JP2663741B2 - Material testing machine with learning function - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention generates an optimum setting signal based on past learning data when data on a specimen and data on a test condition including a load pattern are input. The present invention relates to a material testing machine with a learning function that corrects a setting signal so that an actual load pattern obtained by the setting signal becomes an input load pattern, and stores and updates the correction data as the learning data.

[0002]

2. Description of the Related Art Conventionally, as shown in FIG. 9, an actuator 1 for loading a specimen and a setting signal output unit 2 for generating and outputting a setting signal based on input data indicating a load pattern of the specimen. A load cell 3 for detecting a load caused by deformation of the specimen due to the load, and a drive control unit 4 for driving the actuator so that the setting signal and the detection signal match.
There is known a material testing machine having the following. Reference numeral 5 denotes an amplifier for an output signal of the load cell 3.

It is considered that a specimen is loaded by such a material testing machine in a load pattern of a predetermined frequency having an upper limit Lmax and a lower limit Lmin as shown in FIG. When an operator inputs an upper limit value Lmax, a lower limit value Lmin, and a predetermined frequency from an operation panel (not shown), a waveform shown in FIG. 10 is output from the setting signal output unit 2 and the actuator 1 is driven to load the specimen. The load acting on the specimen is detected by the load cell 3, and the difference from the set signal is obtained by the deviation unit 4a of the drive control unit 4. The actuator 1 is driven by the difference signal via the amplifier 4b. Therefore, the load pattern actually applied to the test piece overshoots, for example, as shown in FIG. 11, and after a while after the start of the test, the waveform becomes the desired upper limit value Lmax and lower limit value Lmin.

[0004]

However, in such a conventional material testing machine, even if the desired upper and lower limits are the same, the upper and lower limits are adjusted according to the material and frequency of the specimen. It is necessary to input, and the fact is that the operation depends on the intuition of the operator. Therefore, in some cases, it may take time from the start of the test until the load pattern converges to the desired upper and lower limit values.

An object of the present invention is to test a desired load pattern on a specimen by simply inputting data relating to the specimen and data representing a test condition including a desired load pattern without relying on the intuition of a skilled person. An object of the present invention is to provide a material testing machine with a learning function provided immediately after the start.

[0006]

The present invention will be described with reference to FIG. 1 showing an embodiment. The present invention relates to an actuator 11 for repeatedly loading a specimen at a predetermined frequency.
Setting signal output means 20 for generating and outputting a setting signal based on input data indicating a load pattern of the specimen;
The present invention is applied to a material testing machine including a detecting unit 12 for detecting a load or a displacement accompanying a deformation of a specimen due to a load, and a drive control unit 4 for driving an actuator 11 so that a setting signal and a detection signal match. . According to the first aspect of the present invention, there is provided a setting signal correction unit for correcting a setting signal so that an actual load pattern matches a desired load pattern, and a test condition including data and a load pattern relating to a specimen. Data and setting signal correction means 10
Storage means 10 for storing the correction data in step 1 in association with each other
2 and the setting signal output means 20 is stored in the storage means 102
This is achieved by calculating the correction data from the input data of the specimen and the test conditions based on the stored contents of the above and generating the setting signal. According to the second aspect of the present invention, a gain correction means 201 for correcting the gain of the drive control means so that an actual load pattern matches a desired load pattern, and data on various test conditions including data on the specimen and gain. Storage means 2 for storing the correction data in the gain correction means in association with each other
02, and is achieved by configuring the drive control means 4 to calculate correction data from the data of the input specimen and test conditions based on the stored contents of the storage means 202 and to set the gain. Is done.

[0007]

According to the first aspect of the present invention, the setting signal is generated by calculating the correction data of the setting signal from the input data of the test condition including the specimen and the load pattern. The setting signal correction unit 101 corrects the setting signal so that the actual load pattern obtained by the setting signal matches the input load pattern. The correction data at this time is stored in the storage means 102. According to the second aspect of the present invention, the gain of the drive control means 4 is set by calculating the gain correction data from the input test condition data including the specimen and the gain. The gain correction means 201 corrects the gain so that the actual load pattern obtained by this gain matches the input load pattern. The correction data at this time is stored in the storage unit 202.

In the section of the means for solving the above-mentioned problems and the operation, which explain the constitution of the present invention, the same reference numerals as those of the embodiments are used to make the present invention easy to understand. The invention is not limited to the embodiments.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a material testing machine with a learning function according to the present invention will be described with reference to FIGS. FIG. 3 shows the overall configuration of a material testing machine with a learning function according to the present invention,
Numeral 0 is a well-known device, and includes an actuator 11 for repeatedly loading the specimen at a predetermined frequency, a load cell 12 for detecting a load acting on the specimen, and a displacement meter 13 for detecting displacement. The controller 20 has a feedback control circuit as shown in FIG. Reference numeral 30 denotes a setting device which is a characteristic part of the present invention.
It can be configured by a computer including M, RAM, and other peripheral circuits, and performs learning control by executing a program described later. Reference numeral 40 denotes a keyboard for inputting data relating to the specimen and data relating to the test conditions.

