Grid interface host unit parallel search method, device and storage medium

Abstract

The present invention relates to the technical field of computational fluid dynamics, and specifically discloses a method, apparatus and storage medium for parallel searching of host cells at grid interfaces, including: partitioning any two paired first grid regions and second grid regions respectively to obtain first sub-grid regions and second sub-grid regions having the same number as the number of processes; determining the first virtual point coordinates of the first interface subset and the second virtual point coordinates of the second interface subset; searching for the first approximate host cell of the first virtual point according to the first virtual point coordinates, and searching for the second approximate host cell of the second virtual point according to the second virtual point coordinates; determining the first host cell of the first interface according to the first approximate host cell, and determining the second host cell according to the second approximate host cell. The method for parallel searching of host cells at grid interfaces provided by the present invention realizes the searching in the parallel mode of host cells at grid interfaces.

Description

Technical Field

[0001] The present invention relates to the technical field of computational fluid dynamics, and in particular, to a method for parallel searching of host cells at grid interfaces, a device for parallel searching of host cells at grid interfaces, and a storage medium. Background Art

[0002] Currently, in engineering designs in fields such as aerospace, ships, and automobiles, the application of simulation calculations is becoming increasingly widespread, and at the same time, the model structure is becoming more and more complex. For complex geometric shapes, especially models containing moving parts, it is necessary to separately model non-moving parts and moving parts, and couple the movement and non-moving parts during simulation calculations. During the coupling process, it is necessary to process the interface between the meshes of the two parts to achieve the transfer of field variables.

[0003] The key to processing the grid interface lies in determining the field variables on the interface. The current processing method is as Figure 1 shown. A point is extended from the centroid celement of the current-side grid cell to the other-side grid region through the centroid cface of the interface. This point is called the virtual point virtual-point of the interface. By determining the position of the virtual point in the other-side grid region, the value at the virtual point is given. The conventional method is to assign the value of the grid cell where the virtual point is located to the virtual point. After assigning a value to the virtual point, the value on the interface can be determined by interpolation, thereby realizing the numerical transfer between the two grid regions.

[0004] In the above processing method, the most crucial thing is to find the grid cell where the virtual point is located, that is, the host cell of the virtual point or the interface. Currently, the conventional methods include the barycentric coordinate method. In the serial mode, due to the integrity of the grid, the grid cell containing the virtual point can be accurately found, ensuring the accuracy of the fluid field numerical transfer. However, when using the parallel mode to improve the calculation speed, after the grid is partitioned, there will be a situation where the search path is interrupted by the process grid boundary, as Figure 2 shown.

[0005] Therefore, how to achieve the search of the host cell at the grid interface in the parallel mode has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0006] The present invention provides a method for parallel searching of host cells at grid interfaces, a device for parallel searching of host cells at grid interfaces, and a storage medium, which solve the problem in the related art that the search of the host cell at the grid interface in the parallel mode cannot be achieved.

[0007] As the first aspect of the present invention, a method for parallel searching of host cells at grid interfaces is provided, which includes:

[0008] Partition the first grid region and the second grid region of any two paired regions respectively to obtain first sub-grid regions and second sub-grid regions with the same number as the number of processes. Among them, the first grid region includes a first set of interface surfaces, the second grid region includes a second set of interface surfaces, and the first grid region and the second grid region are connected through the first set of interface surfaces and the second set of interface surfaces. Each process includes a first sub-grid region and a second sub-grid region. The first sub-grid region of at least one process includes a first sub-set of interface surfaces, and the second sub-grid region of at least one process includes a second sub-set of interface surfaces. The first sub-sets of interface surfaces of all processes form the first set of interface surfaces, and the second sub-sets of interface surfaces of all processes form the second set of interface surfaces;

[0009] Determine the first virtual point coordinates of each first interface surface in the first sub-set of interface surfaces of each process and the second virtual point coordinates of each second interface surface in the second sub-set of interface surfaces respectively;

[0010] For each process, search for the first approximate host cell of the first virtual point in the second grid cells corresponding to the second sub-sets of interface surfaces of all processes according to the first virtual point coordinates of each first interface surface, and search for the second approximate host cell of the second virtual point in the first grid cells corresponding to the first sub-sets of interface surfaces of all processes according to the second virtual point coordinates of each second interface surface;

[0011] For each process, determine the first host cell of the first interface surface according to the first approximate host cell of the first virtual point of each first interface surface, and determine the second host cell of the second interface surface according to the second approximate host cell of the second virtual point of each second interface surface;

[0012] Realize the numerical transfer of the flow field between the first grid region and the second grid region according to the field variables in the first host cells of all processes and the field variables in the second host cells of all processes.

[0013] Furthermore, the first sub-grid region of each process includes a plurality of first grid cells, and the second sub-grid region of each process includes a plurality of second grid cells. Determining the first virtual point coordinates of each first interface surface in the first sub-set of interface surfaces of each process and the second virtual point coordinates of each second interface surface in the second sub-set of interface surfaces respectively includes:

[0014] For each process, calculate the first body center coordinates and first face center coordinates of all first grid cells in the first sub-grid region, and calculate the second body center coordinates and second face center coordinates of all second grid cells in the second sub-grid region;

[0015] Calculate the first virtual point coordinates of each first interface in the first interface subset of the process based on the first centroid coordinates of the first interface and the first centroid coordinates of the first grid cell where the first interface is located, and calculate the second virtual point coordinates of each second interface in the second interface subset of the process based on the second centroid coordinates of the second interface and the second centroid coordinates of the second grid cell where the second interface is located.

[0016] Furthermore, the calculation formulas for the first virtual point coordinates and the second virtual point coordinates are both:

[0017] C virtual-Poin = C face + |C face - C element |e,

[0018] where C face represents the centroid coordinate of the first interface or the second interface, C element represents the centroid coordinate of the first grid cell where the first interface is located or the second grid cell where the second interface is located, C virtual-Point represents the first virtual point coordinate or the second virtual point coordinate, |C face - C element | represents the distance from the centroid of the first interface to the centroid of the first grid cell where it is located, or the distance from the centroid of the second interface to the centroid of the second grid cell where it is located, and e represents the unit vector in the outer normal direction of the first interface or the unit vector in the outer normal direction of the second interface.

