WO2020136976A1

WO2020136976A1 - Method for identifying elastic characteristics of adhesive

Info

Publication number: WO2020136976A1
Application number: PCT/JP2019/031120
Authority: WO
Inventors: 洋輔植木; 澤田　貴彦
Original assignee: 日立化成株式会社
Priority date: 2018-12-27
Filing date: 2019-08-07
Publication date: 2020-07-02
Also published as: JP2020106378A

Abstract

Provided is a method for identifying the elastic characteristics of an adhesive in a state where adherends are mutually adhered by the adhesive. In the method, an adhesive strength test is carried out, the strain distribution of the adhesive is measured in the adhesive strength test, and using the strain distribution, the elastic characteristics of the adhesive are calculated. Due to this configuration, the elastic characteristics of an adhesive which is in a state of adhering can be estimated using data from nothing more than a common adhesive strength test.

Description

Method for identifying elastic properties of adhesives

The present invention relates to a method for identifying elastic properties of an adhesive.

In recent years, the number of cases where adhesive bonding using an adhesive is adopted in mechanical structures is increasing. In particular, in transportation equipment such as automobiles and aircrafts, there is an urgent need to reduce the weight to improve fuel efficiency.

For these structural members, it is effective to use not only conventional steel members, but also various metals such as aluminum and other light metals, and fiber-reinforced resins in combination. Therefore, a dissimilar material joining technique for joining dissimilar materials is inevitably required.

　Adhesive joining has attracted particular attention because it does not use additional joining members such as bolts, so it has a great effect on weight reduction. Further, even in the case of joining materials of the same kind, it is a suitable joining technique for materials to which welding cannot be applied.

Demand for structural adhesives for mechanical structures is expanding under these circumstances.

Since a large load and stress act on the joint using the structural adhesive, its elastic properties may affect the deformation behavior and rigidity of the joint or the entire mechanical structure. That is, it is necessary to select the structural adhesive in consideration of not only simple adhesive strength but also elastic characteristics of the adhesive itself.

Also, in the adhesive development process, it is necessary to grasp the elastic properties of the adhesive and feed it back to the development efficiently.

Non-Patent Document 1 describes a single lap joint test (single lap joint test, hereinafter referred to as “SLJ test”) that is widely used as a test for evaluating the adhesive strength of an adhesive joint.

Non-Patent Document 2 describes a GR theoretical model which is one of stress distribution models of a test piece used for the SLJ test (hereinafter also referred to as “SLJ test piece”).

In Patent Document 1, it is possible to directly compare the strength determination result of the adhesive with respect to the adhesive strength determination of the numerical analysis model of the structure having the adhesive, regarding the adhesive structure in which two adherends are adhered by the adhesive. In order to achieve the above, a method of calculating the adhesive property by a finite element analysis model that models the adhesive as a beam element is disclosed.

There are the following publicly known examples regarding the physical property distribution of the rubber-like elastic body having the adhesive surface, although it is not the physical property distribution of the adhesive.

Patent Document 2 discloses a rubber-like elastic composite in which the degree of deformation of a rubber-like elastic body is measured, and the physical property value in the vicinity of the adhesive interface is calculated from the ratio of the physical property distributions in the vicinity of the adhesive interface and the part away from the adhesive surface. The method for measuring the physical property distribution in the vicinity of the adhesive interface is disclosed.

JP, 2009-99132, A Japanese Patent Laid-Open No. 2000-304666

Usually, in order to understand the elastic properties of the adhesive, it is necessary to perform a tensile test or compression test of the adhesive alone.

　To obtain the elastic properties of the adhesive, it is necessary to test the adhesive alone. For that purpose, it is necessary to manufacture a test piece composed of an adhesive alone, but for that purpose, the adhesive must be molded into an arbitrary shape. There is a problem that it is difficult to manufacture such a test piece in the first place with an adhesive having a strong tackiness.

