Patent Application
20110313742
The present invention relates to a method for computer simulation of an article made of a viscoelastic material, wherein when the article model is deformed, the viscoelastic property already defined for elements constituting an article model are renewed correspondingly to the strain amounts obtained when the article model is deformed.
In recent years, in order to save time and cost for developing various articles made of viscoelastic materials as typified by high-polymer materials such as rubber and elastomer (for example, vehicle tires, balls used in sports, rollers used in printing machines, etc.), computerized deformation simulations have become widely employed to evaluate various performances and/or characteristics, for example, dynamic performances of a tire during running, rebound characteristics of a golf ball at impact and the like.
In the computerized deformation simulations, the article as an analysis object is modeled by a finite number of elements to define an article model, and values of a viscoelastic property are defined for the elements correspondingly to the materials represented by the elements, then deformation simulation or deformation calculation is carried out on the article model under specified conditions.
In general, when a viscoelastic material is deformed and the strain amount of the viscoelastic material is changed, then values of a viscoelastic property thereof also makes a change, which means that the viscoelastic property has dependency on strain amount as well as dependency on strain rate.
In the heretofore known simulation algorithms, however, such strain amount dependencies are not taken into consideration. In other words, the elements of the article model are provided with fixed values of the viscoelastic property determined at a given strain amount. Accordingly, the heretofore known simulation algorithms have their own limits in the betterment of simulation accuracy.
It is therefore, a primary object of the present invention to provide a method for simulating an article made of a viscoelastic material by which the article is simulated with accuracy and thereby it is possible to accurately estimate performances and/or properties of the article.
In this invention, the term “article” means every thing having a definite shape as the analysis object. In particular, finished products supplied to the users (for example, golf ball, rubber roller, tire, etc.) as well as unfinished products including their component parts (for example, a core of a golf ball and the like) are included.
The article is made of one or more kinds of viscoelastic materials.
The term “viscoelastic material” means every material having viscosity and elasticity. For example, thermoplastic resins, thermo-setting resins, various elastomers, rubber and the like are included.
The viscoelastic property include at least the relation between stress σ and strain ε of the viscoelastic material.
The values of the viscoelastic property can be determined from measurement of the viscoelastic material, using various test procedures, for example, creep test, dynamic viscoelastic property test, split Hopkinson bar test, relaxation test, ultrasonic test and the like.
According to one aspect of the present invention, a method for computerized simulation of an article made of a viscoelastic material comprises:
a step of defining an article model of the article made with a finite number of elements;
a step of defining the value of the viscoelastic property for each of the elements of the article model;
a step of making a deformation simulation of the article model;
the following three steps carried out at least once:
a step of obtaining a strain amount of each of the elements from the deformation simulation,
a step of renewing the value of the viscoelastic property according to the obtained strain amounts so as to redefine the article model, and
a step of again making the deformation simulation on the redefined article model; and
a step of outputting a simulation result of the deformation simulation.
Therefore, the strain amount dependency of the viscoelastic material is reflected in deformation calculation of each of the elements of the article model, and highly accurate simulations become possible.
According to another aspect of the present invention, a method for computerized simulation of an article made of a viscoelastic material comprises:
a step of defining an article model of the article made with a finite number of elements;
a step of defining the value of the viscoelastic property for each of the elements of the article model;
a step of making a deformation simulation of the article model; and
a step of outputting a simulation result of the deformation simulation, wherein
the value of the viscoelastic property defined for each of the elements is that obtained from a previous measurement of the viscoelastic material at a prospective strain amount expected in the deformation simulation.
Therefore, the strain amount dependency of the viscoelastic material is reflected in deformation calculation of each of the elements of the article model, and highly accurate simulations become possible.
a) and
Embodiments of present invention will now be described in detail in conjunction with accompanying drawings.
First, an article model is defined. The article model is a model of an article modeled by a finite number of elements (e) so that numerical analysis method, e.g. for example finite element method and the like can be used to make calculations.
In order to reduce the computational cost, it is possible to reduce the number of the elements by utilizing a geometrical symmetry of the article. In the case of
The article model 2 is entered and stored in the computer 1. More specifically, numerical data on indexes of the elements (e) constituting the article model 2 and coordinate values of node points thereof are entered in the storage device of the computer 1. This operation is performed by an operator through the input devices.