FIG. 7 shows an example of various data stored in the storage device of the setting device 30. For a specimen having a specific material, tensile strength, and cross-sectional area, the frequency of the load pattern and the upper limit value are shown. , The upper limit correction value, the lower limit value, and the lower limit correction value are stored in association with each other. As will be described in detail below, the correction values of the upper and lower limits are the desired upper and lower limits, in the example of the figure, +10.0 kg, -10.0 kg.
Is a deviation of a desired upper / lower limit value from an upper / lower limit value of a setting signal actually necessary to load the specimen, and is learned and determined by a procedure described below.

When the program shown in FIG. 4 is started, data on the specimen is read in step S1, and data on the test conditions are read in step S2. These various data are input in advance from the keyboard 40 or the like. In the description of this embodiment, the test condition is set to the frequency C
The description will be made assuming that YC is the load upper limit Lmax and the load lower limit Lmin.

In step S3, a test condition setting process shown in FIG. 6 is performed, and the upper limit values Lmax and Lmin are set to ΔMaxt,
Each is corrected by ΔMint. Details of the processing in FIG. 6 will be described later. In step S4, the start of the test is determined. If the result is negative, the process returns to step S1, and if the result is positive, the process proceeds to step S5. In step S5, the maximum value Lmaxr and the minimum value Lminr of the actual load detected by the load cell 12 are read, and it is determined in step S6 whether the flag is 1. Flag is 1
If not, the flow advances to step S7 to input the upper limit value Lma
x is set to the register Lmax1, and the lower limit Lmin is set to the register Lmin1.
Respectively. If the flag is 1, step S7 is skipped.

In step S8, the difference ΔLmax between the maximum value Lmaxr of the actual load read by the load cell and the set load upper limit Lmax, and the difference ΔLmax between the minimum value Lminr of the actual load and the set lower limit Lmin of the load are set. Find the difference ΔLmin. Next, in step S9, it is determined whether or not ΔLmax and ΔLmin are equal to or smaller than predetermined values. In step S9, ΔLmax or ΔLm
If it is determined that in is not within the predetermined value, step S10
The flag is set in step and the process proceeds to step S11. In step S11, (ΔLmax / 2) is subtracted from the upper limit value Lmax, which is the current setting signal, to obtain (ΔLmin /
2) is added, and the subtraction result and the addition result are set as new setting signals Lmax and Lmin, respectively, and the process returns to step S5.

In step S9, ΔLmax and ΔLmin
Is determined to be within the predetermined value, the process proceeds to step S12, and the difference ΔMaxt between the input upper limit Lmax1 stored in the register and the currently set load upper limit Lmax, and the input lower limit Lmin1 stored in the register. , And the difference ΔMint between the currently set load lower limit value Lmin and the currently set load lower limit value Lmin. In step S13, the upper and lower limit values finally corrected and set are correlated with the input data on the specimen and the test condition data. Is stored in the storage unit shown in FIG. 7 as correction value data ΔMaxt and ΔMint. If the deviation has already been stored as correction value data for the same condition, the stored value is obtained this time and updated with the correction value. Thereafter, the flag is reset in step S14.

Next, the learning processing procedure shown in FIG. 6 will be described. First, in step S21, the correction value is expressed as a function of frequency based on the data shown in FIG. 7, that is, the past learning data. In step S22, the correction value ΔMaxt, Δ
Find Mint. Then, the process proceeds to step S23, in which the upper and lower limit set values Lmax and Lmin are obtained, ΔMaxt and ΔMint.
To fix.

For example, as shown in FIG. 7, a frequency CYC4 having no past correction value data, an upper / lower limit + 10.0Kg,
-10.0 kg is used as data of test conditions, and the correction value ΔM
The case where the setting signal is corrected by obtaining axt4 and ΔMint4 will be described. The frequencies CYC1-3 and CYC5
Correction value ΔMaxt as a function of frequency from the correction data of
4, ΔMint4 is obtained.

The upper and lower limit values Lma set in this way
x and Lmin are supplied from the setting device 30 to the controller 20, and the controller 20 outputs a load pattern setting signal determined by the upper and lower limit values and the frequency. This setting signal is supplied to the actuator 11 via the deviation unit 4a and the amplifier 4b, and the actuator 11 is driven to load the specimen. The load applied to the specimen is detected by the load cell 12, the difference from the set signal is obtained by the deviation unit 4a, and the actuator 11 is driven and controlled in accordance with the difference to load the specimen in a desired load pattern.

