JP3301247B2 - Method and apparatus for estimating true density of material, and method for estimating cavity amount of material - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for estimating a true density of a material, and a method and an apparatus for estimating a true density from a chemical component of a material. Further, the present invention relates to a method for estimating the amount of cavities in a material using the estimated true density.

[0002]

2. Description of the Related Art Heretofore, there have been almost no specific reports on methods for estimating the true density of a material whose components change due to differences in various conditions at the time of manufacture. inside that,
One, "Quantitative evaluation of shrinkage cavitation tendency of various aluminum alloys", Chen Yoshinori, Doctoral Dissertation, Graduate School of Engineering, Tohoku University,
PP 38-43, (1990) has been reported. In this paper,
A wide range of 13 alloys (ADC1, ADC3, ADC
4, ADC6, ADC10, ADC12, AC2A, A
C2B, AC4A, AC4B, AC4C, AC7A, 3
For 90R), a chemical regression analysis is performed on the chemical composition value and the true density data of each of the samples to obtain an equation for estimating the true density from the chemical composition.

[0003]

However, the method for estimating the true density disclosed in the above-mentioned paper merely obtains a regression coefficient by simply performing multiple regression analysis on all the obtained data. However, with this method, the accuracy of estimating the true density is poor. In other words, only an attempt was made to formulate an estimation formula from a mere calculation value of a regression coefficient, no device for improving the estimation accuracy is disclosed, and no estimation error is sufficiently evaluated. Therefore, the present estimation method has a problem that the estimation accuracy and the reliability of the obtained predicted value are low and cannot be applied practically.

For example, in the field of metal materials, for example, in recent years, various methods for producing highly reliable aluminum alloy castings have been developed. For example, from the viewpoint of weight reduction of important safety parts of automobile parts, iron-based castings have been developed. → Replacement with aluminum alloy castings is under consideration. At this time, the shrinkage cavity is measured as one standard of the quality of the aluminum alloy casting. In this case, the “true density value” is used when obtaining the cavity amount of the material from the density, such as the shrinkage cavity amount in the casting. Even if the amount of cavities is about 1% by volume, the properties such as the strength of the material are considerably affected. Therefore, a more accurate true density value is desired for quantification. For example, 1
% Void volume ratio with accuracy within 10%, error
True density values within 0.1% are required. However, it is difficult for the above-mentioned conventional true density estimation method to secure such accuracy, and at present, it is not used.

[0005] Therefore, usually, the density value of the material is merely used as a reference value indicating the void volume. But,
Since the "density" varies depending on the chemical composition of the material, there is a problem that only a relative comparison can be made within the same lot. Therefore, in order to compare the absolute amount of cavities between materials, it is desired to accurately determine the true density value of the materials. However, in order to determine the true density, it was necessary to prepare a sound sample of the same composition and without cavities each time and measure the density, which was only performed in some limited research . That is, there is a problem that it is difficult to produce a true density sample having no defect, and this operation is very troublesome.

The inventors of the present invention have conducted intensive studies to solve the problems of the prior art as described above, and as a result of repeating various systematic experiments, the present invention has been accomplished.

(Object of the Invention) It is an object of the present invention to provide a method and an apparatus for estimating the true density of a material which can easily and accurately obtain the true density of the material from chemical components. Another object of the present invention is to provide a method for easily and accurately estimating the amount of cavities of a material using the estimated true density.

The present inventors have focused on the following with respect to the above-mentioned problems of the prior art. That is, the density of the material is a value inherent to the material, and is theoretically determined from the composition of the material, the densities of the respective constituent elements constituting the structure, and the composition ratio thereof according to the compound rule. However, practical materials have a complex structure, and the constituents of the structure and the proportions of the structure vary in various ways depending on the manner of coagulation and trace amounts of impurity components.
In addition, since the density of each component is often unknown, such a method can be applied only to a limited material composed of only elements having a clear structure and a known density.

On the other hand, even when an element having a high density forms a compound with another element, it is considered that the density of the compound also tends to increase because the atomic structure of the element does not change.
Therefore, the inventors focused on the fact that the approximate true density of the material can be estimated only from the composition ratio of the component elements without considering the complexity of the structure. However, in order to estimate the density of the material with high accuracy, it is necessary to consider the characteristics of the tissue. However, it is almost impossible to estimate the structure of a complex alloy system from the chemical composition alone and further estimate the density of each component of the structure. Therefore, information having an organizational feature is included in the measured value of the true density of the material, and by analyzing the data consisting of the chemical components and the measured true density values corresponding to the components in more detail, The present invention has been made by focusing on the fact that the true density can be estimated with high accuracy in consideration of the structural characteristics of the target material.

[0010]

[Means for Solving the Problems]

(Constitution of the First Invention) The method for estimating the true density of a material according to the present invention comprises:
Preparing a plurality of healthy standard samples containing at least two or more types of chemical components constituting the material to be estimated, and having a cavity volume ratio containing all of the chemical components in any one thereof is equal to or less than a predetermined value, Based on the information obtained in the standard sample data creation step of obtaining the true density actual measurement value information and the chemical composition information of the sample, based on the information obtained in the standard sample data creation step, contribute to the estimation of the true density from the chemical components constituting the target material to be estimated. A true density estimation information generating step of obtaining information for estimating the true density of a material from the selected chemical component; a true density estimation information obtained from the true density estimation information creating step; A true density estimating step of estimating the true density of the target material based on the chemical component information of the target material corresponding to the density estimation information and the composition information of the chemical component.

(Constitution of the Second Invention) The apparatus for estimating the true density of a material according to the present invention comprises: a knowledge base having information on actual measured values of the true density of a material to be estimated and information on chemical composition corresponding to the actual measured values;
A standard chemical component extracting means for extracting a standard chemical component of the material to be estimated; and a knowledge volume based on at least one of the standard chemical components extracted by the standard chemical component extracting means. Based on the information obtained by the actual data of the true density of a plurality of healthy materials, and standard data selecting means for selecting chemical composition information corresponding to the actual measured values to obtain standard data, and the standard data selecting means. A true density estimation information creating means for creating information for estimating a chemical component contributing to the estimation of the true density of the material to be estimated from the standard chemical component and the true density of the material to be estimated, and the true density estimation information creating means Based on the obtained true density estimation information, the chemical component information of the target object material corresponding to the true density estimation information, and the composition information of the chemical component, the density of the target object material is calculated. A true density estimating means for estimating a, characterized in that it consists of.

(Constitution of the Third Invention) The method for estimating the amount of cavities in a material according to the present invention comprises a cavity containing at least two or more types of chemical components constituting the material to be estimated, and any one of which contains all of the chemical components. A plurality of healthy standard samples whose volume ratios are equal to or less than a predetermined value are prepared, and a standard sample data generating step of obtaining true density measured value information and chemical composition information of the standard sample, and a standard sample data generating step A true density estimation information generating step of selecting a chemical component that contributes to the estimation of the true density from the chemical components constituting the material to be estimated based on the information, and obtaining information for estimating the true density of the material from the selected chemical component; Based on the true density estimation information obtained from the true density estimation information creating step, the chemical component information of the material to be estimated corresponding to the true density estimation information, and the composition information of the chemical component. Intimate A true density estimating step of estimating, and a cavity amount estimating step of estimating a cavity amount of the estimated material from the estimated true density obtained in the true density estimating step and the density of the estimated material. I do.

[0013]

The mechanism by which the method and apparatus for estimating the true density of the present invention and the method for estimating the amount of cavities of a material can obtain a high-accuracy true density or amount of cavities easily is not necessarily clear. Think like.

(Operation of the First Invention) The method for estimating the true density of a material according to the present invention is as follows.
A plurality of sound standard samples are prepared which contain at least two or more chemical components constituting the material to be estimated, and which have a cavity volume ratio that is below a predetermined value and contain any of the chemical components. Next, the true density and the chemical composition of the prepared standard sample are measured to obtain the true density measured value information and the chemical composition information of the standard sample. Thereby, standard sample data necessary for estimating the true density of the material can be obtained.

