US 20030125644 A1 Abstract A numerical model related to fluttering of an insect, when an equivalent model of actual structure of a wing of the insect is moved in the air in accordance with a model of fluttering motion of the wing of the insect is calculated by fluid-structure interactive analysis, in which behavior of the wing and behavior of the air are given as numerical values, including interaction therebetween. Thereafter, a method of controlling a fluttering robot, wing shape and the like are determined by modifying numerical models of fluttering of an insect prepared by fluid-structure interactive analysis, in accordance with sensitivity analysis. Accordingly, a method of preparing numerical models of wing and air considering the behavior of the wing of the insect in the air is provided and, in addition, a method of manufacturing a fluttering robot utilizing the numerical model prepared by this method of preparing numerical model can be provided.
Claims(23) 1. A method of preparing a fluid-structure interactive numerical model, for preparing, when an actual structure as a living organism performs a prescribed motion in a fluid, a numerical model related to said fluid and a numerical model related to said structure, comprising:
actual structure measuring step of measuring physical values related to an actual structure of said structure; the step of preparing equivalent numerical model of actual structure that can be regarded as equivalent to said actual structure, in which the physical values related to said actual structure measured in said actual structure measuring step are given as numerical values; manner of motion measuring step causing said actual structure make a prescribed motion and measuring physical values related to a manner of said prescribed motion; motion model preparing step of preparing a motion model in which the physical values related to said manner of prescribed motion are given as numerical values; and fluid-structure interactive analysis step of calculating, when said equivalent numerical model of actual structure performs said prescribed motion represented by said motion model in a pre-set virtual fluid for analysis, a numerical model related to the fluid of said virtual fluid (numerical fluid model) and a numerical model related to the structure of said equivalent numerical model of actual structure, by performing fluid-structure interactive analysis in which behavior of said fluid and behavior of said structure are given as numerical models including interaction therebetween. 2. The method of preparing fluid-structure interactive numerical model according to said structure includes a first structure and a second structure different from but of the same type as said first structure; said actual structure measuring step includes
the first measuring step of measuring physical values related to the actual structure of said first structure, and
the second measuring step of measuring physical values related to actual structure of a featured portion of second structure;
said step of preparing equivalent numerical model of actual structure includes
the step of preparing a numerical model of detailed figure in which physical values related to the actual structure of said first structure measured in said first measuring step are given as numerical values, and
converting step of preparing said equivalent model of actual structure, by performing a prescribed conversion on said numerical model of detailed figure, using the physical values related to the actual structure of featured portion of said second structure measured in said second step; and
in said manner of motion measuring step, physical values related to the manner of prescribed motion of said actual second structure are measured.
3. The method of preparing fluid-structure interactive numerical model according to in said first measuring step, shape, mass and rigidity are measured as physical values related to the structure of said first structure; and in said second measuring step, shape and rigidity are measured as physical values related to the structure of a featured portion of said second structure. 4. The method of preparing fluid-structure interactive numerical model according to measurement of said prescribed motion is performed only for a featured portion that enables identification of an attitude of the actual said structure. 5. The method of preparing fluid-structure interactive numerical model according to 6. The method of preparing fluid-structure interactive numerical model according to said structure is a wing of an insect; said fluid is air; and said prescribed motion is a fluttering motion. 7. The method of preparing fluid-structure interactive numerical model according to in said actual structure measuring step, physical values related to the actual structure of the wing of an insect are measured, assuming that the actual structure of said wing of the insect consists of a collection of shell structures. 8. The method of preparing fluid-structure interactive numerical model according to in said the motion model preparing step, said motion model is prepared by using values obtained by smoothing time-differentiated values of the physical values related to said manner of motion of said wing actually measured. 9. The method of preparing fluid-structure interactive numerical model according to said smoothing is performed separately on translational motion and rotational motion of said wing. 10. The method of preparing fluid-structure interactive numerical model according to said motion model consists of time-sequential position data of prescribed three points among actual portions of said wing of the insect, when the prescribed three point change during said prescribed motion. 11. The method of preparing fluid-structure interactive numerical model according to said prescribed three points are selected from a portion among actual portions of the wing of said insect which is not deformed by the motion based on said motion model. 12. The method of preparing fluid-structure interactive numerical model according to said prescribed three points are selected from a potion near a root of the actual wing of said insect. 13. The method of preparing fluid-structure interactive numerical model according to a triangle formed by said prescribed three points is a right triangle. 14. The method of preparing fluid-structure interactive numerical model according to said actual structure measuring step includes a bending rigidity measuring step of measuring bending rigidity of the actual wing of said insect; and said numerical model of detailed figure of the actual wing of the insect has its thickness determined to have bending rigidity equivalent to the bending rigidity of said actual wing of the insect measured in said bending rigidity measuring step. 15. A method of manufacturing a fluttering robot, including
a wing for a fluttering motion, a wing driving apparatus for driving said wing in accordance with a numerical model of driving force, and a wing drive control apparatus controlling said wing driving apparatus by applying said numerical model of driving force to said wing driving apparatus, comprising the step of
determining said numerical model of driving force by a numerical model of driving force prepared in accordance with the method of preparing fluid-structure interactive numerical model according to
16. A method of manufacturing a fluttering robot including
a wing for a fluttering motion, a wing driving apparatus driving said wing, and a wing drive control apparatus controlling the wing driving apparatus, comprising the step of
determining a numerical model of a structure of said wing by modifying said equivalent numerical model of actual structure used in the method of preparing fluid-structure interactive numerical model according to
17. The method of manufacturing a fluttering robot according to change in numerical model of lift force with respect to change in a prescribed numerical model is used as the sensitivity in said sensitivity analysis. 18. A method of manufacturing a fluttering robot including
a wing for a fluttering motion, a wing driving apparatus driving the wing, and a wing drive control apparatus for controlling the wing driving apparatus, comprising the step of
determining manner of fluttering motion of said wing by modifying said motion model used in the method of preparing fluid-structure interactive numerical model according to
19. The method of manufacturing a fluttering robot according to change in numerical model of lift force with respect to change in a prescribed numerical model is used as the sensitivity in said sensitivity analysis. 20. A method of manufacturing a fluttering robot, comprising:
the step of preparing numerical model of artificial wing, preparing a numerical model related to a structure of an artificial wing; the step of preparing numerical model of detailed figure, which will be a reference of interpolation; the step of preparing detailed numerical motion model, preparing detailed numerical motion model corresponding to the manner of motion of said numerical model of detailed figure; the step of detailed fluid-structure interactive analysis, calculating a numerical model related to the structure of said numerical model of detailed figure and a numerical model related to fluid of said numerical model of detailed figure, by performing fluid-structure interactive analysis, using said numerical model of detailed figure and said detailed numerical motion model; the step of preparing first numerical model of interpolated structure, for preparing a first numerical model of interpolated structure by interpolating the numerical model related to the structure of said artificial wing and said numerical model of detailed figure with a prescribed interpolation ratio; the step of preparing a first numerical motion model of the interpolated structure corresponding to said first numerical model of the interpolated structure, by changing said detailed numerical motion model such that change in a specific numerical model is smaller than change in the specific numerical model of the numerical model related to the structure and the numerical model related to fluid resulting from fluid-structure interactive analysis using said first numerical model of the interpolated structure and said detailed numerical motion model; the step of first fluid-structure interactive analysis calculating a numerical model related to a structure of said first numerical model of interpolated structure and a numerical model related to fluid of said first numerical model of interpolated structure, by performing fluid-structure interactive analysis, using said first numerical model of interpolated structure and said first numerical motion model of the interpolated structure; the step of preparing a second numerical model of interpolated structure by interpolating the numerical model related to the structure of said artificial wing and said first numerical model of interpolated structure with a prescribed interpolation ratio; the step of preparing a second numerical motion model of interpolated structure corresponding to said second numerical model of interpolated structure, by changing said first numerical motion model of the interpolated structure, such that change in a specific numerical model is smaller than change in the specific numerical model of the numerical model related to the structure and the numerical model related to fluid resulting from fluid-structure interactive analysis using said second numerical model of interpolated structure and said first numerical motion model of the interpolated structure; and the step of second fluid-structure interactive analysis calculating a numerical model related to a structure of said second numerical model of interpolated structure and a numerical model related to fluid of said second numerical model of interpolated structure, by performing fluid-structure interactive analysis, using said second numerical model of interpolated structure and said second numerical motion model of the interpolated structure; wherein
while successively updating respective said models until the numerical model of interpolated structure is matched or approximated with the numerical model related to the structure of said artificial wing, accumulatively repeating steps similar to said step of preparing second numerical model of interpolated structure, said step of preparing the second numerical motion model of the interpolated structure and said second fluid-structure interactive analysis step and thereafter, using the numerical motion model of the interpolated structure corresponding to the numerical model of interpolated structure that matches or is approximated with the numerical model related to the structure of said artificial wing, the method of controlling a wing driving apparatus driving the artificial wing is determined.
