JP6998243B2 - Impact test analyzer, impact test analysis system, impact test analysis method and program - Google Patents

Description

The present invention relates to an impact test analyzer, an impact test analysis system, an impact test analysis method and a program.

When the Charpy impact test is performed using an instrumented Charpy impact tester, a load displacement curve can be obtained. Examples of the load displacement curve are shown in FIGS. 5 to 6. In this load-displacement curve, it is known that the point (change point) where the slope of the tangent line changes significantly is related to the change in the fracture mode that occurs in the test piece. Using this relationship, the change points related to each fracture mode are identified from the load-displacement curve, and the impact force at the specified change points is used to estimate mechanical properties such as fracture toughness and crack propagation stop toughness. The method to do is proposed.

Further, among the regions surrounded by the load displacement curve in the lower figure of FIG. 5 and the X-axis and the Y-axis, for example, the area of the region divided based on the point (point P4) indicating the impact force at the time of unstable fracture, and FIG. It is known that there is a correlation between the area ratios of the ductile fracture surface region 32 in the fracture surface image shown in 1.

As a related technique, in Patent Document 1, the displacements from the load displacement curve obtained by the Charpy impact test to the maximum load and the maximum load are obtained, and the crack length is measured from the fracture surface of the test piece after the test, and these values are obtained. A method for evaluating brittle fracture propagation stop performance is described.

Japanese Unexamined Patent Publication No. 2017-3377

However, there is no clear rule on the method of specifying the change point of the fracture mode (points P1 to P5 in the lower figure of FIG. 5) from the load displacement curve. In addition, the load-displacement curve often contains noise that hinders the identification of the change point. Therefore, depending on the skill and subjectivity of the operator, there is a possibility that the specified change points may differ even if the load displacement curve is the same. For this reason, there is a demand for a method of eliminating the subjectivity of the operator and improving the accuracy of the method of specifying the change point of the load displacement curve which is the basis of the mechanical characteristics.

Therefore, it is an object of the present invention to provide an impact test analyzer, an impact test analysis system, an impact test analysis method and a program capable of solving the above-mentioned problems.

According to one aspect of the present invention, the impact test analyzer has a test result acquisition unit that acquires a load displacement curve showing the relationship between the load and displacement obtained by the impact test, and a fracture surface of the test piece of the impact test. An image acquisition unit that acquires a captured fracture surface image, an image analysis unit that analyzes the fracture surface image and detects a plurality of image regions having different characteristics from each other, and each of the plurality of image regions are on the fracture surface. The ratio of the area occupied and the area of the load displacement characteristic region corresponding to each of the image regions formed by dividing the region surrounded by the load displacement curve and the coordinate axes based on a predetermined change point in the load displacement curve. Based on the first estimation model that defines the relationship, the load-displacement curve to be analyzed, and the image area to be analyzed, the load-displacement characteristic from the area surrounded by the load-displacement curve and coordinate axes to be analyzed. It is provided with a first estimation unit that calculates the area of the region and calculates the change point based on the calculated area of the load displacement characteristic region.

According to one aspect of the present invention, the impact test analyzer is based on the load displacement curve obtained by the past impact test and the fracture surface image, and the area occupied by each of the image regions in the fracture surface. A learning unit that learns the relationship between the ratio of , Further prepare.

According to one aspect of the present invention, the image analysis unit detects the image region corresponding to each of the ductile fracture surface, the brittle fracture surface, and the shear fracture surface based on the brightness of the pixel.

According to one aspect of the present invention, the change point corresponds to the impact force at the time of unstable fracture occurrence.

According to one aspect of the present invention, the change point corresponds to the impact force at the time of stopping the crack propagation.

According to one aspect of the present invention, the impact test analyzer has a second estimation model that defines a combination of the shape of the load displacement curve and the position of the change point in the load displacement curve, and the load displacement to be analyzed. Based on the curve, the second estimation unit that estimates the change point from the load displacement curve to be analyzed, and the estimation result by the second estimation unit, the necessity of estimation by the first estimation unit is determined. It further includes an estimation necessity determination unit for determination.

According to one aspect of the present invention, the impact test analysis system includes a Charpy impact tester, a camera for photographing the test piece of the Charpy impact test, and the impact test analyzer.

According to one aspect of the present invention, the impact test analysis method includes a step of acquiring a load displacement curve showing the relationship between the load and the displacement obtained by the impact test, and a fracture in which the fracture surface of the test piece of the impact test is photographed. A step of acquiring a surface image, a step of analyzing the fracture surface image to detect a plurality of image regions having different characteristics from each other, a ratio of the area occupied by each of the plurality of image regions to the fracture surface, and the load. The first estimation that defines the relationship with the area of the load displacement characteristic region corresponding to each of the image regions formed by dividing the region surrounded by the displacement curve and the coordinate axes based on a predetermined change point in the load displacement curve. Based on the model, the load-displacement curve to be analyzed, and the image area to be analyzed, the area of the load-displacement characteristic region is calculated from the area surrounded by the load-displacement curve and the coordinate axes to be analyzed. It has a step of calculating the change point based on the calculated area of the load displacement characteristic region.