Next, in the true density estimation information creating step,
Based on the information obtained in the standard sample data creation step, a chemical component that contributes to the estimation of the true density is selected from the chemical components constituting the target material. Next, information for estimating the true density of the material is obtained from the selected chemical components. This makes it possible to obtain effective information for estimating the true density of the material.

Next, in the true density estimating step, the true density estimating information obtained in the true density estimating information creating step, the chemical component information of the material to be estimated corresponding to the true density estimating information, and the chemical component The true density of the material to be estimated is estimated based on the composition information. Thereby, it is considered that the true density of the material can be accurately estimated. From the above, it is considered that the true density of a material can be easily and accurately obtained from a chemical component by the method for estimating the true density of a material of the present invention.

(Function of the Second Invention) In the apparatus for estimating the true density of a material according to the present invention, first, the knowledge base has actual value information of the actual density of the material to be estimated and chemical composition information corresponding to the actual value. . Further, the standard chemical component extracting means extracts the standard chemical component of the material to be estimated.

Next, in a standard data selecting means, the measured values of the true densities of a plurality of healthy materials having a cavity volume ratio of not more than a predetermined value, including at least the extracted standard chemical components in any one of the knowledge bases. The information and the chemical composition information corresponding to the measured value are selected. Thereby, standard sample data necessary for estimating the true density of the material can be obtained.

Next, in the true density estimation information creating means,
Based on the information obtained by the standard data selection means,
From the standard chemical components, information for estimating the chemical components contributing to the estimation of the true density of the material to be estimated and the true density of the material to be estimated are created. Thereby, the true density of the material can be accurately estimated.

Next, the true density estimating means, the true density estimating information obtained by the true density estimating information creating means, the chemical component information of the material to be estimated corresponding to the true density estimating information, and the chemical component information of the chemical component. The true density of the material to be estimated is estimated based on the composition information. Thereby, it is considered that the true density of the material can be accurately estimated. As described above, it is considered that the true density of a material can be easily and accurately obtained from a chemical component by the true density estimating apparatus of the present invention.

(Operation of the Third Invention) The method for estimating the amount of voids in a material according to the present invention first estimates the true density of the material. That is, in the standard sample data creation step, at least two or more types of chemical components constituting the material to be estimated are included, and
A plurality of sound standard samples whose cavity volume ratios containing all of the chemical components are equal to or less than a predetermined value are prepared, and information on actual measured true density and information on chemical compositions of the standard samples are obtained. Next, in the true density estimation information creating step, a chemical component contributing to the estimation of the true density is selected from the chemical components constituting the material to be estimated based on the information obtained in the standard sample data creating step, and the selected Obtain information for estimating the true density of a material from chemical components. Next, in the true density estimation step, the true density estimation information obtained in the true density estimation information creating step, the chemical component information of the material to be estimated corresponding to the true density estimation information, and the composition information of the chemical component are obtained. Based on this, the true density of the target material is estimated. Thereby, the true density of the material can be accurately estimated.

Next, in the cavity amount estimation step, the cavity amount of the material to be estimated is estimated from the estimated true density obtained in the true density estimation step and the apparent density of the material to be estimated. Thus, it is considered that the estimated true density is a value almost coincident with the actual true density, so that the cavity amount of the material can be accurately estimated. It is considered that the cavity amount estimation method described above can easily and accurately estimate the cavity amount of a material using the estimated true density.

[0023]

【The invention's effect】

(Effect of the First Invention) The true density of a material can be easily and accurately obtained from chemical components by the method for estimating the true density of a material according to the present invention.

(Effect of the Second Invention) The true density of a material according to the present invention can easily and accurately obtain the true density of a material from chemical components.

(Effect of the Third Invention) The method for estimating the amount of voids in a material according to the present invention makes it possible to easily and accurately estimate the amount of voids in a material.

[0026]

【Example】

(Specific Description of the Invention) Hereinafter, the method and apparatus for estimating the true density and the method for estimating the amount of cavities of a material according to the present invention will be described with reference to more specific and limited inventions (other inventions / specific examples). explain.

(Material to be Estimated) The material to be estimated in the present invention can be applied to any material having a chemical component such as a metal material, an inorganic material, and an organic material. Among them, it is preferable that the chemical composition and the structure of the material have a close relationship because the true density of the material can be accurately estimated.

When the material to be estimated is a metal material, the physical properties of atoms are easily reflected, and it is considered that there is a good correlation between the density of the component elements and the density of the alloy. Can be accurately estimated. Further, when the material to be estimated is aluminum or an aluminum alloy material, since the main component, aluminum, is a light element, the effect of the relatively heavy additive element tends to appear clearly. The true density can be accurately estimated from the quantity. Practically, in aluminum alloy castings in which Al-Si alloys are mainly used, cavities called shrinkage cavities pose a problem. Therefore, it is very important to know the true density and the amount of cavities. It is suitable as.

[Method of Estimating True Density] In the method of estimating true density of a material according to the present invention, in the standard sample data creation step, the chemical composition is within the range of the chemical composition of the material to be estimated and It is a step of preparing a plurality of sound standard samples including all chemical components constituting the material and having a cavity volume ratio of a predetermined value or less, and obtaining true density measured value information and chemical composition information of the standard samples. preferable. Thus, the standard sample data created in the standard sample data creating step becomes more appropriate data, and the accuracy of estimating the true density can be improved.

The method for estimating the true density of a material according to the present invention is a method for extracting a standard chemical component of a material to be estimated.
And extent, either to include standard chemical components that are at least the extraction, the true density measured value information of a given value or less healthy plurality of materials void volume ratio, and chemical composition information corresponding to the measured value Based on the information obtained in the obtained standard data creating step and the standard data selecting step, the chemical components contributing to the estimation of the true density of the material to be estimated and the true density of the material to be estimated are estimated from the standard chemical components. Create information
The true density estimation information creating step , the true density estimation information obtained from the true density estimation information creating step, and the chemical component information of the target object material and the composition information of the chemical component corresponding to the true density estimation information. A true density estimating step of estimating the true density of the material to be estimated.

(Standard Chemical Component Extraction Step) In the standard chemical component extraction step of the present invention, standard chemical components of the material to be estimated are extracted. At this time, it is preferable that the standard chemical components extracted in the standard chemical component extraction step include at least the main chemical components of the material to be estimated. This allows
Since the main components constituting the material can be almost completely selected, the estimation accuracy can be improved. For example, when the estimation target material is an aluminum alloy of a metal material, eleven main elements (Si, Cu, Mg, Zn, Fe, M) contained in the aluminum alloy as standard chemical components are used.
n, Ti, Cr, Ni, Sn, Pb). Further, it is preferable that the standard chemical component extracted in the standard chemical component extraction step includes at least a chemical component contained in the material to be estimated at a predetermined value or more. As a result, it is possible to select a chemical component having a relatively large contribution to the true density of the material, so that the estimation accuracy can be improved.

(Standard Data Creation Step) In the standard data creation step of the present invention, the true densities of a plurality of healthy materials having at least the above-mentioned extracted standard chemical components and having a void volume fraction of a predetermined value or less are measured. Value information and chemical composition information corresponding to the actually measured value are created. According to the estimation method of the present invention, it is considered that the information having the organizational characteristics of the material is included in the true density of the material. It is desirable to select a sample in advance. That is, it is desirable that the number of data is as large as possible. Further, it is preferable to select data having a chemical composition in which the contained elements and the contents are appropriately dispersed within the applicable composition range.

If the information included in the standard data is insufficient, a result that the estimation accuracy is poor is obtained in the true density estimation information creating step. In this case, the result is fed back. By increasing the number of data or limiting the range of application composition, standard data is re-created so that the structural characteristics of the material to be estimated can be sufficiently expressed, and analyzed again to improve the estimation accuracy. Can be improved.

In the method for estimating the true density of a material according to the present invention, the standard chemical component extracting step is a step of extracting a standard component within a chemical composition range of the material to be estimated, and the standard data creating step is: Preparing a plurality of healthy standard samples having a cavity volume ratio equal to or less than a predetermined value, including all the chemical components constituting the material to be estimated, and measuring the true density measured value information and the chemical composition information of the standard sample. Preferably, it is a step of obtaining. Thereby, the standard components extracted in the standard chemical component extraction step and the standard data created in the standard data creation step can be made more appropriate, and the accuracy of estimating the true density can be improved.