As described above, if the correction values of the upper and lower limits corresponding to the data on the specimen and the data on the test conditions already exist, the input upper and lower limits are corrected and set using the correction values. . Even when such a correction value does not exist, if there is some nearby data, a correction value is calculated from these nearby data and the upper and lower limit values are corrected and set according to the procedure of FIG. When there is no nearby correction value data, the test is performed with the input upper and lower limit values, and the upper and lower limit values are obtained until the desired upper and lower limit values are obtained in steps S5 to S11 shown in FIGS. Modify the settings,
The deviation between the corrected upper and lower limit value and the desired upper and lower limit value is stored in a predetermined location of the storage unit shown in FIG. 7 as correction values ΔMaxt and ΔMint.

In the above, the case where the frequency and the upper and lower limit values are used as the test conditions has been particularly described. However, the present invention is not limited to such specific test conditions, and can be similarly applied to various test conditions. Therefore, the method of correcting and setting the input test conditions, the method of calculating the correction value, etc.
Since the methods differ depending on the test conditions used and the methods themselves are well known, only the method of optimizing the gain by learning control will be described here.

The gain referred to here is the gain of the amplifier 4b in FIG. 9, and conventionally, the operator appropriately sets the gain before the test based on past intuition. For example, when the test specimen is to be loaded with the load rising characteristic shown by the solid line in FIG. 8A, if the actual rising characteristic is as shown by the broken line in FIG. Perform the operation to increase the gain before the test. If the actual rise characteristic is as shown by the broken line in FIG. 8B, an operation of reducing the gain is performed prior to the second and subsequent tests.

Therefore, if the gain setting depending on such intuition is set by learning control according to the above-described method, the gain setting is extremely simplified. Such a gain setting can be performed in the same manner as described above. That is,
This can be realized by preparing a table for storing the gain in association with the data on the specimen and the data on the test conditions, and storing or updating the gain sequentially for each test.

In the configuration of the above embodiment, the controller 20 controls the setting signal output section, the load cell controls the detecting means,
The drive control unit 4 constitutes the drive control means, the storage device of the setting device 30 constitutes the storage means, and step S11 of the setting device 30 corresponds to the correction means.
23 corresponds to a specific method of the setting signal setting procedure of the setting signal output means.

[0023]

As described above in detail, according to the present invention, the setting based on the setting signal and the gain correction value stored in advance in association with the data on the test specimen and the test condition set in the past. Signals and gains are set, and the correction values are newly stored or updated for each test. It is greatly simplified.

[Brief description of the drawings]

FIG. 1 is a diagram corresponding to claims of the present invention.

FIG. 2 is a diagram corresponding to claims of the present invention.

FIG. 3 is an overall configuration diagram showing one embodiment of a material testing machine with a learning function according to the present invention.

FIG. 4 is a flowchart showing a part of a procedure for correcting and setting input test conditions.

FIG. 5 is a flowchart continued from FIG. 4;

FIG. 6 is a flowchart showing details of step S3 in FIG. 4;

FIG. 7 is a diagram showing a table in which various test conditions and correction values relating to a specific specimen are associated with each other;

FIG. 8 is a characteristic diagram illustrating an example of a load pattern of a lamp load.

FIG. 9 is a block diagram illustrating an example of a feedback control circuit of the material testing machine.

FIG. 10 shows a load pattern.

FIG. 11 shows a load pattern.

[Explanation of symbols]

 Reference Signs List 4 drive control unit 4a deviation device 10 material testing machine main body 11 actuator 12 load cell 20 controller 30 setting device 40 keyboard 101 setting signal correction means 102 storage means 201 gain correction means 202 storage means

Claims (2)

(57) [Claims] An actuator for repeatedly loading a specimen at a predetermined frequency, setting signal output means for setting and outputting a setting signal indicating a load pattern of the specimen, and a load or displacement accompanying deformation of the specimen due to the load In a material testing machine comprising: a detecting means for detecting the load signal; and a drive control means for driving the actuator so that the setting signal and the detection signal coincide with each other, such that an actual load pattern matches a desired load pattern. Setting signal correction means for correcting the setting signal, and storage means for storing data relating to the test specimen and data relating to various test conditions including the load pattern and correction data in the setting signal correction means in association with each other, The setting signal output unit is configured to perform the correction based on the input data of the specimen and the test condition based on the storage content of the storage unit. A material testing machine with a learning function, wherein data is calculated to generate the setting signal. 2. An actuator for loading a specimen, setting signal output means for setting and outputting a setting signal indicating a load pattern of the specimen, and detecting means for detecting a load or displacement accompanying deformation of the specimen due to the load. And a drive control unit that drives the actuator so that the setting signal and the detection signal match, wherein the drive control unit controls the drive control unit so that an actual load pattern matches a desired load pattern. Gain control means for correcting a gain, and storage means for storing data relating to a specimen, data relating to various test conditions including the gain, and correction data in the gain correction means in association with each other, and the drive control means Calculating the correction data from the input data of the specimen and the test condition based on the storage contents of the storage means, A material testing machine with a learning function, characterized by setting