In addition, an adhesive that is actually in an adhered state with the adherend may exhibit different elastic characteristics from the single body in the adhered state due to the chemical interaction with the adherend and the effect of residual stress. ..

In the SLJ test described in Non-Patent Document 1, it is possible to use a general-purpose tensile tester because the test piece is relatively easy to manufacture.

The main function of the adhesive is to maintain the bond. Therefore, also in the selection and development of the adhesive, the adhesive strength test represented by the SLJ test is usually preferentially performed.

However, the mechanical properties obtained by the SLJ test are only the nominal shear adhesive strength obtained by dividing the fracture load value of the adhesive joint by the adhesive area, and a separate test is required to obtain the elastic properties. There is.

Since the method described in Patent Document 1 uses the finite element method, the calculation load is large and the calculation time is long. Therefore, it is difficult to adopt from the viewpoint of improving the efficiency and speed of the adhesive development process.

The method described in Patent Document 2 measures the physical property distribution of the rubber in the vicinity of the adhesive interface of the rubber-like elastic composite, not the physical property distribution of the adhesive body.

The present invention aims to estimate the elastic properties of an adhesive in an adhesive state by using data of only a general adhesive strength test.

The present invention includes a plurality of means for solving the above problems, but if one example is given, it is a method for identifying elastic properties of an adhesive in a state where adherends are adhered to each other, and an adhesive strength test is performed. It is carried out, the strain distribution of the adhesive in the adhesive strength test is measured, and the elastic characteristic of the adhesive is calculated using the strain distribution.

According to the present invention, it is possible to estimate the elastic property of the adhesive in the adhesive state by using the data of only the general adhesive strength test.

It is a schematic perspective view which shows the test piece of the single lap joint test which is an example of an adhesive strength test. It is a flowchart which shows the identification method of the elastic characteristic of an Example. It is a graph which shows the shear strain distribution of the adhesive agent in the single lap joint test obtained by the finite element analysis. It is a graph which shows the vertical strain distribution of the adhesive agent in the single lap joint test obtained by the finite element analysis. It is a graph which shows an example of the regression result of the strain distribution model with respect to the strain distribution measured value in an Example.

The present invention is a method for identifying elastic properties of an adhesive in a state where adherends are adhered to each other, performing an adhesive strength test, measuring a strain distribution of the adhesive during the adhesive strength test, and a strain distribution thereof. Is used to calculate the elastic characteristics such as the elastic coefficient of the adhesive.

Note that the present invention can be applied not only to an adhesive but also to a load test on a member having an adhesive portion to measure strain distribution. That is, the present invention is not limited to the identification of the elastic properties of the adhesive and includes the following inventions.

A method of identifying elastic properties of a member having an adhesive portion, in which a load test (strength test) such as a tensile test or a compression test is performed on the member, and the strain distribution generated in the member at that time is measured. Then, the strain distribution is used to calculate the elastic characteristic of the member, which is an elastic characteristic identifying method.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a schematic perspective view showing a test piece of an SLJ test (single lap joint test) which is an example of an adhesive strength test.

As shown in this figure, the SLJ test piece has a structure in which two plate-like adherends 1 are partially overlapped and bonded with an adhesive. In this case, the length in the longitudinal direction of the adherend 1 at the bonded portion (bonding portion 2) is the wrap length L. L can be arbitrarily set and tested. W is the length (plate width) in the lateral direction (width direction).

In the SLJ test, both ends of the test piece are gripped by a tensile tester, and a tensile load F is applied to apply a shearing force to the adhesive portion 2.

Normally, when a fracture occurs, the load value at that time is divided by the area (L×W) of the adhesive portion 2 to obtain the nominal shear stress, and this is a test that is performed.

In the present invention, when the SLJ test, which is an adhesive strength test, is performed, the strain distribution 5 generated in the adhesive portion 2 is experimentally measured. The strain distribution 5 is a distribution of strain in the L direction on the center line 3 (shown by a chain line in the figure) on the front surface of the adhesive portion 2.