(True density estimation information creation step) In the true density estimation information creation step of the present invention, the true density of the target material is estimated from the standard chemical components based on the information obtained in the standard data selection step. Information for estimating the contributing chemical components and the true density of the material to be estimated is created.

The true density estimation information creating step according to the present invention includes a statistical value representing the contribution of each chemical component such as a standard partial regression coefficient to the true density measured value, and a statistical value between each chemical component and the true density measured value. Based on the characteristics of the distribution of the data shown by a graph such as a partial regression plot, so that the statistical value indicating the goodness of fitting of the expression represented by the contribution ratio of the estimation expression, etc. is optimal,
It is preferable to select a chemical component that contributes to the estimation of the true density and determine a function form of the estimation formula. This makes it possible to obtain true density estimation information that reflects the structural characteristics of the material included in the standard data to the maximum.

A preferred configuration of the true density estimation information creating step of the present invention is a true density estimation information model for creating a true density estimation information model including a predetermined functional form having the composition information of the standard chemical component as an element.
A density estimation information model creation step, and a tentative true density estimation information system for creating a tentative true density estimation information system based on the true density estimation information model and the standard data obtained in the standard data creation step.
Based on the constant information system creation process and the provisional true density estimation information system,
A chemical component selecting step of selecting a chemical component required for estimating the true density from the standard chemical component and information indicating the relationship between the true density measured value and / or the estimated true density; True density estimation information to create a true density estimation information system based on the value and / or information indicating the relationship with the estimated true density.
Report creation process .

((True Density Estimation Information Model Creation Step)) In the true density estimation information model creation step, a true density estimation information model including a predetermined functional form having the composition information of the standard chemical component as an element is created. This step is, specifically,
This is a model expression of a true density estimation expression as exemplified in Expression 1.

The model formula is preferably expressed in the form of a sum of terms indicating the influence of the amount of each chemical component so that statistical analysis such as multiple regression analysis can be easily performed. Also,
It is preferable to make the function form arbitrary so that the structural characteristics of the material and the weighting of the influence of each component element can be expressed. Although an example of the model formula is shown in Formula 1, the formula is not limited to the model formula, and any formula that can perform statistical analysis such as multiple regression analysis and can weight the influence of each constituent element on the structural characteristics of the material and each component element can be used. , Any model formula may be used. It is preferable that the model formula is a model formula that can make use of features such as the structural characteristics of the material to be estimated and the degree of relation to the chemical composition.

((Provisional True Density Estimation Information System Creation Step)) In the provisional true density estimation information system creation step, a temporary true density estimation information system is created based on the provisional true density estimation information model and the standard data obtained in the standard data creation step. Create a density estimation information system. This temporary true density estimation information system preferably includes at least information indicating how much each chemical component contributes to the true density estimation.

Examples of the index indicating how much each chemical component contributes to the estimation of the true density include a standard partial regression coefficient and a standard error. For example, a chemical component having a large standard partial regression coefficient is determined to have a large contribution to the estimation of the true density. On the other hand, a chemical component having a large standard error is determined to have a large contribution to the true density estimation error, that is, a small contribution to the true density estimation.

((Chemical Component Selection Step)) In the present chemical component selection step, information indicating the relationship between the standard chemical component and the true density actual measurement value and / or the estimated true density based on the provisional true density estimation information system. From, the chemical components required for estimating the true density are selected.

A preferred configuration of the chemical component selection step of the present invention is to select and remove a chemical component that reduces the estimation accuracy from information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density. And a step of selecting and optimizing a chemical component necessary for estimating the true density from information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density. It is preferable to select the chemical components necessary for estimating the true density by appropriately repeating the above steps. As a result, it is possible to select a chemical component necessary and sufficient for estimating the true density, and to select only a chemical component that effectively contributes to the estimation of the true density.

Regarding the selection and removal of chemical components, a component that does not contribute to the estimation of the true density is selected and removed, or a component that contributes to the estimation of the true density is selected and left, and other components are removed. Do. Examples of the selection and removal of chemical components include a method of selecting a chemical component having a large standard partial regression coefficient and a small standard error, and removing a chemical component having a small standard partial regression coefficient and a large standard error.

As for the optimization of chemical components, it is preferable to repeat the selection and removal of chemical components so as to obtain the best fit of the model. Examples of indices for determining the fit of the model include a variance ratio F ₀ , a contribution ratio R ² , and Akaike's information criterion AIC. Where F ₀ and R
^It is determined that the larger the value of ^2, the better the fit of the model, and the smaller the AIC, the better the fit of the model.

Another preferred configuration of the chemical component selecting step of the present invention may include a standard chemical component changing step of checking the validity of the standard chemical components and changing the standard chemical components as necessary. preferable. Thus, for example, even when the number of standard data is equal to or less than the number of chemical components included in the estimation target material, it is possible to accurately select a chemical component that contributes to the estimation of the true density. In other words, when the number of standard data is small as described above, it is not possible to select all of the chemical components contained in the material to be estimated as the standard chemical components, and thus it is not possible to select a chemical component having a large contribution in one analysis. there is a possibility. By providing a standard chemical component change step, problems in such a case can be eliminated,
The accuracy of estimating the true density can be improved.

Another preferred configuration of the chemical component selection step of the present invention is to remove abnormal data from the standard data based on the standard chemical component and information indicating the relationship between the true density measured value and / or the estimated true density. It is preferable to include a step of removing abnormal data. In the case where the number of standard data is small and individual data greatly affects the accuracy of the estimation formula, it is particularly preferable to remove abnormal data in this step. As a result, it is possible to prevent the accuracy of the estimation formula from being deteriorated due to the influence of the abnormal data, and the accuracy of the true density estimation is further increased.

The specific criteria for selecting the removal data are as follows. Regardless of the selection / removal of the chemical component, when the standard residual is always abnormally large (significance level 5%), it is determined to be abnormal data. By removing some data, F ₀ , R ² , AI
If the fit of the model indicated by C or the like is clearly improved, the data is determined to be abnormal data. In the partial regression plot, data that clearly departs from other data trends is determined to be abnormal data. Predicted value-residual plot is far apart, and standard residual and student residual are abnormally large (significance level 5%
The data shall be abnormal data. It is preferable that the abnormal data be removed only once.

((True density estimation information system creation step)) In the true density estimation information system creation step, the true density estimation information system creation step is performed based on the information indicating the relationship between the selected chemical component and the true density actual measurement value and / or the estimated true density. Create a density estimation information system.

In the true density estimation information system creation step, the relationship between the selected chemical component and the true density actual measurement value and / or the estimated true density is examined in more detail, and a higher level is taken into account in view of the structural characteristics of the material. This is a step of estimating the true density with high accuracy.

A preferred configuration of the true density estimation information system creating step is based on the selected chemical component and information indicating the relationship between the true density measured value and / or the estimated true density, and uses the composition information of the selected chemical component as an element. A function form determining step of determining a predetermined function form; information indicating a relationship between a selected chemical component including the determined function form and a true density actual measurement value and / or an estimated true density;
A true density estimation information system determining step of determining a true density estimation information system based on information obtained by removing abnormal data from the standard data.

The determination of the functional form is preferably carried out in consideration of how the selected chemical component affects the true density. Specifically, there is a method of determining an optimal function form from the relationship between each chemical component amount and the true density measured value and / or the estimated true density by performing a partial regression plot and removing the influence of other chemical components. No. The index for determining whether or not it is optimal includes F ₀ , R ² , AIC, and the like described above.

As a specific example, each of the optimized explanatory variables x
Partial regression plots the relationship between _i and the objective variable ρ ₀ , and the variance ratio F
_The function form f _i (x _i ) is determined so that ₀ and the contribution ratio R ² are maximum and the AIC value is minimum. As a result of examining the function form based on the partial regression plot, if, for example, a tendency that the estimated value decreases in a region where the amount of chemical components is large is recognized as a whole, the function form is set as a predetermined exponential function.
Thereby, the fitting property of the equation can be improved.