[0019] Furthermore, for each process, search for the first approximate host cell of the first virtual point in the second grid cells corresponding to the second interface subset of all processes according to the first virtual point coordinates of each first interface, including:

[0020] Collect the second centroid coordinates of each second grid cell where each second interface in the second interface subset of all processes is located to form a second centroid coordinate set;

[0021] For each process, find the second centroid coordinate closest to the first virtual point in the second centroid coordinate set according to the first virtual point coordinates of each first interface;

[0022] Determine the target second grid cell where the second centroid coordinate closest to the first virtual point is located as the first approximate host cell, determine the process where the target second grid cell is located as the process where the first approximate host cell is located, and determine the grid number of the target second grid cell as the grid number of the first approximate host cell.

[0023] Further, for each process, searching for the second approximate host cell of the second virtual point in the first grid cells corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface includes:

[0024] Collecting the first centroid coordinates of the first grid cells where each first interface in the first interface subsets of all processes is located to form a first centroid coordinate set;

[0025] For each process, finding the first centroid coordinate closest to the second virtual point in the first centroid coordinate set according to the second virtual point coordinates of each second interface;

[0026] Determining the target first grid cell where the first centroid coordinate closest to the second virtual point is located as the second approximate host cell, determining the process where the target first grid cell is located as the process where the second approximate host cell is located, and determining the grid number of the target first network cell as the grid number of the second approximate host cell.

[0027] Further, for each process, determining the first host cell of each first interface according to the first approximate host cell of the first virtual point of each first interface includes:

[0028] For each process, taking the first approximate host cell of the first virtual point of each first interface as the starting cell, and performing host cell search in the second sub-grid region of the process where the first approximate host cell is located according to the centroid coordinate method to obtain a first search result;

[0029] Judging whether there is a real first host cell according to the first search result;

[0030] If there is no real first host cell, determining the first approximate host cell as the first host cell of the first interface.

[0031] Further, for each process, determining the second host cell of each second interface according to the second approximate host cell of the second virtual point of each second interface includes:

[0032] For each process, taking the second approximate host cell of the second virtual point of each second interface as the starting cell, and performing host cell search in the first sub-grid region of the process where the second approximate host cell is located according to the centroid coordinate method to obtain a second search result;

[0033] Judging whether there is a real second host cell according to the second search result;

[0034] If there is no real second host cell, determining the second approximate host cell as the second host cell of the second interface.

[0035] Further, the numerical transfer of the flow field between the first grid region and the second grid region is realized according to the field variables in the first host unit of all processes and the field variables in the second host unit of all processes, including:

[0036] Assign values to the first virtual points according to the field variables of the first host unit of each process, and assign values to the second virtual points according to the field variables of the second host unit of each process;

[0037] Send the assignments of the first virtual points and the assignments of the second virtual points to the processes where the corresponding interface is located, to realize the numerical transfer of the flow field between the first grid region and the second grid region.

[0038] As another aspect of the present invention, there is provided a grid interface host unit parallel search device for implementing the grid interface host unit parallel search method described above, which includes:

[0039] A partitioning module for partitioning any two paired first grid regions and second grid regions respectively to obtain first sub-grid regions and second sub-grid regions having the same number as the number of processes. Wherein, the first grid region includes a first interface set, the second grid region includes a second interface set, the first grid region and the second grid region are connected through the first interface set and the second interface set. Each process includes a first sub-grid region and a second sub-grid region. The first sub-grid region of at least one process includes a first interface subset, and the second sub-grid region of at least one process includes a second interface subset. The first interface subsets of all processes form the first interface set, and the second interface subsets of all processes form the second interface set;

[0040] A virtual point coordinate determination module for respectively determining the first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset of each process;

[0041] A search module for, for each process, searching for the first approximate host unit of the first virtual point in the second grid cells corresponding to the second interface subsets of all processes according to the first virtual point coordinates of each first interface, and searching for the second approximate host unit of the second virtual point in the first grid cells corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface;

[0042] A host unit determination module, which is configured to, for each process, determine the first host unit of each first interface according to the first approximate host unit of the first virtual point of each first interface, and determine the second host unit of each second interface according to the second approximate host unit of the second virtual point of each second interface;

[0043] A flow field numerical transfer module, which is configured to implement the flow field numerical transfer between the first grid region and the second grid region according to the field variables in the first host units of all processes and the field variables in the second host units of all processes.

[0044] As another aspect of the present invention, a storage medium is provided, which is configured to store computer instructions, and when the computer instructions are loaded and executed by a processor, the grid interface host unit parallel search method described above is implemented.

[0045] The grid interface host unit parallel search method provided by the present invention determines virtual points in the interface set of the paired grid regions to determine approximate host units, and determines host units according to the approximate host units, so as to implement the flow field numerical transfer of the paired grid regions under multiple processes, thereby realizing the parallel search of the grid interface host units. Therefore, the grid interface host unit parallel search method provided by the present invention can realize the search in the parallel mode of the grid interface host units. In addition, the grid interface host unit parallel search method of the present invention can ensure that when the approximate host unit is used as the host unit because the real host unit is not finally found in the interface grid of the paired grid, the numerical values at the virtual points will not deviate too much. Description of the Drawings

[0046] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used to explain the present invention together with the following specific embodiments, but do not constitute a limitation to the present invention.

[0047] Figure 1 A schematic diagram of a grid interface processed by the prior art.

[0048] Figure 2 A schematic diagram showing that the host unit search path of the virtual point is interrupted by the process grid boundary after the grid is partitioned by the prior art.

[0049] Figure 3 A flowchart of the grid interface host unit parallel search method provided by the present invention.

