JP2004301654A - Method and device of estimating effective internal friction angle of powder, and method and device of designing storage tank for powder - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To acquire an effective internal friction angle of a powder without spending time and effort. <P>SOLUTION: A neutral network model for defining cause-and-effect relations between the fundamental physical properties and the effective internal friction angle of the powder is developed based on actually measured data on the physical properties and the friction angle of the powder. The friction angle of the powder is estimated/calculated by inputting the physical properties of the powder into the developed network model. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method and an apparatus for estimating an effective internal friction angle of a powder, and a method and an apparatus for designing a storage tank for a powder.
[0002]
[Prior art]
In designing equipment for handling powder, not only basic properties such as density, particle size distribution, and liquid content of the powder layer, but also mechanical properties such as shear properties and adhesion properties of the powder layer are grasped. There is a need.
[0003]
For example, in a powder storage tank facility such as a hopper or a silo, when powders approach each other and an external force is applied thereto, a blockage occurs near a discharge port, which may cause a problem that the powder cannot be discharged. . In many cases, the mechanical properties of the powder layer are deeply involved in the problem.
[0004]
Therefore, the designer of the equipment that handles the powder evaluates the mechanical properties of the powder layer of the powder to be handled, and based on the evaluation, performs appropriate design of the equipment so that the above-described problem hardly occurs. .
[0005]
As a method for evaluating the mechanical properties of the powder layer, an evaluation method based on a shear test is known. The shear test is a test method for evaluating the relationship between the normal stress σ acting on the collapse surface and the shear stress τ when the static powder layer changes to a dynamic state due to collapse, and evaluating the fracture characteristics of the static powder layer. Say. For a powder layer having a predetermined porosity ε, an experiment was performed while changing the vertical stress σ, and the shear stress τ for each vertical stress σ was plotted on the σ−τ plane to obtain a fracture envelope YL as shown in FIG. Can be obtained.
[0006]
Here, Non-Patent Document 1 states that when the powder is a highly adherent powder that easily causes a problem such as blockage at the discharge port, the destructive envelope YL conforms well to the following equation. Is described.
(Τ / C) ⁿ = (Σ + T) / T
(However, T is the tensile breaking strength, C is the adhesive force, and n is the shear index.)
The above equation is called the Warren-Spring equation or the Farley-Valentin equation.
[0007]
Further, the relationship between the normal stress σ acting on the collapse surface and the shear stress τ in the steady state after the change to the dynamic state is obtained by a shear test, and plotted on the σ-τ plane, as shown in FIG. Such a limit state line CSL can be obtained. As shown, the limit state line CSL is a straight line passing through the origin on the σ-τ plane.
[0008]
Then, as shown in FIG. 12, a molding circle passing through the intersection of the destruction envelope YL and the limit state line CSL and in contact with the destruction envelope YL can be obtained. The tangent from the origin of the molding circle is the effective destruction envelope EYL. The inclination angle δ of the effective fracture envelope EYL corresponds to the friction angle of the powder under pressure, and is called the effective internal friction angle (see Non-Patent Document 2).
[0009]
As described above, the fracture envelope YL and the critical state line CSL can be obtained by the shear test, and the tensile rupture strength T, the adhesive force C, the shear index n, the effective internal friction angle can be obtained from the fracture envelope YL and the critical state line CSL. Mechanical properties such as δ can be obtained.
[0010]
Further, Non-Patent Document 1 describes a method for evaluating fluidity and quality control of a powder and obtaining a diameter and an inclination angle of an outlet of a storage tank using a fracture envelope YL obtained by a shear test. ing.
[0011]
[Non-patent document 1]
Kimihiro Hong, "Measurement Method of Mechanical Properties of Powder and Its Utilization", Powder and Industry, October 1, 1992, Vol. 24, No. 10, p. 61-74
[0012]
[Non-patent document 2]
Handbook of Powder Equipment and Apparatus, Editorial Committee, "Powder Equipment and Equipment Handbook", First Edition, Nikkan Kogyo Shimbun, May 30, 1995, p. 63-68
[0013]
[Problems to be solved by the invention]
As described above, in order to obtain the fracture envelope YL and the limit state line CSL by the shear test, it is necessary to repeat the shear test while changing the vertical stress σ, which takes time and effort. Furthermore, when designing a powder storage facility such as a hopper or silo, it is necessary to repeat the shear test by changing the porosity ε by applying pre-consolidation to the powder layer, which results in additional time and labor. I was
[0014]
In addition, since the shear test needs to be repeated, a large amount of powder sample (at least about 500 to about 1000 g) is required. Therefore, for example, when the powder is expensive, or when the powder is toxic or corrosive and it is difficult to handle the powder in large quantities, it is difficult to repeat the shear test, and the destructive envelope As a result, the line YL could not be obtained, and equipment for handling powder could not be designed.
[0015]
On the other hand, among the basic physical properties of the powder, the density, particle size distribution, liquid content, and the like can be measured more quickly and easily than the mechanical properties, and are measured in a small amount (about 5 to about 10 g). be able to. For this reason, if the mechanical properties can be evaluated from the basic physical properties that can be easily measured, the mechanical properties can be easily obtained by measuring the basic physical properties. However, at present, a method for evaluating mechanical properties from basic physical properties has not yet been known.
[0016]
The present invention has been made to solve the above problems, and an object of the present invention is to provide a method capable of obtaining an effective internal friction angle of a powder without taking time and labor.
[0017]
[Means for Solving the Problems]
In order to solve the above problems, the method of estimating the effective internal friction angle of the powder of the present invention is based on measured data of the basic physical properties of the powder and the effective internal friction angle of the powder, A construction step of constructing a neural network model (hereinafter abbreviated as “NN model”) that defines a causal relationship between the basic physical properties of the powder and the effective internal friction angle of the powder. I have.