In the true density estimation information system determination step, when the statistical value indicating the difference between the predicted value based on the estimation formula such as the standard residual and the student residual and the actually measured value is equal to or larger than a predetermined value, the standard data corresponding to this is used. It is preferable that the true density estimation information system is determined based on the standard data reconstructed by removing the abnormal data as abnormal data. As a result, it is possible to prevent the accuracy of the estimation formula from deteriorating due to the influence of a few abnormal data due to measurement errors and the like.

(True Density Estimation Step) In the true density estimation step, the true density estimation information obtained in the true density estimation information creating step, the chemical component information of the material to be estimated corresponding to the true density estimation information and The true density of the material to be estimated is estimated based on the composition information of the chemical component.

[Method of Estimating Cavity of Material] The method of estimating the cavity of a material according to the present invention estimates the true density of the material to be estimated by the above method of estimating the true density of the material of the present invention, and performs the above-mentioned cavity amount estimating step. The cavity amount of the estimated material is estimated from the estimated true density and the density of the estimated material.

The method of estimating the amount of cavities includes a method of obtaining the image from a cross-sectional structure photograph by image processing and a method of obtaining the apparent density from the apparent density of the sample. The latter method requires only measurement of the density. The amount of cavities can be easily estimated. Further, in the present invention, since the true density of a material can be estimated with high accuracy and simply from only the chemical composition, it is possible to easily and accurately obtain the cavity amount from the estimated true density and the apparent density of the sample. Further, when a structural change occurs due to a heat treatment or the like, the amount of cavities can be estimated with higher accuracy by correcting a change in the true density due to the change.

As a specific example of the estimation of the cavity amount, an example of the estimation of the shrinkage cavity amount of the metal material will be briefly described. That is,
With respect to a plurality of castings having different shrinkage cavities, respective densities ρ are obtained by the Archimedes method, and using the true density value ρ ₀ obtained from the true density estimation formula, the shrinkage porosity P is calculated by the following equation.
Is calculated. P = [100 / [rho] - [100 / [rho _0] (cm ³ / 100g) Confirmation of the validity of shrinkage cavities amount obtained from this is, for example,
It can be performed by obtaining the shrinkage cavity amount by image processing based on the cross-sectional structure photographs of the plurality of castings.

In the method for estimating the true density of a material according to the present invention,
If the estimation accuracy is insufficient, it is preferable to perform the following method. That is, according to the present invention, an optimal true density estimation formula can be derived using the true density estimation information included in the standard data. However, if the accuracy of the obtained estimation formula is insufficient, the cause is considered to be that the true density estimation information included in the standard data is insufficient,
Increasing or optimizing the number of standard data,
The standard data is recreated by limiting or optimizing the application composition range, and the analysis is performed again, whereby a more accurate estimation formula can be obtained.

In the method for estimating the true density of a material according to the present invention,
In the case where the number of standard data is small, for example, abnormal data may affect the results of chemical component optimization or functional form optimization, resulting in insufficient estimation accuracy. In such a case, it is desirable to perform the optimization of the chemical component and the optimization of the function form again after removing the abnormal data. In addition, when the function form is significantly different from the initially assumed equation (for example, a linear equation), the estimation accuracy may be insufficient. In such a case, it is desirable to re-execute the optimization of chemical components and the removal of abnormal data after determining the function form.

The present invention can be carried out by appropriately changing the execution order of each step according to the situation.

In the present invention, some or all of the steps can be performed by arithmetic processing using a computer. Further, the obtained result can be displayed by the computer display means.

[Material true density estimation device]

(Knowledge Base) The knowledge base of the present invention has at least actual value information of the true density of the material to be estimated and chemical composition information corresponding to the actual value. The form of the information that this knowledge base has is a unique form that expresses the relationship between these numerical information even if the actual value of the true density of the material to be estimated and the information of the chemical composition corresponding to the actual value are provided as numerical information. Or any other form. The standard data selection means may take any form as long as the information can be extracted in a necessary form.

(Standard chemical component extracting means / standard data selecting means) In the apparatus for estimating the true density of a material according to the present invention, the standard chemical component extracting means extracts a standard component within the chemical composition range of the material to be estimated. And the standard data creating means prepares a plurality of sound standard samples having a cavity volume ratio of not more than a predetermined value, including all chemical components constituting any of the materials to be estimated, and It is preferably a means for obtaining true density measured value information and chemical composition information of the sample. Thus, the standard chemical components extracted by the standard chemical component extraction unit and the standard data created by the standard data selection unit can be made more appropriate, and the accuracy of true density estimation can be improved.

(True Density Estimation Information Creation Means) A preferred configuration of the true density estimation information creation means is a true density estimation information model for creating a true density estimation information model including a predetermined functional form having the composition information of the standard chemical component as an element. Estimating information model creating means ; temporary true density estimation information system creating means for creating a temporary true density estimation information system based on the true density estimation information model and the standard data obtained in the standard data creating step; A chemical component selecting means for selecting a chemical component necessary for estimating the true density from information indicating the relationship between the standard chemical component and the estimated true density based on the true density estimation information system; A true density estimation information system creating means for creating a true density estimation information system based on information indicating a relationship with the density . As a result, highly accurate true density estimation information reflecting the structural characteristics of the material included in the standard data can be obtained.

((Chemical Component Selection Means)) A preferred configuration of the chemical component selection means is a chemical composition for lowering the estimation accuracy from information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density. Means for selecting and removing components; and means for selecting and optimizing chemical components necessary for estimating the true density from information indicating the relationship between the standard chemical components and the true density measured value and / or the estimated true density. It is characterized in that the chemical components necessary for estimating the true density are selected by appropriately repeating these steps. As a result, it is possible to select a chemical component necessary and sufficient for estimating the true density, and to select only a chemical component that effectively contributes to the estimation of the true density.

Another preferred configuration of the chemical component selecting means is characterized in that it includes a standard chemical component changing means for checking the validity of the standard chemical components and changing the standard chemical components as necessary. . Thus, for example, even when the number of standard data is equal to or less than the number of chemical components included in the estimation target material, it is possible to accurately select a chemical component that contributes to the estimation of the true density. In other words, when the number of standard data is small as described above, it is not possible to select all of the chemical components contained in the material to be estimated as the standard chemical components, and thus it is not possible to select a chemical component having a large contribution in one analysis. there is a possibility. By providing the standard chemical component changing means, the problem in such a case can be eliminated, and the accuracy of estimating the true density can be improved.

Another preferred configuration of the chemical component selection means of the present invention is to remove abnormal data from the standard data based on the standard chemical component and information indicating the relationship between the true density measured value and / or the estimated true density. And an abnormal data removing means. In the case where the number of standard data is small and individual data greatly affects the accuracy of the estimation formula, it is necessary to remove abnormal data in this process.
Particularly preferred. As a result, it is possible to prevent the accuracy of the estimation formula from being deteriorated due to the influence of the abnormal data, and the accuracy of the true density estimation is further increased.

((True Density Estimation Information System Creation Means)) A preferred configuration of the true density estimation information system creation means is based on the selected chemical component and information indicating the relationship between the true density actual measurement value and / or the estimated true density. Function form determining means for determining a predetermined function form having the composition information of the selected chemical component as an element, information indicating a relationship between the selected chemical component including the determined function form and the estimated true density, and the standard data And a true density estimation information system determining means for determining a true density estimation information system based on information obtained by removing abnormal data from the information. Thus, by analyzing the relationship between the selected chemical component and the true density actual measurement value and / or the estimated true density in more detail, it is possible to perform highly accurate true density estimation in consideration of the structural characteristics of the material.

[0072]

Embodiments of the present invention will be described below.

(First Embodiment) This embodiment is an example in which the true density estimation target material is an Al-Si alloy as a metal material.

The true density estimating apparatus of this embodiment will be described with reference to FIG. FIG. 1 is a schematic explanatory view of an example of an apparatus suitable for estimating the true density of a material.

