FOUP internal AMC flow characteristic simulation method based on mechanism model

Abstract

The invention provides an FOUP internal AMC flow characteristic simulation method based on a mechanism model, and relates to the technical field of semiconductors, and the method comprises the steps: building an FOUP mechanism model, and carrying out the grid processing of the FOUP mechanism model through a topological mapping segmentation method; selecting an initial AMC simulation point through grid importance analysis, and optimizing the initial AMC simulation point by using an ecological niche theory to obtain a simulation sampling point layout; and setting a boundary condition and an initial condition in combination with the actual working condition of the FOUP, and carrying out analogue simulation on the flow characteristics of the AMC in the FOUP based on the simulation sampling point layout to obtain an analogue simulation result. The FOUP internal AMC flow characteristic simulation method not only can effectively solve the simulation problem of FOUP internal AMC flow characteristics, but also can provide scientific basis and technical support for pollution control in the semiconductor manufacturing field, and has important engineering value and application prospect.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and more particularly, to a method for simulating the internal AMC flow characteristics of a FOUP based on a mechanism model. Background Art

[0002] In the semiconductor manufacturing industry, the Front Opening Unified Pod (FOUP), as a standardized portable container, is widely used for storing and transporting wafers between different process equipment. Its core function is to provide a clean and stable microenvironment for wafers to avoid contamination of the wafer surface by external pollutants (such as airborne molecular contaminants, AMC), thereby ensuring the yield and reliability of chip manufacturing. During the operation of the FOUP, air management and pollutant diffusion are crucial factors. Especially during the accumulation and removal of pollutants inside the FOUP, factors such as air flow, air velocity, temperature, and humidity directly affect the distribution of pollutants. Particularly in semiconductor manufacturing, the management of airborne particles is of utmost importance, and the diffusion of any tiny particle or pollutant may lead to production defects. However, despite the strict sealing and filtering measures in the design of the FOUP, there are still potential flow and contamination problems, which may gradually accumulate during long-term transportation and affect the final wafer quality.

[0003] Although existing FOUP designs can control the diffusion of pollutants to a certain extent, in practical applications, the air flow characteristics and pollutant distribution inside the FOUP are still uneven. Most current simulation methods mainly focus on rough estimations of air flow and pollutant distribution, but these methods often ignore the influence of subtle flows and local pollution sources, resulting in simulation results that cannot fully and accurately predict the pollutant diffusion pattern inside the FOUP.

[0004] In view of the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] In view of this, the present invention provides a method for simulating the internal AMC flow characteristics of a FOUP based on a mechanism model to solve the above-mentioned problems.

[0006] To solve the above problems, the specific technical solution adopted by the present invention is as follows: A method for simulating the internal AMC flow characteristics of a FOUP based on a mechanism model, comprising the following steps: S1. Obtain the structural parameter information of the FOUP and establish a mechanism model of the FOUP, and perform grid processing on the mechanism model of the FOUP by using a topological mapping segmentation method; S2. Based on the mechanism model of the FOUP after grid processing, select the initial AMC simulation points through grid importance analysis, and optimize the initial AMC simulation points using the niche theory to obtain the layout of the simulation sampling points; S3. Set the boundary conditions and initial conditions in combination with the actual working conditions of the FOUP, and simulate the internal AMC flow characteristics of the FOUP based on the layout of the simulation sampling points to obtain the simulation results.

[0007] Preferably, for obtaining the structural parameter information of the FOUP and establishing the mechanism model of the FOUP, the grid processing of the mechanism model of the FOUP using the topological mapping segmentation method includes the following steps: S11. Collect the structural parameters of the FOUP based on the FOUP design document, and use modeling software to construct the mechanism model of the FOUP, including the air inlet, exhaust port, tray position, and air flow channel; S12. Determine the internal space of the FOUP based on the mechanism model of the FOUP, and use the Gaussian mixture model to divide the internal space of the FOUP to obtain regular-shaped regions and irregular-shaped regions; S13. For the regular-shaped regions and irregular-shaped regions, use the topological mapping segmentation method for grid processing respectively; S14. For the regular-shaped regions and irregular-shaped regions after grid processing, connect the grids of the regular-shaped regions and irregular-shaped regions through the grid fusion method.