[0018]
According to the above method, the NN model is constructed from the measured data of the basic physical properties of the powder and the effective internal friction angle. By inputting the basic physical properties of the powder to the constructed NN model, the estimated data of the effective internal friction angle is output from the NN model. Therefore, since the effective internal friction angle of the powder can be estimated from the basic physical properties of the powder, the effective internal friction angle of the powder can be obtained without taking much time and effort.
[0019]
An NN model was constructed by selecting various basic physical properties, an effective internal friction angle was estimated using the constructed NN model, and the accuracy of the estimated data with respect to the actually measured data was examined. It was found that when the liquid content of the body was not included, the accuracy was significantly reduced. Therefore, it is desirable that the basic physical properties include at least the liquid content of the powder.
[0020]
It was also found that good accuracy was obtained when the average particle size, particle density, liquid content, and porosity of the powder were selected as basic physical properties. Therefore, it is desirable that the basic physical properties include at least the average particle diameter, particle density, liquid content, and porosity of the powder.
[0021]
Further, it has been found that when the standard deviation of the particle size of the powder is added to the basic physical properties, even better accuracy can be obtained. Therefore, it is desirable that the basic physical properties further include the standard deviation of the particle size of the powder.
[0022]
Further, the method of estimating the effective internal friction angle of the powder according to the present invention is the method as described above, wherein a plurality of data that can be input to the NN model constructed by the construction step are provided for each basic physical property of the powder. A first table creation step of creating a first table composed of a plurality of data selected for each of the basic physical properties by discretely selecting the data; and a data generation method for each of the basic physical properties selected from the first table. Calculating the data of the effective internal friction angle by inputting the basic physical property data set into the NN model, and selecting all of the basic physical property data sets selectable from the first table. Until selected, the repeating step of repeating the calculation step, from the data of the effective internal friction angle corresponding to each data set of the basic physical properties A second table creating step of creating a second table, and a search program for outputting the effective internal friction angle data corresponding to the input data set of basic physical properties by searching the first table and the second table And generating a.
[0023]
In general, the constructed NN model can often be used only by dedicated software for constructing the NN model. For this reason, it is impossible to estimate an effective internal friction angle corresponding to basic physical properties on a computer on which dedicated software is not installed.
[0024]
On the other hand, according to the above-described method, when the search program is executed and the data set of the basic physical properties is input, the search program searches the first table and the second table to obtain the input basic physical properties. The data of the effective internal friction angle corresponding to the data set of is output. Therefore, even with a computer on which dedicated software is not installed, the provision of the first and second tables and the search program makes it possible to estimate an effective internal friction angle corresponding to basic physical properties.
[0025]
The intervals of the data discretely selected in the first table creation step may be equal intervals. However, based on the distribution of the basic physical properties, for example, the intervals of the data are narrowed in a range having a large frequency, In a small range, the data interval can be widened.
[0026]
Further, the method for designing a powder storage tank of the present invention uses an effective fracture envelope determined from the effective internal friction angle estimated by the above method, and a fracture envelope and a wall fracture envelope determined by a shear test. Thus, the diameter and the inclination angle of the outlet in the powder storage tank are obtained.
[0027]
According to the above-described method, it is not necessary to perform a shear test when obtaining an effective destruction envelope, so that a powder storage tank can be quickly designed.
[0028]
Further, the device for estimating the effective internal friction angle of the powder of the present invention is an input means for inputting the basic physical properties of the powder and the effective internal friction angle of the powder, and an input by the input means. The causal relationship between the basic physical properties of the powder and the effective internal friction angle of the powder is defined based on the measured data of the basic physical properties of the powder and the effective internal friction angle of the powder. Means for constructing an NN model to be performed, and inputting the basic physical properties of the powder inputted by the input means to the NN model constructed by the constructing means, thereby obtaining an effective internal friction of the powder. And estimating means for estimating the angle.
[0029]
According to the above configuration, the construction means constructs the NN model from the measured data of the basic physical properties of the powder and the effective internal friction angle, and the estimation means computes the basic NN model of the powder with respect to the constructed NN model. By inputting physical properties, estimated data of the effective internal friction angle is output from the NN model. Therefore, since the effective internal friction angle of the powder can be estimated from the basic physical properties of the powder, the effective internal friction angle of the powder can be obtained without taking much time and effort.
[0030]
Further, the apparatus for estimating the effective internal friction angle of the powder according to the present invention, in the above-described configuration, relates to a plurality of data that can be input to the NN model constructed by the construction means for each basic physical property of the powder. A first table creating means for creating a first table composed of a plurality of data selected for each of the basic physical properties by discretely selecting the basic physical properties; and data from each of the basic physical properties selected from the first table. By inputting a data set of basic physical properties into the estimating means, the effective internal friction angle data is obtained from the estimating means, and the data of the basic physical properties can be selected from a first table. By repeating until all of the sets are selected, a second table is created which includes the effective internal friction angle data corresponding to each data set of the basic physical properties. Table creation means, and search program creation means for creating a search program for outputting a data set of an effective internal friction angle corresponding to the input data set of basic physical properties by searching the first table and the second table. Is further provided.
[0031]
According to the above configuration, the first table, the second table, and the first table, the second table created by the search program creating unit, and the first program are utilized by using the first program. And by searching the second table, it is possible to output effective internal friction angle data corresponding to the input data set of basic physical properties. Therefore, even with a computer in which dedicated software is not installed, the provision of the first and second tables and the search program makes it possible to estimate the destructive characteristics corresponding to the basic physical properties.