[0050] Figure 4 A flowchart of the specific implementation process of the grid interface host unit parallel search method provided by the present invention.

[0051] Figure 5Flow chart for determining the coordinates of virtual points provided by the present invention.

[0052] Figure 6 Flow chart for the first approximate host cell search provided by the present invention.

[0053] Figure 7 Flow chart for the second approximate host cell search provided by the present invention.

[0054] Figure 8 Schematic diagram of the principle of the centroid coordinate method provided by the present invention.

[0055] Figure 9a Schematic diagram of the infinite long cavity roof-driven square cavity flow model provided by the present invention.

[0056] Figure 9b Schematic diagram of the grid of the infinite long cavity roof-driven square cavity flow provided by the present invention.

[0057] Figure 10a For Figure 9a Schematic diagram of the model after the model is partitioned.

[0058] Figure 10b For Figure 9b Schematic diagram of the grid after the grid is partitioned.

[0059] Figure 11 Schematic diagram of the distribution of virtual points, approximate host cells and host cells at the interface between the inner and outer square regions provided by the present invention.

[0060] Figure 12 Schematic diagram of the situation where the host cell search is blocked by the process boundary provided by the present invention.

[0061] Figure 13a Schematic diagram of the 10-process parallel calculation result provided by the present invention.

[0062] Figure 13b Schematic diagram of the non-interface serial calculation result provided by the present invention.

[0063] Figure 14 Schematic diagram of the propeller open water model provided by the present invention.

[0064] Figure 15 Schematic diagram of the propeller open water grid provided by the present invention.

[0065] Figure 16a Pressure contour map of the propeller blade surface after partition processing provided by the present invention.

[0066] Figure 16b Pressure contour map of the propeller blade surface without partition processing in the prior art.

[0067] Figure 17 It is a structural block diagram of the grid interface host unit parallel search device provided by the present invention.

[0068] Figure 18 It is a structural block diagram of the electronic device provided by the present invention. Detailed implementation manners

[0069] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments may be combined with each other. The present invention will be described in detail below with reference to the drawings and in combination with the embodiments.

[0070] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0071] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so as to implement the embodiments of the present invention described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily need to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0072] In this embodiment, a grid interface host unit parallel search method is provided. Figure 3 It is a flowchart of the grid interface host unit parallel search method provided by the embodiment of the present invention. As Figure 3 shown, it includes:

[0073] S100. Partition the first grid region and the second grid region of any two paired grid regions respectively to obtain first sub-grid regions and second sub-grid regions with the same number as the number of processes. Among them, the first grid region includes a first interface set, the second grid region includes a second interface set, the first grid region and the second grid region are connected through the first interface set and the second interface set. Each process includes a first sub-grid region and a second sub-grid region. The first sub-grid region of at least one process includes a first sub-interface set, and the second sub-grid region of at least one process includes a second sub-interface set. The first sub-interface sets of all processes form the first interface set, and the second sub-interface sets of all processes form the second interface set;

[0074] It should be noted that in the embodiments of the present invention, the set of interfaces with the interface type in the first grid region forms the first interface set. Similarly, the set of interfaces with the interface type in the second grid region forms the second interface set. Since both the first grid region and the second grid region are divided into the same number of first sub-grid regions and second sub-grid regions according to the number of processes, and the set of interfaces with the interface type in the first sub-grid region forms the first sub-interface set, and the set of interfaces with the interface type in the second sub-grid region forms the second sub-interface set. It should also be understood here that the first sub-interface sets of all processes form the first interface set, and the second sub-interface sets of all processes form the second interface set.

[0075] Specifically, for example, if the number of processes is 10, the paired first grid region and second grid region can be partitioned to obtain 10 first sub-grid regions and 10 second sub-grid regions. The first grid region and the second grid region are connected through interfaces. The first grid region includes a first interface set, and the second grid region includes a second interface set. The first interface set of the first grid region is connected to the second interface set of the second grid region. After partitioning the above 10 processes into 10 first sub-grid regions and 10 second sub-grid regions, each process includes a first sub-grid region and a second sub-grid region. When the interface types of all the first sub-grid regions and second sub-grid regions after partitioning include interfaces, the first sub-grid regions and second sub-grid regions of all processes include sub-interface sets; of course, after partitioning, there may still be cases where the interface types of the first sub-grid regions and second sub-grid regions in some processes do not include interfaces. It can be understood that this situation will not affect the subsequent search for host units.

[0076] S200. Determine the first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset;

[0077] In the embodiments of the present invention, it is necessary to determine the first virtual point coordinates of each first interface in the first interface subset of each process, and similarly determine the second virtual point coordinates of each second interface in the second interface subset of each process.

[0078] It should be understood that for each process, the first sub-grid region includes a plurality of first grid cells, and the second sub-grid region includes a plurality of second grid cells. For the plurality of first interfaces included in the first interface subset, each first interface in the first interface subset belongs to a first grid cell, and not all first grid cells have a first interface. Similarly, for the plurality of second interfaces included in the second interface subset, each second interface in the second interface subset belongs to a second grid cell, and not all second grid cells have a second interface.

[0079] S300. For each process, search for the first approximate host cell of the first virtual point in the second grid cells corresponding to the second interface subsets of all processes according to the first virtual point coordinates of each first interface, and search for the second approximate host cell of the second virtual point in the first grid cells corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface;

[0080] In the embodiments of the present invention, for each process, after determining the first virtual point coordinates of each first interface, search for the first approximate host cell of the first virtual point in the second grid cells corresponding to the second interface subsets of all processes. Specifically, it can be understood that, for example, there are 10 processes, and each process has a second interface subset. Therefore, the 10 processes correspond to 10 second interface subsets, and each second interface subset corresponds to a plurality of second grid cells. Therefore, by searching all the second grid cells corresponding to the 10 second interface subsets, the first approximate host cell of the first virtual point is determined; similarly, the second approximate host cell of the second virtual point can be determined by searching the plurality of first grid cells corresponding to the 10 first interface subsets of the 10 processes.

