Establishing method of turbomachinery grid division database, computing device and application of computing device

Abstract

A method for establishing a turbomachine grid division database comprises the following steps: providing a plurality of blade profile samples, and providing blade profile characteristic parameters and grid control parameters of the blade profile samples; normalizing the blade profile characteristic parameters and constructing a distance function; performing k-means clustering analysis on the leaf profile sample according to the distance function, and dividing the leaf profile sample into k clusters; and clustering iteration is carried out until a clustering result meeting an iteration condition is obtained, and a turbomachine grid division database including the grid control parameters of the blade profile samples, the classification result and the clustering center is obtained. According to the method, a database for quantitatively comparing the similarity degree of the blade profile samples can be established, and the method is used for performing rapid grid division on the blade profiles. The invention further provides a calculation device and a turbomachinery grid division method.

Description

Technical Field

[0001] The present invention belongs to the field of aeroengines, and in particular relates to a method for establishing a turbomachinery grid partitioning database, a computing device and an application thereof. Background Art

[0002] In the fluid simulation calculation of impeller machinery, HOH structured grid is generally used to generate the computational domain grid. First, a suitable grid topology is given, but the grid topology of the blade generally has more control parameters. In the grid generation process, it is necessary to generate a high-quality grid by adjusting these control parameters. When generating grids for different blade types, the control parameters also change greatly. The change in control parameters is often related to geometric parameters such as the inlet structural angle and outlet structural angle of the blade. Therefore, when changing the blade type, it is necessary to manually adjust the grid control parameters to achieve a higher grid quality. This process is cumbersome and time-consuming, which increases the grid generation time of the simulation calculation. Therefore, providing an impeller machinery grid division database to achieve rapid division of impeller machinery grids has a high practical value for improving the computational efficiency of simulation. Summary of the invention

[0003] The purpose of the present invention is to provide a method for establishing a turbomachinery mesh partitioning database, so as to provide a database for fast mesh partitioning. The present invention also provides a turbomachinery mesh partitioning method.

[0004] An embodiment of one aspect of the present invention provides a method for establishing a turbomachinery grid partitioning database, using the k-means method to perform cluster analysis on known blade profile samples. K-means is one of the typical algorithms for unsupervised machine learning, and is a partitioning method that iteratively searches for k clusters so that the loss function corresponding to the clustering result is minimized. The embodiment of the present invention classifies blade profiles with similar structural features into one category through k-means clustering, and quantifies the similarity between different blade profile samples through a component distance function. The method specifically includes the following steps:

[0005] A plurality of blade profile samples are provided, and blade profile characteristic parameters and grid control parameters of the blade profile samples are provided; the blade profile characteristic parameters are normalized and a distance function is constructed; a k-means clustering analysis is performed according to the distance function, and a blade profile database is divided into k clusters given a k value; an iteration condition is set, and when the blade profile characteristic parameters and the grid control parameters in the same cluster do not meet the iteration condition, clustering is repeated until the blade profile characteristic parameters and the grid control parameters in the same cluster meet the iteration condition; and an impeller machinery grid division database including the blade profile characteristic parameters of the samples, the grid control parameters, the classification results and the cluster centers is obtained.

[0006] This method can be used to quantitatively classify different blade types and establish a blade mesh database with different cluster centers. When a new blade type needs to be meshed, the new blade type can be classified using the same quantitative method, and the blade type sample that is closest to the structural characteristics of the new blade type can be quickly determined from the database. Based on the mesh control parameters of the blade type sample, automatic meshing can be quickly achieved, thereby effectively improving meshing efficiency and reducing the time consumed by meshing during simulation analysis.

[0007] Furthermore, in some embodiments, the blade profile characteristic parameters include a plurality of combinations of an inlet geometry angle, an outlet geometry angle, a blade rotation direction, an inlet twist angle, an outlet twist angle, a grid fillet size, and whether or not it is an end-bend blade.

[0008] Furthermore, in some embodiments, the iteration conditions include: the blade rotation direction is consistent, whether the end-bend blades are consistent, the inlet geometric angle difference is not greater than 12°, the outlet geometric angle difference is not greater than 12°, the inlet torsion angle difference is not greater than 25°, and the outlet torsion angle difference is not greater than 25°.

[0009] Furthermore, in some embodiments, the iteration conditions include: the difference in the number of grid nodes in the grid control parameters does not exceed 10, and the difference in the inclination angle at the same position in the grid control parameters does not exceed 20°.

[0010] Further, in some embodiments, the normalized calculation method for extracting the characteristic parameters of the blade profile is: inew =W i (X i -X imin ) / (X imax -X imin ), where X inew is the normalized data, i represents different characteristic parameters, X imin is the minimum value of the characteristic parameter i, X imax is the maximum value of the characteristic parameter i, W i is the given weight value.