[0008] Preferably, for determining the internal space of the FOUP based on the mechanism model of the FOUP and using the Gaussian mixture model to divide the internal space of the FOUP to obtain regular-shaped regions and irregular-shaped regions, it includes the following steps: S121. Perform particle discretization processing on the internal space of the FOUP to obtain a meshless particle set; S122. For each particle in the meshless particle set, according to its spatial position, use a preset search radius to find adjacent particles to form a neighborhood set, and extract particle features from the neighborhood particles of each particle. The particle features include particle coordinates, coordinate variance, and local particle density; S123. According to the particle features, use two-dimensional singular value decomposition and Gaussian mixture model to perform clustering analysis on each particle in the meshless particle set to obtain a clustering result; S124. Based on the clustering result, divide the internal space of the FOUP to obtain regular-shaped regions and irregular-shaped regions.

[0009] Preferably, according to the particle characteristics, clustering analysis is performed on each particle in the meshless particle set by using two-dimensional singular value decomposition and Gaussian mixture model, and the clustering result includes the following steps: S1231. Sort out the particle characteristics to form a particle characteristic matrix, and construct a covariance matrix according to the particle characteristic matrix; S1232. Decompose the covariance matrix by two-dimensional singular value decomposition method to obtain eigenvectors and eigenvalues, and use the eigenvectors to transform the particle characteristic matrix to obtain a reduced-dimensional characteristic matrix; S1233. Flatten the reduced-dimensional characteristic matrix of each particle into a one-dimensional vector, and construct a new characteristic data set; S1234. Use the Gaussian mixture model to perform clustering analysis on the new data set to obtain a clustering result.

[0010] Preferably, the grid processing of the regular-shaped region and the irregular-shaped region by using the topological mapping segmentation method respectively includes the following steps: S131. For the regular-shaped region and the irregular-shaped region, respectively construct an initial computational domain homeomorphic to a hexahedron; S132. Adopt the planning strategy in the topological mapping segmentation method to decompose the initial computational domain into multiple sub-regions, and perform boundary parameterization processing on each sub-region to determine the parameter range; S133. Based on the geometric shape characteristics of the regular-shaped region and the irregular-shaped region, calculate the boundary function, and calculate the surface node coordinate values of the sub-regions according to the boundary function; S134. Connect the surface node coordinates of each sub-region to form a quadrilateral mesh, and perform spatial splicing on the quadrilateral mesh to generate a hexahedron mesh.

[0011] Preferably, the internal space of the FOUP is divided into a regular-shaped region and an irregular-shaped region based on the clustering result, including: Calculate the statistical characteristics of each cluster, including the mean and covariance matrix; Based on the statistical characteristics of each obtained cluster, divide the internal space of the FOUP into a regular-shaped region and an irregular-shaped region.

[0012] Preferably, based on the mechanism model of the FOUP after grid processing, initial AMC simulation points are selected through grid importance analysis, and the initial AMC simulation points are optimized by using the niche theory to obtain the layout of simulation sampling points, including the following steps: S21. Collect the AMC historical data of the FOUP, and combine the emission characteristics of the AMC to identify and locate the pollution source position; S22. Based on the location of the pollution source, use the grid-processed FOUP mechanism model to analyze the importance of the grids, and select the initial AMC simulation points according to the analysis results; S23. Use the niche theory to optimize the distribution of the initial AMC simulation points and generate the final layout of the simulation sampling points.

[0013] Preferably, the collecting the AMC historical data of the FOUP, combining with the emission characteristics of the AMC, identifying and locating the pollution source location includes the following steps: S211. Based on the pre-configured monitoring system of the FOUP, obtain the AMC historical data of the preset time period and perform preprocessing, and the preprocessing includes data cleaning, missing value filling and outlier processing; S212. According to the preprocessed AMC historical data, identify and mark the exceeding-standard events by analyzing the change of the pollutant concentration; S213. Based on the mechanism model of the FOUP, use computational fluid dynamics to simulate the air flow path inside the FOUP, and combine with the exceeding-standard events to analyze the diffusion path of the pollutants and judge the diffusion source of the pollutants; S214. Combine the equipment layout of the FOUP and the diffusion source of the pollutants, and determine the pollution source location through the positioning algorithm.

[0014] Preferably, the based on the pollution source location, using the grid-processed FOUP mechanism model to analyze the importance of the grids, and selecting the initial AMC simulation points according to the analysis results includes the following steps: S221. Determine the influence range of the pollution source on the internal environment of the FOUP based on the pollution source location; S222. Analyze the importance of each grid unit in the FOUP according to the influence range of the pollution source on the internal environment of the FOUP; S223. According to the analysis results of the grid importance, select the key grid units as the initial AMC simulation points.