21. The method of manufacturing a fluttering robot according to in said step of preparing the second numerical motion model of the interpolated structure, said first numerical motion model of the interpolated structure is changed such that change in a specific numerical model of the numerical model related to the structure and the numerical model related to fluid is made zero. 22. The method of manufacturing a fluttering robot according to 23. The method of manufacturing a fluttering robot according to said numerical model of detailed figure is said equivalent numerical model of actual structure used in the method of preparing fluid-structure interactive numerical model according to Description [0001] 1. Field of the Invention [0002] The present invention relates to a method of preparing a fluid-structure interactive numerical model for preparing a numerical model related to a fluid and a numerical model related to a structure when an actual structure as a living organism performs a prescribed motion in the fluid, as well as to a method of manufacturing a fluttering robot using the same. [0003] 2. Description of the Background Art [0004] Conventionally, attempts have been made to prepare a numerical model of the manner of fluttering flight that mimics the flight of an insect as an example of a structure of a living organism, with the manner of flight given as numerical expression, by a computer. Further, a simulator or a game of fluttering flight has been known that utilizes the thus prepared numerical model, in which the manner of fluttering flight mimicking an insect is displayed on a display. [0005] In this field of art, it is very useful to analyze a prescribed motion of a structure as a living organism in a fluid, that is, fluttering motion of an actual insect in the air, to know the mechanism of fluttering flight and to use the knowledge for controlling a fluttering robot that flies fluttering. The reason for this is that there is an almost infinite combinations of the manner of fluttering and the wing shapes of insects, and hence a formidable time is necessary to optimize the manner of fluttering or the wing shape of the insect, to meet a requested specification of the fluttering flight, and hence such an approach is impractical. In the course of evolution, insects have their manner of fluttering or wing shapes optimized. Therefore, as can be seen from the fact that bird wings are considered in designing glider wings, the method of obtaining basic principle of fluttering flight from insects is very efficient as compared with other methods. [0006] Further, it is very useful in the industry to prepare a new numerical model by modifying the numerical model related to the air and the numerical model related to the wing structure of an insect obtained from the insect so as to feed back the influence of the modification for better suited wing structure and analysis of the manner of fluttering. [0007] Therefore, in preparing a numerical model of the manner of fluttering, a method has been considered in which wing motion is extracted from images of fluttering flight of an insect picked-up by a high speed camera and obtain data of the extracted wing motion successively by image processing, for example. [0008] In order to analyze air behavior associated with fluttering, a method has been used in which an insect flies in a wind tunnel in which a trace such as a smoke flows and the behavior of the trace is monitored. [0009] Recently, an experiment has been conducted in which a scaled model of a wing is moved in a fluid having high viscosity such as mineral oil, so as to prepare an environment in which the Reynolds number of the fluid is made equal to the Reynolds number of the air, though the manner of fluttering is far moderate than the actual manner of fluttering of an insect, so as to facilitate measurement of velocity when the insect flies fluttering at high speed. [0010] In the above described method, however, it is impossible to simultaneously obtain the behavior of the air as the fluid and the behavior of the insect as a structure of a living organism. Therefore, it has been impossible to prepare a numerical model that involves interaction between the air and the wing of the insect, that is, to prepare a fluid-structure interactive numerical model. [0011] In the following, conventional methods of preparing numerical models will be specifically described. [0012] (1) Method of Getting Numerical Expression of Wing Deformation Using High Speed Camera [0013] According to A. Azuma, “The Biokinetics of flying and swimming” Springer-Verlag, Tokyo, 1992, a high speed camera is used, and positions of a line marker marked on a wing of an insect are captured to obtain attitude of the wing. Given the image processing capability of presently used computers, it is possible to obtain numerical expression, including change in shape, the manner of fluttering of a wing, by observing featured portion of the pattern of the entire wing, using video images obtained by the high speed image-pickup. [0014] Though it is possible by this method to grasp the wing behavior, it is not at all possible to grasp fluid behavior. Therefore, when an image is to be displayed on a computer using a numerical model related to the fluttering flight of an insect, for example, it is possible to express fluttering motion, where influence of air flow made by the wing on other structures cannot be calculated. [0015] (2) Fluid Flux Observation Using Trace [0016] A method in which a trace such as smoke is caused to flow in a wind tunnel to visualize the fluid flux of the trace has been long used for analyzing not only the fluttering flight of an insect but general behavior of a fluid. [0017] What is obtained by this method, however, is the shape of the fluid flux of the trace and not the velocity of the flowing trace. Therefore, it has been unsuccessful to obtain numerical model of the fluid behavior. [0018] As a similar method using a trace, a method has been recently applied to measure velocity in a pump, in which colored particles having approximately the same density as the fluid in the pump are caused to flow in the fluid, the manner of movement of the particles is detected by image processing, and the manner of movement is time-differentiated, to measure the velocity. [0019] It is noted, however, that only extremely small particles can float and not fall down, over a long period of time in a fluid having very small density such as air. Therefore, considering the capability of identification by a camera, it is impossible to use this method when the fluid is air. [0020] Consider analysis of fluttering of a dragon fly for one period. The fluttering period of a dragon fly is about 30 Hz, and the velocity caused by the fluttering is about 10 m/sec. Therefore, one particle of the smoke moves by the distance of 30 cm in this period. Therefore, in order to measure velocity of one period of fluttering of a dragon fly, that is, in order to measure the manner of movement of a certain trace, it is necessary to pick-up an image of at least an area of 30 cm×30 cm. [0021] When the area of 30 cm×30 cm is picked-up by a CCD camera having 1000 pixels×1000 pixels, it follows that an area of 300 μm×300 μm is picked-up by one pixel. The diameter of a particle that can float over a long period of time in the air, such as a pollen, is about 3 μm. The seeming area of the particle 3 μm in diameter is only {fraction (1/10000)} of the area picked-up by 1 pixel. Assuming that the light reflectance is the same, the trace floating in the air has the luminance of {fraction (1/10000)} of the luminance of an ordinary object having such an area that can be captured by a plurality of pixels. [0022] The inventors of the present invention conducted an experiment. An object having sufficiently large area is picked-up by a high speed camera under the illumination of 40000 lux. Even in this experiment, only 1024 frames at most could be picked-up in 1 second, using a microlens of 105 mmf 2.8 and a CCD camera comparable to ISO100. When an object moving with the period of about 30 Hz is to be captured with high accuracy as video images, the number of frames as high as 1024 is still insufficient. In order to capture the movement of a particle of 3 μm has video images, it follows that the product of CCD camera sensitivity and the luminance of illumination must be multiplied 10000 times. Therefore, such a method is considered impractical. [0023] (3) Measurement of Fluid Using Scaled Model [0024] M. Dickinson et al. (SCIENCE 1999 Vol. 284 pp. 1954-1960 “Wing Rotation and the Aerodynamic Basis of Insect Flight”) noted that fluid having the same Reynolds number behave equivalently, and found a method in which a large scaled wing model is moved at a velocity of at most several Hz in a mineral oil having high viscosity, whereby movement of particles mixed in a fluid that is equivalent to the fluttering of a fly and corresponds to the particles described in item (2) above is detected by image processing, enabling measurement of the movement. [0025] It is truth that when the Reynolds number is the same, fluid of different types behave in the same manner. The behavior of particle structures moving in fluid of different types but having the same Reynolds number, however, differ considerably. Therefore, by this method, it is impossible to correctly grasp deformation of particle structure moving in the fluid. [0026] For example, the relation of velocity F=s×f always holds where s represents scale and f represents velocity of the air in which an actual fly moves fluttering. As to the deformation D of actual fly wing, wing deformation D=s×d holds for some deformation d, while it does not hold for other deformation d. [0027] As described above, it has been unsuccessful through conventional methods to prepare an interactive numerical model between fluid and structure. For example, it has been impossible to properly represent a movement of a petal when a butterfly rests on a flower. [0028] By any of the above described methods, it has been impossible to calculate torque for driving the wing, for example, for a fluttering robot that mimics motion of a wing of an insect, since in methods (1) and (2), actual measurement of physical parameters is impossible and in method (3), force used for deformation of the wing is not considered as the wing deformation is different from the actual structure. Therefore, it has been difficult to obtain numerical model of a driving force for driving the wing, which is most important in forming a control mechanism controlling the fluttering flight. [0029] In short, by the numerical models prepared in accordance with the prior art, it has been difficult to obtain numerical expression of motion of a structure, which is a living organism, in a fluid, including fluid-structure interaction. [0030] Further, when a robot mimicking the structure as a living organism is to be manufactured, it has been necessary to design the driving force, allowing for a margin of the driving force, as it has been difficult to calculate the numerical model of driving force for driving the structure as a living organism. [0031] By any of the conventional methods, it is impossible to prepare a numerical model of the manner of fluttering flight of the robot mentioned above. Attempts have been made to actually fabricate the fluttering flight robot as described above, to drive the fluttering robot in various different manners of driving, and to prepare a numerical model of the manner of fluttering drive through trial and error. When the manner of fluttering drive that brings about the manner of motion such as a turn or a change in attitude is to be studied, it is first of all necessary that conditions to lift the fluttering robot are satisfied. Therefore, such a study can be made only under very limited conditions. Therefore, either by the method using a numerical model or by the method through experiment, it has been difficult to efficiently find the manner of fluttering drive of a fluttering robot. [0032] An object of the present invention is to provide a method of preparing a fluid-structure interactive numerical model including interaction between fluid behavior and behavior of a structure as a living organism, when the actual living structure performs a prescribed motion in a fluid, and to provide a method of manufacturing a fluttering robot using the same. [0033] The present invention provides a method of preparing a fluid-structure interactive numerical model that prepares a numerical model related to the fluid and the numerical model related to the structure, when the structure as a living organism performs a prescribed motion in a fluid. [0034] The method of preparing the fluid-structure interactive numerical model of the present invention includes the actual structure measuring step for measuring physical values related to the actual structure, and the step of preparing equivalent numerical model of actual structure that can be regarded as equivalent to the actual structure, in which the physical values related to the actual structure measured in the actual structure measuring step are given as numerical values. [0035] Further, the method of preparing fluid-structure interactive numerical model of the present invention includes the manner of motion measuring step in which an actual structure is caused to perform a prescribed motion and physical values related to the prescribed manner of