[0081] S400. For each process, determine the first host cell of the first interface according to the first approximate host cell of the first virtual point of each first interface, and determine the second host cell of the second interface according to the second approximate host cell of the second virtual point of each second interface;

[0082] In an embodiment of the present invention, for each process, based on the first approximate host unit of the first virtual point of each first interface, it is further determined whether the first approximate host unit is the first host unit of the first interface. Similarly, based on the second approximate host unit of the second virtual point of each second interface, it is further determined whether the second approximate host unit is the second host unit of the second interface.

[0083] S500. Implement numerical transfer of the flow field between the first grid region and the second grid region according to the field variables in the first host units of all processes and the field variables in the second host units of all processes.

[0084] After determining the first host unit and the second host unit of each process as described above, the field variables of the first host unit are sent to the corresponding first interface, and the field variables of the second host unit are sent to the corresponding second interface, thereby implementing numerical transfer of the flow field between the first grid region and the second grid region.

[0085] In summary, the grid interface host unit parallel search method provided by the present invention determines virtual points in the interface set of the paired grid regions to determine approximate host units, and determines host units based on the approximate host units, realizing numerical transfer of the flow field in the paired grid regions of multiple processes, thereby realizing parallel search of the grid interface host unit. Therefore, the grid interface host unit parallel search method provided by the present invention can achieve search in the parallel mode of the grid interface host unit. In addition, when the grid interface host unit parallel search method of the present invention searches for host units in the interface grid of the paired grid, it can ensure that when the true host unit is not finally found and the approximate host unit is used as the host unit, the values at the virtual points will not deviate too much.

[0086] In an embodiment of the present invention, as Figure 4 shown, for two grid regions G 1 and G 2 , the first interface set of the first grid region G 1 is If 1 , and the second interface set of the second grid region G 2 is If 2 . If 1 and If 1 are connected to each other to realize the connection of the grid blocks of G 1 and G 2 . The subscripts 1 and 2 are grid region numbers.

[0087] Perform parallel computing using MPI (Message Passing Interface), and use a hierarchical partitioning algorithm (such as the Metis partitioning algorithm) to partition G according to the number of processes 1 and the grid G 2 respectively. After partitioning, each process contains a first sub-grid region G 1 and a second sub-grid region G s1 of G 2 respectively. The first sub-grid region of G s1 . The first interface subset contained in G s1 is If s1 , and the second interface subset contained in G s2 is If s2 .

[0088] In the embodiment of the present invention, the first sub-grid region of each process includes a plurality of first grid units, and the second sub-grid region of each process includes a plurality of second grid units. Determine the first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset respectively, as Figure 5 shown, including:

[0089] S210. For each process, calculate the first body center coordinates and first face center coordinates of all first grid units in the first sub-grid region, and calculate the second body center coordinates and second face center coordinates of all second grid units in the second sub-grid region;

[0090] In the embodiment of the present invention, for each process, calculate the body center coordinates of all grid body units of G s1 and G s2 and the face center coordinates of all grid face units. The body center coordinates of all grids of the obtained G s1 constitute the first body center coordinates of the first sub-grid and the first face center coordinates of the first sub-grid, and the body center coordinates of all grids of the obtained G s2 constitute the second body center coordinates of the second sub-grid and the second face center coordinates of the second sub-grid. The specific calculation of the body center coordinates can be understood as taking the average of the sum of the face center coordinates of all faces that make up the body, and the specific calculation of the face center coordinates can be understood as taking the average of the sum of all point coordinates that make up the face. The specific calculation methods are well known to those skilled in the art and will not be elaborated here.

[0091] S220. Calculate the first virtual point coordinates of each first interface in the first interface subset of the process based on the first centroid coordinates of the first interface and the first centroid coordinates of the first grid cell where the first interface is located, and calculate the second virtual point coordinates of each second interface in the second interface subset of the process based on the second centroid coordinates of the second interface and the second centroid coordinates of the second grid cell where the second interface is located.

[0092] In the embodiments of the present invention, calculate the first virtual point coordinates of each first interface based on the first centroid coordinates of the first grid cell and the first centroid coordinates of the first interface in the first interface subset determined above. Similarly, calculate the second virtual point coordinates of each second interface based on the second centroid coordinates of the second grid cell and the second centroid coordinates of the second interface in the second interface subset.

[0093] Specifically, the calculation formulas for the first virtual point coordinates and the second virtual point coordinates are both:

[0094] C virtual-Point = C face + |C face - C element |e,

[0095] where, C face represents the centroid coordinates of the first interface or the second interface, C element represents the centroid coordinates of the first grid cell where the first interface is located or the second grid cell where the second interface is located, C virtual-P represents the first virtual point coordinates or the second virtual point coordinates, |C face - C element | represents the distance from the centroid of the first interface to the centroid of the first grid cell where it is located, or the distance from the centroid of the second interface to the centroid of the second grid cell where it is located, and e represents the unit vector in the outer normal direction of the first interface or the unit vector in the outer normal direction of the second interface.

[0096] In the embodiments of the present invention, for each process, search for the first approximate host cell of the first virtual point in the second grid cells corresponding to the second interface subset of all processes according to the first virtual point coordinates of each first interface. As Figure 6 shown, it includes:

[0097] S310. After collecting the second centroid coordinates of each second grid cell where each second interface in the second interface subset of all processes is located, form a second centroid coordinate set;

[0098] In the embodiments of the present invention, it should be understood that, in order to search for the first approximate host unit, the second centroid coordinates of each second grid unit where each second interface in the second interface subset of all processes are collected to form a second centroid coordinate set. For example, if there are 10 processes in total, the second centroid coordinates of each second grid unit where each second interface in the second interface subsets within the 10 processes are collected.