[0032]
Further, the powder storage tank design device of the present invention, the effective fracture envelope determined from the effective internal friction angle estimated by the device having the above configuration, the fracture envelope and the wall fracture envelope determined by the shear test. It is characterized in that the diameter and the inclination angle of the outlet in the powder storage tank are obtained by using the method.
[0033]
According to the above configuration, it is not necessary to perform a shear test when obtaining an effective destruction envelope, so that a powder storage tank can be quickly designed.
[0034]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to FIGS. FIG. 1 shows an outline of a powder storage tank design system according to the present embodiment. The system includes an actual measurement data acquisition device 10, an NN model construction device 11, and a storage tank design device 12. These devices 10 to 12 have the same configuration as a general computer.
[0035]
The measurement data acquisition device 10 acquires the measurement data of the basic physical properties and the effective internal friction angle δ of various powders used for constructing the NN model by measuring with various measuring devices, and acquires the NN model. Send it to the construction device 11. In addition, the measurement data acquisition device 10 sends measurement data of basic physical properties of the powder specified by the storage tank design device 12 to the storage tank design device 12.
[0036]
The NN model construction device 11 constructs an NN model based on the actual measurement data received from the actual measurement data acquisition device 10 using the actual measurement data of the basic physical properties as an input variable and the actual measurement data of the effective internal friction angle δ as an output variable. Further, the NN model construction device 11 creates a table (first table) of basic physical properties, and uses the created table and the constructed NN model to create a table of effective internal friction angles (second table). Create
[0037]
In addition, the NN model construction device 11 creates a search program that outputs an effective internal friction angle δ corresponding to the input basic physical property by searching a table of basic physical properties and a table of effective internal friction angles. The created search program is sent to the storage tank designing device 12 together with the basic physical property table and the effective internal friction angle table.
[0038]
The storage tank design device 12 specifies the powder to be stored in the storage tank to be designed, and requests and receives the measured data of the specified powder from the measured data acquisition device 10. Next, the storage tank design device 12 executes the search program received from the NN model construction device 11 and inputs the actual measurement data of the specified powder received from the actual measurement data acquisition device 10 to thereby obtain the effective internal friction of the specified powder. Output the angle δ. Then, the storage tank designing device 12 uses the effective internal friction angle δ of the designated powder output from the search program and other mechanical characteristics to design a storage tank in which clogging hardly occurs.
[0039]
In the above configuration, the NN model construction device 11 constructs an NN model based on the basic physical properties of the powder acquired from the measurement data acquisition device 10 and the actual measurement data of the effective internal friction angle δ. By inputting the basic physical properties of the powder into the constructed NN model, the effective internal friction angle δ of the powder is output from the NN model.
[0040]
Therefore, since the effective internal friction angle δ of the powder can be estimated from the basic physical properties of the powder, the effective internal friction angle δ of the powder can be obtained without taking much time and effort.
[0041]
The NN model construction device 11 creates a table of basic physical properties, a table of effective internal friction angles, and a search program, and sends them to the storage tank design device 12, and the storage tank design device 12 executes the search program, When the measured data of the designated powder is input, the effective internal friction angle δ of the designated powder is output. Therefore, even when the dedicated software for constructing the NN model is not installed in the storage tank design device 12, the storage tank design device 12 includes a table of basic physical properties, a table of effective internal friction angles, and a search program. The effective internal friction angle δ corresponding to the basic physical properties can be estimated.
[0042]
In addition, the storage tank designing device 12 can measure the basic physical properties of the powder when obtaining the effective internal friction angle δ of the powder, and it is not necessary to perform a shear test. Can be done.
[0043]
The transfer of data and the like between the devices can be performed via a communication medium or a recording medium. Further, the powder storage tank design system of the present embodiment has a configuration including three devices 10 to 12 as shown in FIG. 1, but any two devices are combined into one device. It can be integrated, or all three devices can be integrated into one device.
[0044]
Hereinafter, each configuration will be described in more detail. There are many basic physical properties of the powder obtained by the measurement data obtaining apparatus 10, such as shape, particle size, particle size distribution, and density. In the present embodiment, the average particle size D50 of the powder, the geometric standard deviation SD, particle density Dens, liquid content Liq%, and porosity ε were used.
[0045]
The average particle size D50 and the geometric standard deviation SD can be obtained by measuring the particle size distribution. Examples of the method for measuring the particle size distribution include a sieve method, a sedimentation method, a light scattering method, an inertia method, a diffusion method, and an electrostatic classification method. In this example, the average particle diameter and the geometric standard deviation SD were measured using a laser light scattering particle size distribution analyzer (Mastersizer S manufactured by Malvern Instruments, UK). Although there are various definitions for the average particle diameter, in this embodiment, the average particle diameter D50 is NMD (number median diameter), that is, the central cumulative value based on the number.
[0046]
Particle density Dens refers to the ratio of the total mass to the total volume occupied by the particles alone. In this example, the particle density Dens was measured using a gas displacement type density meter (micromeritics AccuPyc 1330 manufactured by Shimadzu Corporation).
[0047]
The liquid content Liq% refers to a percentage of the mass of the liquid component removed by drying the particles with respect to the total mass of the particles including the liquid component. Usually, the liquid content Liq% is measured by a drying loss method. When the liquid component is water, the water content may be measured by the Karl Fischer method. In this example, the liquid content Liq% was measured by the Karl Fischer method with respect to a sample having a low water content, usually by a drying loss method.
[0048]
The porosity ε refers to the ratio of the volume of the space to the apparent total volume occupied by the particles. The porosity ε is the same pre-consolidation as when measuring the shear stress τ in the shear test, and when the powder is filled into the filling container, the apparent volume occupied by the powder, the volume of the filling container, Is calculated from the filling mass.
