CN115455794A - LBM Parallel Optimization Method, Device and Storage Medium Based on Connected Pores to Divide Calculation Area - Google Patents

Abstract

The invention provides a LBM parallel optimization method and device based on a connected pore division calculation area, which can divide the sub-domain number in a balanced manner according to the calculation node number, so that each calculation node can perform load balancing and efficient processing, and the calculation processing efficiency is improved. The method comprises the following steps: step 1, determining the total number N of subdomains to be decomposed according to the number N of calculation nodes in the system; step 2, decomposing the calculation domain: dividing the basin into n along the x-axis _x Regions having the same or similar number of cells, and dividing each region into n along the y-axis _y Sub-regions with the same or similar number of cells, and finally dividing n _x ×n _y Sub-region dividing n along z-axis _z Then, obtaining N groups ofSubdomains with the same or similar unit numbers, wherein each subdomain is a three-dimensional region consisting of a plurality of pore units; and the unit number difference between the sub-domain with the maximum unit number and the sub-domain with the minimum unit number after decomposition should not exceed one thousandth of the total unit number; and 3, distributing a calculation task.

Description

LBM parallel optimization method, device and storage based on connected pores to divide computing area medium

technical field

The invention belongs to the technical field of pore structure simulation calculation, and in particular relates to an LBM parallel optimization method, device and storage medium for dividing calculation areas based on connected pores.

Background technique

Lattice Boltzmann Method (Lattice Boltzmann Method, referred to as LBM) is an effective means to simulate fluid flow in the pore structure of porous media at the mesoscopic scale. In order to make the simulation results representative, the size of the simulated sample should be large enough, and due to the complexity of the fluid flow in the pore structure, the pore-scale LBM often faces a large number of computing resources and storage space requirements during numerical simulation. Therefore, large-scale LBM simulation needs to be optimized in parallel, and its execution time and memory requirements depend on the amount of data and the parallel optimization method.

The current basic idea of domain decomposition for parallel computing is to decompose the computing domain into several sub-domains, and then restore the interface and perform load balancing. But for complex porous materials, this step often takes a lot of time and iterative steps to achieve an acceptable load balance. Among them, the recursive bisection method (Recursive Bisection Method) is one of the most commonly used computational domain decomposition schemes, which divides the computational domain into two sub-domains evenly, and then performs the same division on the sub-domains. Finally, after R times of recursive total Divide 2 ^R sub-domains. This partition scheme can well balance the workload of regular or irregular structures in homogeneous media, but there are three problems: 1) It is still difficult to achieve good workload balance for non-uniform and highly heterogeneous porous structures; 2 ) The number of sub-domains divided by this method must be 2 ^R , and the number of processors (computing nodes) is generally not exactly equal to it, resulting in some processors being unable to be effectively used; 3) When there are many sub-domains divided, repeated recursion is required The number of divisions R will be large and the efficiency will be low.

Contents of the invention

The present invention is carried out to solve the above problems, and the purpose is to provide an LBM parallel optimization method, device and storage medium based on connected pores to divide the computing area, which can divide the number of sub-domains in a balanced manner according to the number of computing nodes, so that each computing node can Efficient processing of load balancing to improve computing processing speed.

In order to achieve the above object, the present invention adopts the following scheme:

<method>

The invention provides an LBM parallel optimization method based on connected pores to divide the calculation area, which is characterized in that it comprises the following steps:

Step 1. Determine the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determine the total number N of sub-domains that should be decomposed according to the number N of computing nodes in the system;

Step 2. Decompose the computational domain:

Divide the watershed into n _x areas with the same or similar number of units along the x-axis, then divide each area into n _y sub-areas with the same or similar number of units along the y-axis, and finally divide this n _x ×n _y The sub-regions are divided n _z times along the z-axis to obtain n _x ×n _y ×n _z =N sub-regions with the same or similar number of units, each sub-region is a three-dimensional region composed of multiple pore units; and, after decomposition The difference in the number of units between the subdomain with the maximum number of units M _max and the subdomain with the minimum number of units M _min should not exceed one thousandth of the total number of units; the number of units between subdomains can be further reduced according to the accuracy requirements of the calculation domain division The allowable difference value;