According to one aspect of the present invention, the program is a means for obtaining a load displacement curve showing the relationship between the load and the displacement obtained by the impact test, and the fracture surface obtained by photographing the fracture surface of the test piece of the impact test. Means for acquiring an image, means for analyzing a fracture surface image to detect a plurality of image regions having different characteristics, a ratio of an area occupied by each of the plurality of image regions to the fracture surface, a load displacement curve, and the like. A first estimation model that defines the relationship with the area of the load displacement characteristic region corresponding to each of the image regions formed by dividing the region surrounded by the coordinate axes based on a predetermined change point in the load displacement curve. Based on the load-displacement curve to be analyzed and the image area to be analyzed, the area of the load-displacement characteristic region is calculated from the area surrounded by the load-displacement curve and the coordinate axes to be analyzed. It functions as a means for calculating the change point based on the area of the load displacement characteristic region.

According to the impact test analyzer, impact test analysis system, impact test analysis method and program of the present invention, the mechanical characteristics can be estimated accurately from the load displacement curve obtained by the Charpy impact test.

It is a block diagram of the impact test analysis system in each embodiment of this invention. It is a figure explaining the test method and the method of taking a fracture surface image in each embodiment of this invention. It is a figure which shows an example of the fracture surface image in 1st Embodiment of this invention. It is a figure explaining the load displacement characteristic region in the 1st Embodiment of this invention. It is the first figure which shows an example of a load displacement curve. It is the 2nd figure which shows an example of a load displacement curve. It is a flowchart which shows an example of the analysis process by 1st Embodiment of this invention. It is a flowchart which shows an example of the analysis process by 2nd Embodiment of this invention. It is a figure which shows an example of the hardware composition of the impact test analyzer in each embodiment of this invention.

<First Embodiment>
Hereinafter, the impact test analysis method according to the first embodiment of the present invention will be described with reference to FIGS. 1 to 7. FIG. 1 is a block diagram of an impact test analysis system according to each embodiment of the present invention.
FIG. 2 is a diagram illustrating a test method and a method of photographing a fracture surface image in each embodiment of the present invention.
As shown in FIG. 1, the impact test analysis system includes an instrumented Charpy impact tester 1, a camera 2, and an impact test analyzer 10.
The impact test analyzer 10 is an apparatus used for estimating the change point appearing in the load displacement curve obtained as a result of the test by the instrumented Charpy impact tester 1. The impact test analyzer 10 is a computer including a CPU (Central Processing Unit) and the like. It is known that the change point is related to the change in mechanical properties (fracture mode) that occurs in the test piece, and if the change point can be estimated, it is based on the load (impact force) corresponding to the change point. Therefore, the mechanical properties can be calculated by various methods.

The instrumentation Charpy impact tester 1 applies a load to the test piece 20 by a hammer 3 as shown in the left figure of FIG. A strain sensor or the like is attached to the hammer 3, and the instrumented Charpy impact tester 1 measures the strain when a load is applied to the test piece 20 by the hammer 3 and estimates the impact force and displacement. The load displacement curves illustrated in FIGS. 5 to 6 are output.

When the impact test is performed, the test piece 20 is destroyed as shown in the right figure of FIG. 2, and the test piece 20 is divided into the test piece 20a and the test piece 20b. The camera 2 photographs the fracture surface of the test piece 20a or the test piece 20b. For the camera 2, for example, an optical microscope or an electron microscope can be used. For example, when photographing the test piece 20a, the camera 2 photographs the fracture surface of the test piece 20a from the direction of the arrow in the right figure of FIG. The camera 2 outputs the captured image of the fracture surface (fracture surface image) to the impact test analyzer 10.

Next, the function of the impact test analyzer 10 will be described. As shown in FIG. 1, the impact test analysis device 10 includes a test result acquisition unit 11, an image acquisition unit 12, an image analysis unit 13, a learning unit 14, an estimation necessity determination unit 15, and a first estimation unit 16. A second estimation unit 17, an input / output unit 18, and a storage unit 19 are provided.
The test result acquisition unit 11 acquires the load displacement curve output by the instrumentation Charpy impact tester 1.
The image acquisition unit 12 acquires a fracture surface image such as a test piece 20a taken by the camera 2.
The image analysis unit 13 analyzes the fracture surface image and detects a plurality of image regions having different characteristics from each other. Here, the analysis of the fracture surface image will be described with reference to FIG.