[0049]
In this example, as a device for performing a shear test, POWDER BED TESTER PTHN-13BA manufactured by Sankyo Piotech Co., Ltd., which is a shear test device of Jenike, is used. This product can measure the shear stress τ by a shear test, can measure the tensile rupture strength T by a tensile rupture method, and can measure the porosity ε when measuring the tensile rupture strength T. Therefore, in this example, the tensile breaking strength T and the porosity ε were measured using POWDER BED TESTER PTHN-13BA manufactured by Sankyo Piotech.
[0050]
On the other hand, regarding the effective internal friction angle δ, which is a mechanical property of the powder, as shown above, the normal stress σ and the shear stress τ measured by the shear test are plotted on the σ-τ plane, as shown in FIG. The destruction envelope YL and the limit state line CSL to be obtained are obtained, the effective destruction envelope EYL is obtained from the destruction envelope YL and the limit state line CSL, and the effective internal friction angle δ is obtained from the effective destruction envelope EYL. In this example, the effective internal friction angle δ was measured using the above-described shear test apparatus.
[0051]
The above measurement was carried out for one sample of powder by changing the porosity ε to two or three types. FIG. 2 shows the measured basic physical properties of the powder and the statistics of the effective internal friction angle δ in tabular form.
[0052]
In FIG. 2, the statistics of the average particle diameter D50, the geometric standard deviation SD, the particle density Dens, the liquid content Liq%, the porosity ε, and the effective internal friction angle δ are described in order from the top in each row. I have. In each column, in order from the left, a row number Col, a property name, a unit, a variable name Variable, a total number Total, a valid number Valid, a minimum value Min, a maximum value Max, an average value Mean, and a standard deviation Std Dev are displayed. Each is listed.
[0053]
The statistics shown in FIG. 2 may be used when the NN model construction device 11 creates a basic physical property data table.
[0054]
356 actual measurement data of the basic physical properties of the powder and the effective internal friction angle δ measured as described above are sent to the NN model construction device 11. As described above, the NN model construction device 11 constructs an NN model based on the measured data received from the measured data acquisition device 10.
[0055]
Recently, application software for constructing an NN model based on actual measurement data on a computer having a general configuration has been developed and sold. For example, “MATLAB (registered trademark)” of Cyber Systems and “JMP (registered trademark)” of SAS Institute are known as general scientific software, and “Insights (registered trademark)” of Pavilion Technologies is known. Also, "Aspen IQ (trademark)" of Aspen Tech is known as software specialized for chemistry. In the present embodiment, the NN model was constructed using “Insights (registered trademark)”.
[0056]
FIG. 3 shows a comparison between an estimated value calculated by an NN model constructed using various powders and an actually measured value. In the figure, each column includes, in order from the left, a sample (sample) name, an actually measured value of an input variable (actually measured data on the input side), an actually measured value of an output variable (actually measured data on the output side), and an output variable. Two estimates are listed.
[0057]
Note that the estimated value on the left side is an estimated value when an NN model is constructed using five input variables (NMD, geometric standard deviation, particle density, liquid content, and porosity). On the other hand, the estimated value on the right side is an estimated value when an NN model is constructed by removing the porosity ε from the input variables.
[0058]
In the sample, antioxidant A was 6- [3- (3-t-butyl-4-hydroxy-5-methylphenyl) propoxy] 2,4,8,10-tetra-t-butyldibenz [d , F] [1,3,2] dioxaphosphepin, and the antioxidant B is 2,2′-methylene-bis- (4-methyl-6-t-butylphenol).
[0059]
Referring to FIG. 3, when the NN model is constructed using five input variables (NMD, geometric standard deviation, particle density, liquid content, and porosity), there is a small variation in the estimated value with respect to the actually measured value. Can understand. On the other hand, when the NN model is constructed by excluding the porosity from the five input variables, it can be understood that the dispersion of the estimated value with respect to the actually measured value is large.
[0060]
Then, when one or two were excluded from the five input variables and the NN model was constructed, it was confirmed how the accuracy of the estimated value with respect to the actually measured value changes. As shown in FIG. In the figure, the standard deviation is obtained by dividing the sum of squares of the difference between the actually measured value and the estimated value by the number of data. The smaller the standard deviation, the smaller the variation of the estimated value with respect to the actually measured value.
[0061]
Referring to FIG. 4, it can be understood that the accuracy is not significantly deteriorated even if the geometric standard deviation SD is excluded from the input variables. Therefore, the effective internal friction angle δ can be accurately determined from the four basic physical properties of the average particle diameter D50, the particle density Dens, the liquid content Liq%, and the porosity ε, and the geometric standard deviation SD must be added. It can be understood that the accuracy can be obtained with higher accuracy.
[0062]
Also, referring to FIG. 4, it can be understood that excluding the liquid content Liq% from the input variables significantly deteriorates the accuracy. Therefore, it can be understood that at least the liquid content Liq% needs to be included in the input variable in order to accurately determine the effective internal friction angle δ from the basic physical properties of the powder.
[0063]
Next, the processing operation in the NN model construction device 11 will be described with reference to FIGS. First, as described above, a data set of basic physical properties (average particle diameter D50, geometric standard deviation SD, particle density Dens, liquid content Liq%, and porosity ε) of the powder, and the effective internal friction angle δ An NN model expressing the relationship with the data is constructed based on the actually measured data (Step S10, hereinafter sometimes simply referred to as “S10”. The same applies to other steps).
[0064]
Next, for each basic physical property, a plurality of data that can be input to the constructed NN model are discretely selected, so that a basic physical property data table composed of a plurality of data selected in each basic physical property is obtained. It is created (S11). FIG. 6 shows an example of the created basic physical property data table.