Step 3. Assign computing tasks:

When computing in parallel, assign N subdomains to N computing nodes one by one; when computing subdomains, computing nodes only consider all connected pore units, and renumber these connected pore units, and then They are respectively stored in the one-dimensional array p _i , where i is the number of the subfield; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the pores that store the ordinal number of the pore unit and the corresponding coordinates are obtained An array of cells, each pore cell is tracked by its unique ordinal number and coordinates;

For each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derivative arrays are obtained; the momentum and local fluid density data of the pore unit Corresponding to the ordinal number of the pore unit, it is stored as a one-dimensional array to obtain a series of second derived arrays.

Specifically, the LBM parallel optimization method based on connected pores to divide the calculation area provided by the present invention may also have the following characteristics: in step 2, the maximum number of units M _max :

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively.

Preferably, the LBM parallel optimization method based on connected pores to divide the calculation area provided by the present invention can also have the following characteristics: in step 3, the main data structure of each pore unit is its fluid particle distribution function and equilibrium fluid particle distribution function ; The momentum and local fluid density data before and after the calculation of the pore unit and the ordinal number of the pore unit are correspondingly stored as a one-dimensional array, and four second derived arrays are obtained.

Preferably, the LBM parallel optimization method based on connected pores to divide the calculation area provided by the present invention may also include: Step 4. Setting the communication mode between sub-domains: when dealing with adjacent or diagonal pore units between different sub-domains , extend the interface layer on the interface of the current subdomain with an additional layer of units to cover adjacent and diagonal pore units, so that the communication mode of these expanded pore units is consistent with that of the interface units in the current subdomain.

<device>

Further, the present invention also provides a device for automatically implementing the above <method>, characterized in that it includes:

The determining part determines the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determines the total number of sub-domains N to be decomposed according to the number N of computing nodes in the system;

Calculate the domain decomposition part, divide the watershed into n _x regions with the same or similar number of units along the x-axis, then divide each region into n _y sub-regions with the same or similar number of units along the y-axis, and finally divide this n _x ×n _y sub-regions are divided n _z times along the z-axis to obtain n _x ×n _y ×n _z =N sub-domains with the same or similar number of units; the decomposed sub-domains should meet the condition: have the largest number of units The difference in the number of units between the subdomain of M _max and the subdomain with the minimum number of units M _min shall not exceed one thousandth of the total number of units;

The calculation task allocation department assigns N sub-domains to N computing nodes one by one for parallel computing; when the computing nodes process sub-domains, they only consider all connected pore units, and recalculate these connected pore units. number, and then store them in the one-dimensional array p _i respectively, where i is the subfield number; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the ordinal number of the stored pore unit is obtained and the pore unit array of the corresponding coordinates; for each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derivative arrays are obtained; the pore unit The momentum and local fluid density data of the unit are stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of second derived arrays are obtained;

The control part communicates with the determination part, the calculation domain decomposition part and the calculation task distribution part to control their operation.

Preferably, the LBM parallel optimization device for dividing the calculation area based on connected pores provided by the present invention may also have such a feature: in the calculation domain decomposition part, the maximum number of units M _max is:

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively.

Preferably, the LBM parallel optimization device based on connected pores to divide the calculation area provided by the present invention may also include: a communication mode setting part, connected to the control part in communication, and processing adjacent or diagonal pore units between different subdomains When , the interface layer on the interface of the current subdomain is extended with an additional layer of units to cover the adjacent and diagonal pore units, so that the communication mode of these expanded pore units is consistent with that of the interface units in the current subdomain.

Preferably, the LBM parallel optimization device for dividing the calculation area based on connected pores provided by the present invention may further include: an input display unit, together with a determination unit, a calculation domain decomposition unit, a calculation task allocation unit, a communication mode setting unit and a control unit It is connected by communication, and is used for allowing the user to input operation instructions and display them accordingly.

Preferably, the LBM parallel optimization device based on connected pores to divide the calculation area provided by the present invention may also have such a feature: the input display unit can determine the sample pore unit information and connectivity and the number of calculation nodes determined by the determination unit according to the operation instruction Or the total number N of sub-domains is displayed, the decomposition of the computing domain decomposition part is displayed, and the distribution of the computing task distribution part is displayed accordingly.