motion are measured, and the motion model preparing step for preparing a motion model in which the physical values related to the prescribed manner of motion are expressed as numerical values. [0036] In the method of preparing fluid-structure interactive numerical model of the present invention further includes the fluid-structure interaction analysis step, in which, in a preset numerical fluid model for analysis, a prescribed motion represented by a motion numerical model is performed by an equivalent numerical model of actual structure, the numerical model related to the model fluid and the numerical model related to the equivalent numerical model of actual structure are processed to numerical models including interaction between the fluid behavior and the structural behavior. [0037] According to this method, the fluid-structure interactive analysis used for analyzing behavior of a structure other than the living organism is applied to analysis of behavior when a structure as a living organism performs a prescribed motion in a fluid, whereby it becomes possible to prepare the fluid-structure interactive numerical model including interaction between the behavior of the fluid and the behavior of the structure as the living organism. [0038] In the method of preparing the fluid-structure interactive numerical model of the present invention, preferably, the structure includes a first structure and a second structure which is different from but of the same type (species) as the first structure. [0039] Further, the actual structure measuring step includes the first measuring step for measuring physical values related to the actual structure of the first structure, and the second measuring step for measuring physical values related to a featured portion of the actual structure of the second structure. [0040] Further, the step of preparing equivalent numerical model of actual structure includes the step of preparing reference structure numerical model, in which physical values related to the actual structure of the first structure measured in the first step are expressed as numerical values, and the converting step for performing a prescribed conversion on the reference structure numerical model, using physical values related to the actual structure of the featured portion of the second structure measured in the second measuring step, to prepare the equivalent numerical model of actual structure. [0041] Further, in the manner of motion measuring step, physical values related to the manner of motion of the actual second structure are measured. [0042] By such a method, it is possible to have the first structure from which physical values related to the structure are measured and the second structure for which physical values related to the manner of a prescribed motion are measured, as separate structures. Therefore, it becomes possible to use such a method in that the first structure is broken and physical values related to the structure are measured, and the second structure is not broken and the physical values related to the structure are measured. According to such a method, it is possible to measure without damaging at least the second structure. Therefore, such an approach is advantageous in view of environmental protection, as the structure as a living organism is not damaged. [0043] In the method of preparing fluid-structure interactive numerical model of the present invention, preferably, in the first measuring step, shape, mass and rigidity are measured as physical values related to the structure of the first structure, and in the second measuring step, shape and rigidity are measured as physical values related to the structure of the featured portion of the second structure. [0044] According to this method, mass of the second structure is not measured. Therefore, at least the second structure can be measured without breaking or damaging the same. Therefore, an exemplarity method can be provided which is advantageous in view of environment as mentioned above, that minimizes damage to the structure as a living organism. [0045] In the method of preparing fluid-structure interactive numerical model of the present invention, measurement of a prescribed motion may be performed only on a featured portion of the actual structure. [0046] In this method, only the physical values related to the manner of prescribed motion on a featured portion from which the attitude of the actual structure as a living organism can be identified, are measured. Therefore, as compared with an example in which physical values related to the manner of prescribed motion at every portion of the actual structure are measured, the measurement of physical values related to the manner of prescribed motion can be simplified. [0047] In the method of preparing fluid-structure interactive numerical model of the present invention, preferably, the object is a virtual space where unsteady flow is generated, when an actual structure moves therein. By this method, the amount of calculation necessary for fluid-structure interactive analysis can be reduced. [0048] In the method of preparing fluid-structure interactive numerical of the present invention, the structure may be a wing of an insect, the fluid may be air, and the prescribed motion may be a fluttering motion. [0049] Here, it becomes possible to prepare a numerical model related to the air and a numerical model related to the wing structure, including interaction between the behavior of the air and the behavior of the wing of an insect. [0050] In the method of preparing fluid-structure interactive numerical model of the present invention, physical values related to the actual structure of the wing of an insect may be measured, assuming that the actual structure of the wing of an insect mentioned above consists of a collection of shell structures, in the actual structure measuring step. [0051] In this method, utilizing the fact that deformation experienced by the wing of an insect is dominantly bending deformation, a combination of beam and film structures, that is the actual structure of the wing of an insect, is approximated as a collection of shell structures. Accordingly, the amount of calculation for preparing the equivalent numerical model of actual structure can be reduced. Further, measurement of physical values related to the actual structure of the wing of an insect can be simplified, and modeling accuracy of the equivalent numerical model of actual structure can substantially be improved. [0052] In the method of preparing fluid-structure interactive numerical model of the present invention, in the motion model preparing step, the motion model may be prepared by using values obtained by smoothing time-differentiated physical values related to the actually measured manner of motion of the wing. [0053] By this method, it is possible to prevent physical values (velocity, acceleration and the like) of any abnormal manner of motion of the wing from being included in the motion model because of quantization error at the time of measuring the manner of motion. [0054] Further, the aforementioned smoothing may separately be performed on translational motion and rotational motion. Accordingly, it becomes possible to prevent, change due to smoothing in distance between prescribed two points on the wing as the object of measurement of the manner of motion. [0055] In the method of preparing fluid-structure interactive numerical model of the present invention, the aforementioned motion model may consist of position data represented as time sequence, when prescribed three points among actual portions of the wing of an insect change during the prescribed motion. [0056] The wing of an insect is approximately flat, and therefore, when prescribed three points of the actual wing of an insect are replaced by a triangular plane and the manner of prescribed motion is measured, there would be no problem. Utilizing this fact, in the present method, measurement of numerical values representing actual position or attitude of the wing of an insect necessary for preparing the motion model can be simplified. [0057] In the method of preparing fluid-structure interactive numerical model of the present invention, the aforementioned prescribed three points may be selected among portions that are not subjected to deformation by the motion based on the motion model, from actual portions of the wing of an insect. By this method, it is possible to prevent error involved in the measurement of position and attitude of the prescribed three points, which error caused by the deformation of the wing. [0058] Further, the prescribed three points may be selected from portions close to the root of the actual wing of an insect. By this method, it becomes possible to measure position and attitude of the wing at a portion which is relatively free of deformation. [0059] More preferably, a triangle formed by the prescribed three points is a right triangle. In that case, the position and attitude of the prescribed three points can be measured with higher accuracy than when the prescribed three points form other triangle. [0060] In the method of preparing fluid-structure interactive numerical model of the present invention, the actual structure measuring step includes the step of measuring flexural rigidity of the actual wing of an insect. In the numerical model of detailed figure of the actual wing of the insect, thickness may be determined such that flexural rigidity of the model is equivalent to the flexural rigidity of the actual wing measured in the flexural rigidity measuring step. [0061] By this method, it is possible to determine the thickness of the equivalent numerical model of actual structure which deforms equivalently as the actual wing, in an easy and reasonable manner, in the equivalent numerical model of actual structure having the shell structure. [0062] According to an aspect, the present invention provides a method of manufacturing a fluttering robot that includes a wing for fluttering motion, a wing driving apparatus driving the wing in accordance with a numerical model of driving force, and a wing drive control apparatus controlling the wing driving apparatus by applying the numerical model of driving force to the wing driving apparatus. [0063] In the method of manufacturing the fluttering robot in accordance with an aspect, the numerical model of driving force is determined by the numerical model of driving force prepared by the method of preparing fluid-structure interactive numerical model of air and the wing of an insect described above. [0064] By this method, the wing driving apparatus can be controlled by applying numerical model of driving force for driving the wing, instead of a numerical model of the displacement angle of a wing portion, or the numerical model of displacement of the wing. Therefore, design of the method of controlling driving force to the wing driving apparatus is facilitated. [0065] Further it is unnecessary to ensure large margin of the driving force for the wing driving apparatus. Thus, the wing driving apparatus can be reduced in size, and the energy supplied to the wing driving apparatus can be reduced. As a result, the fluttering robot can be made light-weight. [0066] According to another aspect, the present invention provides a method of manufacturing a fluttering robot that includes a wing for fluttering motion, a wing driving apparatus driving the wing, and a wing drive control apparatus controlling the wing driving apparatus. [0067] In the method of manufacturing the fluttering robot in accordance with this aspect, the numerical model of the wing structure of the above described fluttering robot is determined by modifying, in accordance with sensitivity analysis, the equivalent numerical model of actual structure used in the method of preparing fluid-structure interactive numerical model between the air and the wing of an insect described above. [0068] By this method, it becomes possible to manufacture a fluttering robot having a wing structure different from that of an insect, with high efficiency, by modifying the wing structure of an insect as a base. [0069] According to another aspect, the present invention provides a method of manufacturing a fluttering robot that includes a wing for fluttering motion, a wing driving apparatus driving the wing, and a wing drive control apparatus controlling the wing driving apparatus. [0070] In the method of manufacturing a fluttering robot in accordance with this aspect, the manner of fluttering motion of the wing of the fluttering robot mentioned above is determined by modifying, in accordance with sensitivity analysis, the motion model used in the method of preparing fluid-structure interactive numerical model between the air and the wing of an insect described above. [0071] By this method, it is possible to manufacture a fluttering robot that flies in a manner of fluttering different from that of an insect, efficiently, by modifying the manner of fluttering flight of an insect as a base, without the necessity of studying, one by one, the innumerable manners of fluttering flight. [0072] In the method of manufacturing a fluttering robot in accordance with the above described aspects of the present invention, change in numerical model of lift force relative to the change in a prescribed numerical model, may be used as the sensitivity for sensitivity analysis. [0073] According to this method, when the lift force as one of the most