The true density estimating apparatus of the present invention uses the knowledge base 1
Standard chemical component extracting means 2 and standard data selecting means 3
And true density estimation information creating means 4 and true density estimation means 5. The true density estimation information creating unit 4 includes a true density estimation information model creating unit 41, a provisional true density estimation information system creating unit 42, a chemical component selecting unit 43, and a true density estimation information system creating unit 44. True density estimation information system creating means 44
Is composed of a function determining unit 441, a true density estimation information system determining unit 442, and a fit checking unit 443.

First, the knowledge base 1 is constructed. That is, a total of 78 types of samples were prepared in three groups in order to prepare standard sample data for use in estimating the true density of the Al-Si alloy. First, the first group is a compound alloy sample prepared by mixing pure aluminum and various mother alloys, and is shown in Table 1 (sample numbers: 1 to 30). Next, the second group is a commercially available Al-Si based alloy ingot sample such as ADC12, AC4C, AC2B, etc., which is shown in Table 2 (sample numbers: 31 to 61). The third group is Al-Si based alloy samples prepared by re-melting a practical casting, and is shown in Table 3 (sample numbers: 62 to 78). Tables 1 to 3 respectively
Table 2 shows the chemical analysis values of the Al-Si alloy sample. In addition,
The chemical composition was determined by the IC for the commercially available alloy ingot samples of the second group and the remelted samples of practical castings of the third group.
It was determined by P emission analysis and the like. For the compound alloys of the first group, the values calculated from the analysis values of the master alloy and the pure aluminum ingot were substituted. This calculated value has been empirically confirmed to be in good agreement with the actual chemical analysis value.
It should be noted that the chemical composition is the main one contained in the aluminum alloy.
One element (Si, Cu, Mg, Zn, Fe, Mn, Ti,
Cr, Ni, Sn, Pb).

Next, the true density of the sample was determined. That is, first, various alloys were melted, subjected to a deoxidation treatment by a flux and a vacuum degassing treatment, and then cast into a water-cooled copper mold only at a lower portion to prepare a sample having almost no shrinkage cavities. The density of this as-cast material was measured by the Archimedes method, and this was used as the actual measured value of the true density. The results are shown in Tables 1 to 5. These data (data shown in Tables 1 to 5) were constructed as a database in the knowledge base 1.

[0078]

[Table 1]

[0079]

[Table 2]

[0080]

[Table 3]

[0081]

[Table 4]

[0082]

[Table 5]

Next, in the standard chemical component extracting means 2,
The standard chemical components of the estimated material were extracted. In the present embodiment, eleven main elements (Si,
Cu, Mg, Zn, Fe, Mn, Ti, Cr, Ni, S
n, Pb) were extracted as standard chemical components.

Next, the standard data selecting means 3 outputs a plurality of sound volumes from the knowledge base 1 including at least one of the standard chemical components extracted by the standard chemical component extracting means and having a cavity volume ratio of a predetermined value or less. Information on the actual measured value of the true density of the material and the chemical composition information corresponding to the actual measured value were selected to obtain standard data for creating true density estimation information. In this embodiment, since the data corresponding to the standard data is selected as the data forming the knowledge base 1, the data forming the knowledge base 1 is used as the standard data.

Next, creation of true density estimation information (estimation of a true density estimation formula) was performed in the following steps. This flowchart is shown in FIG.

(First Step) First, the true density estimation information model creation means 41 creates a model equation of the true density estimation equation [P1: true density estimation information model creation step]. In this embodiment, the true density of a material having a similar texture is pure A
term represents the true density a ₀ for l, the effect of each element a _i f _i (x
_i ) was considered to be approximated by the sum, and Equation 1 was used as a model equation. Ρ ₀ : true density (target variable), x _i : component amount [% by weight] of each element (explanatory variable).

[0087]

(Equation 1)

(Second Step) Next, the provisional true density estimation information system creating means 42 assumes the function form of the true density estimation formula and assumes the true density estimation formula [P2: Creation of temporary true density estimation information system] Process]. In this embodiment, within a limited compositional range, the influence of each element amount is considered to be approximated by a linear equation, a function form, it was assumed that _{f i (x i) = x} i ··· (2) . Next, for the sample (n = 78),
An objective variable [rho _0, the relationship between the explanatory variable x _i, determined by multiple regression analysis. Thereby, a provisional true density estimation formula shown in FIG. 3 was obtained. The regression coefficient and the analysis of variance table are also shown in FIG.

Next, based on the temporary true density estimation formula, the chemical component selecting means 43 obtains information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density.
A chemical component necessary for estimating the true density is selected (chemical component selection step). In this example, the chemical components were selected by the following third and fourth steps.

(Third Step) First, in the result (FIG. 3) of the (second step), an explanatory variable having a small standard partial regression coefficient and a large standard error is selected and deleted [P3]. That is, first, titanium (Ti), chromium (Cr), nickel (Ni), tin (Sn), and lead (Pb)
Regarding the quantity, these explanatory variables were deleted because the standard partial regression coefficient was as small as -0.007 to +0.02 and the standard error was as large as one to more than 12 to 59.

(Fourth Step) Next, the above (third step) is repeated until the variance ratio F ₀ and the contribution ratio R ^{2 are} larger and the Akaike's information criterion (AIC) is smaller.
The type and number of explanatory variables are optimized [P4].

Here, AIC (Akaike's Information)
Criterion) is one of the criteria for the goodness of the model. In the case of a multiple regression model having p explanatory variables, AIC (p) = n (log _e 2π + 1) + n · log _e [S _e
(P) / n] +2 (p + 2). Note that S _e (p) is the residual sum of squares.
The smaller the AIC, the better. AIC minimization is asymptotically equivalent to Mallow's C _p , another measure of model goodness, when n is large enough, and they choose 2.0 as the limit in the regression coefficient test. Is known to asymptotically correspond to

Further, in the above (third step),
As an explanatory variable to be deleted, Mg was selected and removed because the standard partial regression coefficient was small and the standard error was large.
As a result, the adjusted degree of freedom R ² becomes 0.993, and the AI
The C value was almost the same as 381.7, and the dispersion ratio F ₀ became maximum at 2334, satisfying the condition of the fourth step. Note that Si, Cu, Zn, Fe, and Mn were not deleted because F ₀ and R ² decreased and AIC increased due to the removal.

(Fifth Step) Next, the function form determining means 4
According to 41, the function form of the true density equation is determined [P5: function determination step]. In the present embodiment, each optimized explanatory variable x _i
And the objective variable ρ ₀ are plotted in partial regression, and the variance ratio F ₀
And the function form f _i (x _i ) was determined such that the contribution ratio R ² was the maximum and the AIC value was the minimum.

As a result of examining the function form based on this partial regression plot, a tendency was observed that the estimated value became smaller as a whole in a region where the amount of Mn was large.
It was (wt%)] ^1/2. Thereby, as shown in FIG. 5, the partial regression plot was distributed substantially on a straight line, and the fitting property of the equation was improved. This is considered to be a result corresponding to the fact that Mn has an effect of changing the shape of the Fe-based compound, and the effect on the microstructure is slightly changed when the amount of Mn is small and large.

FIGS. 6 to 9 show partial regression plots for each explanatory variable. In each case, a linear relationship was observed between the explanatory variable and the predicted value, and it was confirmed that the function form was appropriate.

(Sixth Step) Next, the true density estimation information system determination means 442 removes abnormal data and determines a true density estimation formula [P6: true density estimation information system determination step].
In the present embodiment, data that is far apart in the predicted value-residual plot and in which the standard residual and the student residual are abnormally large (significant level 5%) are removed as abnormal data. That is, five sets of data indicated by ** in Tables 1 to 5 were removed, multiple regression analysis was performed again, and the true density estimation formula was determined. The obtained true density estimation formula, regression coefficient, and analysis of variance table are shown in FIG.