<storage medium>

In addition, the present invention also provides a computer-readable storage medium storing a program for realizing the above <method>.

Function and Effect of Invention

The LBM parallel optimization method, device and storage medium based on connected pores to divide the computing area provided by the present invention can randomly and evenly divide the total number of sub-domains according to the number of computing nodes. The total number of sub-domains is not limited to 2 ^R , and the present invention can effectively reduce The communication complexity caused by the tortuousness of the interface balances the workload of all sub-domains in the process of water domain decomposition, without the need for secondary optimization such as iteration, reduces system memory consumption, and simplifies the calculation task allocation process, and the present invention The communication between the sub-domains is efficient and easy to implement, and the communication time is reduced; therefore, the present invention can significantly improve the calculation efficiency, and is especially suitable for the processing of pore-scale simulation of porous materials with a large amount of calculation and large memory consumption, and opens up new possibilities. The calculation area decomposition and processing ideas of this paper have great promotion value.

Description of drawings

Fig. 1 is a schematic diagram of the decomposition process of the computational domain in an embodiment of the present invention, wherein (a) is decomposed along the x-axis, (b) is decomposed along the y-axis, and (c) is decomposed along the z-axis;

Fig. 2 is a schematic diagram of the process of marking the fluid units in the porous medium with a one-dimensional array after decomposing and obtaining the sub-domains in the embodiment of the present invention;

3 is a schematic diagram of the interface between the 0 ^# subdomain and the 1 ^# subdomain in the embodiment of the present invention (white is a pore unit, and gray is a solid unit);

Fig. 4 is a D3Q19 discrete speed model diagram related to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an information exchange process of diagonal units between different sub-domains according to an embodiment of the present invention.

detailed description

The LBM parallel optimization method, device and storage medium based on connected pore division calculation area involved in the present invention will be described in detail below with reference to the accompanying drawings.

<Example>

The LBM parallel optimization method based on connected pores to divide the calculation area provided in this embodiment includes the following steps:

Step 1. Determine the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determine the total number of sub-domains N to be decomposed according to the number of calculation nodes N in the system; Materials and pore distribution data can be directly defined by programming language, and X-CT technology can be used to obtain pore distribution data for non-uniform and highly heterogeneous porous structures.

Step 2. Decompose the computational domain:

Divide the watershed of the porous media sample along the x-axis into n _x regions with the same or similar number of units, then divide each region into n _y sub-regions with the same or similar number of units along the y-axis, and finally divide the n The _x ×n _y sub-regions are divided n _z times along the z-axis to obtain n _x ×n _y ×n _z =N sub-regions with the same or similar number of units; and, after decomposition, the sub-domain with the largest number of units M _max and The difference in the number of units between subdomains with the minimum number of units M _min should not exceed one thousandth of the total number of units.

Maximum number of units M _max :

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively.

Specifically, as shown in Fig. 1, firstly, scan continuously along the z, y and x directions, traversing each unit in the calculation area. It should be noted that the scanning order of z→y→x is to traverse in the order of z first and then y on the x _i section. Figure 1(b) is based on Figure 1(a), each sub-domain is divided into two small areas along the y-axis, specifically, the number of pore units in the sub-domain is traversed in the order of x→z→y And store it into a one-dimensional array, and then divide the one-dimensional array associated with each sub-domain into two parts, and at this time the watershed is divided into 3×2×1 sub-regions. Figure 1(c) On the basis of the previous division, use the same method to traverse the pore units of each subdomain in the scanning order of x→y→z and repeat the division process. Finally, the simulation area is divided into 3×2×2 subdomains . Next, as shown in Figure 2, the connected pore units are screened and numbered and stored in a one-dimensional array, and there are N pore units in the watershed. As shown in Fig. 3, all connected pore units are divided into multiple groups of equal size, each group is associated with a subdomain, and the total number of groups is equal to the number of subdomains after decomposition. Then, divide the one-dimensional array into three groups, each group contains N/3 or N/3+1 pore units, and the last unit numbers of the three groups are N ₀ , N ₁ and N respectively. Finally, the interface between two adjacent subdomains is recovered, as shown in Fig. 3 and Fig. 1(a), the interface contains the last pore unit of each partition.