[0099] S320. For each process, find the second centroid coordinate closest to the first virtual point coordinate of each first interface in the second centroid coordinate set;

[0100] Find the second centroid coordinate closest to the first virtual point according to the nearest distance method in the above second centroid coordinate set.

[0101] S330. Determine the target second grid unit where the second centroid coordinate closest to the first virtual point is located as the first approximate host unit, determine the process where the target second grid unit is located as the process where the first approximate host unit is located, and determine the grid number of the target second grid unit as the grid number of the first approximate host unit.

[0102] Take the target second grid unit where the second centroid coordinate closest to the first virtual point is found as the first approximate host unit.

[0103] In the embodiments of the present invention, for each process, search for the second approximate host unit of the second virtual point in the first grid units corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface, as Figure 7 shown, including:

[0104] S340. Collect the first centroid coordinates of each first grid unit where each first interface in the first interface subsets of all processes to form a first centroid coordinate set;

[0105] S350. For each process, find the first centroid coordinate closest to the second virtual point coordinate of each second interface in the first centroid coordinate set;

[0106] S360. Determine the target first grid unit where the first centroid coordinate closest to the second virtual point is located as the second approximate host unit, determine the process where the target first grid unit is located as the process where the second approximate host unit is located, and determine the grid number of the target first network unit as the grid number of the second approximate host unit.

[0107] Similarly, similar to finding the first approximate host unit, when finding the second approximate host unit, a first set of body-centered coordinates is first formed, and the nearest distance method is used to find the first body-centered coordinate in the first set of body-centered coordinates that is closest to the second virtual point, and the target first grid unit where the found first body-centered coordinate closest to the second virtual point is located is used as the second approximate host unit.

[0108] It should be understood that each process searches for the approximate host unit of each virtual point in If s1 and If s2 For the search of the approximate host unit of the virtual point in If s1 : Collect the body-centered coordinates of the body grids where If s2 is located in all processes. The collected set of body-centered coordinates is If 2 Vc, and then If 2 Vc is sent to each process. For each interface virtual point in the process in If s1 , find the body-centered coordinate in If 2 Vc that is closest to the virtual point. The process where the grid represented by this body-centered coordinate is located and the grid cell number are the process and grid number where the approximate host unit is located. Similarly: For the search of the approximate host unit of the virtual point in If s2 : Collect the body-centered coordinates of the body grids where If s1 is located in all processes. The collected set of body-centered coordinates is If 1 Vc, and then If 1 Vc is sent to each process. For each virtual point in the process in If s2 , find the body-centered coordinate in If 1 Vc that is closest to the virtual point. The process where the cell represented by this body-centered coordinate is located and the grid cell number are the process and grid number where the approximate host unit is located.

[0109] In the embodiment of the present invention, for each process, determining the first host unit of each first interface according to the first approximate host unit of the first virtual point of each first interface includes:

[0110] For each process, using the first approximate host unit of the first virtual point of each first interface as the starting unit, perform a host unit search in the second sub-grid area of the process where the first approximate host unit is located according to the centroid coordinate method to obtain a first search result;

[0111] Judge whether there is a real first host unit according to the first search result;

[0112] If there is no real first host unit, then determine the first approximate host unit as the first host unit of the first interface.

[0113] In an embodiment of the present invention, for each process, determining a second host unit of each second interface according to a second approximate host unit of a second virtual point of each second interface includes:

[0114] For each process, using the second approximate host unit of the second virtual point of each second interface as a starting unit, searching for a host unit within a first sub-grid region of the process where the second approximate host unit is located according to the barycentric coordinate method to obtain a second search result;

[0115] Judging whether there is a true second host unit according to the second search result;

[0116] If there is no true second host unit, determining the second approximate host unit as the second host unit of the second interface.

[0117] It should be understood that within each process, according to the virtual point coordinates and the corresponding approximate host unit, using the approximate host unit as a starting unit, searching for a host unit within the sub-grid of the process where the approximate host unit is located according to the barycentric coordinate method. If no true host unit is found when searching to the grid boundary and the process boundary, the approximate host unit is used as the host unit.

[0118] In an embodiment of the present invention, the principle of the barycentric coordinate method is as Figure 8 shown: For a face composed of n 1 , n 2 , n 3 , the vector from n 1 to n 2 is n 1 to n 3 the vector is n 1 to the vector from the virtual point virtual-point is n 1 to the vector from the center of the unit body center is Calculate the position judgment quantity temp of the point and the polyhedron:

[0119]

[0120] If for all faces of a certain grid, the position judgment quantity temp of the point and the polyhedron is greater than or equal to 0, it means that the virtual point is inside the unit, and then this unit is the host unit of the interface virtual point; if there is 1 or more faces where the position judgment quantity temp of the point and the polyhedron is less than 0, it indicates that the virtual point is outside these faces. Randomly select one of these faces as the next unit to be searched until a grid unit containing the virtual point is searched. Performing such cyclic operations forms a search path for the host unit.

[0121] It should be understood that in the embodiments of the present invention, by using the nearest distance as the determination scheme, the probability that the approximate host unit and the true host unit (the virtual point is within the unit) are in the same process (the same sub-grid) is increased. After finding the approximate host unit, within the sub-grid where the approximate host unit is located, according to the virtual point coordinates, with the approximate host unit as the starting search unit, the barycentric coordinate method is used to find the true host unit within the current sub-grid. When the true host unit is found or the search path reaches the grid boundary or the process boundary, the search stops. If the true host unit is still not found after the search stops, the approximate host unit is used as the host unit. At this time, the finally determined host unit is either the approximate host unit or the true host unit.

[0122] In the embodiments of the present invention, the numerical transfer of the flow field between the first grid region and the second grid region is realized according to the field variables in the first host unit of all processes and the field variables in the second host unit of all processes, including:

[0123] Assigning values to the first virtual points according to the field variables of the first host unit of each process, and assigning values to the second virtual points according to the field variables of the second host unit of each process;

[0124] Sending the assignments of the first virtual points and the assignments of the second virtual points to the processes where the corresponding interfaces are located, to realize the numerical transfer of the flow field between the first grid region and the second grid region.