[0015] Preferably, the using the niche theory to optimize the distribution of the initial AMC simulation points and generate the final layout of the simulation sampling points includes the following steps: S231. For the initial AMC simulation points, use the niche overlap method to calculate the overlap degree of every two initial AMC simulation points on the influence range of the internal environment of the FOUP, and use the niche dominance method to calculate the niche dominance of each initial AMC simulation point relative to other initial AMC simulation points; S232. Combine the results of niche overlap and niche dominance to analyze the redundancy and competitive advantages between the simulation points; S233. Based on the analysis results, adjust the layout of the simulation points to obtain the final layout of the simulation sampling points.

[0016] The beneficial effects of the present invention are as follows: The present invention uses the topological mapping segmentation method to perform grid processing on the FOUP, decomposing the complex geometric structure into multiple small units, facilitating refined analysis of each local area, capable of improving the spatial resolution of the simulation, making the results more accurate. By performing importance analysis on the grid-based model, the key areas with the greatest impact on pollution diffusion are identified, thereby selectively choosing the initial simulation points in a targeted manner, reducing unnecessary waste of computing resources. By capturing the physical laws inside the FOUP through the mechanism model and simultaneously using numerical simulation technology to simulate the flow characteristics, the leap from theory to practice is achieved. It can not only effectively solve the problem of simulating the AMC flow characteristics inside the FOUP, but also provide a scientific basis and technical support for pollution control in the semiconductor manufacturing field, having important engineering value and application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings: Figure 1 is a flowchart of a method for simulating the AMC flow characteristics inside the FOUP based on a mechanism model according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only some embodiments of this application, rather than all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0019] According to an embodiment of the present invention, a method for simulating the AMC flow characteristics inside the FOUP based on a mechanism model is provided.

[0020] Now, the present invention will be further described in conjunction with the drawings and specific embodiments. As Figure 1 shown, a method for simulating the AMC flow characteristics inside the FOUP based on a mechanism model according to an embodiment of the present invention includes the following steps: S1. Obtain the structural parameter information of the FOUP and establish a mechanism model of the FOUP, and perform grid processing on the mechanism model of the FOUP using the topological mapping segmentation method; As a preferred embodiment, obtaining the structural parameter information of the FOUP and establishing a mechanism model of the FOUP, and performing grid processing on the mechanism model of the FOUP by using the topological mapping segmentation method includes the following steps: S11. Collect the structural parameters of the FOUP based on the FOUP design document, and use modeling software to construct a mechanism model of the FOUP, including an air inlet, an air outlet, a tray position, and an air flow channel; It should be noted that key structural parameters are extracted from the design document, including: the position, size, and shape of the air inlet; the position, size, and shape of the air outlet; the size, quantity, and relative position of the trays; the size and layout of the air flow channel.

[0021] S12. Determine the internal space of the FOUP based on the mechanism model of the FOUP, and use a Gaussian mixture model to divide the internal space of the FOUP into regions, obtaining regular-shaped regions and irregular-shaped regions; As a preferred embodiment, determining the internal space of the FOUP based on the mechanism model of the FOUP and using a Gaussian mixture model to divide the internal space of the FOUP into regions, obtaining regular-shaped regions and irregular-shaped regions includes the following steps: S121. Perform particle discretization processing on the internal space of the FOUP to obtain a meshless particle set; It should be noted that performing particle discretization processing on the internal space of the FOUP is a method of converting the internal space of the FOUP into particle representation. The meshless method is a discretization technology that does not rely on traditional grids, but represents the space by distributing particles in the space. The particles are freely distributed in the space and can adapt to different physical phenomena in irregular geometries.

[0022] Each particle not only represents a position during discretization, but also has physical properties, such as: the three-dimensional coordinate position of the particle, the flow velocity, density, temperature, etc. of the particle.