FIG. 3 is a diagram showing an example of a fracture surface image according to the first embodiment of the present invention.
As shown in FIG. 3, the fracture surface image 30 includes a notch region 31, a ductile fracture surface region 32, a brittle fracture surface region 33, and a shear fracture surface region 34. Each of the notch region 31, the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34 is composed of pixels having different brightness from each other, and is in the range of brightness corresponding to each image region 31 to 34. Is preset. Therefore, the image analysis unit 13 divides the fracture surface image 30 into image regions having the same brightness by image analysis processing. The image analysis unit 13 corresponds to each of the divided image areas and the image areas 31 to 34 with each of the notch, ductile fracture surface, brittle fracture surface, and shear fracture surface previously recorded in the storage unit 19. Correspond by referring to the information indicating the range of the attached brightness. As a result, the image analysis unit 13 detects a plurality of image regions (notch region 31, ductile fracture surface region 32, brittle fracture surface region 33, shear fracture surface region 34). Here, it is decided to detect a plurality of image areas based on the brightness, but for example, not only the brightness but also the color and the state of the surface of the fracture surface reflected in the image (for example, particles of a predetermined size are captured). Each image area may be detected by (such as). Further, regarding the characteristics such as brightness, the threshold value and the range thereof may be calculated by using machine learning.

The learning unit 14 divides the area surrounded by the ratio of the area occupied by each of the plurality of image areas to the fracture surface image 30, the load displacement curve, and the coordinate axes (X-axis and Y-axis) based on the change points P4 and P5. A first estimation model is created that defines the relationship with the area of the formed region (called the load-displacement characteristic region). (A1) First, the learning unit 14 calculates the area ratio of the image region based on the analysis result by the image analysis unit 13. For example, the learning unit 14 calculates the ratio of the area of the ductile fracture surface region 32 in the image to the area of the entire fracture surface image 30 for the ductile fracture surface region 32. The area of the entire fracture surface image 30 is, for example, the total area of each portion of the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34. The learning unit 14 calculates the area ratios of the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34, respectively. The learning unit 14 records each calculated area ratio in the storage unit 19.
(A2) Further, the learning unit 14 calculates the area of the load displacement characteristic region. Here, the load displacement characteristic region will be described with reference to FIG.

FIG. 4 is a diagram illustrating a load displacement characteristic region according to the first embodiment of the present invention.
The learning unit 14 divides the region surrounded by the load displacement curve L1 and the X-axis and the Y-axis by a line L4 passing through the change point P4 and perpendicular to the X-axis and a line L5 passing through the change point P5 and perpendicular to the X-axis. do. The range surrounded by the load-displacement curve L1 and the X-axis, the Y-axis, and the line L4 is the load-displacement characteristic region r1. The vertically long portion surrounded by the load-displacement curve L1 and the X-axis, the line L4, and the line L5 is the load-displacement characteristic region r2. The horizontally long portion surrounded by the load-displacement curve L1 and the X-axis and the line L5 is the load-displacement characteristic region r3. The learning unit 14 divides the load-displacement characteristic regions r1 to r3 based on the load-displacement curve L1 and the coordinate information of the change points P4 and P5, and calculates the area of each of the load-displacement characteristic regions r1 to r3. Here, the coordinate information of the change points P4 and P5 is given together with the load displacement curve L1. That is, the load displacement curves L1 and the change points P4 and P5 are given as teacher data, and the learning unit 14 uses the load displacement curves L1 and the change points P4 and P5 given as the teacher data as correct answers and the load displacement characteristic regions r1 to r3. Calculate the area of. The learning unit 14 records the calculated area of the load displacement characteristic regions r1 to r3 in the storage unit 19.

(A3) Next, the learning unit 14 learns the relationship between the area ratio of (A1) and the area of the load displacement characteristic region of (A2). (A3-1) Specifically, the learning unit 14 calculates the ratio (referred to as ductility ratio) between the area ratio of the ductile fracture surface region 32 and the area of the load displacement characteristic region r1. Similarly, the learning unit 14 calculates the ratio of the area ratio of the brittle surface region 33 to the area of the load displacement characteristic region r2 (referred to as the brittle ratio), and the area ratio of the shear fracture surface region 34 and the load displacement characteristic region r3. Calculate the area ratio (called the shear ratio). (A3-2) The learning unit 14 acquires the fracture surface image and the load displacement curve obtained from a plurality of test results for the test piece 20 of the same material, and obtains (A1), (A2), (A3-1). The process of is repeated, and a large number of calculation results of ductility ratio, brittleness ratio, and shear ratio are recorded in the storage unit 19. Alternatively, the learning unit 14 records in the storage unit 19 the values of the area ratio of the ductile fracture surface region 32 and the load displacement characteristic region r1 obtained from the respective test results. (A3-3) The learning unit 14 learns each relationship from the calculation results of the ductility ratio, the brittleness ratio, and the shear ratio accumulated in large numbers, and creates a first estimation model. For example, the learning unit 14 learns the relationship between the area ratio of the ductile fracture surface region 32 and the load displacement characteristic region r1 area from the ductility ratio accumulated in large numbers. As the simplest example, the learning unit 14 calculates the average value of the ductility ratios accumulated in large numbers, and uses the average value as the first estimation for the area ratio of the ductile fracture surface region 32 and the load displacement characteristic region r1 area. Use as a model. Further, for example, the learning unit 14 obtains an approximate straight line from the area ratio of the ductile fracture surface region 32 and the load displacement characteristic region r1 accumulated in large numbers, or uses a predetermined statistical method or machine learning to obtain a relationship between the two (ducity ratio). Is calculated. Similarly, the learning unit 14 has a relationship between the area ratio of the brittle fracture surface region 33 and the area ratio of the load displacement characteristic region r2 (brittleness ratio), and a relationship between the area ratio of the shear fracture surface region 34 and the area of the load displacement characteristic region r3. Calculate (shear ratio). The first estimation model 19a contains the calculated information defining these three relationships. The learning unit 14 records the first estimation model 19a in the storage unit 19.