[0125] It should be understood that the field variables in the host unit are used to assign values to the virtual points, and the values on the virtual points are sent to the processes where the interfaces are located to realize the numerical transfer of the flow field between G 1 and G 2 .

[0126] In summary, for the grid interface host unit parallel search method provided by the present invention, first, the grid unit closest to the virtual point is found in the interface grid of the paired grids as the approximate host unit, which can ensure that when the real host unit is not finally found and the approximate host unit is used as the host unit, the value at the virtual point will not deviate too much. Second, the nearest distance is used as the determination scheme, which increases the probability that the approximate host unit and the real host unit are within the same process (the same sub-grid). After finding the approximate host unit, within each process, according to the virtual point coordinates, starting from the approximate host unit as the search unit, the barycentric coordinate method is used to find the real host unit within the sub-grid where the approximate host unit is located. When the real host unit is found, or the search path reaches the grid boundary or the process boundary, the search stops. If the real host unit is still not found after the search stops, the approximate host unit is used as the host unit. At this time, the finally determined host unit is either the approximate host unit or the real host unit. Therefore, the grid interface host unit parallel search method of the present invention realizes cross-process search in parallel mode.

[0127] The following Figures 9a to 13b illustrates the specific implementation process of the grid interface host unit parallel search method according to the embodiments of the present invention.

[0128] As Figure 9a and Figure 9b shown, for the numerical calculation of the infinitely long roof-driven cavity flow, the grid is a three-dimensional solid network, there is only one layer of grid in the z direction and the z-direction surface is set as a symmetry plane (Symmetry) to represent the infinitely long cavity. The overall model is a square nesting, and the inner and outer square regions are connected through an interface (Interface). The type of the top surface is a velocity inlet (Velocity-In), and the types of the surrounding surfaces are walls (Wall).

[0129] In this embodiment, 10 processes are used for parallel partitioning. As Figure 10a and Figure 10b shown, for the partitioned model and grid, when performing parallel partitioning, both the inner and outer square calculation domains are divided into 10 blocks, and each process has one block of the inner and outer domains. It can be seen from Figure 10a that the interface Interface between the inner and outer square regions, as well as the partition boundaries Process-Boundary after the parallel partitioning of the inner and outer square regions.

[0130] According to the method for calculating the virtual point coordinates mentioned above, the virtual point coordinates of the interfaces of each grid unit within each sub-grid region after partitioning are calculated, and the approximate host unit for each virtual point is found. After finding the approximate host unit, starting from the approximate host unit as the search unit, the barycentric coordinate method is used to find the real host unit.

[0131] As shown Figure 11 in the figure, it is the distribution of virtual point coordinates (red dots), approximate host unit centroid coordinates (blank squares), and host unit centroid coordinates (blue triangles). Here, it can be understood that Figure 11 if the virtual point in

[0132] is in the same grid cell as the host unit, it means finding the true host unit on the surface, that is, when the red dot and the blue triangle are in the same grid cell, it means finding the true host unit; for the case with a blank square, it represents the position of the approximate host unit. Figure 11 In Figure 12 , there is a situation where the initial host unit (i.e., the approximate host unit) of Grid1 is different from the final host unit (i.e., the true host unit), which indicates that it is necessary to use the centroid coordinate method to search for the host unit after obtaining the approximate host unit in the embodiment of the present invention. In the green dotted circle in Figure 13a and Figure 13b , it can be seen that the search path of the virtual point is blocked by the process boundary. At this time, the approximate host unit is used as the final host unit. Since the approximate host unit is adjacent to the unit where the virtual point is located, it will not affect the calculation results. Figure 13a and Figure 13b are the comparison of the calculation results ([[]] Figure 13a and Figure 13b ) under the 10 processes of Embodiment 10 of the present invention and the calculation results ([[]]

[0133] ) under the serial condition without interface when the Reynolds number (Re) is 1000. Figures 14 to 16b The following combines

[0134] to illustrate the feasibility of the grid interface host unit parallel search method of the embodiment of the present invention. Figure 14 and Figure 15 In this embodiment, the open water calculation of the propeller, that is, the propeller rotates alone in uniform water flow. The model and grid are as shown in 3 , the outer domain grid is water area, the inner domain grid contains the propeller, the outer domain grid does not move, the inner domain grid rotates, and the inner and outer domain grids are connected by an interface. The relevant parameter settings are as follows: the water incoming flow velocity is 1.8 m / s, the propeller rotation speed n = 18 rad / s, the density of water ρ = 996.76 kg / m 2 .

[0135] In this embodiment, the calculation uses 8 processes, and the grid interface host unit parallel search method of the embodiment of the present invention is used to process the interface, and the calculation results are compared with the unpartitioned processing method in the prior art. Figure 16a and Figure 16bIt is a comparison schematic diagram of the pressure cloud map on the propeller surface and the processing method without zoning in the prior art. Table 1 and Table 2 are respectively the thrust coefficient k T and the torque coefficient K Q of the propeller and the comparison of the model test results with the processing method without zoning in the prior art.

[0136] Table 1 Comparison of Thrust Coefficient and Model Test Results in the Prior Art

[0137] Calculation result Model test Error 0.214 0.210 1.9％

[0138] Table 2 Comparison of Torque Coefficient and Model Test Results in the Prior Art

[0139] Calculation result Model test Error 0.026 0.02599 0.038％

[0140] According to the comparison of this embodiment, the grid interface host unit parallel search method of the present invention realizes the search in the grid interface host unit parallel mode, and the results are approximate compared with the prior art by means of model tests, that is, it proves that the grid interface host unit parallel search method of the present invention not only has feasibility, but also can obtain accurate search results.