important parameters for fluttering flight is made approximately the same as the lift force of an actual insect, it becomes easier to realize stable fluttering flight of the artificial fluttering robot. [0074] According to a still further aspect, the present invention provides a method of manufacturing a fluttering robot, including the step of preparing numerical model related to a structure of an artificial wing, the step of preparing a numerical model of detailed figure that will be a reference for interpolation, and the step of preparing detailed numerical motion model that corresponds to the manner of motion of the numerical model of detailed figure. [0075] Further, according to another aspect, the present invention provides a method of manufacturing a fluttering robot including the step of detailed fluid-structure interactive analysis step for calculating a numerical model related to the structure of the numerical model of detailed figure and the numerical model related to fluid of the numerical model of detailed figure, by performing fluid-structure interactive analysis, using the numerical model of detailed figure and the detailed numerical motion model. [0076] According to a still further aspect, the method of manufacturing a fluttering robot includes the step of preparing a first numerical model of interpolated structure, by interpolating, with a prescribed interpolation ratio, the numerical model related to the structure of the artificial wing and the numerical model of detailed figure. [0077] According to a still further aspect, the method of manufacturing a fluttering robot includes the step of preparing a first numerical motion model of the interpolated structure that corresponds to the first numerical model of the interpolated structure, by modifying the detailed numerical motion model such that change in a specific numerical model becomes smaller than the change in the specific numerical model among the numerical models related to the structure and to the fluid when fluid-structure interactive analysis is performed using the first numerical model of the interpolated structure and the detailed numerical motion model. [0078] According to a still further aspect, the method of manufacturing a fluttering robot includes the step of first fluid-structure interactive analysis, for calculating a numerical model related to the structure of the first numerical model of the interpolated structure and a numerical model related to the fluid of the first numerical model of the interpolated structure, by performing fluid-structure interactive analysis using the first numerical model of the interpolated structure and first numerical motion model of the interpolated structure. [0079] Further, according to a still further aspect, the method of manufacturing a fluttering robot includes the step of preparing a numerical model of a second interpolated structure by interpolating, with a prescribed interpolation ratio, the numerical model related to the structure of the artificial wing and the first numerical model of the interpolated structure. [0080] According to a still further aspect, the method of manufacturing a fluttering robot includes the step of preparing a second numerical motion model for the interpolated structure that corresponds to the second numerical model of the interpolated structure, by changing the first numerical motion model of the interpolated structure such that change in a specific numerical model becomes smaller than the change in the specific numerical model among the numerical models related to the structure and to the fluid when fluid-structure interactive analysis is performed using the second numerical model of the interpolated structure and the first numerical motion model for the interpolated structure. [0081] A method of manufacturing a fluttering robot in accordance with a still further aspect includes the step of second fluid-structure interactive analysis for calculating a numerical model related to the structure of the second numerical model of the interpolated structure and a numerical model related to the fluid of the second numerical model of the interpolated structure, by performing fluid-structure interactive analysis using the second numerical model of the interpolated structure and second numerical motion model of the interpolated structure. [0082] In the method of manufacturing a fluttering robot in accordance with a still further aspect of the present invention, steps similar to the step of preparing second numerical model of the interpolated structure, the step of preparing second numerical motion model of the interpolated structure and the step of second fluid-structure interactive analysis are accumulatively repeated while each of the numerical models described above is updated successively until the numerical model of the interpolated structure matches or is approximated to the numerical model related to the structure of the artificial wing and, thereafter, using the numerical motion model of the interpolated structure that corresponds to the numerical model of the interpolated structure matching or being approximated to the numerical model related to the structure of the artificial wing, the method of controlling the wing driving apparatus is determined for driving the artificial wing. [0083] According to this method, as the numerical model of interpolated structure is interpolated to be closer to the numerical model related to the structure of artificial wing, it becomes possible to realize a manner of fluttering flight that is close to the manner of fluttering flight of the numerical motion model corresponding to the numerical model of detailed figure (for example, the equivalent numerical model of actual structure). As a result, it becomes possible to determine the method of controlling wing driving apparatus driving the artificial wing such that even the wing of the artificial fluttering robot is driven in the manner of fluttering flight that is close to the manner of fluttering flight of the detailed figure (for example, an actual insect). [0084] In the fluttering robot in accordance with a still further aspect of the present invention, preferably, in the step of preparing second numerical motion model of the interpolated structure, the first numerical motion model of the interpolated structure is changed such that change in a specific numerical model of the numerical model related to the structure and the numerical model related to the fluid becomes zero. [0085] By this method, it becomes possible to determine the method of controlling the wing driving apparatus driving the artificial wing such that even the wing of the artificial fluttering robot assumes approximately the same manner of fluttering flight of a detailed figure (for example, an actual insect), as regards a specific numerical model. [0086] More preferably, in the method of manufacturing a fluttering robot in accordance with another aspect of the present invention, the specific numerical model mentioned above is a numerical model of lift force. When the lift force which is one of the most important parameters for fluttering flight is made approximately equal to the lift force of the actual detailed figure (for example, an actual insect), it becomes easier to realize more stable fluttering flight of the fluttering robot having artificial wings. [0087] In the method of manufacturing a fluttering robot in accordance with another aspect of the present invention, the numerical model of detailed figure may be the equivalent numerical model of actual structure used for the method of preparing fluid-structure interactive numerical model described above. [0088] As a result, it becomes possible to determine the method of controlling wing driving apparatus for driving artificial wings such that even the artificial wings of the fluttering robot assume the manner of fluttering flight that is close to the actual manner of fluttering flight of a structure represented by the equivalent numerical model of actual structure, that is, the actual structure of an insect. [0089] The foregoing and other objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings. [0090]FIG. 1 represents the method of preparing fluid-structure interactive numerical model and a process for manufacturing a fluttering robot in which the fluid-structure interactive numerical model is modified by sensitivity analysis. [0091]FIG. 2 represents beam and film structures of a wing of an actual insect in accordance with the first embodiment. [0092]FIG. 3 represents a method of measuring shape of the wing in accordance with the first embodiment. [0093]FIG. 4 represents wing thickness distribution related to the first embodiment. [0094]FIG. 5 represents an approximately flat area of a wing, related to the first embodiment. [0095]FIG. 6 represents a method of measuring rigidity at a featured portion of a wing in accordance with the first embodiment. [0096]FIG. 7 represents another method of measuring rigidity at a featured portion of a wing in accordance with the first embodiment. [0097]FIG. 8 illustrates a method of preparing fluid mesh in accordance with the first embodiment. [0098]FIG. 9 represents an outline of the fluid mesh in accordance with the first embodiment. [0099]FIG. 10 represents fluid behavior of the fluid-structure interactive numerical model in accordance with the first embodiment. [0100]FIG. 11 represents transition of fulcrum reaction on a fulcrum of a wing, when the fluid-structure interactive numerical model flutters, in accordance with the first embodiment. [0101]FIG. 12 represents a schematic configuration of the fluttering robot in accordance with the first embodiment. [0102]FIG. 13 is an illustration related to a method of sensitivity analysis in accordance with the first embodiment. [0103]FIG. 14 is an illustration related to the method of interpolating wing shape in accordance with the first embodiment. [0104] FIGS. [0105]FIG. 20 is a graph representing driving torque exerted on the equivalent numerical model of actual structure in accordance with the first embodiment. [0106]FIG. 21 is an illustration representing the method of preparing fluid mesh in accordance with the second embodiment. [0107]FIG. 22 represents an outline of the fluid mesh in accordance with the second embodiment. [0108]FIG. 23 represents transition of fulcrum reaction on the fulcrum of a wing, when the fluid-structure interactive numerical model flutters in accordance with the second embodiment. [0109]FIGS. 24 and 25 represent specific examples of equivalent numerical model of actual structure in accordance with the second embodiment. [0110]FIG. 26 is a graph representing driving torque exerted on the equivalent numerical model of actual structure in accordance with the second embodiment. [0111] (First Embodiment) [0112] First, referring to Tables 1 to 6 and FIGS. [0113] The method of preparing fluid-structure interactive numerical model of the present embodiment is to prepare a numerical model related to air as the fluid and a numerical value related to the wing structure, obtained by analyzing structure of the wing of an insect and the manner of fluttering flight of the insect, when the insect flies fluttering in the air. [0114] The method of manufacturing a fluttering robot in accordance with the present embodiment is to manufacture a fluttering robot mimicking the structure of the wings of an insect and mimicking the manner of fluttering flight of the insect, using a numerical model prepared by the method of preparing fluid-structure interactive numerical model described above. [0115] Specifically, the numerical model related to the fluid refers to a numerical model of velocity and pressure of the fluid. Further, the numerical model related to the structure mainly refers to a numerical model of the manner of motion such as movement and the deformation of the structure in contact with the fluid, as well as a numerical model of force such as internal stress acting on the structure. [0116] In the method of preparing fluid-structure interactive numerical model of the present embodiment, description will be given on an example in structure of the wing (hereinafter referred to as “equivalent numerical model of actual structure”) that represents shape and rigidity of the wing in numerical values, and through process 5, a numerical model of the manner of fluttering flight of the wing, that is, the driving force model of the wing (hereinafter referred to as “numerical model of fluttering motion”) are prepared. [0117] Therefore, it is assumed for simplicity of description that the body model of process 6 is only the shape and mass distribution represented as numerical values. Specifically, the body model is used only for the function of providing boundary condition for the fluid, that is, to provide inertia for rotation or translational movement at the fulcrum of the wing. [0118] Actually, it may be possible to change the position, attitude and the shape of the numerical model of the body. Even when these aspects are considered, the numerical models can be prepared by applying the similar method as preparing the fluid-structure interactive numerical model of the wing, for the change in position, attitude and the shape. [0119] More specific procedure for preparing the numerical model is as follows. [0120] First, a wing of a dragon fly (hereinafter referred to as “Sample A”) is separated, for example, as needed, to precisely measure physical