is detected.

By applying a load to the sample S as described above, when a displacement of the sample S exceeding the minimum resolution of the displacement meter 6 is detected, at that time,
For example, the proportional gain value and the target value at that time are kept constant. As a result, the load (load) is frozen without applying an excessive load to the sample S. Then, in this state, the system waits for the stabilization of the displacement generated in the sample S as described above. The time required for the stabilization of the displacement is schematically shown in FIG. 7 as the displacement stabilization time at the time of the step response.
Since it differs depending on the spring constant KL of the sample S, as described above, in the range of the spring constant KL covering the test range in the material testing machine, the time is set in consideration of the time required for stabilizing the displacement when the spring constant KL is the smallest. I do.

After waiting for the stabilization of the displacement of the sample S as described above, the displacement amount and the load value are periodically collected over a plurality of samples, and the average is obtained. By averaging the plurality of samples, variations due to disturbance and other factors are absorbed. Thereafter, the calculation of [average load value / average displacement amount] is performed to estimate the spring constant KL of the sample S. That is, by applying a light load (load) that does not cause a load on the sample S, a minute displacement is generated in the sample S, and the spring constant KL of the sample S is estimated from the relationship between the load and the displacement at that time. I do.

After the spring constant KL of the sample S has been estimated in this way, the control gain corresponding to the estimated spring constant KL is calculated back according to the table described above. Thereby, the optimum control gains KP and KI to be set in the control system are obtained.
Can be obtained, and the automatic setting of the optimization control gain when testing the sample S whose spring constant KL is unknown can be realized.

FIG. 8 shows an example of a processing procedure in the control section 8 for executing the automatic control gain setting processing as described above. Briefly describing this processing procedure, for example, first, the control system is set to the load control mode [Step S1], and the gain characteristic for the automatic gain adjustment is shown in FIG. 5 (a) or FIG. 6 (a). [Step S2]. The target load pattern for adjustment is set so as to have a ramp-hold waveform [Step S3]. Then, after setting an initial gain (minimum value) for the control system [Step S4], load / displacement control is started under the above gain characteristics and target load pattern [Step S5].

In this state, it is monitored whether or not the minimum displacement detectable by the displacement meter 6 is detected from the displacement of the sample S detected by the displacement meter 6 [Step S6]. And sample S
When the minimum displacement is detected, the proportional gain and the control target value at that time are frozen, and the load (load) applied to the sample S thereafter is kept constant [Step S7]. Then, it waits for the elapse of the predetermined time T2 for stabilizing the displacement described above [Step S8]. At this time, the load (load) in the above control system
Is continuously performed [Step S9].

When the displacement of the sample S is stabilized [Step S10], the load value and the displacement amount at that time are periodically repeated over a plurality of samples, and the average thereof is determined [Step S11]. Then, the spring constant KL of the sample S is calculated from the relationship between the average load and the average displacement [Step S12], and the spring constant KL of the sample S estimated by the calculation is calculated from the table data as shown in FIG. , An optimum control gain suitable for the spring constant KL of the sample S, that is, a proportional gain KP and an integral gain KI are obtained, and these are set in the control system [Step S1]
3]. These control gains are obtained for each of the load system and the displacement system.

If the control gains KP and KI determined for each control system as described above are set in accordance with the estimated spring constant KL of the sample S, the control gain is determined by, for example, examining the step response. It is determined whether or not has been optimized, and a series of control gain setting processing ends [step S14]. Thereafter, a test of the sample S, specifically, a fatigue test, a creep test, a reclamation test, a tensile test, and a compression test are executed under the automatically set control gain.

Thus, according to the material tester which functions as described above and automatically optimizes and sets the control gain for the servo control system, the rigidity (spring constant) of the test sample S is completely unknown. However, the optimum control gain according to the properties of the sample S can be set very simply and efficiently in a short time. In addition, under a simple control algorithm, the spring constant KL can be estimated without applying an excessive load (load) to the sample S, and the optimum control gain can be set. In particular, wait for the stabilization of the displacement in a state where a slight load is applied to the sample S, and sample the load value and the amount of displacement applied to the sample S in a stable state over a plurality of samples. Since the spring constant KL is estimated based on this, the estimation accuracy can be made sufficiently high. Then, the control gain optimal for the control system is calculated backward from the estimated spring constant and set in the control system, so that the control gain can be automatically set efficiently without imposing an extra burden on the sample S. There is an advantage to say.