The estimation necessity determination unit 15 determines the necessity of estimation by the first estimation unit 16 based on the estimation result by the second estimation unit 17. The processing of the estimation necessity determination unit 15 will be described with reference to FIG. 8 (second embodiment).

In the first estimation unit 16, the image analysis unit 13 detects the first estimation model 19a, the load displacement curve obtained as a result of performing a Charpy impact test on the test piece 20 to be analyzed, the test piece 20a to be analyzed, and the like. Based on the plurality of image regions (ductile fracture surface region 32, etc.), the change points corresponding to the points P4 to P5 in FIG. 5 are estimated.
(B1) First, the first estimation unit 16 performs the same processing as in (A1) above to calculate the area ratio of the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34. do. The first estimation unit 16 records the calculated area ratio in the storage unit 19.
(B2) Next, the first estimation unit 16 changes in the load displacement curve obtained as a result of performing a Charpy impact test on the test piece 20 to be analyzed using the result of (B1) and the first estimation model 19a. The coordinate information of the change point corresponding to the points P4 to P5 is estimated. First, the first estimation unit 16 calculates the area of the load displacement characteristic regions r1 to r3 using the result of (B1) and the first estimation model 19a. For example, the first estimation model 19a defines the ductility ratio after learning. As a result of (B1), since the area ratio of the ductile fracture surface region 32 of the test piece 20a to be analyzed is known, the first estimation unit 16 determines the area of the load displacement characteristic region r1 from this area ratio and the ductility ratio. Can be calculated. The first estimation unit 16 similarly calculates the area of the load displacement characteristic region r2 and the area of the load displacement characteristic region r3.

(B3) Next, the first estimation unit 16 applies the calculated areas of the load displacement characteristic regions r1 to r3 to the load displacement curve of the test piece to be analyzed, and calculates points P4 and P5. For example, in FIG. 4, the first estimation unit 16 is the load displacement characteristic region r1 calculated by the area of the load displacement curve L1, the X axis, the Y axis, and the portion surrounded by the line L4 with the Y axis as a reference. The line L4 like this is calculated. That is, while moving the line L4 from the position where the value of the x coordinate becomes 0 in the positive direction of the X axis, the area of the load displacement curve L1, the X axis, the Y axis, and the portion surrounded by the line L4 is the load displacement characteristic. The x-coordinate of the line L4 when it becomes the region r1 is calculated. Next, the first estimation unit 16 calculates the intersection of this line L4 and the load displacement curve L1. This intersection is the change point P4. Next, the first estimation unit 16 performs the same processing with the calculated line L4 as a reference, and the area of the portion surrounded by the load displacement curve L1 and the X-axis, the line L4, and the line L5 becomes the load displacement characteristic region r2. The line L5 like this is calculated. Then, the first estimation unit 16 calculates the intersection (change point P5) of the calculated line L5 and the load displacement curve L1.

Alternatively, the first estimation unit 16 may calculate the change points P4 and P5 with reference to the line H1 perpendicular to the X axis passing through the intersection of the load displacement curve L1 and the X axis. For example, the first estimation unit 16 moves the line L5 in the negative direction of the X-axis with the line H1 as a reference, and the area of the load displacement curves L1 and the X-axis and the portion surrounded by the line L5 is the load displacement. The line L5 when it becomes the characteristic region r3 is calculated, and the change point P5 is calculated. Subsequently, the first estimation unit 16 moves the line L4 in the negative direction of the X-axis with the line L5 as a reference, and the area of the portion surrounded by the load displacement curves L1 and the X-axis, the line L5, and the line L4. Is the load displacement characteristic region r2, the line L4 is calculated, and the change point P4 is calculated. The first estimation unit 16 outputs the calculated coordinate information of the change points P4 and P5 as the estimated values of the change points P4 and P5 in the load displacement curve to be analyzed.

