The present invention relates to a computerized method for creating a two-dimensional or three-dimensional finite element model of a rubber composite comprising a rubber matrix and fillers, capable of improving the accuracy of computer simulations.
In recent years, a finite element method is widely employed in various computer simulations. In such computer simulation, an analysis object is discretized into a finite number of elements, and a finite element model is created by the use of the elements. The finite element model is provided with various characteristics, and a deformation calculation is made by calculating displacements of nodes of the elements.
In the case that the analysis object is a rubber composite (b) comprising a rubber matrix (c) and fillers (d) dispersed therein, a finite element model (a) including a model (g) of the rubber matrix (c) and models (h) of the fillers (d) is defined, for example, by using rectangular elements (e) as shown in
In such finite element model (a), due to the rectangles, the surface of a filler model (h) has remarkable concavity and convexity (i) as shown in
Therefore, in the part (or elements) of the rubber matrix model abutting the filler models (h), the rigidity of the rubber matrix model (g) is calculated as if it is higher than it really is, and thereby the simulation accuracy is decreased. This is also true in the case of a three-dimensional model using hexahedron elements as shown in
It is therefore, an object of the present invention to provide a computerized method for creating a finite element model of a rubber composite capable of improving the accuracy of computer simulations.
According to one aspect of the present invention, the computerized method is for creating a two-dimensional finite element model of a rubber composite and comprises
creating a primary finite element model of the rubber composite by defining a rubber matrix model of the rubber matrix and filler models of the fillers by the use of rectangular elements having square shapes and/or oblong shapes, and
creating a second finite element model as the two-dimensional finite element model by moving boundary nodes positioned at boundaries between the rubber matrix model and filler models in the primary finite element model so that concavity and convexity of the surface of each filler model becomes less, whereby, in the second finite element model, some of the rectangular elements including the moved boundary nodes are deformed. and the rest of the rectangular elements are not deformed.
According to another aspect of the present invention, the computerized method is for creating a three-dimensional finite element model of a rubber composite and comprises
creating a primary finite element model of the rubber composite by defining a rubber matrix model of the rubber matrix and filler models of the fillers by the use of hexahedron elements, and
creating a second finite element model as the tree-dimensional finite element model by moving boundary nodes positioned at boundaries between the rubber matrix model and filler models in the primary finite element model so that concavity and convexity of the surface of each filler model becomes less, whereby, in the second finite element model, some of the hexahedron elements including the moved boundary nodes are deformed. and the rest of the hexahedron elements are not deformed.
Therefore, in the computerized methods according to the present invention, the surfaces of the filler models or the boundaries between the filler models and rubber matrix model become smoother and the contact range (area) becomes decreased, As a result, the restraint by the filler models is reduced, and the overestimation of the rigidity in the abutting part is prevented. Thereby, the simulation accuracy can be improved. Further, in the finite element models created by the methods according to the present invention, some of the elements are deformed, but the rest of (most of) the elements are not deformed, therefore, the processing time can be greatly reduced in comparison with a finite element model, all elements of which are reduced in the sizes in order to smoothen the surfaces or boundaries.
a) shows a two-dimensional space.
b) is a closeup thereof.
a) is a closeup of
b) is a closeup of
a) and
a) shows hexahedron elements extracted from the model shown in
b) shows deformed hexahedron elements extracted from the model shown in
a) shows a conventional finite element model.
b) is a closeup thereof.
Embodiments of the present invention will now be described in detail in conjunction with the accompanying drawings.
In the method according to the present invention, a two-dimensional or three-dimensional finite element model (2, 20) of a rubber composite (b) comprising a rubber matrix (c) and fillers (d) dispersed therein is created by the use of a computer 1 in order to simulate deformation of the rubber composite (b) by the use of a computer 1.
As to the simulation method, various methods may be employed regardless of known or unknown types, therefore, the description is omitted here.
The fillers (d) are fine particles having rigidity higher than the rubber matrix (c), for example, carbon black, silica and the like, and each filler (d) is treated as having a round shape. In the case of two-dimensional model, the “round shape” means a circle, oval and the like. In the case of three-dimensional model, the “round shape” means a sphere, spheroid and the like.