The present invention is not limited to the above embodiment. For example, since the conditions of the load and displacement applied to the sample S vary depending on the test conditions of the sample S, the gain characteristics and the target pattern set when estimating the spring constant of the sample S, particularly the maximum load value, etc. You may make it set according to it. Specifically, when a fatigue test is performed, the load applied to the sample S is changed by a triangular wave as shown in FIG. 9A, and when a creep / reclamation test is performed, FIG. This is performed by increasing the load into a ramp waveform as shown in b) and then holding the load constant. Further, when a tensile / compression test is performed, as shown in FIG. 9C, the load or the displacement is increased by a constant gradient. Therefore, according to such a difference in the test conditions, for example, the relationship between the displacement and the load in a state where the sample S is stable at about 30 to 50% of the target maximum load value to be applied to the sample S in the present test is obtained. What is necessary is just to estimate a spring constant.

When a minute displacement exceeding the minimum resolution of the displacement meter 6 is detected, when the load is fixed, the detected displacement itself tends to be buried in noise. Therefore, while reading the displacement amount at every predetermined cycle, for example, the displacement data is filtered using a method such as a moving average, and the detection value (displacement) is determined.
Furthermore, the point in time when a predetermined amount of displacement is detected a plurality of times in succession may be determined to be a point at which displacement has occurred due to a load (load).

Further, when the amount of table data indicating the relationship between the spring constant and the control gain shown in FIG. 4 is small, the average load value and the average displacement are calculated based on an approximate expression determined based on the data. The control gain may be calculated by an approximate calculation according to the spring constant KL estimated from the amount. Further, according to the spring constant and the gain characteristic of the strain control system in the sample S, the control can be similarly performed. In addition, the present invention can be variously modified and implemented without departing from the gist thereof.

[0032]

As described above, according to the present invention, even when the rigidity of a sample is unknown, it is possible to efficiently set an optimum control gain of a control system for testing the sample. Can be. Moreover, the load (load) that is unwilling to the sample
Estimate the spring constant with high accuracy in a short time without adding, and automatically set the control gain suitable for executing the test based on the relationship between the spring constant collected as the actual result and the optimal control gain can do. Therefore, not only can the test efficiency be improved, but also practical effects such as easy execution of the test on the sample can be obtained without knowledge of various samples. .

[Brief description of the drawings]

FIG. 1 is a diagram showing a schematic configuration of a material testing machine according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a control system in the electrohydraulic servo type material testing machine according to the present invention.

FIG. 3 is a diagram showing a concept of an automatic setting process of a control gain in the present invention as a time course.

FIG. 4 is a diagram illustrating a relationship between a spring constant of a sample unique to a control system and an optimum control gain.

FIG. 5 is a diagram showing a relationship between a spring constant and a proportional gain at the time of a step response in a load control system, and a temporal increase characteristic of the proportional gain.

FIG. 6 is a diagram illustrating a relationship between a spring constant and a proportional gain at the time of a step response in a displacement control system, and a temporal increase characteristic of the proportional gain.

FIG. 7 is a diagram showing a change in a time required for displacement stabilization during a step response with respect to a spring constant.

FIG. 8 is a diagram showing an example of a control gain automatic setting procedure executed in the control unit of the material testing machine according to one embodiment of the present invention.

FIG. 9 is a diagram showing control characteristics of a load or a displacement applied to a sample according to a test condition.

[Explanation of symbols]

 S Sample (test piece) 1 Load cell 2 Frame 3 Actuator 4 Hydraulic source 5 Servo valve 6 Displacement gauge 7 Strain gauge 8 Control unit 9 Servo amplifier

Claims (3)

[Claims] 1. A material testing machine for measuring a load, displacement or strain generated in a sample by feedback-controlling a load applied to the sample via a hydraulic servo system, wherein the sample is gradually applied to the sample prior to the start of the test of the sample. First means for applying a load to monitor a change occurring in the sample, and second means for freezing the load applied to the sample when a displacement of the gradually loaded sample is detected; After the load is frozen, a third means for detecting the load applied to the sample and the displacement of the sample after waiting for the stabilization of the displacement, and detecting the load and the displacement of the sample based on the detected load and displacement. A material testing machine comprising: fourth means for estimating a spring constant; and fifth means for initially setting a control gain for the hydraulic servo system according to the estimated spring constant. 2. The method according to claim 1, wherein the first means monitors a change occurring in the sample while gradually increasing a load applied to the sample by gradually increasing a control gain for the hydraulic servo system from a minimum value with time. 2. The material testing machine according to claim 1, wherein the second unit freezes a load applied to the sample when a displacement exceeding a minimum resolution of the displacement meter is detected. 3. 3. The control device according to claim 1, wherein the fifth means obtains a control gain to be set for the hydraulic servo system in accordance with a table indicating a relationship between a spring constant and a control gain obtained in advance. Material testing machine.