The second estimation unit 17 changes corresponding to points P4 to P5 in FIG. 5 based on the second estimation model 19b and the load displacement curve obtained as a result of performing a Charpy impact test on the test piece 20 to be analyzed. Estimate the point. The second estimation model 19b is information that combines a load displacement curve such that there is little noise and the coordinates of the points P1 to P5 are clear, and the coordinate information of the points P1 to P5 in the load displacement curve. For example, the second estimation unit 17 compares the load-displacement curve with less noise of the second estimation model 19b with the load-displacement curve to be analyzed, and the second estimation model has a similarity between the two within a predetermined range. The load displacement curve of 19b is selected. Then, the points corresponding to the coordinate information of the change points P4 and the change points P5 set for the load-displacement curve of the second estimation model 19b are specified on the load-displacement curve to be analyzed. For example, the second estimation unit 17 is the point closest to the point P4 among the points where the slope of the tangent line of the load displacement curve to be analyzed changes by a predetermined value or more, and the distance between the two points is within a predetermined range. Is estimated as a point corresponding to the change point P4 in the load displacement curve to be analyzed. The same applies to the change point P5. The second estimation unit 17 outputs the coordinate information of the points estimated as corresponding points as the estimated values of the change points P4 and P5 in the load displacement curve to be analyzed.

The input / output unit 18 receives input of information from the user to the impact test analyzer 10. Further, the estimation result is output by the first estimation unit 16 and the second estimation unit 17.
The storage unit 19 stores various data. For example, the storage unit 19 stores the first estimation model 19a and the second estimation model 19b. Further, the storage unit 19 stores a fracture surface image and a load displacement curve obtained when a Charpy impact test was performed in the past.

Next, points P1 to P5 of the load displacement curve will be described.
FIG. 5 is a first diagram showing an example of a load displacement curve. FIG. 6 is a second diagram showing an example of a load displacement curve.
The upper figure of FIG. 5 is an example of a load displacement curve with relatively little noise and easy identification of a change point. In the conventional method, the operator performing the analysis performs noise removal on the upper figure of FIG. 5 to generate the curve L2 in the lower figure of FIG. Then, the operator identifies the point where the inclination of the tangent line of the curve L2 changes, and identifies the points P1 to P5. Here, the points P1 correspond to the yield impact force in order from the one with the smallest displacement amount. The point P2 corresponds to the impact force at the time of ductile crack generation. The point P3 is a point corresponding to the maximum impact force. The point P4 corresponds to the impact force at the time of unstable fracture. The point P5 corresponds to the impact force when the crack propagation is stopped. As described above, in the present embodiment, the first estimation unit 16 and the second estimation unit 17 have the coordinates of the point P4 corresponding to the impact force at the time of unstable fracture and the point P5 corresponding to the impact force at the time of stopping the crack propagation. Estimate the information.

With the load-displacement curve as illustrated in FIG. 5, even with the conventional method, there is no room for the subjectivity of the operator to enter, and there is a high possibility that accurate points P1 to P5 can be specified. However, in the case of the load displacement curve as shown in FIG. 6, it is difficult to specify the points P1 to P5 by the conventional method. FIG. 6 shows an example of a load displacement curve whose change point is difficult to identify by the conventional method. In the upper figure of FIG. 6, it is difficult to distinguish between noise and the change point, and in the lower figure of FIG. 6, it is difficult to identify the change point because the waveform of the portion where the displacement amount is small is disturbed. On the other hand, according to the method of the present embodiment, the area ratio of the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34 based on the fracture surface image of the test piece 20a or the like to be analyzed, and the first. Based on the estimation model 19a, the area of the load displacement characteristic regions r1 to r3 in the load displacement curve to be analyzed can be calculated, and the points P4 and P5 are estimated based on the area, so that the shape of the load displacement curve. Points P4 and P5 can be estimated without being directly affected by the state of noise and noise. As a result, it is possible to eliminate the subjectivity of the operator performing the analysis, improve the accuracy, and reduce the time and effort required to identify the points P4 and P5.

Next, the flow of the analysis process of this embodiment will be described.
FIG. 7 is a flowchart showing an example of the analysis process according to the first embodiment of the present invention.
As a premise, the load displacement curve obtained when the Charpy impact test was performed on the test piece 20 of a certain material in the past, and the points P1 to P5 (at least the points P4 and P5) in the load displacement curve. Coordinate information and fracture surface image information exist. These are supervised learning data in which the coordinate information of points P1 to P5 is the correct answer.