(Seventh Step) Next, the fitting property confirming means 443 checks the fitting property of the true density estimation formula [P7]. In the present example, a scatter diagram of a normal probability point plot and a predicted value was created for the residual, and the normality and the uniform variance of the residual were confirmed. That is, FIG. 10 shows a histogram of the residuals. From the figure, the normality of the residual was confirmed.
FIG. 11 shows an actual measurement value-prediction value plot. From the figure, it is understood that the measured value and the predicted value of the true density are in good agreement over the whole. FIG. 12 shows a predicted value-residual plot. The residuals were uniformly dispersed with respect to the predicted values, and the uniformity of the residuals was confirmed.

The true density estimation equation obtained as described above is shown in FIG.
Shown in Table 6 shows standard regression coefficients and confidence intervals of each explanatory variable.

[0100]

[Table 6]

(Test for Estimating Shrinkage Void of Metal Material) Using the obtained equation for estimating true density, a test for estimating shrinkage cavities of Al-Si based alloy castings was conducted.

First, the true density estimating means 5 estimated the true density value of the material to be estimated (true density estimation step).
The alloy structure of the material to be estimated is Al-6% Si-
3.5% Cu-0.3% Mg-0.25% Fe. As a result, a true density value ρ ₀ = 2.753 was obtained.

Next, the shrinkage nest amount was estimated. That is, for three types of castings having different shrinkage cavities, the respective densities ρ are obtained by the Archimedes method, and using the true density value ρ ₀ = 2.753 obtained from the true density estimation formula shown in FIG. And the shrinkage nest amount P was calculated. P = [100 / [rho] - [100 / [rho _0] (cm ³ / 100g) As a result, shrinkage cavities amount, respectively, 0.1 (cm ³ /100g),0.5(c
m ³ was estimated /100g),1.1(cm ³ / 100g) and.

Next, the shrinkage cavities were determined by image processing based on the cross-sectional structure photographs of the three types of castings. As a result, the shrinkage nest amount almost coincided with the estimated shrinkage nest amount.

Next, the true density of the Al-Si alloy casting used in the present embodiment was measured. That is, the Al-Si alloy casting used in the shrinkage cavity estimation test was melted, subjected to a deoxidation treatment with a flux and a vacuum degassing treatment, and then cast into a water-cooled copper mold only at the lower part. It was measured by an Archimedes method using an as-cast material having almost no nest, and an actual measured value of the true density was obtained. As a result, the true density measured value was 2.753 g / cm ³ , which was almost the same as the true density estimated value.

Comparative Example 1 For comparison, a sample similar to the above-mentioned example (sample number: 1)
Relative to 78), the true density estimation equation model formula as the equation (1), the functional form of the true density estimation equation assuming _{f i (x i) = x} i, the target for the sample (n = 78) determined by multiple regression analysis of the relationship between variables [rho ₀ explanatory variables x _i, thereby to obtain a true density estimation formula (3). FIG. 13 shows a histogram of the residual in this case. In this case, the number of explanatory variables is very large, and the equation is complicated.However, since many explanatory variables that do not contribute to the estimation are included, many of them have a large residual compared to the above-described embodiment. It can be seen that the accuracy of the true density estimation is low. In addition, it can be seen that the residual is not normally distributed, and is not a reliable estimation formula.

(Second Embodiment) In the present embodiment, the true density estimation target material is an Al-Si alloy, and the true density measuring device similar to that of the first embodiment is used. This is an example in which standard data for creating density estimation information includes all the standard chemical components extracted by the standard chemical component extracting means. The following description focuses on the differences from the first embodiment.

First, in the standard chemical component extraction means 2,
As standard chemical components of the material to be estimated, eleven main elements (Si, Cu, Mg, Zn,
Fe, Mn, Ti, Cr, Ni, Sn, Pb) were extracted.

Next, the data shown in Table 7 was selected by the standard data selection means 3 from the knowledge base 1 already constructed. The method of creating data extracted by the standard data selecting means 3 from the data constructed in the knowledge base 1 is as follows. That is, 1
First, the chemical compositions of the seven JISAC2B alloys were determined for the eleven main elements contained in the aluminum alloy by chemical analysis such as ICP emission analysis. These alloys were melted, subjected to flux deoxidation and vacuum degassing, and then cast into a water-cooled copper mold only at the bottom. The measured value was used as the actual measured value of the true density.

[0110]

[Table 7]

Next, creation of true density estimation information (estimation of true density estimation formula) was performed in the same manner as in the first embodiment. The obtained true density estimation formula and regression coefficient are shown in FIG.
Shown in

In this embodiment, the standard regression coefficient is small, and the standard error is large, such as Mg, Ti, Ni, Cr,
By removing each element of Sn and Pb, the applicability of the model determined by the degree of freedom corrected contribution, the AIC value, and the like was improved. Further, the data of sample numbers 80 and 95 in Table 5 were judged and deleted as abnormal data. Further, in the same manner as in the first embodiment, the normality and equal variance of the residual and the appropriateness of the function form were confirmed.

Comparative Example 2 For comparison, a simple multiple regression analysis was performed on the same samples as those in the second example (sample numbers: 79 to 95). FIG. 15 shows the obtained true density estimation formula for comparison, the regression coefficient, and the like. As a result, in the case of this example, it was confirmed that the regression equation had a large degree of freedom-corrected contribution and a small estimation error despite the fact that the number of explanatory variables was small and the equation was simple compared to the comparative example. Was done.

(Third Embodiment) In this embodiment, the target material of the true density estimation is an Al-Si alloy of JIS ADC12, and the same true density measuring device as that of the first embodiment is used. This is an example in which the standard data for generating the true density estimation information includes all the standard chemical components extracted by the standard chemical component extracting means. The following description focuses on the differences from the first embodiment.

The true density estimating apparatus of this embodiment will be described with reference to FIG. FIG. 16 is a schematic explanatory view showing another example of a device suitable for estimating the true density of a material.

The true density estimating apparatus of the present invention uses the knowledge base 1
Standard chemical component extracting means 2 and standard data selecting means 3
And true density estimation information creating means 4 and true density estimation means 5. The true density estimation information creation unit 4 includes a true density estimation information model creation unit 41, a provisional true density estimation information system creation unit 42, a chemical component selection unit 43, a true density estimation information system creation unit 44, and a standard chemical component. And changing means 45.
The true density estimation information system creating means 44 includes a function form determining means 44
1; a true density estimation information system determining unit 442;

First, in the standard chemical component extraction means 2,
The standard chemical components of the estimated material were extracted. In the present embodiment, since the number of explanatory variables is larger than the number of data, Ni, T, which is slightly contained at 0.2% by weight or less in any of the samples.
i, Cr, Sn, and Pb are excluded from the standard chemical components, and the above-mentioned 5 out of the 11 main elements contained in the aluminum alloy are excluded.
6 elements excluding the elements (Si, Cu, Mg, Zn, Fe,
Mn) was extracted as a standard chemical component.

Next, the data shown in Table 8 was selected by the standard data selecting means 3 from the knowledge base 1 already constructed. The method of creating data extracted by the standard data selecting means 3 from the data constructed in the knowledge base 1 is as follows. That is, 9
First, the chemical composition of the JIS ADC12 alloy is IC
By chemical analysis such as P emission analysis, the above-mentioned 11 main elements (Si, Cu, Mg, Zn, F
e, Mn, Ti, Cr, Ni, Sn, Pb). These alloys were melted, subjected to flux deoxidation and vacuum degassing, and then cast into a water-cooled copper mold only at the bottom. The measured value was used as the actual measured value of the true density.

[0119]

[Table 8]

Next, creation of true density estimation information (estimation of true density estimation formula) was performed in the following steps. This flowchart is shown in FIG.

(Eleventh Step) First, the true density estimation information model creation means 41 creates a model equation of the true density estimation equation [P11: true density estimation information model creation step]. In this embodiment, as in the first embodiment, Equation 1 is used as a model equation.

(Twelfth Step) Next, the provisional true density estimation information system creating means 42 assumes the function form of the true density estimation formula and assumes the true density estimation formula [P12: Creation of temporary true density estimation information system. Process]. In this embodiment, similarly the functional form of the first preferred embodiment, it was assumed that _{f i (x i) = x} i. Then, for the sample (n = 9), an object variable [rho _0, the relationship between the explanatory variable x _i, determined by multiple regression analysis.