In this case, the measuring means used for measuring the strain distribution 5 is not limited at all, but for example, a digital image correlation method (DIC method) capable of measuring the spatial strain distribution of the measurement region based on the image data. ) Is preferred. The DIC method is a kind of image measurement.

Next, consider a strain distribution model that theoretically expresses the strain distribution 5 that occurs in the test piece in the state where the tensile load F is applied in the SLJ test.

Here, the strain distribution model is a function that represents the strain distribution 5 of the adhesive portion 2, and is represented by the following equation (1). The strain distribution model is also called a "model formula".

Where ε is strain, x is the position (coordinate) of the wrap length in the L direction, F is the tensile load, E _adhesive is the elastic modulus of the adhesive, E _adherent is the elastic modulus of the adherend, and ν _adhesive is the adhesive's elastic modulus. Poisson's ratio, ν _adherent is the Poisson's ratio of the adherend. Further, g is a parameter set including a lap length L that defines the geometrical shape of the SLJ test piece, a plate width W, a thickness of an adherend and an adhesive, and the like.

The strain distribution may be of two types, shear strain and vertical strain, but the present invention is not limited to either strain distribution.

As such a strain distribution model, for example, a finite element model can be considered, but when estimating the elastic properties, it is necessary to repeatedly calculate the strain distribution using the model while changing the parameters. (Details below). Therefore, it is difficult to recommend a finite element model that requires numerical calculation from the viewpoint of calculation load.

Therefore, when the present invention is actually applied, it is preferable to use a one-dimensional strain distribution model. For example, various stress distribution models of SLJ test pieces have been proposed, including the model described in Non-Patent Document 2 (GR theoretical model). Here, the stress distribution model has a form in which the left side of the above equation (1) is replaced with the stress distribution from the strain distribution. Therefore, the stress distribution model can be converted into a strain distribution model by a stress-strain constitutive law such as Hooke's law.

Since the purpose of the present invention is to estimate the elastic properties of the _adhesive , E _adhesive and ν _adhesive are unknown parameters in the strain distribution model. Other parameters such as tensile load, elastic properties of the adherend, dimensions of the SLJ test piece, etc. can be easily acquired or measured in normal cases.

On the other hand, the strain distribution, which is the left side of the above formula (1), can be expressed by using the strain distribution of the adhesive in the state where the tensile load acts on the SLJ test piece. This can be experimentally acquired by strain measurement by the above-mentioned DIC method or the like.

Therefore, when the above-mentioned unknown parameter is changed and the difference between the strain distribution represented by the strain distribution model and the strain distribution obtained experimentally becomes small, the unknown parameter becomes a value closer to the true value. Can be said. That is, the elastic characteristic of the adhesive, which is an unknown parameter, can be identified by regressing the strain distribution model on the strain distribution measurement value.

The strain distribution that occurs in the adhesive when the tensile load F is applied is not uniform. In the case of shear strain, it is known that the strain becomes large at both ends of the adhesion region as shown in strain distribution 5 in FIG.

The strain distribution of the adhesive along the lap length L direction is the shape of the SLJ test piece such as the tensile load value, the elastic characteristics of the adhesive or the adherend, the thickness of the adherend or the adhesive, and the lap length. It is determined by characteristics, physical properties, etc., and is represented by a curve. Therefore, the strain distribution model (model formula) is expressed as a non-linear function. Therefore, it is desirable to use an algorithm classified as a non-linear regression analysis for the regression to the strain distribution measurement value. The model formula is a function representing the relationship between the position in the tensile load application direction in the lap region of the test piece used for the SLJ test and the strain generated in the adhesive.

The present invention does not limit the algorithm at all, but it is preferable to use the Levenberg-Markt method (LM method) or the like used as the nonlinear regression analysis.