[0023] S122. For each particle in the meshless particle set, find adjacent particles within a preset search radius according to its spatial position to form a neighborhood set, and extract particle features from the neighboring particles of each particle. The particle features include particle coordinates, coordinate variances, and local particle densities; Specifically, the search radius is a preset spatial range, indicating the maximum distance that a particle can perceive in the space. S123. According to the particle features, perform clustering analysis on each particle in the meshless particle set by using two-dimensional singular value decomposition and a Gaussian mixture model to obtain a clustering result; As a preferred embodiment, the clustering analysis of each particle in the meshless particle set according to the particle characteristics by using two-dimensional singular value decomposition and Gaussian mixture model, and the obtained clustering results include the following steps: S1231. Sort out the particle characteristics to form a particle characteristic matrix, and construct a covariance matrix according to the particle characteristic matrix; It should be noted that the covariance matrix is an important tool for measuring the mutual relationship between various characteristics, and is usually used to describe the distribution of the data set and the correlation between various characteristics. Before constructing the covariance matrix, it is usually necessary to standardize the particle characteristic matrix. The purpose of standardization is to ensure that the contribution of each characteristic to the covariance matrix is equal, and to avoid some characteristics with larger scales (such as coordinates) from dominating the covariance calculation.

[0024] S1232. Decompose the covariance matrix by two-dimensional singular value decomposition method to obtain eigenvectors and eigenvalues, and use the eigenvectors to transform the particle characteristic matrix to obtain a dimensionality-reduced characteristic matrix; The transformation of the particle characteristic matrix is to use the selected eigenvectors to transform the original particle characteristic matrix. Specifically, it is to multiply the original characteristic matrix by the transpose of the selected eigenvector matrix. Project the original high-dimensional feature space into a low-dimensional space to obtain a dimensionality-reduced characteristic matrix.

[0025] S1233. Flatten the dimensionality-reduced characteristic matrix of each particle into a one-dimensional vector, and construct a new feature data set; It should be noted that after processing by two-dimensional singular value decomposition (SVD) or other dimensionality reduction methods, each particle will correspond to a dimensionality-reduced characteristic matrix. This characteristic matrix is usually a low-dimensional representation, where each row (or column, depending on the specific dimensionality reduction method and implementation) of the characteristic matrix represents the characteristics of a particle. In addition, the flattening operation is to arrange all the elements of the matrix in row (or column) order into a long vector. For example, if the dimensionality-reduced characteristic matrix of a particle is a 2×3 matrix, then the flattened one-dimensional vector will contain 6 elements.

[0026] S1234. Use the Gaussian mixture model to perform clustering analysis on the new data set to obtain clustering results.

[0027] S124. Based on the clustering results, divide the internal space of the FOUP into regular-shaped regions and irregular-shaped regions.

[0028] Specifically, after clustering by the Gaussian mixture model (GMM), the cluster (regular region or irregular region) to which each particle belongs has been obtained. The clustering results not only give the cluster to which each particle belongs, but also can further describe the characteristics of each cluster through information such as the mean and covariance of the cluster, so as to provide a basis for region division.

[0029] In addition, the internal space of the FOUP is divided into regions according to the clustering results. Specifically, it includes: (1) Calculate the statistical characteristics of each cluster, including the mean and covariance matrix; (2) Based on the characteristics (mean and covariance matrix) of each obtained cluster, divide the internal space of the FOUP into two types of regions: Regular-shaped region: Consider the clusters with concentrated means and small covariance matrices in the clustering results as regular-shaped regions. The particle distribution is uniform and the region shape is relatively regular.

[0030] Irregular-shaped region: Consider the clusters with dispersed means and large covariance matrices in the clustering results as irregular-shaped regions. The particle distribution is sparse and the region shape is asymmetric or complex.

[0031] S13. For the regular-shaped region and the irregular-shaped region, use the topological mapping segmentation method to perform grid processing respectively; As a preferred implementation manner, the use of the topological mapping segmentation method to perform grid processing on the regular-shaped region and the irregular-shaped region respectively includes the following steps: S131. For the regular-shaped region and the irregular-shaped region, construct an initial computational domain homeomorphic to a hexahedron respectively; Specifically, the regular-shaped region usually has a simple geometric boundary (such as a cuboid, a cylinder, etc.), and its topological structure directly corresponds to that of a hexahedron. When constructing the initial computational domain, it can be directly mapped to a standard hexahedron. The construction steps include: (1) Parameterize the boundary of the regular region into the six faces of a hexahedron. For example: The cuboid region can be mapped to the parameter space of [0,1]×[0,1]×[0,1]. The cylinder region can be parameterized by polar coordinates and then mapped to a hexahedron (such as mapping the radial direction to [0,1], the axial direction to [0,1], and the circumferential direction is unfolded into a plane).

[0032] (2) Use a linear transformation (such as an affine transformation) to map the physical coordinates to the parameter coordinates.