The golf ball model 2a is made up of three-dimensional hexahedral elements, most of the node points of which are coupled with one another. Thus, the spherical golf ball is modeled as a pseudo-spherical polyhedron. Incidentally, it is also possible that the golf ball model 2a is made with other types of elements than the above-mentioned hexahedral elements.
The golf ball as the analysis object is made of viscoelastic materials only, and composed of a spherical core and a thin cover surrounding the core. The core and cover are made of different viscoelastic materials suitable for the respective functions.
<step S2>
Secondly, for each of the elements (e) of the article model 2 (golf ball model 2a), a value of a viscoelastic property of the viscoelastic material which the element represents is defined.
The viscoelastic material has a strain amount dependency and shows different stress-strain relation and hysteresis loss depending on the strain amount (maximum strain amount).
In
Each of the stress-strain curves loops, and the area enclosed in the loop corresponds to the hysteresis loss.
As for how to define the value of the viscoelastic property (initial value) for each of the elements (e) of the article model 2 (golf ball model 2a), two examples follow.
For each of the materials used in the article, a value of the viscoelastic property of the material measured at a certain strain amount is defined.
In the case of the golf ball, for each of the elements of the golf ball model 2a constituting the cover, a value of the viscoelastic property of the viscoelastic material of the cover measured at a certain strain amount x is defined.
And for each of the elements of the golf ball model 2a constituting the core, a value of the viscoelastic property of the viscoelastic material of the core measured at the same strain amount x as in the cover is defined.
For each of the elements of the golf ball model 2a, a value of the viscoelastic property corresponding to a strain amount which is expected in the undermentioned deformation simulation is defined.
For example, in the case of a golf ball, the strain amount in each part of the golf ball may be assumed from previously obtained actual test data.
Therefore, if the deformation simulation is for a rebound characteristic simulation in which the golf ball model is hit by a golf club model at a certain hitting speed, a value of the viscoelastic property measured at the expected strain amount is used as the initial value of the viscoelastic property of the golf ball material.
As explained, the values of the viscoelastic property are defined for the elements of the golf ball model 2a (article model 2) corresponding to the core and cover, and entered and stored in the computer 1. Thereby, the article model 2 is defined.
In this relation, it is preferable that the measuring of the viscoelastic material is carried out at a strain rate corresponding to the deformation rate in the deformation simulation, whereby it is possible to include a strain rate dependency in the initial value of the viscoelastic property.
Thirdly, based on a finite element method, a deformation simulation is made on the article model 2 (golf ball model 2a).
The deformation simulation includes various kinds of simulations.
For example, the deformation simulation can include a rebound evaluation simulation in which rebound characteristics of a golf ball are evaluated.
Aside from such dynamic simulations, static simulations (for example, a compression test in which the golf ball model 2a or article model 2 is compressed to a plane by a specified force) are also included in the deformation simulation.
In the deformation simulation, deformed state of the article model 2 is calculated at small time intervals from the beginning of the deformation (beginning of the hit of the golf ball model 2a), and the necessary historical data on the change in the physical value are stored in the computer 1.
In the course of the deformation calculation, in order to duplicate the viscoelastic property of each of the elements, a viscoelasticity model can be employed in the deformation calculation of the element. Maxwell model, Kelvin-Voight model and generalized Maxwell model are typical examples of the viscoelasticity model which can be used suitably.
This kind of the deformation simulation can be made, for example, by using a general-purpose explicit solver software “LS-MYNA” of Livermore Software Technology Corp. on the computer.
b) shows a visualization of the deformed state at a certain moment during the impact of the golf ball model 2a and impact tool model 3.
As shown in this figure, in each of the core and cover of the golf ball model 2a, the strain amount becomes larger on the impact tool model 3 side, but becomes smaller on the opposite side. In actuality, the strain amount is broken up into several levels to indicate in different colors, therefore, the level of the strain amount is readily recognizable.
As readily understandable from the above explanation, if the fixed value of the viscoelastic property is uniformly defined for the elements of the cover and the fixed value of the viscoelastic property is uniformly defined for the elements of the core, then it is impossible to make highly accurate simulation due to the strain amount dependency.
All of the elements (e) constituting the article model 2 (golf ball model 2a) are classified into a plurality of groups based on their strain amount. This process is automatically carried out by the computer 1.
With respect to each of the elements (e), the strain amount during one deformation cycle from the beginning of the deformation to the dissipation of the deformation is determined from the above-mentioned deformation simulation.
As to the strain amount, the maximum value of the strain amount of each element (e) during one deformation cycle is used in this embodiment.