When using such a non-linear regression analysis, the strain distribution model is iteratively calculated while giving temporary elastic properties (elastic coefficient, Poisson's ratio, etc.) to the strain distribution model. At this time, if a finite element model having a large degree of freedom is used as the strain distribution model, the calculation load may increase explosively. Therefore, as described above, by using a model such as GR theory that does not involve numerical integration, it is possible to identify the elastic coefficient within a practically acceptable time range.

FIG. 2 is a flow chart showing the above-mentioned elastic property identification method.

In the figure, assuming that the adhesive strength test S110 is performed, the strain distribution measurement S120 of the adhesive is performed at that time to obtain the strain distribution measurement value. On the other hand, the test condition of the adhesive strength test S110 is input to the strain distribution model (S130), and the theoretical value of the strain distribution is obtained.

By performing regression analysis S140 so that this theoretical value approximates the measured value, the elastic characteristics of the adhesive such as the elastic coefficient and Poisson's ratio are finally obtained.

Hereinafter, in the above method, the case where the DIC method is used as the method of the strain distribution measurement S120 and the GR theory is used as the strain distribution model will be described in more detail.

Normally, the DIC method uses a visible light camera to measure the strain distribution on the surface of an object that has been randomly patterned. Therefore, in the case of the SLJ test piece, the strain distribution of the adhesive in the part exposed on the side surface of the test piece is measured.

On the other hand, part of the strain distribution model for the SLJ test including the GR theory does not consider the strain distribution in the width W direction of the SLJ test piece. Therefore, an error may become large between the strain distribution based on the strain distribution model and the strain distribution on the side surface of the test piece based on the DIC method.

FIG. 3 is a graph showing the shear strain distribution of the adhesive in the SLJ test calculated based on the finite element method. That is, it is the strain distribution in the L direction of FIG.

The horizontal axis is the position in the L direction with the central part of the adhesive portion 2 in FIG. 1 as the zero point, and the vertical axis is the shear strain at each position. The ◯ mark is the shear strain in the central portion of the adhesive portion 2 in FIG. 1 in the plate width direction (the direction orthogonal to the front surface (side surface portion) of the adhesive portion 2 ). The symbol Δ indicates the shear strain at the center line 3 in front of the adhesive portion 2. The broken line is the shear strain calculated based on the GR theory, and corresponds to the same position as the center of the circle in the plate width direction.

FIG. 4 also shows the result of calculating the vertical strain acting in the direction of peeling the adhesive (thickness direction) at each position in the L direction in the same manner as in FIG. The circles, triangles, and broken lines are values at the same positions as in FIG. 4 differs from FIG. 3 only in the direction of strain.

Below, explanation will be given by comparing FIG. 3 and FIG.

The shear strain distribution in Fig. 3 is smaller than the vertical strain distribution in Fig. 4 in most areas where the difference in strain in the strip width direction is small. In other words, the area where the ◯ mark, the Δ mark and the broken line are close to each other is wide.

In the vertical strain distribution of FIG. 4, the ◯ mark and the Δ mark have an overlapping region in the central portion in the L direction, but the difference between the ◯ mark and the Δ mark is large at both end portions in the L direction. .. In addition, comparing the ◯ mark, which is the vertical strain at the same position, with the broken line, it can be seen that there is an overlapping region at both ends in the L direction, but there is a clear difference at the center in the L direction.

Since the value measured by the DIC method is the strain distribution on the side surface of the plate width, when combining the DIC method and the strain distribution model that does not consider the plate width like GR theory, the triangle mark and the broken line are used. It will be compared. Therefore, it is possible to identify the elastic characteristic with higher accuracy by using the measured value of the shear strain distribution of FIG. 3 instead of the measured value of the vertical strain distribution of FIG.

Also, the shear strain distribution based on the finite element method shown in Fig. 3 is the strain distribution in the center of the adhesive in the thickness direction. As shown in the figure, in the GR theory (broken line), the end of the lap region has the maximum value. In the finite element method (marked with Δ), which is closer to the actual condition, the shear strain is extremely small at the ends. In other words, the curve connecting the triangles is convex upward near the ends.