[0033] The boundary curvature of the irregular-shaped region (such as complex mechanical parts, biological tissues) changes greatly and cannot be directly mapped to a hexahedron. It needs to be decomposed into multiple parameterizable sub-regions through piecewise homeomorphic mapping or geometric deformation. The construction steps include: Step 1: Divide the irregular region into multiple sub-regions, and each sub-region is approximated as a regular shape (such as a quadrilateral or a hexahedron block). Step 2: Construct a mapping homeomorphic to a hexahedron for each sub-region. For example, map the sub-region boundary to a unit square or cube through a radial basis function (RBF) or harmonic mapping. Step 3: Minimize the mapping distortion (such as the Jacobian determinant approaching 1) to avoid distortion during mesh generation.

[0034] S132. Adopt the planning strategy in the topological mapping segmentation method to decompose the initial computational domain into multiple sub-regions, and perform boundary parameterization on each sub-region to determine the parameter range; It should be noted that after constructing the initial computational domains of the regular region and the irregular region, the next step is to further decompose these regions into multiple sub-regions and perform boundary parameterization on each sub-region. Among them, perform an equi-proportional mesh division on the regular region to generate uniformly distributed sub-regions (for example, equally spaced division). For the irregular region, adopt an adaptive division and perform dynamic division based on the boundary curvature or geometric features within the region. In places with a larger curvature within the region, the mesh division is denser; while in places with a smaller curvature, the mesh division is sparser.

[0035] S133. Calculate the boundary function based on the geometric shape features of the regular-shaped region and the irregular-shaped region, and calculate the surface node coordinate values of the sub-region according to the boundary function; As a preferred implementation manner, the calculation formula for calculating the boundary function based on the geometric shape features of the regular-shaped region and the irregular-shaped region and calculating the surface node coordinate values of the sub-region according to the boundary function is: ; ; ; In the formula, X(r, s), Y(r, s), and Z(r, s) represent the XYZ coordinate values of the nodes on the surface of the sub-region in the Cartesian coordinate system, r and s represent two parameters in the natural coordinate system, which are used to represent the relative position of any point on the surface, a1(r) and a2(r) respectively represent the boundary functions related to the r direction on the surface boundary, a1(s) and a2(s) respectively represent the boundary functions related to the r direction on the surface boundary, X(0, 0), X(1, 0), X(0, 1), X(1, 1) respectively represent the X coordinate values of the four corner points of the quadrilateral in the Cartesian coordinate system, Y(0, 0), Y(1, 0), Y(0, 1), Y(1, 1) respectively represent the Y coordinate values of the four corner points of the quadrilateral in the Cartesian coordinate system, and Z(0, 0), Z(1, 0), Z(0, 1), Z(1, 1) respectively represent the Z coordinate values of the four corner points of the quadrilateral in the Cartesian coordinate system.

[0036] S134. Connect the surface node coordinates of each sub-region to form a quadrilateral mesh, and perform spatial splicing on the quadrilateral mesh to generate a hexahedral mesh.

[0037] It should be noted that the surface of each sub-region obtains node coordinates through boundary parameterization, and these node coordinates are known in the local coordinate system. Connect the surface nodes of each sub-region to form a quadrilateral mesh. Once the quadrilateral mesh nodes of all sub-regions are calculated, the next step is to splice the meshes of these sub-regions to finally generate a hexahedral mesh.

[0038] S14. For the regular-shaped region and irregular-shaped region after meshing, connect the meshes of the regular-shaped region and irregular-shaped region through the mesh fusion method.

[0039] After meshing, for the regular-shaped region and irregular-shaped region, the mesh fusion method (Mesh Merging Method) can be used to connect the meshes of these two different shapes, so as to form a coherent overall mesh.

[0040] S2. Based on the mechanism model of the FOUP after meshing, select the initial AMC simulation points through mesh importance analysis, and optimize the initial AMC simulation points using the niche theory to obtain the layout of the simulation sampling points. As a preferred implementation manner, the mechanism model of the FOUP based on meshing, selecting the initial AMC simulation points through mesh importance analysis, and optimizing the initial AMC simulation points using the niche theory to obtain the layout of the simulation sampling points includes the following steps: S21. Collect the AMC historical data of the FOUP, and combine with the emission characteristics of the AMC to identify and locate the source of pollution. As a preferred implementation manner, the collecting the AMC historical data of the FOUP, combining with the emission characteristics of the AMC to identify and locate the source of pollution includes the following steps: S211. Based on the pre-configured monitoring system of the FOUP, obtain the AMC historical data of the preset time period and perform preprocessing, and the preprocessing includes data cleaning, missing value filling, and outlier processing. It should be noted that the historical data of the AMC is collected through the monitoring system of the FOUP. These data include the concentration changes of pollutants, air flow paths, temperature and humidity, etc. in the internal environment of the FOUP. The monitoring system can include gas sensors, flow meters, temperature and humidity sensors, etc.