First, the operator inputs the learning data to the impact test analyzer 10. The test result acquisition unit 11 acquires a supervised load displacement curve included in the learning data and records it in the storage unit 19. The image acquisition unit 12 acquires a fracture surface image included in the learning data and records it in the storage unit 19 (step S11).

Next, the image analysis unit 13 detects the notch region 31, the ductile fracture surface region 32, the brittle fracture surface region 33, and the shear fracture surface region 34 from the fracture surface image. Next, the learning unit 14 performs the above processes (A1) to (A3) to create the first estimation model 19a (step S12). The learning unit 14 records the first estimation model 19a in the storage unit 19.

Next, the worker performs an impact test on the test piece 20 to be analyzed with the instrumentation Charpy impact tester 1, and takes a fracture surface image with the camera 2. The worker inputs the load displacement curve and the fracture surface image of the test piece 20 to be analyzed into the impact test analyzer 10. The test result acquisition unit 11 acquires the load displacement curve to be analyzed and records it in the storage unit 19. The image acquisition unit 12 acquires the fracture surface image to be analyzed and records it in the storage unit 19 (step S13).

Next, the first estimation unit 16 performs the above processes (B1) to (B3) to estimate the coordinate information of the change points P4 and P5 for the analysis target data (step S14). The first estimation unit 16 records the load-displacement curve to be analyzed and the coordinate information of the change points P4 and P5 in association with each other in the storage unit 19. The first estimation unit 16 outputs the estimated coordinate information to the input / output unit 18. Next, the input / output unit 18 displays the coordinate information of the change points P4 and P5 on the display (step S15).

Here, the worker determines whether or not the estimation result by the first estimation unit 16 is correct, and if it is not appropriate, inputs the correct coordinate information to the impact test analyzer 10. The input / output unit 18 receives the input of this correction information and records (overwrites) it in the storage unit 19 in association with the load displacement curve to be analyzed (step S16).

Next, the operator inputs a signal instructing the update of the first estimation model 19a to the impact test analyzer 10 as needed. Then, the learning unit 14 adds the load-displacement curve analyzed this time, the coordinate information of the change points P4 and P5, and the fracture surface image to the training data, and updates the first estimation model 19a (step S17). The learning unit 14 records the updated first estimation model 19a in the storage unit 19. In particular, when the correction information is input in step S16, the first estimation model 19a may be immature, but the number of training data is increased by adding the analysis target data to the given training data. By updating the first estimation model 19a with more learning data, it can be expected that the accuracy and reliability of the first estimation model 19a will be improved.

According to this embodiment, it is possible to estimate the change points P4 and P5 with high accuracy and in a short time by excluding the subjectivity of the operator. Thereby, the mechanical properties (fracture toughness value, crack propagation stopping toughness) of the test piece 20 can be efficiently estimated.

<Second embodiment>
In the first embodiment, if the test piece 20 is made of the same material, the area ratio of the ductile fracture surface region 32 and the area of the load displacement characteristic region r1 of the fracture surface image, the area ratio of the brittle fracture surface region 33 and the load displacement characteristic region are The change points P4 and P5 were estimated based on the fact that there is a certain relationship between the area of r2, the area ratio of the shear fracture surface region 34, and the area of the load displacement characteristic region r3. However, as illustrated in FIG. 5, there are cases where the change points P4 and P5 can be estimated without relying on the estimation of the first estimation unit 16. In the second embodiment, when such a load-displacement curve is obtained, the change point is estimated without using the fracture surface image.

FIG. 8 is a flowchart showing an example of the analysis process according to the second embodiment of the present invention.
First, the operator inputs the learning data to the impact test analyzer 10. The test result acquisition unit 11 acquires a supervised load displacement curve included in the learning data and records it in the storage unit 19. The image acquisition unit 12 acquires a fracture surface image included in the learning data and records it in the storage unit 19 (step S21). Next, the learning unit 14 creates the first estimation model 19a (step S22) in the same manner as in step S12 of FIG. 7, and records it in the storage unit 19. Next, the learning unit 14 creates a second estimation model 19b by associating the load displacement curve with the coordinate information of the points P1 to P5, and records the second estimation model 19b in the storage unit 19 (step S23).

Next, the worker performs an impact test on the test piece 20 to be analyzed with the instrumentation Charpy impact tester 1, and takes a fracture surface image with the camera 2. The worker inputs only the load displacement curve of the test piece 20 to be analyzed into the impact test analyzer 10. The test result acquisition unit 11 acquires the load displacement curve to be analyzed and records it in the storage unit 19 (step S24).