[0065]
It is desirable that the selection range of the data of each basic physical property is slightly wider than the range from the minimum value to the maximum value of the actually measured data shown in FIG. The data may be selected at equal intervals, or may be selected at different intervals according to the frequency distribution of the actually measured data.
[0066]
For example, when the frequency distribution (not shown) of the measured data of the basic physical properties was examined, the frequency distribution of the average particle diameter D50, the geometric standard deviation SD, and the liquid content Liq% was steep. On the other hand, the frequency distribution of the particle density Dens and the porosity ε was a gentle distribution.
[0067]
From the above, in the basic physical property data table shown in FIG. 6, the interval between data selected in the average particle diameter D50, the geometric standard deviation SD, and the liquid content Liq% is narrowed in a range where the frequency is large, The range is small and wide. On the other hand, the intervals of the data selected for the particle density Dens and the porosity ε are equal.
[0068]
Next, a data set of basic physical properties selected from the basic physical property data table created in step S11 is input to the NN model constructed in step S10, and data of the effective internal friction angle δ is calculated. (S12). This step S12 is performed for all data sets that can be selected from the basic physical property data table (S13), so that the effective internal friction angle δ data corresponding to each basic physical property data set is obtained. A corner table is created (S14).
[0069]
That is, when the basic physical property data table is as shown in FIG. 6, a basic physical property data set of 17 × 9 × 17 × 18 × 16 ≒ 750,000 is input to the NN model, thereby An effective internal friction angle table including data of 750,000 effective internal friction angles is created. When the geometric standard deviation SD is excluded from the input variables of the NN model, the total number of effective internal friction angle data is 17 × 17 × 18 × 16 ≒ 83,000.
[0070]
When selecting a data set of basic physical properties in step S12, it is preferable that the data set is selected according to a predetermined rule. In this case, it is possible to determine which data set in the effective internal friction angle table the data set selected from the basic physical property data table corresponds to based on a predetermined rule. Therefore, it is not necessary to store a data set of basic physical properties corresponding to the data of the effective internal friction angle in the effective internal friction angle table, and the scale of the effective internal friction angle table can be reduced.
[0071]
Examples of the predetermined rule include the following. That is, in the basic physical property data table shown in FIG. 6, first, the average particle diameter D50, the standard deviation SD, the particle density Dens, and the liquid content Liq% are selected from the first row, and the porosity ε is set to the initial value. Select in order from the line. When the porosity ε is selected from the last row, in the next selection, the liquid content Liq% is selected from the next row, and the porosity ε is again selected in order from the first row.
[0072]
Hereinafter, every time the porosity ε is selected from the last row, in the next selection, the liquid content Liq% is selected and repeated from the next row. When the liquid content Liq% and the porosity ε are selected from the last row, the next selection is to select the particle density Dens from the next row and the liquid content Liq% and the porosity ε again from the first row as described above. Repeat in order. Hereinafter, the same is repeated for the particle density Dens, the geometric standard deviation SD, and the average particle diameter D50.
[0073]
Next, a search program for outputting data of the effective internal friction angle δ corresponding to the input data set of basic physical properties is created (S15). Details of the search program will be described later. Then, the basic physical property data table created in step S11, the effective internal friction angle table created in step S14, and the search program created in step S15 are sent to the storage tank design device 12, and the NN model construction device is sent. The processing operation of No. 11 ends.
[0074]
Next, a processing operation in the storage tank design device 12 will be described with reference to FIGS. As described above, the storage tank design device 12 outputs the effective internal friction angle δ of the powder by executing the search program received from the NN model construction device 11.
[0075]
FIG. 7 shows the processing operation of the search program. In the search program, first, a search sheet and a search button as shown in FIG. 8A are displayed on the screen 20 of the display (S20).
[0076]
The search sheet is composed of two upper and lower sheets. On the upper sheet, as shown in FIG. 8A, the average particle diameter (50% diameter), the geometric standard deviation, the particle density, the liquid content, and the porosity, which are the basic physical properties of the powder, are shown. , The unit, the minimum value, and the maximum value are displayed, and the input value line and the database line are blank. Note that the minimum value and the maximum value are the minimum value and the maximum value in the basic physical property data table shown in FIG. 6, respectively.
[0077]
On the other hand, on the lower sheet, the unit is displayed for the effective internal friction angle δ, and the row of the calculated value is blank. A search button is displayed on the right side of the lower sheet.
[0078]
With respect to the search sheet and the search button displayed on the screen 20 as described above, the user uses the input device such as the keyboard and the mouse to search the measured data of the designated powder received from the measured data acquisition device 10. Input to the input value row of the sheet and press the search button. The input of the input value to the row may be automatically performed by the storage tank designing device 12.
[0079]
Returning again to the description of the processing operation of the search program, the process waits until the user presses the search button (S21). When the search button is pressed, the data closest to the input values for each basic physical property is used as the basic data. A search is made from the physical property data table and displayed in a row of the database of the search sheet (S22). At this time, when the input value of each basic physical property is out of the range from the minimum value to the maximum value, an error message may be displayed and the processing operation may be stopped.
[0080]
Next, the data of the effective internal friction angle δ corresponding to the data set of the basic physical properties displayed in the row of the database in step S22 is searched from the effective internal friction angle table, and as shown in FIG. It is displayed in the row of the calculated value on the search sheet (S23). Thereafter, by returning to step S21, the effective internal friction angle δ can be repeatedly searched.
[0081]
In the above case, the data of the effective internal friction angle δ corresponding to the data set of the basic physical properties closest to the input value is obtained, but by appropriately using the interpolation method, the effective internal friction angle δ corresponding to the input value can be obtained. The data of the internal friction angle δ can also be obtained. For example, a data set of two basic physical properties closest to the input value is obtained, data of two effective internal friction angles δ corresponding to the data sets of the basic physical properties are obtained, and an intermediate value of the data is calculated. It can also be displayed on a line.