[0041] S212. According to the preprocessed AMC historical data, identify and mark the exceeded standard events by analyzing the changes in pollutant concentrations. S213. Based on the mechanism model of the FOUP, use computational fluid dynamics to simulate the airflow path inside the FOUP, and combine with the over-standard events to analyze the pollutant diffusion path and determine the pollutant diffusion source. Specifically, through CFD simulation, the flow situation of the airflow inside the FOUP can be calculated. The CFD model solves the air flow and heat transfer situation through fluid dynamics equations (such as the Navier-Stokes equation). The simulation results will show the airflow path, including the velocity distribution and pressure distribution of the airflow, revealing the airflow characteristics inside the FOUP.

[0042] S214. Combine the equipment layout of the FOUP and the pollutant diffusion source, and determine the pollution source location through the positioning algorithm.

[0043] Specifically, combining the equipment layout of the FOUP and the pollution source diffusion source, determining the final pollution source location includes: first, according to the equipment layout and airflow channel of the FOUP, analyze the influence range of the pollution source in the equipment space; Then, combine the airflow path, pollutant concentration distribution, reverse speculation results and equipment layout of the FOUP to determine the accurate location of the pollution source; finally, through a comprehensive variety of analysis means, finally locate the specific location of the pollution source.

[0044] S22. Based on the pollution source location, use the FOUP mechanism model after grid processing to analyze the importance of the grids, and select the initial AMC simulation points according to the analysis results. As a preferred implementation manner, the step of analyzing the importance of the grids based on the pollution source location, using the FOUP mechanism model after grid processing, and selecting the initial AMC simulation points according to the analysis results includes the following steps: S221. Determine the influence range of the pollution source on the internal environment of the FOUP based on the pollution source location. S222. Analyze the importance of each grid unit inside the FOUP according to the influence range of the pollution source on the internal environment of the FOUP. S223. According to the importance analysis results of the grids, select the key grid units as the initial AMC simulation points.

[0045] S23. Use the niche theory to optimize the distribution of the initial AMC simulation points to generate the final simulation sampling point layout.

[0046] As a preferred implementation manner, the step of using the niche theory to optimize the distribution of the initial AMC simulation points to generate the final simulation sampling point layout includes the following steps: S231. For the initial AMC simulation points, use the niche overlap method to calculate the overlap degree of every two initial AMC simulation points in the influence range of the FOUP internal environment, and use the niche dominance method to calculate the niche dominance of each initial AMC simulation point relative to other initial AMC simulation points; The formula for calculating the overlap degree of every two initial AMC simulation points in the influence range of the FOUP internal environment by using the niche overlap method is: ; The formula for calculating the niche dominance of each initial AMC simulation point relative to other initial AMC simulation points by using the niche dominance method is: ; In the formula, G ij represents the overlap degree of the initial AMC simulation point i and the initial AMC simulation point j in the influence range of the FOUP internal environment, M represents the number of initial AMC simulation points, m represents the index value of the initial AMC simulation point, C i represents the influence range of the initial AMC simulation point i in the FOUP internal environment, C j represents the influence range of the initial AMC simulation point j in the FOUP internal environment, H represents the number of grids of the FOUP mechanism model, h represents the index value of the grid of the FOUP mechanism model, P ij represents the niche dominance of the initial AMC simulation point i relative to the initial AMC simulation point j, L i>j,hn represents the situation where the influence range of the initial AMC simulation point i in the FOUP internal environment is larger than that of the initial AMC simulation point j in the FOUP internal environment; S232. Combine the results of niche overlap and niche dominance to analyze the redundancy and competitive advantages between simulation points; It should be noted that redundancy reflects the repeatability and ineffectiveness of two or more initial AMC simulation points in the same area. If the influence ranges of two simulation points overlap greatly and their contributions to the pollutant diffusion path are similar, there is redundancy.

[0047] S233. Based on the analysis results, adjust the layout of the simulation points to obtain the final layout of the simulation sampling points.