[0141] As another embodiment of the present invention, a grid interface host unit parallel search device 10 is provided for implementing the grid interface host unit parallel search method described above, wherein, as Figure 17 shown, it includes:

[0142] A zoning module 100 for respectively zoning any two paired first grid regions and second grid regions to obtain first sub-grid regions and second sub-grid regions with the same number as the number of processes. Among them, the first grid region includes a first interface set, the second grid region includes a second interface set, the first grid region and the second grid region are connected through the first interface set and the second interface set. Each process includes a first sub-grid region and a second sub-grid region. The first sub-grid region of at least one process includes a first interface subset, and the second sub-grid region of at least one process includes a second interface subset. The first interface subsets of all processes form the first interface set, and the second interface subsets of all processes form the second interface set;

[0143] A virtual point coordinate determination module 200 for respectively determining the first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset of each process;

[0144] A search module 300, configured to, for each process, search for a first approximate host cell of a first virtual point in second grid cells corresponding to a second interface subset of all processes according to first virtual point coordinates of each first interface, and search for a second approximate host cell of a second virtual point in first grid cells corresponding to a first interface subset of all processes according to second virtual point coordinates of each second interface;

[0145] A host cell determination module 400, configured to, for each process, determine a first host cell of each first interface according to the first approximate host cell of the first virtual point of each first interface, and determine a second host cell of each second interface according to the second approximate host cell of the second virtual point of each second interface;

[0146] A flow field numerical transfer module 500, configured to implement flow field numerical transfer between the first grid region and the second grid region according to field variables in the first host cells of all processes and field variables in the second host cells of all processes.

[0147] The grid interface host cell parallel search device provided by the present invention determines virtual points in the interface set of the paired grid regions to determine approximate host cells, and determines host cells according to the approximate host cells, so as to implement flow field numerical transfer of the paired grid regions of multiple processes, thereby realizing parallel search of the grid interface host cells. Therefore, the grid interface host cell parallel search device provided by the present invention can implement search in the parallel mode of the grid interface host cells. In addition, when the grid interface host cell parallel search device of the present invention searches for host cells in the interface grid of the paired grids, it can ensure that when the true host cell is not finally found and an approximate host cell is used as the host cell, the numerical value on the virtual point will not deviate too much.

[0148] For the specific working process of the grid interface host cell parallel search device provided by the present invention, reference may be made to the description of the grid interface host cell parallel search method above, which will not be elaborated here.

[0149] As another embodiment of the present invention, a storage medium is provided, which is used to store computer instructions, and when the computer instructions are loaded and executed by a processor, the grid interface host cell parallel search method described above is implemented.

[0150] In an embodiment of the present invention, a non-transitory computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions, and these computer-executable instructions can execute the grid interface host unit parallel search method in any of the above method embodiments. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (abbreviation: HDD), or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0151] As another embodiment of the present invention, an electronic device is provided. Among them, it includes a memory and a processor. The memory and the processor are communicatively connected. The memory is used to store computer instructions, and the processor is used to load and execute the computer instructions to implement the grid interface host unit parallel search method described above.

[0152] As Figure 18 shown, the electronic device 80 may include: at least one processor 81, such as a CPU (Central Processing Unit, central processor), at least one communication interface 83, a memory 84, and at least one communication bus 82. Among them, the communication bus 82 is used to realize the connection and communication between these components. Among them, the communication interface 83 may include a display screen (Display), a keyboard (Keyboard). Optionally, the communication interface 83 may further include a standard wired interface and a wireless interface. The memory 84 may be a high-speed RAM memory (Random Access Memory, volatile random access memory), or a non-volatile memory, such as at least one disk memory. Optionally, the memory 84 may further be at least one storage device located far from the aforementioned processor 81. Among them, an application program is stored in the memory 84, and the processor 81 calls the program code stored in the memory 84 to execute any of the above method steps.

[0153] Among them, the communication bus 82 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 82 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 18It is represented by only one thick line, but it does not mean that there is only one bus or one type of bus.

[0154] Among them, the memory 84 may include volatile memory, such as random-access memory (RAM); the memory may also include non-volatile memory, such as flash memory, hard disk drive (HDD) or solid-state drive (SSD); the memory 84 may further include a combination of the above types of memories.

[0155] Among them, the processor 81 may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP.

[0156] Among them, the processor 81 may further include a hardware chip. The above hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The above PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0157] Optionally, the memory 84 is further used to store program instructions. The processor 81 may call the program instructions to implement the grid interface host unit parallel search method as shown in the embodiments of the present invention. Figure 3 Embodiments shown in the grid interface host unit parallel search method.

[0158] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present invention, but the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also regarded as the protection scope of the present invention.

Claims

1. A parallel search method for grid interface host units, characterized in that: include: Partition any two paired first grid areas and second grid areas to obtain first sub-grid areas and second sub-grid areas that are the same number as the number of processes, wherein the first grid area includes a first interface set, the second grid area includes a second interface set, the first grid area and the second grid area are connected with the second interface set through the first interface set, each process includes a first sub-grid area and a second sub-grid area, the first sub-grid area of ​​at least one process includes a first interface subset, the second sub-grid area of ​​at least one process includes a second interface subset, the first interface subsets of all processes form the first interface set, and the second interface subsets of all processes form the second interface set; Determine the first virtual point coordinates of each first interface in the first interface subset and the second virtual point coordinates of each second interface in the second interface subset of each process respectively; For each process, searching for a first approximate host unit of the first virtual point in the second grid cells corresponding to the second interface subsets of all processes according to the first virtual point coordinates of each first interface, and searching for a second approximate host unit of the second virtual point in the first grid cells corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface; For each process, determining a first host unit of each first interface according to a first approximate host unit of a first virtual point of the first interface, and determining a second host unit of each second interface according to a second approximate host unit of a second virtual point of the second interface; Implementing flow field value transfer between the first grid area and the second grid area according to the field variables in the first host unit of all processes and the field variables in the second host unit of all processes; The method of implementing the flow field value transfer between the first grid area and the second grid area according to the field variables in the first host unit of all processes and the field variables in the second host unit of all processes includes: Assigning a value to a first virtual point according to the field variable of the first host unit of each process, and assigning a value to a second virtual point according to the field variable of the second host unit of each process; The assignment result of the first virtual point and the assignment result of the second virtual point are both sent to the process where the corresponding interface is located, so as to realize the flow field value transfer between the first grid area and the second grid area.