As for the computer 1, a general-purpose computer such as workstation can be used. For example, as shown in
firstly, a primary finite element model 2A, 20A including a rubber matrix model 4, 22 of the rubber matrix (c) and a plurality of filler models 5, 23 of the fillers (d) is created by the use of finite elements 6, 24 (first step S1), and then, based on the primary finite element model 2A, 20A, by deforming only elements 6, 24 which include boundary nodes 7A, 25A at the boundaries between the filler models and the rubber matrix model, a second finite element model 2B, 20B is created as the target model 2, 20 (second step S2).
In the method according to the first embodiment of the present invention, a two-dimensional finite element model 2 (as shown in
In this embodiment, the rubber composite is treated as a two-dimensional space 3 for example having a square shape of which each side has a length L1 of about 50 to 150 nm.
The two-dimensional space is meshed with rectangular elements 6 having square shape and/or oblong shape (in this example square elements only). Thus, each element 6 has four nodes 7 and four straight sides 8 extending between the four nodes 7. The length L2 of each side 8 is for example set in a range of about 0.1 to 3 nm.
In the first step S1 in this embodiment, as shown in
Specifically, the contours 9 (or their coordinates stored beforehand) of the fillers (d) are superimposed on the elements 6 (or their coordinates) in the form of mesh as shown in
Then, the filler model 5 of each filler (d) is defined by some of the elements 6 whose centroids 6c are within the contour 9 of the filler (d) concerned as shown in
And by the rest of the elements 6, the rubber matrix model 4 of the rubber matrix (c) is defined.
The computer 1 stores the data such as coordinates of each element in the filler's domain or rubber's domain in the memory. Thereby, as shown in
In the second step S2 in this embodiment, with the computer 1, some boundary nodes 7A (shown in
Specifically, as shown in
By the movements of the boundary nodes 7A at the boundaries, only the elements 6 including the boundary nodes 7A are deformed, and the second finite element model 2B as shown in
Then, the computer 1 updates the coordinates of the respective elements 6 including the moved boundary nodes 7A.
In the second finite element model 2B, therefore, the concavity and convexity 13 of the surface of the filler model 5 is lessened. Further, the rest of the elements 6 none of the nodes 7 of which are positioned at the boundaries are not deformed or each maintain the same shape as in the primary finite element model 2A (in this embodiment, square shape). Thus, it is not necessary to update the coordinates, therefore, the processing time can be greatly reduced.
On each element 6 of the rubber matrix model 4 and filler models 5 of the second finite element model 2B, information necessary for numerical analyses made by a computer simulation is defined. The numerical analyses means a numerical analysis method such as finite element method.
The information necessary for the analyses includes at least the coordinates and indexes (identification numbers) of the nodes 7 of each element 6.
Further, on each element 6 of the rubber matrix model 4 and filler models 5 of the second finite element model 2B, material characteristics (physical properties) of the part corresponding to the concerned element 6 are defined. Specifically, material constants corresponding to the physical properties of the rubber matrix (c) and
material constants corresponding to the physical properties of the filler (d) are respectively defined.
The computer 1 stores these data in the memory.
On the other hand, there is a possibility that, as shown in
In this case, if the boundary nodes 7A are moved to the intersecting point 12 of the straight line 11 and the contour 9 of the filler (d) in the second step S2, then these boundary nodes 7A are overlapped.
Such overlap of the nodes crushes the element 6 for example from the quadrilateral to a triangle, and it becomes impossible to compute the rigidity.
Therefore, in order to avoid such undesirable overlap, before moving the boundary nodes 7A at the boundaries to the intersecting points 12 in the second step S2,
the computer 1 checks, with respect to each boundary, whether a plurality of the boundary nodes 7A are located on one straight line 11 and there is a possibility of their overlap. (Overlap check step)
If it is decided that there is a possibility of overlap, as shown in
the distances L3 of the boundary nodes 7A at the boundary concerned from the intersecting point 12 are calculated, and the distances L3 are modified by multiplying a coefficient in a range of more than 0 and less than 1 in order to obtain the modified distances L3′. (Distance modify step)
Then, the boundary nodes 7A are moved toward the intersecting point 12 to different positions on the straight line 11 by the respective modified distances L3′ in stead of the original distances L3. Accordingly, as shown in
Thus, the target two-dimensional finite element model is created. As a result, the accurate simulation is possible.