The computational domain decomposition scheme directly divides the soil sample into n _x × _ny ×n _z sub-domains, and the difference in workload between sub-domains is the difference in the number of pore units on the adjacent interface between each sub-domain. Subsequent simulations are all based on the divided one-dimensional array, and the simulation area is from 1 to _nxi , where _nxi is the size of the subdomain in the x direction (fluid flow direction) associated with processor i.

Step 3. Assign computing tasks:

When computing in parallel, assign N subdomains to N computing nodes one by one; when computing subdomains, computing nodes only consider all connected pore units, and renumber these connected pore units, and then They are respectively stored in the one-dimensional array p _i , where i is the number of the subfield; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the pores that store the ordinal number of the pore unit and the corresponding coordinates are obtained An array of cells, each pore cell is tracked by its unique ordinal number and coordinates.

For each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derivative arrays are obtained; the momentum and local fluid density data of the pore unit Corresponding to the ordinal number of the pore unit, it is stored as a one-dimensional array to obtain a series of second derived arrays.

The main data structure of each pore unit is its fluid particle distribution function and equilibrium fluid particle distribution function; the momentum and local fluid density data before and after the calculation of the pore unit are stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and four The second derived array.

As shown in Fig. 4, according to the D3Q19 model in LBM, for each connected pore unit x, there are at most 18 adjacent pore units stored in 18 one-dimensional arrays. Each array corresponds to a component. The main data structure of each pore unit is its 19 fluid particle distribution functions and 18 equilibrium fluid particle distribution functions, each function is described by a one-dimensional array, and the other four one-dimensional arrays are used to store the Momentum and local fluid density.

Step 4. Set the communication between sub-domains (extraction and transfer of calculation data): as shown in Figure 5, when dealing with adjacent or diagonal pore units (d and d' in the figure) between different sub-domains , extend the interface layer on the interface of the current subdomain with an additional layer of units to cover adjacent and diagonal pore units, so that the communication mode of these expanded pore units is consistent with that of the interface units in the current subdomain.

In each subfield of this implementation, the overlapping regions are stored in 18 one-dimensional arrays, and no redundant or repeated calculations are performed. The particle distribution function and equilibrium state distribution function of fluid cells from 1 to _nxi in each subdomain are stored in two different sets of one-dimensional arrays (18 one-dimensional arrays in each set). The pointer is chosen to locate the first porosity unit of the interface, and only the porosity unit data in the interface layer is exchanged to the adjacent partition. Since only all pore units including the interface layer in each subdomain are stored, and the adjacent units of each pore unit are stored in 18 one-dimensional arrays, the adjacent units of each pore unit can be Easy OK. In each iteration, the fluid particle distribution function of the corresponding sub-domain is calculated on each processor, and then the data of the sub-domain interface is exchanged between the processors. In data exchange, only 5 of the 18 distribution function components of the fluid unit on the interface layer need to be transferred to the adjacent processors. For example, in the x direction, only the fluid particle distribution functions f ₃ [i], f ₈ [i], f ₉ [i], f ₁₂ [i] and f ₁₃ [i] on the interface nx _i need to be moved from the right The subfield of the left subfield is passed to the left subfield, and only f ₁ [i], f ₇ [i], f ₁₀ [i], f ₁₁ [i], and f ₁₄ [i] need to be passed from the left subfield to right subdomain.

Through the above method, the present invention can arbitrarily and evenly divide the total number of sub-domains according to the number of computing nodes, make full use of all computing nodes for computing and processing, and can effectively reduce the communication complexity caused by the tortuousness of the dividing interface. Balance the workload of all sub-domains without secondary optimization such as iteration, reduce system memory consumption, simplify the calculation task allocation process, and effectively reduce the communication time between sub-domains, so that the present invention can significantly improve computing efficiency , the larger the amount of data, the more obvious the advantages of the method of the present invention, therefore, it is especially suitable for the processing of pore-scale simulation of porous materials with a large amount of calculation and large memory consumption.