This difference is due to the fact that the GR theory does not consider the distribution of stress in the plate thickness direction due to the shape effect at the end of the lap region. Therefore, it is desirable to intentionally exclude the strain measurement value at the end of the lap region to identify the elastic characteristic. That is, by setting the exclusion range of the strain measurement value, the identification error of the elastic characteristic can be reduced.

According to the study by the present inventor, elastic characteristics are identified with high accuracy by excluding 1% or more and 8% or less of the lap region total length from both ends with respect to the lap region total length from the data of shear strain distribution (measured value). I know I can. That is, it is desirable that the exclusion range of the strain measurement value be 1% or more and 8% or less. The exclusion range of the strain measurement value is more preferably 3% or more and 7% or less, and particularly preferably 4% or more and 6% or less.

In this example, the error with respect to the measured value of the elastic coefficient is 10% or less when the exclusion range of the strain measurement value is set to 1% or more and 8% or less. Further, by setting the exclusion range of the strain measurement value to 5%, the elastic characteristic could be identified with the highest accuracy.

FIG. 5 is a graph in which the above conditions are actually applied, and shows an example of the regression result of the strain distribution model with respect to the strain distribution measured value. In the figure, ● indicates the measured value of the shear strain of the adhesive in the SLJ test, and the solid curve is the regression curve based on the GR theory.
The excluded area (excluded range of the strain measurement value) is 5% of the entire lap area from both ends.

As shown in this figure, in the applied area, the regression curve is shaped so as to pass through the central part of the band of variation in measured values at each position. From this, it can be seen that the curve regression based on the GR theory is desirable in the applied region. In this case, the accuracy of estimating the elastic modulus of the adhesive was within 5% due to the error with respect to the measured value of the elastic modulus.

1: Adherend, 2: Adhesive part, 3: Center line, 5: Strain distribution.

Claims

A method for identifying elastic properties of an adhesive in a state where adherends are adhered to each other,
Conduct an adhesive strength test,
Measuring the strain distribution of the adhesive in the adhesive strength test,
A method for identifying an elastic property of an adhesive, wherein the elastic property of the adhesive is calculated using the strain distribution.
In the calculation of the elastic characteristics, using a model formula representing the strain distribution in a state in which a load is applied to the adhesive,
The method for identifying elastic properties of an adhesive according to claim 1, wherein the calculation of the elastic properties is performed by regression of the model formula with respect to a strain distribution measurement value obtained by the measurement.
The method for identifying elastic properties of an adhesive according to claim 2, wherein the adhesive strength test is a single lap joint test.
The elasticity of the adhesive according to claim 3, wherein the model formula is a function representing the relationship between the position in the tensile load loading direction in the lap region of the test piece used for the single-lap joint test and the strain generated in the adhesive. How to identify characteristics.
The method for identifying elastic properties of an adhesive according to claim 4, wherein the strain is shear strain.
The method for identifying the elastic property of the adhesive according to any one of claims 1 to 5, wherein the measurement is image measurement.
The method for identifying elastic properties of an adhesive according to claim 2, wherein only a part of the strain distribution measurement values is used for the calculation of the elastic properties.
The method for identifying elastic properties of an adhesive according to claim 2, wherein the regression is a non-linear regression analysis.
The method for identifying an elastic property of an adhesive according to claim 7, wherein the strain distribution measurement value is used for the calculation of the elastic property by excluding a range of a predetermined length from both ends of the measurement region.
The method for identifying elastic properties of an adhesive according to claim 9, wherein the range is 1% or more and 8% or less of the total length of the region.
The method for identifying an elastic property of an adhesive according to claim 10, wherein the elastic property is an elastic coefficient or a Poisson's ratio.
The method for identifying elastic characteristics of an adhesive according to claim 11, wherein an error of the elastic coefficient with respect to an actually measured value is 10% or less.