Next, the second estimation unit 17 estimates the change point based on the second estimation model 19b (step S25). For example, the second estimation unit 17 compares the load-displacement curve to be analyzed with the plurality of load-displacement curves of the second estimation model 19b, and selects the one having the highest degree of similarity and correlation coefficient. Then, the coordinates of the points corresponding to the coordinates of the change points P4 and P5 in the load displacement curve of the selected second estimation model 19b are calculated with respect to the load displacement curve to be analyzed, and these points are calculated as the change points P4. It is an estimated value of P5. Next, the estimation necessity determination unit 15 determines whether or not estimation by the first estimation model is necessary (step S26). For example, in the estimation necessity determination unit 15, the second estimation unit 17 corresponds to the change points P4 and P5 in the load displacement curve of the second estimation model 19b, which has a high degree of similarity to the load displacement curve to be analyzed. If it can be estimated, it is determined that the estimation by the first estimation model is unnecessary. On the other hand, for example, when there is no similar load-displacement curve, or when there is little change in the slope of the tangent line around the points corresponding to the change points P4 and P5 in the load-displacement curve of the second estimation model 19b having a high degree of similarity. , The estimation necessity determination unit 15 determines that estimation by the first estimation model is necessary.

When estimation by the first estimation model is required (step S26; Yes), the input / output unit 18 displays a message prompting the input of the fracture surface image. On the other hand, the worker inputs the fracture surface image to be analyzed into the impact test analyzer 10. The image acquisition unit 12 acquires the fracture surface image to be analyzed and records it in the storage unit 19 (step S30). Next, the first estimation unit 16 estimates the change points P4 and P5 in the same manner as in step S14 (step S31).

Next, the input / output unit 18 outputs the estimation result (step S27). When the estimation by the first estimation model 19a is unnecessary (step S26; No), the input / output unit 18 displays the estimation result by the second estimation unit 17. When the estimation is performed by the first estimation model, the input / output unit 18 displays the estimation result by the first estimation unit 16.

Next, the worker determines whether or not the estimation result is correct as in the first embodiment, and if not, inputs the correct coordinate information to the impact test analyzer 10. The input / output unit 18 receives the input of the correction information, and records (overwrites) the corrected coordinate information of the change points P4 and P5 in the storage unit 19 in association with the load displacement curve to be analyzed (step S28).

Next, the operator inputs a signal instructing the update of the first estimation model 19a and the second estimation model 19b to the impact test analyzer 10 as needed. The learning unit 14 adds the load-displacement curve analyzed this time, the coordinate information of the change points P4 and P5, and the fracture surface image to the training data, and updates the first estimation model 19a and the second estimation model 19b (step S29). The learning unit 14 records the updated first estimation model 19a and second estimation model 19b in the storage unit 19. For example, when the estimation by the first estimation unit 16 is not performed, it is assumed that the load displacement curve to be analyzed this time is desirable (high purity) data as training data, and the first estimation model 19a is weighted higher than usual. May be updated. This may improve the accuracy of the first estimation model 19a.

According to this embodiment, in addition to the effect of the first embodiment, it is possible to improve the efficiency of work by simplifying the input data. Further, the processing load of the impact test analyzer 10 can be reduced. In addition, by accumulating load-displacement curves of various waveforms with less noise and high purity as the second estimation model 19b, the change points P4 and P5 can be estimated without the need to analyze the fracture surface image in the future. By increasing the ratio, it becomes possible to estimate the change points P4 and P5 at higher speed and with less effort.

FIG. 9 is a diagram showing an example of the hardware configuration of the control device according to each embodiment of the present invention.
The computer 900 is, for example, a PC (Personal Computer) or a server terminal device including a CPU 901, a main storage device 902, an auxiliary storage device 903, an input / output interface 904, and a communication interface 905. The impact test analyzer 10 described above is mounted on the computer 900. The operation of each of the above-mentioned processing units is stored in the auxiliary storage device 903 in the form of a program. The CPU 901 reads a program from the auxiliary storage device 903, expands it to the main storage device 902, and executes the above processing according to the program. Further, the CPU 901 secures a storage area corresponding to the storage unit 19 in the main storage device 902 according to the program. Further, the CPU 901 secures a storage area for storing the data being processed in the auxiliary storage device 903 according to the program. The computer 900 may include a processor such as a GPU (Graphics Processing Unit) in place of the CPU 901 or in addition to the CPU 901. In that case, the operation of each of the above-mentioned processing units may be executed by the GPU or the like.

In at least one embodiment, the auxiliary storage device 903 is an example of a non-temporary tangible medium. Other examples of non-temporary tangible media include magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, semiconductor memories, etc. connected via the input / output interface 904. Further, when this program is distributed to the computer 900 by a communication line, the distributed computer 900 may expand the program to the main storage device 902 and execute the above processing. Further, the program may be for realizing a part of the above-mentioned functions. Further, the program may be a so-called difference file (difference program) that realizes the above-mentioned function in combination with another program already stored in the auxiliary storage device 903.

All or part of the test result acquisition unit 11, the image acquisition unit 12, the image analysis unit 13, the learning unit 14, the estimation necessity determination unit 15, the first estimation unit 16, the second estimation unit 17, and the input / output unit 18 It may be realized by using hardware such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), FPGA (Field-Programmable Gate Array).