[0011] Furthermore, in some embodiments, the distance function D is calculated using the Euclidean distance:

[0012] Where j and k represent any two leaf types.

[0013] Further, in some embodiments, the process of giving the k value is as follows: Establishing the silhouette coefficient SC i =(b i -a i ) / max(b i ,ai ), where b i is the minimum distance between leaf type sample i and samples in other clusters, a i is the average distance from i to the samples in this cluster; given an initial k value, clustering is performed and the silhouette coefficient SC of all the leaf profile samples is calculated. i The average value SC is obtained; the k value is adjusted, clustered again and SC is calculated until the k value that satisfies |SC-1|=min|SC-1| is obtained.

[0014] According to an embodiment of another aspect of the present invention, a computing device is provided, which includes a memory and a processor, wherein the memory stores a computing program, and when the processor executes the computing program, the method for establishing a turbomachinery meshing database provided in any of the aforementioned embodiments can be implemented.

[0015] According to an embodiment of another aspect of the present invention, there is provided a method for meshing an impeller machinery, wherein the impeller machinery meshing database provided by the method for establishing the impeller machinery meshing database provided in any of the aforementioned embodiments is used to mesh a blade profile, and the method comprises the following steps: providing a target blade profile to be meshed, extracting and normalizing blade profile characteristic parameters of the target blade profile, and calculating a distance function of the target blade profile to obtain a target distance function; comparing the target distance function values ​​of the target blade profile with those of each cluster center, and assigning the target blade profile to a cluster where the cluster center having the smallest target distance function value is located as a target cluster; calculating the target distance function values ​​of the target blade profile and each blade profile in the target cluster, and assigning the grid control parameters of the blade profile sample corresponding to the minimum value of the target distance function to the target blade profile to complete meshing.

[0016] Furthermore, in some embodiments, the grid division is generated by HOH structured grid calculation.

[0017] The application of the impeller machinery meshing database makes full use of the quantified blade characteristic parameters to classify the target blade shape, and determines the blade shape sample in the database that is closest to the target blade shape. Rapid meshing is achieved by assigning the mesh control parameters of the existing blade shape sample to the target blade shape, which effectively reduces the workload of repeated meshing and improves the computational efficiency of the simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of establishing a turbomachinery mesh partitioning database in one embodiment;

[0019] Figure 2 A flowchart of fast meshing using a turbomachinery meshing database in one embodiment;

[0020] Figure 3 A schematic diagram of blade rotation direction in one embodiment;

[0021] Figure 4 This is a schematic diagram of the cross-sectional structure of a leaf-shaped sample in one embodiment;

[0022] Figure 5 Schematic diagram of leaf-shaped sample grid division topology in one embodiment.

[0023] The purpose of the above drawings is to explain the present invention in detail so that those skilled in the art can understand the technical concept of the present invention, but it is not intended to limit the present invention. For the sake of simplicity, the above drawings only schematically illustrate the structures related to the technical features of the present invention, and do not strictly follow the actual scale to draw the complete structure and all details. DETAILED DESCRIPTION

[0024] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0025] Reference to "embodiment" herein means that a particular feature, structure or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The phrase appearing in various locations in the specification does not necessarily refer to the same embodiment, nor is it limited to mutually exclusive independent or alternative embodiments. It should be understood by those skilled in the art that the embodiments herein may be combined with other embodiments without structural conflicts. In the description herein, "multiple" means at least two.

[0026] An embodiment of one aspect of the present invention provides a method for establishing a turbomachinery meshing database, so as to establish a turbomachinery meshing database that can quickly perform meshing when performing simulation calculations. The process of the method is as follows: Figure 1 As shown, the basic steps include:

[0027] Step 1: Collect blade profile samples, extract blade profile characteristic parameters of the blade profile samples and provide grid control parameters. In a preferred embodiment, the blade profile characteristic parameters include inlet geometry angle, outlet geometry angle, blade rotation direction, inlet twist angle, outlet twist angle, grid fillet size and whether it is an end-bend blade.

[0028] Step 2: Normalize the blade profile characteristic parameters and construct a distance function. In a preferred embodiment, the distance function is calculated using a weighted normalization method: inew =W i (X i -X imin ) / (X imax -X imin ), where X inewis the normalized data, i represents different characteristic parameters, X imin is the minimum value of the characteristic parameter i, X imax is the maximum value of the characteristic parameter i, W i is the given weight value. Distance function Where j and k represent any two leaf types.

[0029] Step 3: Perform k-means clustering analysis based on the distance function. Given a k value, divide the leaf type database into k clusters, and obtain classification results that meet the iteration conditions through multiple clustering.