However, according to the needs of the deformation simulation to be made, an average value of the strain amount over one deformation cycle or a value of the strain amount at a certain moment from the beginning of the deformation can be used instead of the maximum value of the strain amount.
As to the direction of the strain, tensile strain, compressive strain, and/or shear strain can be used according to the needs of the deformation simulation to be made.
In this embodiment, according to the strain amount ε, the elements are classified into five groups G1-G5 as follows.
group G1: 0%≦ε<10%
group G2: 10%≦ε<20%
group G3: 20%≦ε<30%
group G4: 30%≦ε<40%
group G5: 40%≦ε<50%
<Step s5>
The difference between the value of the viscoelastic property already defined for each of the elements in step 2 and the value of the viscoelastic property at the strain amount determined from the deformation simulation in Step 3 is evaluated, and
if needed, the already defined value of the viscoelastic property is renewed by the newly determined appropriate value of the viscoelastic property including the strain amount dependency, and redefined for the same element.
With respect to each of the viscoelastic materials, the values of the viscoelastic property are measured in advance changing the strain amount, and data on the relations between the strain amount and the viscoelastic property are stored in the computer beforehand. Therefore, the appropriate value of the viscoelastic property can be retrieved based on the strain amount obtained from the deformation simulation in step 3.
The obtained strain amount of the elements in each of the groups G1-G5 has a certain range.
Therefore, in this embodiment, for all of the elements in each of the groups G1-G5, an identical strain amount is used to retrieve the appropriate value(s) of the viscoelastic property.
As to the identical strain amount, preferably the intermediate value in the strain amount range is used.
For example, in the case of the group G1, the strain amount ranges from 0 to 10%, therefore, the intermediate value of 5% is preferably used. Namely,
5% is used in group G1,
5% is used in group G2,
25% is used in group G3,
35% is used in group G4,
45% is used in group G5.
It is not necessarily the case that all of the elements in one group represent an identical viscoelastic material.
For example, the group G2 includes the elements of the core and the elements of the cover are included.
Accordingly, in the group G2, for the elements of the core, the value of the viscoelastic property of the core material measured at 15% strain amount is redefined, and for the elements of the cover, the value of the viscoelastic property of the cover material measured at 15% strain amount is redefined. Namely, in group G1, value(s) at 5% strain amount is(are) redefined, in group G2, value(s) at 15% strain amount is(are) redefined, in group G3, value(s) at 25% strain amount is(are) redefined, in group G4, value(s) at 35% strain amount is(are) redefined, in group G5, value(s) at 45% strain amount is(are) redefined.
<step S6>
In this embodiment, on the article model 2 (golf ball model 2a) of which values of the viscoelastic property have been redefined, the above-mentioned deformation simulation is again carried out under the same simulation conditions as those in the previous deformation simulation.
In order to obtain highly-accurate final calculation results, a convergence calculation process is made by repeating the step S5 and step s6 at least once or until the difference between a physical quantity obtained from the present deformation simulation and that obtained from the most recent deformation simulation becomes within the predetermined acceptable range.
In another way, it is possible to end the convergence calculation when the difference between the distribution of the strain amount obtained from the present deformation simulation and the distribution of the strain amount obtained from the most recent deformation simulation becomes within the predetermined acceptable range.
The physical quantity may be all kinds of physical quantity necessary for evaluating the performance and/or characteristic are obtained, for example, the strain amount of each element, compressibility, coefficient of restitution and the like.
For example, in the case that the deformation simulation is for a rebound evaluation simulation in which the coefficient of restitution is evaluated, it is preferable that the steps S5 and S6 are repeated until the difference between the coefficient of restitution obtained from the most recent rebound evaluation simulation and the coefficient of restitution obtained from the rebound evaluation simulation after the values of the viscoelastic property has been redefined becomes not more than 10%, more preferably not more than 5%, still more preferably not more than 3%.
<step S8>
From the last deformation simulation, the physical quantities are obtained, and the obtained quantities are outputted in an appropriate form, for example, as visual information such as printed or displayed graph, image, characters, etc., and/or as electrical data or signals to be used in the subsequent data processing.
As explained, in the second or subsequent deformation simulation, since the calculation results include the influence of the strain amount dependency of the viscoelastic property (in this embodiment, influence of the strain rate dependency is also included), in comparison with the deformation simulation carried out in the previous step S3, more accurate calculation results can be obtained.