[0082]
Next, the processing operation of designing a storage tank in which clogging is unlikely to occur using the determined effective internal friction angle δ in the storage tank designing apparatus 12 will be described with reference to FIGS. Now, as shown in FIG. 9, it is assumed that a storage tank 30 having a tapered lower portion and a lowermost portion serving as a powder discharge port is designed. The minimum discharge diameter at which blockage is unlikely to occur is defined as the minimum diameter B, and the angle between the vertical direction and the tapered surface is defined as the inclination angle α.
[0083]
10 and 11 show a processing operation for obtaining the minimum aperture B and the inclination angle α. First, the measured data of the basic physical properties (average particle diameter D50, geometric standard deviation SD, particle density Dens, liquid content Liq%, and porosity ε) of the powder to be designed are acquired from the measured data acquisition device 10 ( S30).
[0084]
Next, the search program shown in FIG. 7 is executed to input the actual measurement data of the basic physical properties of the powder, thereby outputting the effective internal friction angle δ (S31). Next, an effective fracture envelope EYL shown in FIG. 12 is obtained from the outputted effective internal friction angle δ (S33).
[0085]
Incidentally, in order to obtain the minimum diameter B and the inclination angle α of the storage tank 30, it is necessary to obtain not only the effective destruction envelope EYL but also the destruction envelope YL and the wall destruction envelope WYL shown in FIG. The fracture envelope YL can be obtained by a shear test as described above (S32). In addition, the wall fracture envelope WYL is obtained by changing the movable plate to a wall material to be designed in a parallel plate type shear test device such as the above-described POWDER BED TESTER PTHN-13BA manufactured by Sankyo Piotech. Can be performed (S34).
[0086]
As shown in FIG. 12, the breaking strength fc is obtained from the breaking envelope YL (S40), and the wall friction angle φw is obtained from the wall breaking envelope WYL (S43). Further, as shown in FIG. 12, the molding circle that is in contact with the destruction envelope YL and the effective destruction envelope EYL is the molding circle that passes through the intersection of the destruction envelope YL and the limit state line CSL and is in contact with the destruction envelope YL. The greater the intersection of the molding circle with the σ axis is the maximum principal stress σ1. Therefore, the maximum principal stress σ1 is obtained from the fracture envelope YL and the effective fracture envelope EYL (S42).
[0087]
Next, at least three sets of (fc, δ, σ1) are determined by repeating steps S30 to S33, S40, and S42 while changing the pre-consolidation of the powder (S44). Thus, since at least three points can be plotted on the σ1-δ plane, the relationship between the maximum principal stress σ1 and the effective friction angle δ can be represented by a graph (S45).
[0088]
Similarly, since at least three points can be plotted on the σ1-fc plane, the relationship between the maximum principal stress σ1 and the breaking strength fc can be represented by a graph (S46). Here, the value of σ1 / fc represents the fluidity of the powder and is defined as FF (flow function).
[0089]
Further, a representative value δ0 of the effective friction angle δ is selected from at least three sets of (fc, δ, σ1) obtained in step S44 (S47). An example of the representative value δ0 is an intermediate value of the effective friction angle δ.
[0090]
Next, using the representative value δ0 of the effective friction angle selected in step S47 and the wall friction angle φw obtained in step S43, the critical inclination angle αc and the flow factor ff, which are boundaries between the funnel flow and the mass flow, are calculated. Is obtained (S48). The critical inclination angle αc can be determined using a flow pattern determination diagram (not shown) in which a boundary line between the funnel flow region and the mass flow region on the α-φw plane is shown for each effective internal friction angle δ.
[0091]
In addition, the flow factor ff can be obtained from a formula using the representative value δ0 of the effective friction angle, the wall friction angle φw, and the critical inclination angle αc as variables. Alternatively, the flow factor ff can be determined using a contour diagram of the flow factor ff on the α-φw plane when the effective friction angle δ is the representative value δ0.
[0092]
Here, the flow factor ff represents the fluidity of the flow channel, and ff = σ1 / (σ1). The bar σ1 is the maximum principal stress acting along the arch in the powder layer. Therefore, the flow factor ff is a straight line passing through the origin on the σ1- (bar σ1) plane.
[0093]
Next, an intersection between the graph of the flow function FF on the σ1-fc plane and the graph of the flow factor ff on the σ1- (bar σ1) plane is determined, and the maximum principal stress σ1 at the intersection is determined (S49). Next, the effective friction angle δ at the determined maximum principal stress σ1 is determined using a graph of the effective friction angle δ with respect to the maximum principal stress σ1 represented in step S45, and this is set as a calculated value δ1 (S50). .
[0094]
Next, the steps S47 to S50 are repeated with the calculated value δ1 as the representative value δ0 until δ1 = δ0 ± 1.5, that is, until the calculated value δ1 matches the representative value δ0 within the allowable range (S51). .
[0095]
Next, the maximum principal stress bar σ1 at the intersection of the graph of the flow function FF and the graph of the flow factor ff obtained in step S49 is obtained (S52), and α = αc-4 (S53). The process in step S53 is performed at an angle 4 degrees narrower than the critical inclination angle αc in order to ensure that all powder is discharged by its own weight when the cross section of the discharge port of the storage tank is circular. Is set as the inclination angle α of the design value. When the cross section of the outlet of the storage tank is rectangular, the critical inclination angle αc is set as the designed inclination angle α.