Next, the chemical component selecting means 43 and the standard chemical component changing means 45 derive the true density from the information indicating the relationship between the standard chemical component and the estimated true density information based on the temporary true density estimation formula. A chemical component necessary for estimation is selected (chemical component selection step). In this example, the chemical components were selected by the following thirteenth to eighteenth steps.

(Thirteenth Step) First, in the result of the (twelfth step), an explanatory variable having a small standard partial regression coefficient and a large standard error is selected and deleted [P13]. That is, first, for Mg, the standard partial regression coefficient is 0.02.
And the standard error was 48, which is one order of magnitude larger than that of Si, Cu, Zn, so this explanatory variable was deleted.

(14th step) Next, the (13th step) is repeated until the variance ratio F ₀ and the contribution ratio R ^{2 are} larger and the Akaike information criterion (AIC) is smaller.
The type and number of explanatory variables are optimized [P14]. Thereby, Mn was newly deleted. Si, Cu, Zn, and Fe were not deleted because deleting them deteriorated the fit of the model represented by F ₀ , R ² , and AIC.

(Fifteenth Step) Next, the standard chemical component is changed by the standard chemical component changing means 45, if necessary, in consideration of the validity of the standard chemical component [P15]. In the present embodiment, Ni not selected by the standard chemical component extraction means 2 is first added as a standard chemical component, and the standard chemical component is changed to Si, Cu, Zn, Fe, Ni, and the thirteenth and fourteenth steps are performed. The steps were repeated to optimize the chemical composition. As a result, Ni had a large standard error of 42 and was deleted because F ₀ and R ² became larger and AIC became smaller when it was deleted. Next, Ti, Cr, Sn, and Pb are sequentially increased by 1
As a result of adding each of them and repeating the same analysis, it was found that removing all of these chemical components resulted in larger F ₀ and R ² ,
It was confirmed that the AIC was small, and the formula was well fit. As a result, Si, Cu, Zn, Fe
Was selected.

(16th step) Next, the function determining means 44
According to 1, the function form of the true density equation is determined [P16: function determination step]. In this embodiment, partial regression plots optimized relationship between the explanatory variable x _i and the target variables [rho _0, the maximum contribution R ² is a variance ratio F _0, and as AIC value is minimized, functions to determine the form _{f i (x i) = x} i. As a result of examining the function form using this partial regression plot, a linear relationship was observed between the explanatory variable and the predicted value in each case, confirming that the function form was appropriate.

(Seventeenth Step) Next, according to the above steps, it was confirmed that in this embodiment, the number of data is small and the influence of individual data on the accuracy of the estimation formula is large.
Remove abnormal data [P17]. In the present embodiment, data that is far apart in the predicted value-residual plot and in which the standard residual and the student residual are abnormally large (significant level 5%) are removed as abnormal data. That is, the data of sample number 99 in Table 6 corresponding to the above was removed as abnormal data.

(Eighteenth Step) Next, a true density estimation formula is determined by the true density estimation information system determining means 442 [P1
8: True density estimation information system determination step]. FIG. 18 shows the obtained true density estimation formula and regression coefficients.

(Nineteenth Step) Next, the fitting property confirmation means 443 checks the fitting property of the true density estimation formula [P19]. In the present example, a scatter diagram of a normal probability point plot and a predicted value was created for the residual, and the normality and the uniform variance of the residual were confirmed. As a result, the normality of the residual was confirmed. It was also confirmed that the measured value and the predicted value of the true density were in good agreement over the whole. Furthermore, the uniformity of the residual was also confirmed.

Next, using the obtained true density estimation formula, A
A shrinkage cavity estimation test of the l-Si alloy casting was performed.

First, the true density estimating means 5 estimated the true density value of the material to be estimated (true density estimation step).
Next, the shrinkage nest amount was estimated. That is, with respect to three types of castings having different shrinkage cavities, respective densities ρ are obtained by the Archimedes method, and using the true density values obtained from the true density estimation formula shown in FIG. Nest volume was calculated. Next, the shrinkage cavities were determined by image processing based on the sectional micrographs of the three types of castings. As a result, it was confirmed that the shrinkage nest amount and the estimated shrinkage nest amount substantially matched. Next, the true density of the casting was measured in the same manner as in the first example. As a result, it was confirmed that the true density estimated value almost coincided with the true density measured value.

Comparative Example 3 For comparison, a simple multiple regression analysis was performed on the same samples as those in the third example (sample numbers: 96 to 104). FIG. 19 shows the obtained true density estimation formula for comparison, the regression coefficient, and the like. As a result, in the case of the present example, it was confirmed that the regression equation had a high degree of freedom-corrected contribution and a small estimation error, despite the fact that the number of explanatory variables was small and the equation was simple compared to the comparative example. Was done.

[Brief description of the drawings]

FIG. 1 is a diagram showing an apparatus for estimating a true density of a material used in a first embodiment of the present invention, and is a schematic explanatory view of the apparatus.

FIG. 2 is a flowchart illustrating an outline of a method of generating true density estimation information (a method of estimating a true density estimation formula) performed in the first embodiment of the present invention.

FIG. 3 is a diagram showing a tentative true density estimation formula obtained by a tentative true density estimation information system creating means, a regression coefficient, and an analysis of variance in the first embodiment of the present invention.

FIG. 4 is a diagram showing a true density estimation formula obtained by a true density estimation information system determining means, a regression coefficient, and an analysis of variance in the first embodiment of the present invention.

FIG. 5 is a diagram showing a partial regression plot for Mn (Mn ^1/2 ) obtained by the function determining means in the first embodiment of the present invention.

FIG. 6 is a diagram showing a partial regression plot for Si obtained by the function determining means in the first embodiment of the present invention.

FIG. 7 is a diagram showing a partial regression plot for Cu obtained by the function determining means in the first embodiment of the present invention.

FIG. 8 is a diagram showing a partial regression plot for Zn obtained by the function determining means in the first embodiment of the present invention.

FIG. 9 is a diagram showing a partial regression plot for Fe obtained by the function determining means in the first embodiment of the present invention.

FIG. 10 is a diagram illustrating a histogram of residuals representing normality with respect to the residuals, which is obtained in a process of checking the fit of the true density estimation formula in the first embodiment of the present invention.

FIG. 11 is a diagram illustrating a relationship between an actually measured value and a predicted value obtained in a process of checking the fit of a true density estimation formula in the first example of the present invention.

FIG. 12 is a diagram showing a relationship between a predicted value representing the homogeneity of the residual and a residual obtained in a process of checking the fit of the true density estimation formula in the first embodiment of the present invention.

FIG. 13 is a diagram for confirming the fit of the comparison true density estimation formula obtained in Comparative Example 1, and is a diagram illustrating a histogram of residuals for determining the normality of the residuals.

FIG. 14 is a diagram illustrating a true density estimation formula, a regression coefficient, and the like obtained by a true density estimation information system determination unit in the second embodiment of the present invention.

FIG. 15 is a diagram illustrating a comparison true density estimation formula, a regression coefficient, and the like obtained in Comparative Example 2.

FIG. 16 is a diagram illustrating a device for estimating the true density of a material used in the third embodiment of the present invention, and is a schematic explanatory diagram of the device.

FIG. 17 is a flowchart illustrating an outline of a method of generating true density estimation information (a method of estimating a true density estimation formula) performed in the third embodiment of the present invention.

FIG. 18 is a diagram illustrating a true density estimation formula, a regression coefficient, and the like obtained by a true density estimation information system determination unit according to the third embodiment of the present invention.