values related to the wing structure. Using the measured physical values, numerical model of detailed figure as reference structure numerical model is prepared. Further, physical values of the manner of fluttering flight of another sample (hereinafter referred to as “Sample B”) are measured, and using the measured physical values, a numerical model of fluttering motion is prepared. [0121] Next, for Sample B, physical values related to the structures of featured portions of the wing that can be measured without damaging Sample B are measured. Using the data of the physical values of featured portions, the numerical model related to the structure of Sample A, that is, the numerical model of detailed figure is converted, whereby a numerical model that can be regarded as equivalent to the numerical model of the wing structure of Sample B, that is, the equivalent numerical model of actual structure, is prepared. [0122] The behavior of the fluid and the behavior of the structure when the equivalent numerical model of actual structure of Sample B is driven in the fluttering manner directly measured from Sample B are calculated by fluid-structure interactive analysis. Thus, the fluid-structure interactive numerical model at the time of fluttering motion including the influence from the ambient fluid is prepared for Sample B. [0123] The procedure for manufacturing a fluttering robot based on the fluid-structure interactive numerical model obtained through the above described method is as follows. [0124] Each procedure will be described in detail with reference to FIGS. [0125] (Measurement of Physical Values Related to the Structure of the Numerical Model of Detailed Figure) [0126] First, physical values related to the structure of Sample A are measured for preparing the numerical model of detailed figure. Generally, equation of motion of a structure is given as an expression of external force and acceleration, using a spring and damper, that is, elasticity with respect to displacement and damping ratio with respect to velocity. [0127] Generally, damping of internal stress in a structure occurs when kinetic energy is converted to thermal energy in the structure. This corresponds to occurrence of structural change or destruction. Considering the fact that a dragon fly of which fluttering frequency is 30 Hz continuously flies for more than a week and broken only by outer damage, the structural change in the wing (plastic deformation) in the fluttering motion of several periods can be considered extremely small. Therefore, in the present embodiment, it is assumed that damping of internal stress in the structure is zero. In other words, it is assumed that the wing is subjected to elastic deformation only and not plastic deformation. [0128] Thus, the equation of motion in the structure is given by the elasticity with respect to displacement, that is, rigidity of the wing, mass of the wing and external force. [0129] Namely, there are three parameters necessary for preparing the structural model of the wing, that is, shape, rigidity and mass of the wing. The method of measuring these will be described with reference to FIGS. [0130] (Measurement of the Shape of Numerical Model of Detailed Figure) [0131] First, measurement of the shape of the wing will be described with reference to FIGS. 2 and 3. [0132] As shown in FIG. 2, the wing [0133] For this purpose, the inventors used the following method for measuring the shape. [0134] First, in order to grasp three-dimensional wing shape, a general, commercially available X-Y stage [0135] As shown in FIG. 3, the distance to wing [0136] Alternatively, it is possible to obtain an image of wing [0137] Accordingly, the arrangement is determined where x, y, z of all the beam structures [0138] For example, in most of commercially available laser distance meters [0139] (Measurement of Rigidity of the Numerical Model of Detailed Figure as Reference Structure Numerical Model) [0140] Basically, a general method can directly be applied to the measurement of rigidity of the wing. More specifically, the following method is used. As already described, wing [0141] The method, however, is disadvantageous as it is complexed. Further, in this method, flexural rigidity is determined as a result of interaction between the beam structure and the film structure, and therefore, error of these two is undesirably involved. [0142] Now, deformation of a wing includes bending deformation and tensile deformation. As the wing of an insect is very thin, deformation caused by bending is dominant in the wing deformation. In other words, the wing of an insect has small distortion and large deformation. [0143] Therefore, in the present embodiment, wing [0144] It is well possible that measurement must be performed with wing [0145] (Measurement of Mass Distribution of Numerical Model of Detailed Figure) [0146] As already described with respect to rigidity measurement, when the structure of wing [0147] Actually, when wing [0148] The mass thus obtained is divided by the size of the piece of which mass has been measured, and the resulting value of division is regarded as mass distribution. By dividing the mass of the piece by the area of the piece, the mass per unit area can be calculated, for a shell structure, for example. [0149] Through the above described manner, the shape of wing [0150] (Modeling of Detailed Figure as Reference Structure) [0151] Modeling of the structure of wing [0152] Generally, in a method frequently used for analyzing a structure, the wing [0153] In the present embodiment, the unit structure of the wing is regarded as a shell structure. More specifically, wing [0154] (Shape Modeling) [0155] As for the shape, the position of a node (intersection point of lines representing a mesh) constituting each mesh is applied. Dependent on the method of analysis, attitude (a plurality of point positional data) of each node is applied. [0156] Here, mesh division is desirably performed taking the direction of beam structure [0157] (Rigidity Modeling) [0158] Thereafter, rigidity is applied to each mesh. [0159] The rigidity of a certain mesh, that is, basic parameters that determine modification against a certain external force are Young's modulus, Poisson's ratio and mesh thickness. [0160] As already described, the wing in the present embodiment is characterized in that the distortion is small and the deformation is large. Therefore, deformation of the shell structures is for the most part determined by the product of Young's modulus and the second moment of area. Poisson's ratio has almost no influence on deformation, and therefore a general value of 0.3 is used as Poison's ratio. In view of preparing equivalent numerical model of actual structure, which will be described later, a rough value may be used as the Young's modulus, such as an average of values resulting from measurements at all the featured portions. Using the bending rigidity measured at each featured portion, the thickness of the mesh is reversed-calculated from the result of numerical calculation or theoretical solution. [0161] For example, when a beam having the length 1, height h and width b has one end fixed and a load of w is applied in the height direction on the other end, displacement x of the beam in the height direction is given as (x=w×l [0162] Accordingly, thickness of the mesh at each featured portion is determined, and a value obtained by interpolation thereof is applied to each mesh. Here, the thickness at each featured portion is the thickness for representing bending rigidity of the combined structure of beam structure [0163] (Mass Modeling) [0164] The mass can simply be calculated by interpolating mass per unit area calculated from the result of mass measurement at each featured portion. What is necessary is simply to apply a value obtained by multiplying the mass per unit area at the mesh position by the mesh area. [0165] Through the above described steps, a numerical model related to the structure representing shape, rigidity and mass distribution of wing [0166] (Preparation of Equivalent Numeral Model of Actual Structure) [0167] Referring to FIGS. [0168] (Measurement of Featured Shape) [0169] When feature shape is to be measured, first, distance from a featured structure such as outline of wing [0170] (Measurement of Representative Rigidity) [0171] Here, the method of measuring rigidity of a featured portion of wing [0172] Generally, rigidity is measured by fixing one end of an object, applying a certain load on another portion of the object, and measuring the resulting amount of displacement. Therefore, the object must be fixed firm enough not to cause any variation when the load is applied to another portion. When a living organism is to be measured without physically damaging the same, it is necessary that the load causes minimum possible damage to the wing [0173] Here, the following two methods may be possible to accurately measure the rigidity of featured portions without damaging the wing [0174] One is the method that uses fixing means confirming to the protruded/recessed shape of wing [0175] Therefore, it is desirable to apply the above described method to such a portion at which difference in shape caused by fixation is small. For example, the rear portion at the tip end of wing [0176] Alternatively, a method of fixing a strong portion of beam structure [0177] (Preparation of Equivalent Numerical Model of Actual Structure by Converting Numerical Model of Detailed Figure) [0178] Here, a method will be described in which the equivalent numerical model of actual structure is prepared by converting the numerical model of detailed figure using representative measured values obtained by measuring representative values of the living organism of wing [0179] First, using the ratio α of the shape calculated by the measurement of featured shapes, the numerical model of the detailed figure is enlarged/reduced. The enlarged/reduced model will be referred to as an intermediate model. [0180] Specifically, when the coordinates of the node of each mesh and the thickness of the mesh are represented as P and T, respectively, and when the node coordinates and mesh thickness after converting the values P and T with the ratio α of the shape calculated in accordance with the featured shapes are represented as P′ and T′, there is the relation P′=P×α and T′=T×α. In the intermediate model, the shape and the mass approximately reflect the shape and mass of Sample B. [0181] Thereafter, the same dynamical condition as the condition used for measuring the featured rigidity of Sample B, that is, fixing condition and load condition are applied, as numerical model, to the intermediate model, and the amount of deformation of the structure is calculated by numerical analysis. When displacement under the load equivalent to the rigidity of the featured value of the living organism is β, Young's modulus E for each mesh is changed to E′ which is calculated as E′=E×β. [0182] More specifically, when a value of displacement under a certain dynamical condition in the intermediate model is 0.5 times the value of displacement under the same condition measured from the living organism, the intermediate model is converted to have the relation between load and displacement comparable to that of the living organism by multiplying Young's modulus of the intermediate model by 0.5. [0183] Accordingly, the model resulting from conversion of Young's modulus of the intermediate model comes to have the shape, mass and rigidity almost the same as those of wing [0184] (Measurement of Position of Featured Point of Wing Motion) [0185] Next, a method of preparing a model of the manner of fluttering, that is, the attitude for driving wing [0186] Accordingly, positions of three points that can be picked-up by a high speed camera are measured, which points have substantially the same position and attitude as the fulcrum of the wing, and the attitude of a plane formed by the three points is considered to be the attitudes of these three points respectively. Thus, the attitude of wing [0187] More specifically, a method may be considered in which three point markers are provided at positions with smallest possible deformation on wing [0188] As the method of measuring position of the points, a conventional method of measuring position of a prescribed point of the wing is used. In the method of preparing fluid-structure interactive numerical model of the present embodiment, the positions of the point markers are calculated using positions on the images picked-up by high speed cameras from two directions. It is noted, however, that the images picked-up by the high speed cameras are quantized. Therefore, in the method of preparing fluid-structure interactive numerical model of the present embodiment, smoothing is performed to alleviate position deviation of the prescribed points resulting from quantization error. The force exerted by the fluid is a function of velocity. Therefore, it is desirable to perform smoothing on time history of velocity, which is the time history of the positions of point markers mentioned above differentiated by time. The aforementioned time-history of velocity must be continuous (continuous function) to enable smoothing. When smoothing is performed on time-differentiated value of the time-history of positions in the rectangular coordinates directly, it is possible that the distance between point markers may vary. Therefore, it is more desirable that