[0096]
By the way, in the limit state in which the arch portion is held at the opening of the storage tank, the force balance equation γ × B × g = H (α) × (bar σ1) is established. Here, γ is the bulk density, g is the gravitational acceleration, and H (α) is a correction term in consideration of the inclination angles α at both ends of the arch portion. In the above balance type, the left side corresponds to the weight of the arch part, and the right side corresponds to the force required to support the arch part. Note that the above balance equation is an equation in the case of using an SI unit system as a unit of force. In the case of using a gravitational unit system such as a gram weight or a kilogram weight, it is necessary to omit the gravitational acceleration on the left side. .
[0097]
Therefore, in order to obtain the minimum opening diameter B, H (α) is obtained using α obtained in step S53 (S54), and H (α) obtained and (bar σ1) obtained in step S52 are obtained. B = (bar σ1) × H (α) / (γ × g) is calculated using the particle density Dens and the bulk density γ obtained from the porosity ε (S55).
[0098]
Here, H (α) can be obtained from the approximate expression H (α) = 2 + (α / 60) when the discharge part at the lower part of the storage tank is concentric, and the cross section of the discharge part is rectangular. In this case, it can be obtained from the approximate expression H (θ) = 1 + (θ / 180). Alternatively, H (α) can be obtained from a graph showing the relationship between the inclination angle α and H (α).
[0099]
Then, the inclination angle α obtained in step S53 and the minimum aperture B obtained in step S55 are output (S56), and the processing operation ends.
[0100]
It should be noted that the present invention is not limited to the embodiments described above, and various changes can be made within the scope shown in the claims, and are obtained by appropriately combining the technical means disclosed in the different embodiments. Embodiments are also included in the technical scope of the present invention.
[0101]
For example, in the above embodiment, the average particle size, geometric standard deviation, particle density, liquid content, and porosity are used as the basic physical properties of the powder, but other basic properties such as the powder aspect ratio are used. By adding target physical properties, the WS formula parameters can be estimated with higher accuracy.
[0102]
【The invention's effect】
As described above, the method for estimating the effective internal friction angle of the powder of the present invention is based on the measured data of the basic physical properties of the powder and the effective internal friction angle of the powder. A construction step of constructing an NN model defining a causal relationship between physical properties and an effective internal friction angle of the powder.
[0103]
Thus, by inputting the basic physical properties of the powder to the constructed NN model, the estimated data of the effective internal friction angle is output from the NN model. This has the effect of obtaining an effective internal friction angle.
[0104]
It is desirable that the basic physical properties include at least the liquid content of the powder.
[0105]
Further, it is desirable that the basic physical properties include at least the average particle size, particle density, liquid content, and porosity of the powder.
[0106]
Furthermore, it is desirable that the basic physical properties further include the standard deviation of the particle size of the powder.
[0107]
Further, the method for estimating the effective internal friction angle of the powder according to the present invention is, as described above, in the above method, inputting the basic physical properties of the powder to the NN model constructed by the construction step. A first table creation step of creating a first table consisting of a plurality of data selected for each of the basic physical properties by discretely selecting a plurality of data to be obtained; Calculating a data of the effective internal friction angle by inputting a data set of basic physical properties including data of physical properties into the NN model; and a basic physical property selectable from a first table. Repeating the calculating step until all of the data sets are selected; and the effective internal friction corresponding to each of the data sets of the basic physical properties. A second table creating step of creating a second table composed of the data of the above, and searching the first table and the second table to output data of the effective internal friction angle corresponding to the input data set of basic physical properties Creating a search program to perform the search.
[0108]
Thus, even if the computer does not have the dedicated software installed therein, by providing the first and second tables and the search program, it is possible to estimate the effective internal friction angle corresponding to the basic physical properties.
[0109]
In addition, it is desirable that the intervals of the data discretely selected in the first table creation step are based on the distribution of basic physical properties.
[0110]
In addition, the method for designing a powder storage tank according to the present invention includes, as described above, an effective fracture envelope determined from the effective internal friction angle estimated by the above method, and a fracture envelope and a wall fracture determined by a shear test. This is a method of obtaining the diameter and the inclination angle of the outlet in the powder storage tank using the envelope.
[0111]
This eliminates the need to perform a shear test when obtaining an effective destruction envelope, and thus has the effect of speeding up the design of a powder storage tank.
[0112]
Further, the device for estimating the effective internal friction angle of the powder of the present invention, as described above, the basic physical properties of the powder, the input means for inputting the effective internal friction angle of the powder, Based on the measured data of the basic physical properties of the powder and the effective internal friction angle of the powder input by the input means, the difference between the basic physical properties of the powder and the effective internal friction angle of the powder. Constructing means for constructing an NN model that defines the causal relationship of the above, and inputting the basic physical properties of the powder inputted by the input means to the NN model constructed by the constructing means, Estimating means for estimating the effective internal friction angle of the body.
[0113]
As a result, the effective internal friction angle of the powder can be estimated from the basic physical properties of the powder, so that the effective internal friction angle of the powder can be obtained without taking time and effort.
[0114]
Further, the device for estimating the effective internal friction angle of the powder according to the present invention, as described above, in the above configuration, inputs each of the basic physical properties of the powder to the NN model constructed by the construction means. A first table creating means for creating a first table including a plurality of data selected for each of the basic physical properties by discretely selecting a plurality of data to be obtained; and a basic table selected from the first table. By inputting a data set of basic physical properties comprising data of physical properties to the estimating means, the data of the effective internal friction angle is obtained from the estimating means, and the data can be selected from the first table. By repeating until all of the basic physical property data sets are selected, a second table including the effective internal friction angle data corresponding to each of the basic physical property data sets is obtained. And a search program for outputting a data set of the effective internal friction angle corresponding to the input data set of basic physical properties by searching the first table and the second table. This is a configuration further including a search program creation unit.
