Method for defining fluid/solid boundary for computational fluid dynamics simulations

A method for defining a fluid/solid boundary for computational fluid dynamics simulations making use of

• a coordinate system mesh which models a region of three-dimensional space including a fluid region, and which is defined by a large number of grid points arranged in the region of three-dimensional space, and
• a solid object model which models a solid object, and whose surface is constituted by planes of finite elements, wherein the method comprises
• a step (a) of preparing the solid object model in a computer,
• a step (b) of preparing the coordinate system mesh in the computer,
• a step (c) of defining, in the computer, straight lines which extend across the above-mentioned region of three-dimensional space, passing through the grid points,
• a step (d) of obtaining intersecting points of the straight lines with the surface of the solid object model,
• a step (e) in which, for each of the straight lines having the intersecting points, the grid points positioned on the straight line are searched for a nearest grid point to each of the intersecting points, and based on the searched-out nearest grid points, the grid points positioned on each of the straight lines are each determined as to whether the grid point is positioned inside the solid object model or in the fluid region,
• whereby all of the grid points arranged in the region of three-dimensional space are determined as to whether the grid point is positioned in the fluid region or inside the solid object model and thereby the boundary between the solid object model and the fluid region is defined on the coordinate system mesh,
• wherein, in the search process in the step (e) performed for each of the straight lines having the intersecting points in order to search the grid points positioned on the straight line for a nearest point to each of the intersecting points,
• taking the coordinates of the grid points positioned on the straight line under consideration and the intersecting points thereof as one-dimensional values based on a one-dimensional coordinate system defined along the straight line under consideration, and
• giving the one-dimensional coordinate values of the grid points as a set of real numbers x[i] (i:
• 1 to n), and
• giving the one-dimensional coordinate values of each of the intersecting points as a query point q,
• the search of the grid points for a nearest point to each of the intersecting points is performed by a nearest neighbor search method comprising
• a construction phase in which a database of the real numbers x[i] (i:
• integers from 1 to n, assigned in ascending order of the real numbers) is prepared in a memory of the computer, and
• a search phase in which the computer searches the nearest neighbor to the query point q by the use of the database, wherein
• the database comprises a series of buckets,
• the buckets respectively correspond to small one-dimensional spaces defined by dividing a one-dimensional space between a minimum real number x[1] and a maximum real number x[n] at regular intervals, whereby the number of the buckets and the number of the small one-dimensional spaces are the same integral number m,
• the buckets have unique bucket indexes from the first bucket B[1] to the last bucket B[m],
• each of the buckets comprises data about real number(s) falling in one of the small one-dimensional spaces corresponding to the bucket under consideration,
• each of the buckets further comprises data about its bucket size, the bucket size defined by the number of real number(s) falling in one of the small one-dimensional spaces which corresponds to the bucket under consideration,
• each of the buckets excepting the first bucket B[1] further comprises data about the bucket index of a last filled bucket, the last filled bucket defined as a bucket whose bucket size is not zero and whose bucket index is closest to that of the bucket under consideration, and
• the search phase which comprises:
• a first step of locating one of the buckets corresponding to one of the small one-dimensional spaces in which the query point q falls,
• a second step of checking the bucket size of the located bucket whether the bucket size thereof is zero or not,
• a third step of searching the nearest neighbor by the use of the data of real number(s) included in
• the located bucket if the checked bucket size is not zero,
• or
• a bucket with a bucket index equal to the bucket index of the last filled bucket included in the located bucket if the checked bucket size is zero.