Furthermore, this embodiment also provides a device capable of automatically implementing the above method, and the device includes a determination unit, a calculation domain decomposition unit, a calculation task allocation unit, a communication mode setting unit, an input display unit, and a control unit.

The determining part determines the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determines the total number N of subdomains to be decomposed according to the number N of computing nodes in the system.

The computational domain decomposition part divides the porous medium sample flow domain into n _x regions with the same or similar number of units along the x-axis, and then divides each region into n _y sub-regions with the same or similar number of units along the y-axis, Finally, divide the n _x ×n _y sub-regions along the z-axis n _z times to obtain n _x ×n _y ×n _z =N sub-regions with the same or similar number of units; the decomposed sub-regions should meet the conditions: have The difference in the number of units between the subdomain with the maximum number of units M _max and the subdomain with the minimum number of units M _min should not exceed one thousandth of the total number of units.

Maximum number of units M _max :

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively.

The computing task allocation department assigns N sub-domains to N computing nodes one by one for parallel computing; when the computing nodes process the sub-domains, only all connected pore units are considered, and these connected pore units are renumbered , and then store them in the one-dimensional array p _i respectively, i is the subfield number; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the ordinal number and The pore unit array of the corresponding coordinates; for each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derivative arrays are obtained; the pore unit The momentum and local fluid density data are stored as one-dimensional arrays corresponding to the ordinal numbers of the pore cells, resulting in a series of second derived arrays.

The communication mode setting part is connected with the control part by communication. When dealing with adjacent or diagonal pore units between different subdomains, the interface layer on the interface of the current subdomain is extended by an additional layer of units to cover adjacent and diagonal pore units. Aperture units, so that the communication mode of these expanded aperture units is consistent with that of the interface units in the current sub-domain.

The input display part is used for allowing the user to input operation instructions and display them accordingly. For example, the input display unit can display the sample void unit information and connectivity determined by the determination unit and the number of computing nodes or the total number of sub-domains N according to the operation instructions, display the decomposition of the calculation domain decomposition unit, and display the calculation task allocation unit The allocations are displayed accordingly.

The control unit communicates with the determination unit, the calculation domain decomposition unit, the calculation task distribution unit and the communication mode setting unit to control their operation.

The above embodiments are merely illustrations for the technical solution of the present invention. The LBM parallel optimization method, device and storage medium based on connected pores to divide the calculation area involved in the present invention are not limited to the content described in the above embodiments, but are subject to the scope defined in the claims. Any modifications, supplements or equivalent replacements made by those skilled in the art of the present invention on the basis of the embodiments are within the protection scope of the claims of the present invention.

Claims (10)

1. The LBM parallel optimization method based on connected pore division calculation area, is characterized in that, comprises the following steps: Step 1. Determine the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determine the total number N of sub-domains that should be decomposed according to the number N of computing nodes in the system; Step 2. Decompose the computational domain: Divide the watershed of the porous media sample along the x-axis into n _x regions with the same or similar number of units, then divide each region into n _y sub-regions with the same or similar number of units along the y-axis, and finally divide the n The _x ×n _y sub-regions are divided n _z times along the z-axis to obtain n _x ×n _y ×n _z =N sub-regions with the same or similar number of units; and, after decomposition, the sub-domain with the largest number of units M _max and The difference in the number of units between subdomains with the minimum number of units M _min should not exceed one thousandth of the total number of units; Step 3. Assign computing tasks: When computing in parallel, assign N subdomains to N computing nodes one by one; when computing subdomains, computing nodes only consider all connected pore units, and renumber these connected pore units, and then They are respectively stored in the one-dimensional array p _i , where i is the number of the subfield; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the pores that store the ordinal number of the pore unit and the corresponding coordinates are obtained An array of cells, each pore cell is tracked by its unique ordinal number and coordinates; For each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derivative arrays are obtained; the momentum and local fluid of the pore unit The density data and the ordinal number of the pore unit are correspondingly stored as a one-dimensional array to obtain a series of second derived arrays. 2. the LBM parallel optimization method based on connected pore division calculation area according to claim 1, is characterized in that: Among them, in step 2, the maximum number of units M _max :