the smoothing is performed on the translational velocity and angular acceleration of the plane formed by the three points, that is, the position and attitude of the plane formed by the three points separated. The method of measuring point positions is not limited to the method described above, and any method that can determine positions of certain three points of wing [0189] In the following, the equivalent model of actual structure used for the fluid-structure interactive analysis of the present embodiment will be specifically discussed with reference to FIGS. [0190] In the present embodiment, positions of these three points are given for the element at the root of the wing. FIG. 15 represents an element at the root of the wing shown in FIG. 4, that is, the element at the lower left corner extracted, with node numbers and the element number used for the analysis in accordance with the present embodiment added. Here, hovering state is assumed, and node [0191]FIG. 17 represents the behavior of the wing calculated in accordance with the reference of Azuma et al. mentioned above, in the coordinate system shown in FIG. 16, for a down stroke of the wing, along with time. FIGS. 18 and 19 represent time-history of x, y and z coordinate values of node [0192] (Fluid-Structure Interactive Analysis) [0193] Here, the method of preparing a numerical model by fluid-structure interactive analysis will be described with reference to FIGS. [0194] The method of analysis used by the inventors is the strong coupling method of fluid and structure, in accordance with ALE finite element analysis method proposed by Qun Zhang, Computer methods in applied mechanics and engineering 190 (2001) pp. 6341-6357 “Analysis of fluid-structure interaction problems with structural buckling and large domain changes by ALE finite element method” and (Tokyo University Thesis, 1999, “ALE (Arbitrary Lagrangian-Eulerian Method) Finite Element Analysis of Structure-Fluid Interactive Problem Involving Structural Buckling and Area Deformation”). The method of application thereof will be discussed in the following. Here, ALE refers to a method in which Eulerian notation (fluid) and Lagrangian notation (structure) are handled uniformly, using a reference coordinate system. Further, the finite element method refers to a method of solving an equation provided by dividing the area to be analyzed into a finite number of elements and integrated by approximation within the elements. Different from the difference method, this method enables handling of free shapes, and convergence properties thereof have been proven mathematically. [0195] First, it is necessary to set a virtual space for preparing the numerical model. According to the calculations made by the inventors, when the wing has a length of 4 cm and fluttering frequency of 30 Hz, the air flow is almost a steady flow at a distance of about 20 cm and further. Therefore, when a spherical space having the radius of 20 cm including the structure for preparing the numerical model is considered to be the object of fluid-structure interactive analysis, the amount of calculation can be reduced. [0196] For simplicity of description here, it is assumed that a fluid-structure interactive numerical model is prepared for an example in which a dragon fly is positioned approximately at the center of a cubic case each side of which is 20 cm in length, and the dragon fly is hovering. [0197] In the ALE finite element method, the following four parameters A to D must be set both for the fluid and the structure. In the following, it is assumed that A represents coordinates of each node, B represents connectivity of each node, C represents boundary condition of each node and D represents physical values of each element. [0198] (Structure Element) [0199] The parameters A, B and D are determined at the time point when the equivalent numerical model of actual structure is prepared. [0200] As for the parameter C, position data represented time-sequentially of the markers on wing [0201] (Fluid Element) [0202] There is no mesh prepared for the fluid element, and therefore, it is prepared by using a commercially available mesh preparing software, for example. In the fluid-structure interactive analysis method used by the inventors, it is necessary to use nodes common to the structural mesh also in the fluid mesh. Therefore, first, a mesh of a plane including the wing is prepared by tetragonal elements, and the thus prepared mesh is swept upward and downward, to prepare a hexahedron mesh. Tetragonal meshes other than the wing are deleted after the preparation of hexahedron mesh. These operations are represented in FIGS. 8 and 9. [0203] In this manner, coordinates and connectivity of respective nodes are prepared for fluid element as well. Further, as the boundary condition, a fixation boundary condition is given to the node of the wing and the nodes forming the outer wall of the cubic case and, in addition, mass density, coefficient of viscosity and bulk modulus of the air are given as physical values of each fluid element. [0204] In order to obtain a numerical model related to the manner of motion converged to the steady hovering state, the inventors analyzed a state in which fluttering of one period is repeated for a plurality of times. [0205] (Actual Data) [0206] The data actually used for analysis will be described with reference to Tables 1 to 6. [0207] Coordinates of each node are common to fluid and structure. Therefore, as shown in Table 1, respective nodes are denoted by Node 1, Node 2, . . . , with respective x, y and z coordinates listed together. In Table 1, such data are referred to as NodeCoords.dat.
[0208] Thereafter, it is necessary to designate connectivity of the nodes forming the wing, so as to designate the wing structure. Thus, for Element1 shown in FIG. 15, Node71, Node83, Node142 and Node337 are designated counterclockwise from the lower left corner, to represent the quad shell element. More specifically, respective shells forming the wing are labeled with numbers such as ShellElement1, ShellElement2, . . . as described above. Such data is referred to as shellMesh.dat in Table 2.
[0209] Similarly, connectivity is designated for the mesh prepared for the fluid area. Here, the fluid area is represented as hexahedron element. Further, in order to represent scalar amount of pressure, an additional node, which is referred to as pressure node, is added to each fluid mesh. The order of designating connectivity is as follows. Four points on the lower surface of the hexahedron element are continuously denoted counterclockwise, four points on the upper surface are thereafter designated also counterclockwise, and finally, the pressure node is designated, by the node numbers. Thus, a set of connectivity of a total of 9 points is given as shown in Table 3. In Table 3, such data is referred to as FluidMesh.dat.
[0210] Further, it is necessary to add physical values to fluid and structure, respectively. Table 4 is to apply physical values of the structure, and Table 5 is to apply physical values of the fluid. These are referred to as ShellMaterial.dat and FluidMaterial.dat, respectively. In ShellMaterial.dat, Young Modulus, Poisson's Ratio, mass density (densimeter) and thickness are listed, for each structural element of the wing. In FluidMaterial.dat, viscosity, mass density (densimeter) and bulk modulus are listed for each fluid element.
[0211]
[0212] Further, Table 6 represents time-history of x, y and z coordinates at the nodes of the wing, in ShellMotion.dat.
[0213] The data described above are examples only, and data format and the values are not limited thereto. [0214] [Results of Calculation] [0215] In the following, an example of the numerical model obtained through the above described method will be discussed. The numerical model is obtained from the data listed above, and the numerical model is not limiting. [0216]FIG. 10 represents the results of calculation of the fluid behavior around the wing [0217]FIG. 11 represents the total nodal force in the direction of y axis at nodes 71, 83 and 337 calculated by this method. As the forced displacement is exerted on these three points, the total of nodal forces at these points represent the force exerted on the body. The nodal force, which is irregular in the initial state, eventually converges to a periodic behavior. Specifically, the behavior of the fluid and the behavior of the structure both become periodic at this time point, which are equivalent to the behavior of the fluid and the behavior of the structure at the time of hovering. It is noted, however, that the direction of generation of the lifting force is the negative direction along the y axis, as shown in FIG. 17. According to the calculation made by the inventors, under the condition of gravitational acceleration of 9.8 m/sec [0218]FIG. 20 shows the driving torque exerted on the wing at this time. Here, θy represents torque in the direction of right turn in the positive direction along the y axis, and θz represents a torque in the direction of right turn along the positive direction of the z axis. [0219] The torque is the driving torque itself of the actuator, and therefore, from the data above, we can say that a mass of 0.1 g can be lift up from the ground by one wing, using an actuator having the torque of 1.5 gf·cm to 3.5 gf·cm. [0220] Accordingly, it is possible to prepare a numerical model of flow for steady hovering, and hence it is possible to calculate weight that can be lifted and the torque of the actuator realizing such an operation. [0221] (Application of the Result of Analysis to Manufacturing of a Robot) [0222] Here, the method of applying the numerical model of fluttering including fluid to the method of controlling a robot will be described with reference to FIGS. 1 and 11 to [0223] The numerical model obtained through the above described method can directly be applied to control of a fluttering robot, as will be described later. Alternatively, it may be possible to clarify air dynamic force utilized by an insect from the numerical model itself obtained from the fluid-structure interactive analysis, and to manufacture a fluttering robot utilizing the clarified air dynamic force. In actual manufacturing a fluttering robot, it is very useful industrially, to modify the fluid-structure interactive numerical model using sensitivity analysis, and to determine structure of the wing of the fluttering robot or to determine the manner of fluttering flight, based on numerical model changed in accordance with the sensitivity analysis. [0224] In the following, method of preparing actuator control method, and obtaining manner of fluttering of wings having different shapes, which are particularly useful in preparing the control method of fluttering robot, will be described. [0225] (Method of Preparing Actuator Control Method) [0226]FIG. 12 is a schematic diagram representing a fluttering robot control system. [0227] For simplicity of description, an actuator [0228] Generally, actuator drive is realized by designating power (numerical value of driving force) of the actuator. Therefore, when there is no model of driving force of the actuator, a mechanism is necessary that converts displacement of a wing to the driving force model, by measuring the displacement of the wing. The reason for this is as follow. Even when the driving force (power) of the actuator is constant, displacement of the wing differs dependent on the load acting on the wing. For example, the driving force (power) of actuator [0229] Further, in an actuator of the type that controls feeding-back the displacement of the wing, that is, time-sequential data of a prescribed position of the wing, it is necessary from its nature that the driving force (power) necessary when the wing reaches a target position must be obtained before the time when the wing reaches the target position. Therefore, excessive energy is necessary. As a result, the size of the actuator itself is increased. [0230] The power necessary for controlling actuator [0231] On the contrary, when control of actuator [0232] Therefore, when the fluid-structure interactive numerical model described above is utilized, the torque for driving actuator [0233] According to this method, flight with actuator and energy source that are lighter in weight becomes possible, and therefore, application of this method is quite useful for the fluttering robot. [0234] It is noted, however, that addition of an apparatus for detecting wing position to the fluttering robot and feeding back the time-sequential data of the wing position from the apparatus are useful to absorb error in driving, for example. Therefore, it is not the case that provision of an apparatus for detecting wing position on the fluttering robot is unnecessary. [0235] When a relation between voltage and torque on actuator [0236] (Manufacturing of Fluttering Robot Based on the Numerical Model) [0237] In most cases, an actuator manufactured by men has characteristics different from the muscle of an insect. Therefore, it is not always true that driving identical with the driving by a muscle of the insect is optimum. [0238] Further, it may sometimes be inefficient to manufacture a structure equivalent to the wing [0239] On the contrary, if it is possible to grasp influence on the driving force model when the structure model or the numerical model of fluttering motion is modified, it becomes possible to derive control of