In addition, it is possible to replace the components in the above-described embodiment with well-known components as appropriate without departing from the spirit of the present invention. Further, the technical scope of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention.

1 ... Instrumentation Charpy impact tester 2 ... Camera 3 ... Hammer 10 ... Impact test analyzer 11 ... Test result acquisition unit 12 ... Image acquisition unit 13 ... Image analysis Unit 14 ... Learning unit 15 ... Estimation necessity determination unit 16 ... First estimation unit 17 ... Second estimation unit 18 ... Input / output unit 19 ... Storage unit

Claims (9)

A test result acquisition unit that acquires a load-displacement curve showing the relationship between the load and displacement obtained by the impact test, and
An image acquisition unit that acquires a fracture surface image obtained by photographing the fracture surface of the test piece of the impact test, and
An image analysis unit that analyzes the fracture surface image and detects a plurality of image regions having different characteristics from each other.
Each of the image regions formed by dividing the ratio of the area occupied by each of the plurality of image regions to the fracture surface and the region surrounded by the load-displacement curve and the coordinate axes based on predetermined change points in the load-displacement curve. Based on the first estimation model that defines the relationship with the area of the load displacement characteristic region corresponding to, the load displacement curve of the analysis target, and the image region of the analysis target, the load displacement curve and the analysis target are analyzed. A first estimation unit that calculates the area of the load-displacement characteristic region from the area surrounded by the coordinate axes and calculates the change point based on the calculated area of the load-displacement characteristic region.
Impact test analyzer. Based on the load-displacement curve and the fracture surface image obtained by the past impact test, the ratio of the area occupied by each of the image regions to the fracture surface and the region surrounded by the load-displacement curve and the coordinate axes are described. A learning unit that learns the relationship with the area of the load-displacement characteristic region formed by dividing based on the change point and creates the first estimation model.
The impact test analyzer according to claim 1. The image analysis unit detects the image region corresponding to each of the ductile fracture surface, the brittle fracture surface, and the shear fracture surface based on the brightness of the pixel.
The impact test analyzer according to claim 1 or 2. The change point corresponds to the impact force at the time of unstable fracture.
The impact test analyzer according to any one of claims 1 to 3. The change point corresponds to the impact force when the crack propagation is stopped.
The impact test analyzer according to any one of claims 1 to 4. Based on the second estimation model that defines the combination of the shape of the load-displacement curve and the position of the change point in the load-displacement curve, and the load-displacement curve to be analyzed, the load-displacement curve to be analyzed is described. The second estimation part that estimates the change point and
Based on the estimation result by the second estimation unit, the estimation necessity determination unit that determines the necessity of estimation by the first estimation unit, and the estimation necessity determination unit.
The impact test analyzer according to any one of claims 1 to 5, further comprising. Charpy impact tester and
A camera that shoots the test piece of the Charpy impact test,
The impact test analyzer according to any one of claims 1 to 6.
Impact test analysis system. A step to obtain a load-displacement curve showing the relationship between the load and displacement obtained by the impact test, and
The step of acquiring a fracture surface image obtained by photographing the fracture surface of the test piece of the impact test, and
A step of analyzing the fracture surface image to detect a plurality of image regions having different characteristics from each other.
Each of the image regions formed by dividing the ratio of the area occupied by each of the plurality of image regions to the fracture surface and the region surrounded by the load-displacement curve and the coordinate axes based on predetermined change points in the load-displacement curve. Based on the first estimation model that defines the relationship with the area of the load displacement characteristic region corresponding to, the load displacement curve of the analysis target, and the image region of the analysis target, the load displacement curve and the analysis target are analyzed. A step of calculating the area of the load-displacement characteristic region from the area surrounded by the coordinate axes and calculating the change point based on the calculated area of the load-displacement characteristic region.
Impact test analysis method. Computer,
A means for obtaining a load-displacement curve showing the relationship between load and displacement obtained by an impact test,
A means for acquiring a fracture surface image obtained by photographing the fracture surface of the test piece of the impact test,
A means for analyzing a fracture surface image to detect a plurality of image regions having different characteristics from each other.
Each of the image regions formed by dividing the ratio of the area occupied by each of the plurality of image regions to the fracture surface and the region surrounded by the load-displacement curve and the coordinate axes based on predetermined change points in the load-displacement curve. Based on the first estimation model that defines the relationship with the area of the load displacement characteristic region corresponding to, the load displacement curve of the analysis target, and the image region of the analysis target, the load displacement curve and the analysis target are analyzed. A means for calculating the area of the load-displacement characteristic region from a region surrounded by coordinate axes and calculating the change point based on the calculated area of the load-displacement characteristic region.
A program to function as.

Images