For example, the rubber composite (b) shown in
a stress-elongation curve in solid line obtained through a computer simulation utilizing the second finite element model 2B in this embodiment (Embodiment 1) shown in
a stress-elongation curve in alternate long and two short dashes line obtained through the computer simulation utilizing a finite element model (a) (comparative example 1) shown in
In the method according to the second embodiment of the present invention, a three-dimensional finite element model 20 (as shown in
In this embodiment, the rubber composite is treated as a three-dimensional space 21 where the fillers (d) are three dimensionally dispersed in the rubber matrix (c).
For example, the three-dimensional space 21 is defined as a cube of which each side has a length L6 of about 50 to 150 nm as shown in
The three-dimensional space 21 is meshed with hexahedron element 24. Each hexahedron element 24 has, as shown in
In the first step S1 in this embodiment, as shown in
Then, the filler model 23 of each filler (d) is defined by some of the hexahedron elements 24 whose centroids 24c are within the contour 27 of the filler (d) concerned.
And by the rest of the elements 24, the rubber matrix model 22 of the rubber matrix (c) is defined.
The computer 1 stores the data such as coordinates of each element in the filler's domain or rubber's domain in the memory. Thereby, as shown in
In the second step S2 in this embodiment, as shown in
some boundary nodes 25A of the nodes 25 located at the boundaries between the rubber matrix model 22 and the filler models 23 of the primary finite element model 20A are moved so that the concavity and convexity of the surface of each filler model 23 becomes less to thereby create or redefine the second finite element model 20B (shown in
Specifically, as shown in
Then, the computer 1 updates the coordinates of the deformed elements of the rubber matrix model 22 and the filler models 23 of the second finite element model 20B in the memory.
By the movements of the boundary nodes 25A at the boundaries, only the hexahedron elements 24 including the boundary nodes 25A are deformed, and there is defined the second finite element model 20B where the concavity and convexity of the surface of each filler model 23 is lessened as shown in
On each hexahedron element 24, information necessary for numerical analyses made by a computer simulation such as the coordinates and indexes (identification numbers) of the nodes 25, material constants corresponding to the physical properties of the rubber matrix (c) and material constants corresponding to the physical properties of the filler (d) are respectively defined. The computer 1 stores these data in the memory.
As explained in the former embodiment, when a plurality of the boundary nodes 25A at the boundary are located on one straight line 29, there is a possibility that these boundary nodes 25A are overlapped each other in the second step S2.
In order to avoid such overlap, before moving the boundary nodes 25A at the boundaries to the intersecting points 30 in the second step S2, similarly to the former embodiment, the computer 1 checks, with respect to each boundary, whether a plurality of the boundary nodes 25A are located on one straight line 29 and there is a possibility of their overlap. (Overlap check step)
If it is decided that there is a possibility of overlap, with the computer 1, the distances of the boundary nodes 25A at the boundary concerned from the intersecting point 30 are calculated, and the distances are modified by multiplying a coefficient in a range of more than 0 and less than 1 in order to obtain the modified distances. (Distance modify step)
Then, the boundary nodes 25A at the boundary concerned are moved toward the intersecting point 30 to different positions on the straight line 29 by the respective modified distances in stead of the original distances.
Then, the computer 1 updates the coordinates of the deformed elements as explained above.
Such overlap check step and distance modify step are carried out on all of the possible boundary nodes 7A.
Thus, the target the three-dimensional finite element model is created.
In
a stress-elongation curve in solid line obtained through a computer simulation utilizing the second finite element model 20B in this embodiment (Embodiment 2) shown in
a stress-elongation curve in alternate long and two short dashes line obtained through the computer simulation utilizing the primary finite element model 20A (as comparative example 2) shown in
These simulation results show that the simulation accuracy of Embodiment 2 is superior to that of the comparative example 2.
Number | Date | Country | Kind |
---|---|---|---|
2011-049309 | Mar 2011 | JP | national |
Number | Date | Country |
---|---|---|
2005-351770 | Dec 2005 | JP |
2007-271369 | Oct 2007 | JP |
2009-193339 | Aug 2009 | JP |
2009-276147 | Nov 2009 | JP |
2009-282569 | Dec 2009 | JP |
Number | Date | Country | |
---|---|---|---|
20120232848 A1 | Sep 2012 | US |