the manner of fluttering flight other than the manner of fluttering of a dragon fly using the wing of the dragon fly, for example. Therefore, it becomes possible to realize wide variety of modifications in design and operation of a fluttering robot. [0240] Here, a method of manufacturing a fluttering robot having wings will be described, in which the fluid-structure interactive numerical model of fluttering flight of Sample B obtained through the above described process is used as the base, and applying a method generally referred to as sensitivity analysis to the base, the base numerical model is deformed. [0241] (Various Sensitivity Analysis) [0242] First, sensitivity analysis related to the shape of a wing [0243] For example, assume that the fluid-structure interactive analysis is performed using a numerical value in which the z coordinate of the node Pt shown in FIG. 13 is changed to Pt×(1+δz), and component in the direction of gravitational acceleration Fz of the fulcrum reaction of wing [0244] Sensitivity analysis of the manner of flight, that-is, sensitivity analysis of the manner of fluttering flight will be described. In the numerical model of the manner of fluttering, the attitude of the wing that changes time-sequentially, that is, time-sequential position data of the wing, is slightly changed for sensitivity analysis. Then, the numerical model related to the fluid and the numerical model related to the structure for the manner of fluttering resulting from the above described series of fluid-structure interactive analysis are compared with the numerical model related to the fluid and the numerical model related to the structure for the manner of fluttering before the change. Thus, sensitivity, that is, a change in a certain target parameter with respect to the change in the numerical model of the manner of fluttering is found. In other words, change in a specific numerical model resulting from the change in the numerical model of the manner of fluttering is calculated. [0245] For example, assume that a numerical model, in which the amplitude θw of wing [0246] In the following, lift force of fluttering, which is important for fluttering flight, will be used as the target parameter, unless specified otherwise. More specifically, an average value of the force in the direction opposite to the gravitational acceleration exerted on the fulcrum of wing [0247] (Obtaining Manner of Fluttering of Wings Having Different Shapes) [0248] First, a method of obtaining the manner of fluttering when the wing shape is changed will be described. More specifically, referring to FIG. 14, a method will be described in which the manner of fluttering that ensures the same fluttering lift force as the fluttering lift force obtained by the equivalent numerical model of actual structure, by the numerical model of artificial wing [0249] In the description of sensitivity analysis above, nodes are provided individually. When numerical models related to the wing structure are changed collectively based on a certain parameter, it is possible to find the sensitivity of a specific parameter with respect to the change in the numerical models related to the wing structure as a whole. As an example, the case will be considered in which sensitivity of fluttering lift force with respect to the change in the wing shape, that is, the change in manner of fluttering when the wing size is enlarged or reduced is calculated. When the wing size is magnified by (1+δ1) with the interpolation ratio being δ1, it is assumed that the fluttering lift force of the fluttering robot is magnified by (1+δ2). [0250] Further, assume that when the amplitude θw of fluttering is magnified by (1+δ3), the fluttering lift force of the fluttering robot is magnified to (1+δ4). From these result, it follows that when the wing size is magnified by (1+δ1), fluttering lift force comparative to that before changing the wing size can be attained by magnifying the amplitude θw of fluttering by (1−δ3×δ2/δ4). When the change in the wing shape is replaced by one parameter and the change in lift force with respect to the change in the one parameter is calculated, the method of analysis can be simplified. [0251] Consider that this method is further developed so that all the parameters representing the structure of the equivalent numerical model [0252] Here, when the numerical model of the small change newly prepared by interpolation, that is, the group of parameters Mb_new (i) of the numerical model [0253] More specifically, it becomes possible to calculate the amplitude θw of fluttering that ensures the same lift force as Mn (i), with respect to Mb_new (i). [0254] Further, the parameter group Mb_new (i) obtain as a result of analysis is newly regarded as Mb_old (i), and again the equation (1) is applied, whereby various sensitivity analysis and modification of the amplitude θw of fluttering based thereon are repeated. [0255] By the repetition of such an operation, Mb_new (i) is updated and, accordingly, the amplitude θw of fluttering is modified. Thus, the amplitude θw of fluttering that realizes the same manner of fluttering lift as the original equivalent numerical model of actual structure Mn (i) for the new model Mm (i) can be found. [0256] More specifically, referring to FIG. 14, relative arrangement of a mesh in the numerical model of the structure of artificial wing [0257] Though accuracy decreases, the above described method is also applicable by using some interpolation, even when relative positional relation of meshes is different. [0258] Further, Mn (i) or the like may be a group of parameters of a function that represents wing shape or rigidity distribution. Here, the sensitivity analysis described above is on the premise that it is performed on a sufficiently small δ, where the change in each parameter can be assumed to be linear. [0259] For simplicity of description, only the amplitude θw of fluttering has been described in connection with the change in the manner of fluttering. Actually, however, parameters representing lift other than the fluttering flight force also change, when the amplitude θw of fluttering is changed. [0260] Therefore, actually, sensitivity of various parameters representing the manner of fluttering with respect to the fluttering force (forces related to fluttering other than the fluttering lift force: for example, thrust generated together with the fluttering lift force) is calculated, and a parameter representing such a manner of fluttering that offsets the change in the fluttering force mentioned above is calculated, as the numerical model of fluttering corresponding to the numerical model of the interpolated structure. More specifically, sets of parameters representing fluttering that cancel the change of fluttering force generated by the interpolation of the numerical model of structure are determined by using linear programming, for example, in consideration of the characteristics of actuator [0261] The foregoing descriptions will be summarized. [0262] In the method of manufacturing a fluttering robot described above, first, a numerical model related to the structure of artificial wing [0263] Thereafter, using the numerical model of detailed figure Mn (0) and detailed numerical motion model N (0), fluid-structure interactive analysis is performed, whereby a numerical model S (0) related to the structure of the numerical model of detailed structure Mn (0) and numerical model F (0) related to the fluid of the numerical model of detailed structure Mn (0) are calculated. [0264] Thereafter, a first numerical model of interpolated structure Mb_new (1)=Mb_old (0)+δ5×(Mm (1)−Mn (0)) is prepared, by interpolating, with a prescribed interpolation ratio, the numerical model related to the structure of artificial wing [0265] Consider that the numerical model S (0) related to the structure and the numerical model F (0) related to the fluid are changed to numerical model S (1, 0) related to the structure and the numerical model F (1, 0) related to the fluid resulting from fluid-structure interactive analysis using the first numerical model of interpolated structure Mb_new (1) and the detailed numerical motion model N (0),. The detailed numerical motion model N (0) is changed such that the degree of change of a specific numerical model (for example, lift force) becomes smaller than the degree of change of the specific numerical model resulting from the change caused by the interactive analysis. In this manner, a first numerical motion model of the interpolated structure N (1) that corresponds to the first numerical model of the interpolated structure Mb_new (1) is prepared. [0266] Thereafter, fluid-structure interactive analysis is performed using the first numerical model of the interpolated structure Mb new (1) and the first numerical motion model of the interpolated structure N (1), a numerical model S (1) related to the structure of the first numerical model of the interpolated structure Mb_new (1) and the numerical model F (1) related to the fluid of the first numerical model of the interpolated structure Mb_new (1) are calculated. Thereafter, a second numerical model of the interpolated structure Mb_new (2)=Mb_old (1)+δ5×(Mm (2)−Mn (1)) is prepared by interpolating, with a prescribed interpolation ratio, the numerical model related to the structure of artificial wing [0267] Consider that the numerical model S (1) related to the structure and the numerical model F (1) related to the fluid are changed to numerical model S (2, 1) related to the structure and the numerical model F (2, 1) related to the fluid resulting from fluid-structure interactive analysis using the second numerical model of the interpolated structure Mb new (2) and the first numerical motion model of the interpolated structure N (1). The first numerical motion model of the interpolated structure N (1) is changed such that the degree of change of the specific numerical model (for example, lift force) becomes smaller than the degree of change of the specific numerical model resulting from the change caused by the interactive analysis. Thus, the second numerical motion model of the interpolated structure N (2) corresponding to the second numerical model of the interpolated structure Mb new (2) is prepared. [0268] Thereafter, fluid-structure interactive analysis is performed using the second numerical model of the interpolated structure Mb_new (2) and the second numerical motion model of the interpolated structure N (2), whereby the numerical model S (2) related to the structure of the second numerical model of the interpolated structure Mb_new (2) and the numerical model F (2) related to the fluid of the second numerical model of the interpolated structure Mb_new (2) are calculated. [0269] Thereafter, by successively increasing the value i, respective numerical models described above (numerical motion model of interpolated structure N (i), numerical model S (i) related to the structure and numerical model F (i) related to the fluid) are successively updated, until the numerical model of interpolated structure Mb_new (i) matches or is approximated to the numerical model related to the structure of artificial wing [0270] Thus, the step of preparing numerical model of interpolated structure Mb_new (i), the step of preparing numerical motion model of interpolated structure N (i) and the step of preparing numerical model S (i) related to the structure and numerical model F (i) related to the fluid by performing fluid-structure interactive analysis are accumulatively repeated. [0271] As a result, a numerical motion model of interpolated structure Mb_new (matching or approximated) corresponding to the numerical model of interpolated structure Mb_new (matching or approximated) that matches or is approximated to the numerical model related to the structure of artificial wing [0272] The numerical motion model of interpolated structure N (i) is a numerical motion model that corresponds to the numerical model of interpolated structure Mb_new (i). The numerical model S (i) related to the structure and the numerical model F (i) related to the fluid are numerical models resulting from fluid-structure interactive analysis under the condition that the wing in accordance with the numerical model of interpolated structure Mb_new (i) is caused to fly, fluttering in the manner in accordance with the numerical motion model of the interpolated structure N (i). [0273] By this method, when the numerical model of interpolated structure Mb_new (i) is interpolated to be closer to the numerical model related to the structure of artificial wing [0274] (Optimization of Wing Structure) [0275] In the foregoing, a method has been described in which the wing structure is changed aiming at the structure of artificial wing [0276] More specifically, it is possible to prepare a wing structure that is suitable for a specification required for a certain fluttering robot. Speaking only of the wing shape, sensitivity of parameters of the manner of fluttering flight with respect to movement of each node of the wing in x, y and z directions, respectively, are calculated, each node of the wing is moved in accordance with the sensitivity of the manner of fluttering flight so that the parameters come close to optimal specification, and such operation is repeated. [0277] More specifically, when the lift force by fluttering is to be maximized, sensitivity of fluttering lift force with respect to movement of each