2. The grid interface host unit parallel search method according to claim 1, characterized in that: The first sub-grid area of ​​each process includes a plurality of first grid units, and the second sub-grid area of ​​each process includes a plurality of second grid units. The first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset are determined respectively, including: For each process, calculating first body-centered coordinates and first face-centered coordinates of all first grid cells in the first sub-grid area, and calculating second body-centered coordinates and second face-centered coordinates of all second grid cells in the second sub-grid area; The first virtual point coordinates of the first interface are calculated based on the first face-center coordinates of each first interface in the first interface subset of the process and the first body-center coordinates of the first grid unit where the first interface is located, and the second virtual point coordinates of the second interface are calculated based on the second face-center coordinates of each second interface in the second interface subset of the process and the second body-center coordinates of the second grid unit where the second interface is located.

3. The grid interface host unit parallel search method according to claim 2, characterized in that: The calculation formulas for the first virtual point coordinates and the second virtual point coordinates are both: in, represents the face-centered coordinates of the first interface or the second interface, represents the body center coordinates of the first grid cell where the first interface is located or the second grid cell where the second interface is located, represents the first virtual point coordinates or the second virtual point coordinates, represents the distance from the center of the first interface to the center of the first grid cell, or the distance from the center of the second interface to the center of the second grid cell. A unit vector representing the outer normal direction of the first interface or the outer normal direction of the second interface.

4. The grid interface host unit parallel search method according to claim 1, characterized in that: For each process, searching for a first approximate host unit of the first virtual point in the second grid units corresponding to the second interface subsets of all processes according to the first virtual point coordinates of each first interface includes: The second body-center coordinates of the second grid cells where each second interface in the second interface subsets of all processes is located are collected to form a second body-center coordinate set; For each process, searching for the second body-center coordinates closest to the first virtual point in the second body-center coordinate set according to the first virtual point coordinates of each first interface; The target second grid unit where the second body center coordinate closest to the first virtual point is located is determined as the first approximate host unit, the process where the target second grid unit is located is determined as the process where the first approximate host unit is located, and the grid number of the target second grid unit is determined as the grid number of the first approximate host unit.

5. The grid interface host unit parallel search method according to claim 1, characterized in that: For each process, searching for a second approximate host unit of the second virtual point in the first grid units corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface includes: The first body-center coordinates of the first grid cells where each first interface in the first interface subsets of all processes is located are collected to form a first body-center coordinate set; For each process, searching the first body center coordinates closest to the second virtual point in the first body center coordinate set according to the second virtual point coordinates of each second interface; The target first grid unit where the first body center coordinate closest to the second virtual point is located is determined as the second approximate host unit, the process where the target first grid unit is located is determined as the process where the second approximate host unit is located, and the grid number of the target first grid unit is determined as the grid number of the second approximate host unit.

6. The grid interface host unit parallel search method according to claim 1, characterized in that: For each process, determining the first host unit of each first interface according to the first approximate host unit of the first virtual point of each first interface includes: For each process, taking the first approximate host unit of the first virtual point of each first interface as the starting unit, searching for host units in the second subgrid area of ​​the process where the first approximate host unit is located according to the barycentric coordinate method, and obtaining a first search result; Determining whether there is a real first host unit according to the first search result; If there is no real first host unit, the first approximate host unit is determined as the first host unit of the first interface.

7. The grid interface host unit parallel search method according to claim 1, characterized in that: For each process, determining the second host unit of each second interface according to the second approximate host unit of the second virtual point of the second interface includes: For each process, taking the second approximate host unit of the second virtual point of each second interface as the starting unit, searching for host units in the first subgrid area of ​​the process where the second approximate host unit is located according to the barycentric coordinate method, and obtaining a second search result; Determining whether there is a real second host unit according to the second search result; If there is no real second host unit, the second approximate host unit is determined as the second host unit of the second interface.

8. A parallel search device for host units at a grid interface, used to implement the parallel search method for host units at a grid interface as claimed in any one of claims 1 to 7, characterized in that: include: A partitioning module, used for partitioning any two paired first grid areas and second grid areas to obtain first sub-grid areas and second sub-grid areas that are the same number as the number of processes, wherein the first grid area includes a first interface set, the second grid area includes a second interface set, the first grid area and the second grid area are connected with the second interface set through the first interface set, each process includes a first sub-grid area and a second sub-grid area, the first sub-grid area of ​​at least one process includes a first interface subset, the second sub-grid area of ​​at least one process includes a second interface subset, the first interface subsets of all processes form the first interface set, and the second interface subsets of all processes form the second interface set; A virtual point coordinate determination module, used to respectively determine the first virtual point coordinates of each first interface in the first interface subset of each process and the second virtual point coordinates of each second interface in the second interface subset; A search module is used to search, for each process, a first approximate host unit of the first virtual point in the second grid units corresponding to the second interface subsets of all processes according to the first virtual point coordinates of each first interface, and to search, for each process, a second approximate host unit of the second virtual point in the first grid units corresponding to the first interface subsets of all processes according to the second virtual point coordinates of each second interface; a host unit determination module, configured to determine, for each process, a first host unit of each first interface according to a first approximate host unit of a first virtual point of the first interface, and to determine a second host unit of each second interface according to a second approximate host unit of a second virtual point of the second interface; The flow field value transfer module is used to realize the flow field value transfer between the first grid area and the second grid area according to the field variables in the first host unit of all processes and the field variables in the second host unit of all processes.

9. A storage medium, characterized in that: Used to store computer instructions, which, when loaded and executed by a processor, implement the grid interface host unit parallel search method described in any one of claims 1 to 7.

Images