[0115]
Thus, even if the computer does not have the dedicated software installed therein, by providing the first and second tables and the search program, it is possible to estimate the effective internal friction angle corresponding to the basic physical properties.
[0116]
In addition, the powder storage tank designing device of the present invention is, as described above, an effective fracture envelope determined from the effective internal friction angle estimated by the device having the above configuration, and a fracture envelope and a wall determined by a shear test. In this configuration, the diameter and the inclination angle of the outlet in the powder storage tank are obtained using the destruction envelope.
[0117]
This eliminates the need to perform a shear test when obtaining an effective destruction envelope, and thus has the effect of speeding up the design of a powder storage tank.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an outline of a storage tank design system for powder according to an embodiment of the present invention.
FIG. 2 is a diagram showing, in a table format, statistics of actually measured data regarding basic physical properties of powder used for constructing an NN model and parameters of a WS formula.
FIG. 3 is a diagram showing, in a table format, a comparison between an estimated value calculated by an NN model constructed using various powders and an actually measured value.
FIG. 4 is a diagram showing, in a table format, the accuracy of an estimated value with respect to an actually measured value when an NN model is constructed by excluding one or two input variables.
FIG. 5 is a flowchart showing a processing operation of the NN model construction device.
FIG. 6 is a diagram showing, in a table form, an example of a basic physical property data table created by the NN model construction device.
FIG. 7 is a flowchart showing a processing operation of a search program created in the NN model construction device.
FIG. 8 is a schematic view showing a search sheet and a search button displayed on a display of the storage tank designing apparatus. FIG. 8A shows a state before the search is executed, and FIG. The state of is shown.
FIG. 9 is a schematic diagram showing a storage tank designed by the storage tank designing device.
10 is a flowchart showing a processing operation for obtaining a minimum diameter B and an inclination angle α of the storage tank shown in FIG.
11 is a flowchart showing a processing operation for obtaining a minimum diameter B and an inclination angle α of the storage tank shown in FIG. 9;
FIG. 12 is a graph showing a correspondence between a normal stress and a shear stress obtained by a shear test.
[Explanation of symbols]
10 Measurement data acquisition device
11 NN model construction device
12 Storage tank design equipment
30 storage tanks
B Minimum diameter of storage tank outlet
α Angle of inclination of storage tank outlet

Claims (10)

A method for estimating an effective internal friction angle of a powder,
Neural network that defines a causal relationship between the basic physical properties of the powder and the effective internal friction angle of the powder based on measured data of the basic physical properties of the powder and the effective internal friction angle of the powder. A method for estimating an effective internal friction angle of a powder, comprising a construction step of constructing a network model. The method according to claim 1, wherein the basic physical property includes at least a liquid content of the powder. The method according to claim 2, wherein the basic physical properties further include an average particle diameter, a particle density, and a porosity of the powder. The method according to claim 3, wherein the basic physical property further includes a standard deviation of a particle size of the powder. For each basic physical property of the powder, by discretely selecting a plurality of data that can be input to the neural network model constructed in the constructing step, it is composed of a plurality of data selected for each basic physical property. A first table creation step of creating a first table;
Calculating a data of the effective internal friction angle by inputting a data set of basic physical properties including data of the basic physical properties selected from the first table to the neural network model;
Repeating the calculating step until all of the selectable data sets of basic physical properties are selected from the first table;
A second table creating step of creating a second table including the data of the effective internal friction angle corresponding to each data set of the basic physical properties;
Creating a search program that outputs the effective internal friction angle data corresponding to the input data set of basic physical properties by searching the first table and the second table. Item 1. The method for estimating an effective internal friction angle of a powder according to Item 1. The method for estimating an effective internal friction angle of a powder according to claim 5, wherein intervals of data discretely selected in the first table creation step are based on distribution of basic physical properties. 7. A method using an effective fracture envelope determined from an effective internal friction angle estimated by the method according to any one of claims 1 to 6, and a fracture envelope and a wall fracture envelope determined by a shear test. A method for designing a powder storage tank, comprising determining an outlet diameter and an inclination angle of a body storage tank. A device for estimating the effective internal friction angle of the powder,
Input means for inputting the basic physical properties of the powder and the effective internal friction angle of the powder,
Based on the measured data of the basic physical properties of the powder and the effective internal friction angle of the powder input by the input means, the basic physical properties of the powder and the effective internal friction angle of the powder are compared. Construction means for constructing a neural network model that specifies a causal relationship between
Estimating means for estimating an effective internal friction angle of the powder by inputting the basic physical properties of the powder input by the input means to the neural network model constructed by the constructing means. An apparatus for estimating an effective internal friction angle of a powder, characterized in that: For each basic physical property of the powder, by discretely selecting a plurality of data that can be input to the neural network model constructed by the constructing means, it is composed of a plurality of data selected for each basic physical property. First table creation means for creating a first table;
By inputting a basic physical property data set consisting of data of each basic physical property selected from the first table to the estimating means, the data of the effective internal friction angle is obtained from the estimating means, and A second table comprising the data of the effective internal friction angle corresponding to each data set of the basic physical properties by repeating until all of the data sets of the basic physical properties which can be selected from the first table are selected. Second table creation means for creating
A search program creating means for creating a search program that outputs a data set of an effective internal friction angle corresponding to the input data set of basic physical properties by searching the first table and the second table. The apparatus for estimating an effective internal friction angle of a powder according to claim 8, wherein: An outlet in a powder storage tank using an effective fracture envelope determined from an effective internal friction angle estimated by the apparatus according to claim 8 or 9, and a fracture envelope and a wall fracture envelope determined by a shear test. An apparatus for designing a storage tank for a powder, wherein a diameter and an inclination angle of the powder are determined.

Images