[0048] Specifically, according to the redundancy analysis, delete or rearrange the redundant simulation points with a high overlap degree. By adjusting the positions of these redundant points, ensure the uniform distribution of the simulation points inside the FOUP and avoid unnecessary repeated calculations. In addition, based on the analysis results of redundancy and competitive advantage, an optimization algorithm (such as genetic algorithm, simulated annealing, particle swarm optimization, etc.) is used to further adjust the layout of the simulation points, making them evenly distributed in the internal environment of the entire FOUP, reducing redundancy, and improving the coverage rate.

[0049] S3. Set the boundary conditions and initial conditions in combination with the actual working conditions of the FOUP, and simulate the AMC flow characteristics inside the FOUP based on the layout of the simulation sampling points to obtain the simulation results.

[0050] Among them, the initial condition is the setting of the fluid state inside the FOUP at the beginning of the simulation. Common initial conditions include: the initial air velocity distribution inside the FOUP, the initial concentration distribution of pollutants, the initial temperature and humidity inside the FOUP, the distribution of the initial pressure, and so on.

[0051] In addition, based on the layout of the simulation sampling points, the CFD tool can accurately simulate the air flow characteristics and pollutant diffusion inside the FOUP.

[0052] In summary, with the help of the above technical solutions of the present invention, the present invention uses the topological mapping segmentation method to grid the FOUP, decomposes the complex geometric structure into multiple small units, which is convenient for refined analysis of each local area, can improve the spatial resolution of the simulation, make the results more accurate, identify the key areas that have the greatest impact on pollution diffusion through the importance analysis of the gridded model, so as to select the initial simulation points targeted, reduce the waste of unnecessary computing resources, capture the physical laws inside the FOUP through the mechanism model, and at the same time use the numerical simulation technology to simulate the flow characteristics, realizing the leap from theory to practice. It can not only effectively solve the simulation problem of the AMC flow characteristics inside the FOUP, but also provide a scientific basis and technical support for pollution control in the semiconductor manufacturing field, and has important engineering value and application prospects.

[0053] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can be implemented in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be implemented in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code.

[0054] The specific embodiments described above have further elaborated on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for simulating the flow characteristics of AMC inside a FOUP based on a mechanism model, characterized in that: The following steps are involved: S1, obtaining the structural parameter information of FOUP and establishing the mechanism model of FOUP, and meshing the mechanism model of FOUP by using the topological mapping segmentation method; S2. Based on the mechanism model of the FOUP after gridding, the initial AMC simulation points are selected through grid importance analysis, and the initial AMC simulation points are optimized using the niche theory to obtain the layout of simulation sampling points; S3. Boundary conditions and initial conditions are set in combination with the actual working conditions of the FOUP, and the AMC flow characteristics inside the FOUP are simulated based on the layout of simulation sampling points to obtain simulation results.

2. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 1, characterized in that: The step of obtaining the structural parameter information of the FOUP and establishing the mechanism model of the FOUP, and meshing the mechanism model of the FOUP using the topological mapping segmentation method comprises the following steps: S11. Collect the structural parameters of the FOUP based on the FOUP design file, and use the modeling software to build a mechanism model of the FOUP, including the air inlet, exhaust port, tray position and air flow channel; S12, determining the internal space of the FOUP based on the mechanism model of the FOUP, and dividing the internal space of the FOUP into regions using a Gaussian mixture model to obtain regular shape regions and irregular shape regions; S13, for the regular shape area and the irregular shape area, respectively, a topological mapping segmentation method is used to perform gridding processing; S14, for the regular shape region and the irregular shape region after the meshing process, the meshes of the regular shape region and the irregular shape region are connected by a mesh fusion method.

3. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 2, characterized in that: The method of determining the internal space of the FOUP based on the mechanism model of the FOUP and dividing the internal space of the FOUP into regions using the Gaussian mixture model to obtain regular shape regions and irregular shape regions includes the following steps: S121, performing particle discretization processing on the internal space of the FOUP to obtain a gridless particle set; S122, for each particle in the gridless particle set, find adjacent particles according to its spatial position using a preset search radius to form a neighborhood set, and extract particle features from the neighborhood particles of each particle, wherein the particle features include particle coordinates, coordinate variance, and local particle density; S123, performing cluster analysis on each particle in the gridless particle set by using two-dimensional singular value decomposition and Gaussian mixture model according to the particle characteristics, and obtaining a clustering result; S124. Based on the clustering result, the internal space of the FOUP is divided into regions to obtain regular shape regions and irregular shape regions.

4. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 3, characterized in that: The method of clustering each particle in the gridless particle set by using two-dimensional singular value decomposition and Gaussian mixture model according to the particle characteristics to obtain the clustering result includes the following steps: S1231, sorting out the particle features to form a particle feature matrix, and constructing a covariance matrix according to the particle feature matrix; S1232, decomposing the covariance matrix by two-dimensional singular value decomposition method to obtain eigenvectors and eigenvalues, and transforming the particle characteristic matrix by using the eigenvectors to obtain a characteristic matrix after dimension reduction; S1233, flattening the dimension-reduced feature matrix of each particle into a one-dimensional vector, and constructing a new feature data set; S1234. Perform cluster analysis on the new data set using a Gaussian mixture model to obtain clustering results.

5. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 2, characterized in that: The meshing process for the regular shape area and the irregular shape area using the topological mapping segmentation method comprises the following steps: S131. For the regular shape region and the irregular shape region, construct an initial calculation domain that is homeomorphic to the hexahedron respectively; S132, using the planning strategy in the topological mapping segmentation method, decomposing the initial calculation domain into multiple sub-regions, and performing boundary parameterization processing on each sub-region to determine the parameter range; S133, calculating a boundary function based on geometric shape features of the regular shape region and the irregular shape region, and calculating the coordinate values ​​of the surface nodes of the sub-region according to the boundary function; S134, connecting the surface node coordinates of each sub-region to form a quadrilateral mesh, and spatially splicing the quadrilateral mesh to generate a hexahedral mesh.

6. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 3, characterized in that: The internal space of the FOUP is divided into regions based on the clustering result to obtain regular shape regions and irregular shape regions, including: Calculate the statistical characteristics of each cluster, including the mean and covariance matrix; Based on the statistical characteristics of each cluster, the internal space of the FOUP is divided into regular shape areas and irregular shape areas.

7. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 1, characterized in that: The mechanism model of the FOUP after gridding is used to select the initial AMC simulation points through grid importance analysis, and the initial AMC simulation points are optimized by using the niche theory to obtain the layout of the simulation sampling points, which includes the following steps: S21. Collect the AMC historical data of FOUP and identify and locate the pollution source based on the emission characteristics of AMC. S22. Based on the location of the pollution source, the importance of the grid is analyzed using the FOUP mechanism model after gridding, and the initial AMC simulation point is selected according to the analysis results; S23. Use the ecological niche theory to optimize the distribution of the initial AMC simulation points and generate the final simulation sampling point layout.

8. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 7, characterized in that: The collecting of the AMC historical data of the FOUP and identifying and locating the pollution source position in combination with the emission characteristics of the AMC include the following steps: S211, based on the pre-configured FOUP monitoring system, obtaining AMC historical data of a preset time period and performing preprocessing, wherein the preprocessing includes data cleaning, missing value filling and outlier processing; S212, based on the pre-processed AMC historical data, by analyzing the change of pollutant concentration, identifying and marking the exceeding event; S213. Based on the mechanism model of FOUP, computational fluid dynamics is used to simulate the airflow path inside the FOUP, and combined with the exceeding-standard event, the diffusion path of the pollutants is analyzed to determine the diffusion source of the pollutants; S214. Determine the location of the contamination source by using a positioning algorithm based on the equipment layout of the FOUP and the diffusion source of the contamination.

9. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 7, characterized in that: The method of performing importance analysis on the grid based on the location of the pollution source and selecting the initial AMC simulation point according to the analysis result by using the gridded FOUP mechanism model includes the following steps: S221, determining the impact range of the pollution source on the internal environment of the FOUP based on the location of the pollution source; S222, performing importance analysis on each grid unit in the FOUP according to the impact range of the pollution source on the internal environment of the FOUP; S223. According to the importance analysis result of the grid, select key grid units as initial AMC simulation points.

10. The method for simulating AMC flow characteristics inside a FOUP based on a mechanism model according to claim 7, characterized in that: The method of optimizing the distribution of the initial AMC simulation points by using the ecological niche theory to generate the final layout of the simulation sampling points includes the following steps: S231. For the initial AMC simulation points, the degree of overlap between the influence ranges of each two initial AMC simulation points on the FOUP internal environment is calculated using the niche overlap method, and the niche advantage of each initial AMC simulation point relative to other initial AMC simulation points is calculated using the niche advantage method; S232. Analyze the redundancy and competitive advantage between simulation points by combining the results of niche overlap and niche advantage; S233. Based on the analysis results, the simulation point layout is adjusted to obtain a final simulation sampling point layout.

Images