node of the wing is calculated, and each node is moved by a small amount in accordance with the sensitivity, so that the sensitivity increases. For example, a node of which sensitivity is negative is moved to the negative direction. It is noted that the sensitivity changes non-linearly, and therefore, sensitivity analysis must be newly performed for the change of the shape. Further, it is also possible to perform sensitivity analysis using parameters obtained by reducing characteristics of the wing. For example, it may be possible to prepare a wing structure by calculating ratio of enlargement of a wing that can obtain a certain fluttering lift force at a certain fluttering frequency with respect to the original wing. [0278] (Optimization of the Manner of Fluttering) [0279] The above described sensitivity analysis may be used by itself, to optimize the method of controlling the manner of fluttering flight of the fluttering robot. [0280] For example, lift force in the fluttering of an insect may significantly fluctuate, as shown in FIG. 11. By changing the wing attitude, that is, time-sequential data of wing position based on the sensitivity analysis, it is possible to obtain a manner of fluttering with less fluctuation while maintaining the fluttering lift force. In this manner, it is possible to prepare a control method based on the required specification corresponding to the artificial article. [0281] Further, it is also possible to consider a method of preparing numerical model of the manner of fluttering drive that brings about a prescribed manner of fluttering motion of the fluttering robot. This is attained by performing sensitivity analysis of the change in the manner of motion of the fluttering robot with respect to the change in W (i), where W (i) is a parameter representing the manner of driving the wing, and by changing the manner of fluttering drive based on the result of analysis. [0282] Using such a method of preparing numerical model of the manner of fluttering flight, it is possible to study the nature of the numerical model of the manner of fluttering drive in a simple manner, even when the numerical model of the manner of fluttering drive does not satisfy the condition of lift. For example, in the conventional method through experiment, when the fluttering robot turns to the right, it is necessary to study the manner of fluttering drive that enables a right turn while the robot is in a lifted state. By the method using the numerical model of the manner of fluttering drive prepared in accordance with the fluid-structure interactive analysis of the present embodiment, it is possible to prepare a numerical model of the manner of fluttering flight for a right turn at first, and thereafter, to change the numerical model of the manner of fluttering flight such that the change in lift force resulting from the manner of fluttering drive for a right turn is compensated for. [0283] (Optimization of Actuator Drive) [0284] Similar to the discussion above, actuator drive can also be modified. For example, there may be a case in which driving of actuator [0285] Assume that the number of images obtained per one second by a high speed camera, that is, sampling frequency, is 2×fc. The highest control frequency for driving the wing [0286] By these sensitivities, it becomes possible to obtain the change in shape of wing [0287] (Manufacturing of Fluttering Robot) [0288] By combining the methods using sensitivity analysis described above, it is possible to obtain and prepare the manner of fluttering that can ensure similar fluttering lift force, based on the numerical models described above, even for a fluttering robot of which wing shape [0289] Further, it is possible to prepare the wing structure and control method thereof that take into consideration the driving characteristic of the actuator. Similar method may also be used, not only for the fluid-structure interactive numerical model described above but for the model of nervous system, model of driving muscles or model of information processing of an insect, to prepare a new numerical model based on the specification of an artifact, using the corresponding numerical model of the insect as a base, and to manufacture a fluttering robot based on the thus prepared numerical model. The procedures for manufacturing the fluttering robot are as shown generally in FIG. 1. [0290] (Others) [0291] (Modeling Procedures) [0292] Dragon flies of the same species generally have the wings of the same shape, though there is individual difference. The wings have very complicated shape, and therefore, fine modeling of the structure or rigidity would be time-consuming. [0293] In the fluid-structure interactive analysis of the present embodiment, as a method of optimizing the wing structure or the manner of motion, sensitivity analysis, which is the most explicit, has been used. It is also possible, however, to use different method as the method of optimization. Examples of other effective methods may include a method using learning with neural networks, and a method of optimization using genetic algorithm. [0294] In the method of optimizing wing design using generic algorithm, first, each element of the group of parameters M (i) representing the numerical model of each wing structure described above is regarded as a gene and coded. Fluid-structure interactive analysis is performed on the numerical model of the wing structure represented thereby, performing a prescribed motion in the air. The result thus obtained is evaluated to be satisfactory or not, using some evaluation function, such as magnitude of the lift force. Based on the evaluation, numerical models having genes of satisfactory structure are multiplied (combined), and such a process is repeated, whereby a numerical model of a wing structure that is close to an optimal structure can be prepared. [0295] The method of using gene algorithm can also be applied to preparation of a numerical model of the manner of fluttering motion. As an example, there is a method in which the parameter representing wing motion is given by W (i), and the gene algorithm is applied to W (i), in the same manner as to M (i). In this method, first, the following relations are set, where w is a function representing an angle of the wing with respect to a horizontal plane, T represents time for one period of fluttering, ω=2×π/T, f [0296] It is noted that the wing, when separated from the living organism, loses moisture rapidly, and hence rigidity thereof changes. From the same reason, it is necessary to maintain the physiological condition of the dragon fly while measurement is made. [0297] In view of environmental protection, the dragon fly used for obtaining data should desirably be returned to the nature after the end of measurement, and the physiological condition of the insect should be maintained during measurement. When the measurement takes a long time, it becomes necessary to prepare water and feed, and to maintain temperature and moisture, and therefore, considerable labor and facility are necessary. [0298] From the reasons above, though it is desirable to measure rigidity of the wing without damaging a dragon fly, it is difficult to measure fine shape or rigidity without breaking the wing. Therefore, in the present embodiment, a solution is made by preparing a sample used for fine measurement of the wing, and a separate sample used for measuring wing drive. If it is unnecessary to perfectly satisfy the above described conditions, it is possible to model the wing and the manner of fluttering, through different methods. [0299] For example, if the measurements can be performed in a sufficiently short period of time as compared with the life of the insect for measurement, it is possible to prepare the equivalent numerical model of actual structure of the wing separately for each sample. Therefore, the series of processes for preparing the numerical model of detailed figure of the wing are not essential. [0300] (Method of Fluid-Structure Interactive Analysis) [0301] Various proposals have been made on the fluid-structure interactive analysis. [0302] As the simplest method, it is possible to determine velocity by solving a moving boundary problem of the fluid only, from the video images of the wing picked-up by a high speed camera. The wing structure can be analyzed by deforming the structure by itself. [0303] In this method, however, it is necessary to measure in every step of analysis the movement of every portion (point) of the wing. Therefore, the data amount would be formidable. Further, the analysis is limited to the results of fluttering flight, and therefore, this method cannot be used for an application to the fluttering robot described above. [0304] A method solving a problem having interaction between fluid and the structure, weak coupling method in which determinant equations of fluid and structure are calculated alternately, and strong coupling method in which equations of the entire system including fluid and structure (coupling equations) are calculated at one time have been proposed. According Qun Zhang et al. mentioned above, it is most efficient and optimal to use the strong coupling method, to solve the problem having strong interaction between fluid and the structure, as in the present embodiment. [0305] The fluid-structure interactive structural analysis of Qun Zhang et al. described above has been made on a structure other than a living organism. A fluid-structure interactive structural analysis of the present embodiment has been made on a behavior of a living organism, taking fluttering motion of an insect in the air as an example. It is the feature of the present invention that, by applying the fluid-structure interactive structural analysis to a living organism, manufacturing of a robot mimicking the living organism is facilitated. [0306] (Second Embodiment) [0307] The method of preparing the fluid-structure interactive numerical model in accordance with the second embodiment will be described with reference to Table 7 and FIGS. [0308] The numerical model of the structure used in the method of preparing fluid-structure interactive numerical model of the present embodiment corresponds to the numerical model of the structure used in the method of preparing fluid-structure interactive numerical model of the first embodiment shown in FIG. 4, with the thicknesses of the wing changed, from 0.35 mm, 0.18 mm, 0.15 mm, 0.12 mm, 0.1 mm and 0.05 mm, respectively, to 0.045 mm, 0.012 mm, 0.010 mm, 0.008 mm, 0.006 mm and 0.004 mm. [0309] In the method of preparing fluid-structure interactive numerical model of the first embodiment, the fluttering motion is approximately in the vertical direction, and hence the direction of generation of the lift force is positive along the y axis in FIG. 17. In the method of preparing fluid-structure interactive numerical model of the present embodiment, the direction of fluttering is changed to approximately horizontal direction, and therefore, the lift force is generated in the direction positive along the z axis. Thus, the lift force in this direction is considered to be positive. [0310] In the method of preparing fluid-structure interactive numerical model of the present embodiment, the mesh structure shown in FIGS. 8 and 9 used in the first embodiment are replaced by the mesh structures shown in FIGS. 21 and 22. [0311] Further, in the method of preparing fluid-structure interactive numerical model of the present embodiment, the relation between the lift force and time shown in FIG. 11 obtained by the first embodiment is as shown in FIG. 23. [0312] In the method of preparing fluid-structure interactive numerical model of the present embodiment, the relation between position and time shown in FIGS. 18 and 19 obtained in the first embodiment is changed to the relation shown in FIGS. 24 and 25. [0313] Further, in the method of preparing fluid-structure interactive numerical model of the present embodiment, the relation between torque and time shown in FIG. 20 obtained by the first embodiment is changed to the relation shown in FIG. 26. Here, referring to FIG. 26, Tθ represents a torque for obtaining driving in the θ direction, and Tβ represents torque for driving in the β direction. [0314] In the method of preparing fluid-structure interactive numerical model of the present embodiment, the x and y coordinate values of Node83 of Table 1 used in the first embodiment are changed from 0.001 to 0.000889 and from 0.001 to 0.000889, respectively. [0315] Further, in the method of preparing fluid-structure interactive numerical model of the present embodiment, Young's modulus, mass density and thickness of Table 4 used in the first embodiment are replaced from 1.00E+09 to 0.5E+09, from 1.2E+03 to 0.7E+03, and from 0.35E−03 to 4.5E−05, respectively. [0316] In ShellMotion.dat, time-history of x, y and z coordinates at the nodes of the wing are as shown in Table 7, in place of those shown in Table 6.
[0317] As regards other figures and tables, the configuration and method used in the method of preparing fluid-structure interactive numerical model of the present embodiment are the same as those of the first embodiment. [0318] Although the present invention has been described and illustrated in detail, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation, the spirit and scope of the present invention being limited only by the terms of the appended claims. Classifications
Legal Events
Rotate |