[0030] In the preferred embodiment, step 3.1 is also included: given an initial k value, clustering is performed and the silhouette coefficient SC is calculated. i =(b i -a i ) / max(b i ,a i ) and the average silhouette coefficient SC of all leaf profile samples, where b i is the minimum distance between leaf type sample i and samples in other clusters, a i is the average distance from i to the samples in this cluster; the k value is adjusted until SC is closest to 1, and the determined k value is obtained.

[0031] In a preferred embodiment, the conditions that the iteration should meet include the consistent blade rotation direction, whether the blade is end-bent, the difference in inlet geometry angle is not greater than 12°, the difference in outlet geometry angle is not greater than 12°, the difference in inlet torsion angle is not greater than 25°, and the difference in outlet torsion angle is not greater than 25°. In a preferred embodiment, the conditions that the iteration should meet also include that the difference in the number of grid nodes in the grid control parameters does not exceed 10, and the difference in the inclination angle at the same position in the grid control parameters does not exceed 20°.

[0032] Step 4: Obtain the impeller mesh mechanical partitioning database including cluster centers, classification results, blade profile characteristic parameters and mesh control parameters.

[0033] Another embodiment of the present invention provides a computing device, which can be a general-purpose computer or a special-purpose computing device, and includes a memory and a processor, wherein a computing program is stored in the memory, and when the processor executes the computing program, the method for establishing a turbomachinery meshing database provided in the above embodiment can be implemented. The computing program can be written in a general-purpose computer language, such as Python, Java, etc.

[0034] Another embodiment of the present invention provides an application of a turbomachinery meshing database, which uses the turbomachinery meshing database provided in the above embodiment to quickly mesh the blade profile. The process is as follows: Figure 2 As shown, including:

[0035] Step 5.1: Given the target leaf shape;

[0036] Step 5.2: Extract blade characteristic parameters of the target blade;

[0037] Step 5.3: Normalize the blade profile characteristic parameters and establish the target distance function;

[0038] Step 5.4: Classify the target leaf type according to the target distance function, and take the cluster with the smallest target distance function value as the target cluster;

[0039] Step 5.5: Compare the target distance function values ​​corresponding to each blade sample in the target cluster, take the blade sample corresponding to the minimum value as the blade sample with the highest similarity to the target blade sample, assign the grid control parameters of the blade sample to the target blade sample and complete the grid division. In a preferred embodiment, the grid division process is generated by HOH structure and grid calculation.

[0040] In a preferred embodiment, the process of establishing a turbomachinery meshing database and using the database to perform meshing is as follows:

[0041] First, n blade profile data are collected, such as blade profile characteristic parameters and grid control parameters of the compressor, fan, and turbine, as shown in Table 1.

[0042]

[0043] Table 1 Blade data

[0044] Among them, the blade characteristic parameter a i represents the characteristic parameters of the i-th blade, combined with Figure 4 , subscripts 1-7 represent the inlet geometry angle 6 (α1), outlet geometry angle 7 (α2), blade rotation, inlet twist angle, outlet twist angle, grid fillet size, and whether it is an end-bent blade. Among them, the inlet twist angle is the difference between the inlet structural angle of the blade at the hub and the inlet structural angle of the blade tip, and the outlet twist angle is the difference between the inlet structural angle of the blade at the hub and the outlet structural angle of the blade tip; the blade rotation is as follows Figure 3 As shown in the figure, the rotation characteristic parameter value of blade type a is defined as 1, and the rotation characteristic parameter value of blade type b is defined as 0; the same value assignment method is used for end-bent blades. i represents the grid control parameters of the i-th blade, and the subscripts 1-10 represent the blade type in Figure 5 Different control parameters p1-p in the meshing topology of the leaf sample shown 10 .

[0045] Normalize the characteristic parameters of the leaf shape and construct a weighted distance function: X inew =W i (X i -X imin ) / (X imax -X imin ), where X inew is the normalized data, i takes values ​​of 1-7, representing the above-mentioned different characteristic parameters (inlet geometry angle, outlet geometry angle, blade rotation direction, inlet twist angle, outlet twist angle, grid fillet size, and whether it is an end-bend blade), X imin is the minimum value of characteristic parameter i, X imax is the maximum value of characteristic parameter i, W i is the given weight value. The range of each eigenvalue is shown in Table 2.

[0046]

[0047] Table 2 Range of characteristic values

[0048] The Euclidean distance is used to represent the distance D between samples of different leaf types:

[0049]

[0050] Among them, j and k represent any two leaf shape samples.

[0051] According to the leaf shape characteristic parameters and distance function, k-means distance analysis is performed to divide the leaf shape data into k clusters. The specific parameters of the leaf shape sample characteristics are shown in Table 3.