19 is a diagram illustrating a comparison true density estimation formula, a regression coefficient, and the like obtained in Comparative Example 3. FIG.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-54-53564 (JP, A) JP-A-5-5690 (JP, A) JP-A-5-52735 (JP, A) JP-A-62 194440 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) G01N 9/00 JICST file (JOIS)

Claims (22)

(57) [Claims] 1. A plurality of sound standard samples containing at least two or more kinds of chemical components constituting a material to be estimated, and having any one of them having a cavity volume ratio of not more than a predetermined value. A standard sample data preparation step of preparing and obtaining true density actual measurement value information and chemical composition information of the standard sample; based on the information obtained in the standard sample data preparation step, a true Selecting a chemical component that contributes to the density estimation, and obtaining a true density estimation information obtaining information for estimating the true density of the material from the selected chemical component; and a true density estimation obtained from the true density estimation information generating step. A true density estimating step of estimating the true density of the material to be estimated based on the information and the chemical component information of the material to be estimated and the composition information of the chemical component corresponding to the true density estimation information. Special The method of estimating the true density of a material. 2. A standard sample data creating step, wherein the chemical composition is within the chemical composition range of the material to be estimated and any of the chemical components constituting the material to be estimated is included.
2. The method according to claim 1, further comprising the steps of: preparing a plurality of healthy standard samples having a cavity volume ratio of not more than a predetermined value, and obtaining true density measured value information and chemical composition information of the standard samples. Density estimation method. 3. A standard chemical component extracting step of extracting a standard chemical component of a material to be estimated; and a plurality of sound materials having a cavity volume ratio of not more than a predetermined value, including at least one of the extracted standard chemical components. The true density actual measured value information, a standard data creation step of obtaining the chemical composition information corresponding to the actual measured value, based on the information obtained by the standard data selecting step,
A true density estimation information creating step of creating information for estimating the chemical density and the true density of the estimated target material from the standard chemical component that contributes to the estimation of the true density of the estimated target material; and A true density estimating step of estimating the true density of the target object based on the obtained true density estimation information, the chemical component information of the target object material corresponding to the true density estimation information, and the composition information of the chemical component; A method for estimating the true density of a material, comprising: 4. The standard chemical component extraction step is a step of extracting a standard component within the chemical composition range of the material to be estimated, and the standard data creation step is a step of extracting all chemical components constituting the material to be estimated. 4. A step of preparing a plurality of sound standard samples containing components and having a cavity volume ratio of not more than a predetermined value, and obtaining true density measured value information and chemical composition information of the standard samples. A method for estimating the true density of the described material. 5. The method for estimating the true density of a material according to claim 3, wherein the standard chemical component extracted in the standard chemical component extracting step includes at least a main chemical component of the material to be estimated. 6. The true density estimation of a material according to claim 3, wherein the standard chemical component extracted in the standard chemical component extraction step includes at least a chemical component contained in the material to be estimated at a predetermined value or more. Method. 7. A true density estimation information model creating step of creating a true density estimation information model including a predetermined functional form having the composition information of the standard chemical component as an element, A provisional true density estimation information system creating step of creating a provisional true density estimation information system based on the estimated information model and the standard data obtained in the standard data creation step; A chemical component selection step of selecting a chemical component necessary for estimating the true density from information indicating the relationship between the chemical component and the true density measured value and / or the estimated true density; and 4. A true density estimation information system creating step of creating a true density estimation information system based on information indicating a relationship with the estimated true density. Estimation method. 8. A process for creating a true density estimation information system in the true density estimation information creating process, wherein the composition of the selected chemical component is determined based on the selected chemical component and information indicating the relationship between the true density measured value and / or the estimated true density. A function form determining step of determining a predetermined function form using information as an element; and information indicating the relationship between the selected chemical component and the true density actual measurement value and / or the estimated true density, including the determined function form; 8. The method according to claim 7, further comprising: determining a true density estimation information system based on information obtained by removing abnormal data from the standard data. . 9. The chemical component selection step in the true density estimation information creating step includes the step of:
8. The method according to claim 7, further comprising the step of removing abnormal data from the standard data based on information indicating a relationship with the estimated true density. 10. The chemical component selecting step in the true density estimation information creating step, wherein the standard chemical component and the true density actual measurement value and / or
Or a step of selecting and removing a chemical component that reduces the estimation accuracy based on the information indicating the relationship with the estimated true density, and information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density. Selecting and optimizing the chemical components necessary for estimating the true density based on the above, and selecting the chemical components required for estimating the true density by repeating these steps as appropriate. The method for estimating a true density of a material according to claim 7. 11. The chemical component selecting step in the true density estimation information creating step includes confirming the validity of the standard chemical component and including a standard chemical component changing step of changing the standard chemical component as necessary. 11. The method for estimating a true density of a material according to claim 10, wherein: 12. An estimated true density accuracy determining step of determining the appropriateness of the accuracy of the estimated true density, and when the estimated true density accuracy determining step determines that the accuracy is insufficient, the data number of the standard data is determined. 4. The method according to claim 3, further comprising a standard data changing step of changing the standard data by changing the range of the chemical composition of the material to be estimated. 13. The method according to claim 3, wherein the material to be estimated is a metal material. 14. The material to be estimated is aluminum or an aluminum alloy material.
A method for estimating the true density of the described material. 15. The method according to claim 14, wherein the material to be estimated is an Al—Si alloy material. 16. A knowledge base having actual value information of the true density of the material to be estimated and chemical composition information corresponding to the actual value, a standard chemical component extracting means for extracting a standard chemical component of the material to be estimated, Any one of the standard chemical components extracted from the knowledge base by the standard chemical component extraction means.
Standard data selecting means for selecting actual measured value information of true densities of a plurality of healthy materials whose cavity volume ratio is equal to or less than a predetermined value and chemical composition information corresponding to the measured values to obtain standard data; and the standard data selecting means Based on the information obtained by
A true density estimation information creating means for creating information for estimating a chemical component contributing to the estimation of the true density of the material to be estimated from the standard chemical component and the true density of the material to be estimated, and True density estimating means for estimating the true density of the target object based on the obtained true density estimation information, the chemical component information of the target object material corresponding to the true density estimation information, and the composition information of the chemical component. And a device for estimating the true density of the material. 17. A true density estimation information model creating means for creating a true density estimation information model including a predetermined function form having the composition information of the standard chemical component as an element; Provisional true density estimation information system creating means for creating a provisional true density estimation information system based on the estimated information model and the standard data obtained in the standard data creation step; A chemical component selecting means for selecting a chemical component necessary for estimating the true density from information indicating the relationship between the chemical component and the true density measured value and / or the estimated true density; 17. A true density estimation information system creating means for creating a true density estimation information system based on information indicating a relationship with the estimated true density. Density estimator. 18. A true density estimation information system creating means, based on a selected chemical component and information indicating a relationship between a true density actual measurement value and / or an estimated true density, a predetermined chemical component having the composition information of the selected chemical component as an element. Function form determining means for determining a function form; information indicating a relationship between a selected chemical component including the determined function form and a measured true density value and / or an estimated true density; and removing abnormal data from the standard data. 18. The apparatus according to claim 17, further comprising: a true density estimation information system determining unit that determines a true density estimation information system based on the information. 19. The chemical component selecting means in the true density estimation information creating means, wherein the standard chemical component and the true density actual measurement value and / or
18. The apparatus for estimating the true density of a material according to claim 17, further comprising abnormal data removing means for removing abnormal data from the standard data based on information indicating a relationship with the estimated true density. 20. The chemical component selecting means in the true density estimation information creating means, wherein the standard chemical component and the true density actual measurement value and / or
Or, from the information indicating the relationship with the estimated true density, means for selecting and removing a chemical component that reduces the estimation accuracy, and from the information indicating the relationship between the standard chemical component and the true density measured value and / or the estimated true density, Means for selecting and optimizing a chemical component necessary for estimating the true density, and selecting a chemical component necessary for estimating the true density by repeating these steps as appropriate. 16. An apparatus for estimating a true density of a material according to 16. 21. The chemical component selecting means in the true density estimation information creating means confirms the validity of the standard chemical component and includes a standard chemical component changing means for changing the standard chemical component as necessary. 21. The apparatus for estimating a true density of a material according to claim 20, wherein: 22. A method for estimating a true density of a material to be estimated based on the true density estimation information obtained by claim 1 to claim 15, and calculating a cavity amount of the material to be estimated from the estimated true density and an apparent density of the material to be estimated. A method for estimating the amount of cavities in a material, characterized by being estimated.