In order to introduce the strain amount dependency into the values of the viscoelastic property of each of the elements with precision, it is desirable to increase the number of the groups. In this light, the number of the groups is preferably not less than 3, more preferably not less than 5, still more preferably not less than 10. However, if the groups are too many, the computational time and cost increases, therefore, it is preferable that the number of the groups is not more than 20.
In order to confirm the effects of the simulation method according to the present invention, with respect to the following four kinds of two-piece golf balls A-D, ball-hitting simulation in order to evaluate the amount of spin and rebound velocity was carried out, and further, rebound evaluation simulation in order to evaluate the coefficient of restitution of the core was carried out.
The golf balls A-D each had such a hardness distribution such that the hardness is smoothly increased from the center of the core towards the outside. The values of JIS-C hardness measured at the distances of 0, 5, 10 and 15 mm from the center of the core are shown in the table.
Golf ball models A-D of the golf balls A-D were defined according to their structures.
Each of the core model and the cover model was a generalized Maxwell model.
For each of the elements of the core model, the value of the viscoelastic property (initial value) of the viscoelastic material measured at 15% strain amount was defined.
For each of the elements of the cover model, the value of the viscoelastic property (initial value) of the viscoelastic material measured at 5% strain amount was defined.
The values of the viscoelastic property was determined through a relaxation test using a slab of the viscoelastic material.
[Rebound Evaluation Simulation}
Actual rebound test was carried out according to the initial velocity test procedure of R&A Rules Limited, wherein the core at rest was hit by an aluminum cylindrical body of 198 grams at a velocity (V1) of 40 m/s, and the velocity (V1') of the cylindrical body and the velocity (V2′) of the golf ball after the hit were measured.
From the measured velocities, the coefficient of restitution of the core was determined through the calculation of the following expression:
Coefficient of restitution (COR)=(V2′−V1′)/V1.
Then, the rebound evaluation simulation was carried out under the same testing conditions as above.
In the rebound evaluation simulation, as shown in
wherein the core model was made with hexahedral solid elements the number of which was 15200, and
the impact tool model 3 was made with tetrahedral solid elements the number of which was 39119.
[Ball-Hitting Simulation]
Actual rebound test was carried out as follows.
Using a golf club (driver “XXIO 4” manufactured by SRI Sports Limited, shaft flex R, loft angle 11 degrees) mounted on a swing robot, the golf balls were hit at head speed of 21 m/s, 40 m/s and 50 m/s, and the amount of spin and rebound velocity of the golf ball immediately after the impact were measured.
Further, data on the motion of the club head immediately before the impact (e.g. blow angle, loft angle and shaft inclination angle) and the ball hitting position in the club face were measured to utilize these in the simulation.
Then, the ball-hitting simulation was carried out under the same testing conditions as above.
In the ball-hitting simulation, as shown in
The coefficient of friction between the club head model 4 and the golf ball model 2a was set to 0.15.
In the ball-hitting simulation, the club head model 4 was moved according to the three-dimensional motion measured in the above-mentioned actual rebound test immediately before the impact.
The solver used was LS-DYNA ver.971R4.2.
The computer used was Altix XE340 (CPU: two parallel Intel(R) Xeon(R) processors, Memory: 16 GB) manufactured by SGI Japan Ltd.
The calculation time of the rebound evaluation simulation was about 20 minutes.
The calculation time of the ball-hitting simulation was about 100 minutes.
The maximum value εm of the strain amount of each of the elements was pulled up from the time history of the strain amount of each of the elements calculated in the first simulation.
And according to the maximum value εm of the strain amount, the elements were classified into groups and the values of the viscoelastic property were redefined as follows.
<Core> group (0%≦εm<7.5%): value at 3% strain amount
In Tables 3-9, together with the results of the actual tests, the results in the first simulations are shown as “Reference” where the strain amount dependency was not taken into consideration, and the results in the second simulations are shown as “Embodiment” where the strain amount dependency was taken into consideration.
Further, in
From the test results it was confirmed that, in both of the rebound evaluation simulation and ball-hitting simulation, the test results of the second simulation corresponding to a method according to the present invention, show higher correlation coefficients than those in the first simulation corresponding to a conventional simulation method.
Therefore, according to the present invention, it is possible to quantitatively predict performances and/or properties of the article with accuracy which is very important for the simulations for the purpose of development of articles.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2010-140838 | Jun 2010 | JP | national |