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively. 3. the LBM parallel optimization method based on connected pore division calculation area according to claim 1, is characterized in that: Wherein, in step 3, the main data structure of each pore unit is its fluid particle distribution function and equilibrium fluid particle distribution function; the momentum and local fluid density data before and after the calculation of the pore unit correspond to the ordinal number of the pore unit Stored as a one-dimensional array, resulting in four second derived arrays. 4. the LBM parallel optimization method based on connected pore division calculation area according to claim 1, is characterized in that, also comprises: Step 4. Set the communication mode between subdomains: when dealing with adjacent or diagonal pore units between different subdomains, extend the interface layer on the current subdomain interface with an additional layer of units to cover adjacent and diagonal pore units Aperture units, so that the communication mode of these expanded aperture units is consistent with that of the interface units in the current sub-domain. 5. The LBM parallel optimization device based on connected pores to divide the calculation area, is characterized in that, comprising: The determining part determines the pore unit information and connectivity of the sample according to the pore distribution data of the porous medium sample to be simulated; determines the total number of sub-domains N to be decomposed according to the number N of computing nodes in the system; Computational domain decomposition part, divide the porous media sample flow domain into n _x regions with the same or similar number of units along the x-axis, and then divide each region into n _y sub-regions with the same or similar number of units along the y-axis , and finally divide the n _x ×n _y sub-regions along the z-axis n _z times to obtain n _x ×n _y ×n _z =N sub-regions with the same or similar number of units; the decomposed sub-regions should meet the conditions: The difference in the number of units between the subdomain with the maximum number of units M _max and the subdomain with the minimum number of units M _min should not exceed one thousandth of the total number of units; The calculation task allocation department assigns N sub-domains to N computing nodes one by one for parallel computing; when the computing nodes process sub-domains, they only consider all connected pore units, and recalculate these connected pore units. number, and then store them in the one-dimensional array p _i respectively, where i is the subfield number; each connected pore unit is associated with a coordinate, and stored in the order of the one-dimensional array, and the ordinal number of the stored pore unit is obtained and the pore unit array of the corresponding coordinates; for each pore unit: each function in the main data structure of the pore unit is stored as a one-dimensional array corresponding to the ordinal number of the pore unit, and a series of first derived arrays are obtained; The momentum and local fluid density data of the pore unit are stored as a one-dimensional array corresponding to the ordinal number of the pore unit to obtain a series of second derived arrays; The control unit is communicatively connected with the determination unit, the calculation domain decomposition unit and the calculation task allocation unit, and controls their operations. 6. the LBM parallel optimization device based on connected pore division calculation area according to claim 5, characterized in that: Among them, in the calculation domain decomposition part, the maximum number of units M _max :

Minimum number of units M _min :

In the formula, NX, NY, and NZ are the total number of units containing pores and solids on the x, y, and z axes of the simulation area, respectively. 7. The LBM parallel optimization device based on connected pore division calculation area according to claim 5, is characterized in that, also comprises: The communication mode setting unit is connected to the control unit in communication, and when processing adjacent or diagonal pore units between different subdomains, an additional layer of units is extended to the interface layer on the interface of the current subdomain to cover adjacent and diagonal pore units. The diagonal aperture units make the communication mode of these expanded aperture units consistent with the interface units in the current sub-domain. 8. The LBM parallel optimization device based on connected pore division calculation area according to claim 5, further comprising: The input and display unit is communicatively connected to the determination unit, the calculation domain decomposition unit, the calculation task allocation unit, the communication mode setting unit, and the control unit, and is used to allow the user to input operation instructions and perform are displayed accordingly. 9. The LBM parallel optimization device based on connected pore division calculation area according to claim 5, characterized in that: Wherein, the input display unit can display the sample void unit information and connection status determined by the determination unit and the number of calculation nodes or the total number N of sub-domains determined by the determination unit according to the operation instruction, and display the decomposition status of the calculation domain decomposition unit , and correspondingly display the distribution status of the computing task distribution unit. 10. A storage medium, characterized in that: A program for realizing the LBM parallel optimization method for dividing calculation regions based on connected pores according to any one of claims 1 to 4 is stored.

Images