[0052]

[0053]

[0054] Table 3 Leaf type sample characteristic parameters and classification results

[0055] There are 36 blade samples in total in Table 3. First, the samples can be significantly divided into two categories by blade rotation: moving blades and stationary blades, with a moving blade rotation of 0 and a stationary blade rotation of 1. Among the stationary blades with a rotation of 1, the inlet structural angle and the outlet structural angle of samples 1 and 2 are both near 0, which are significantly different from other stationary blades and can be divided into one category; the inlet structural angle of samples 3 and 4 is close to 0, while the outlet structural angle is near 120, which is significantly different from other stationary blades and can be divided into one category; the inlet structural angle and the outlet structural angle of samples 14 to 16 are both small, and can be divided into one category; among the remaining stationary blade samples, samples 5 to 10 are divided into one category according to the fillet size of 0; samples 17 to 21 have a fillet size of 3 and are end-bend blades, which are divided into one category; samples 11 to 13 and 22 are not end-bend blades, and the fillet size is not 0, which are divided into one category. Similarly, the rotor blades with a rotation direction of 0 are divided into two categories: the rounded corner size in samples 23 to 29 is 0, and the rounded corner size in samples 31 to 36 is not 0. Based on the above analysis, the samples can be roughly divided into 8 categories, that is, the initial k = 8. First, clustering is performed and the silhouette coefficient SC is defined i =(b i -a i ) / max(b i ,a i ), where b i is the minimum distance between leaf type sample i and samples in other clusters, a i is the average distance from i to the samples in this cluster, and the average silhouette coefficient SC of all samples is calculated.

[0056] The clustering model is trained, the k value is increased or decreased, and the quality of the model is evaluated according to the SC value. When SC is closest to 1, that is, |SC-1|=min|SC-1|, the clustering model is considered to be in a good state. The k value is adjusted respectively, and finally min|SC-1| is obtained when k=13, and the given k value in k-means clustering is determined to be 13.

[0057] The clustering results are analyzed. K-means clustering easily converges to the local optimal solution. Multiple clustering iterations are required to determine whether the results meet expectations. The iterative criteria for blade feature parameters include: the judgment basis for similar blade features: 1) The blades of the same category have the same rotation direction; 2) The blades of the same category have rounded corners or no rounded corners; 3) Whether the blades of the same category have the same end-bend blade features; 4) The inlet angle deviation of the blades of the same category is not greater than 12°; 5) The outlet angle deviation of the blades of the same category is not greater than 12°; 6) The inlet twist angle deviation of the blades of the same category is not greater than 25°; 7) The outlet twist angle deviation of the blades of the same category is not greater than 25°.

[0058] At the same time, the similarity of the grid control parameters of the blade samples should also be judged. The grid control parameters of each blade sample are shown in Table 4, where H1, H2, H3, H4, BS1, BP1, BS2, BP2, O1 represent the number of grid nodes at different positions, and Hin, Hout, Sin, Sout represent the inclination angles at different positions. The iterative criterion of the grid control parameters is: 1) The deviation of the number of grid nodes at the same position of the blade of the same category does not exceed 10; 2) The deviation of the inclination angle at the same position of the blade of the same category does not exceed 20.

[0059]

[0060]

[0061] Table 4. Leaf sample grid control parameters

[0062] When the characteristic parameters and grid control parameters of the leaf profile samples in the same cluster meet the above iterative criteria, the clustering results are considered to have met the requirements, and the final clustering results are shown in Table 3. It can be considered that the distance D can more accurately reflect the similarity between the two leaf profiles. The cluster centers in each cluster are shown in Table 5.

[0063]

[0064] Table 5 Cluster centers

[0065] Thus, a turbomachinery meshing database including blade profile characteristic parameters, mesh control parameters, classification results, and cluster centers of blade profile samples is obtained.

[0066] Next, the database is used to mesh a target blade shape, and the characteristic parameters of the target blade shape are shown in Table 6.

[0067]

[0068] Table 6 Target blade profile characteristic parameters

[0069] The blade characteristic parameters of the target blade are normalized, the target distance function is established, and the target distance function values ​​between the target blade and each cluster center are calculated, as shown in Table 7.

[0070]

[0071] Table 7 Distance between target leaf shape and cluster center

[0072] Therefore, it can be determined that the target blade profile is the smallest distance from cluster 12 , and the target blade profile is classified into cluster 12 .

[0073] Next, the distances between each blade profile sample in cluster 12 and the target blade profile are compared, and it is determined that the distance of sample 23 is 0.1346, and the distance of sample 24 is 0.0025. Sample 24 is the blade profile sample most similar to the target blade profile. The grid control parameters of sample 24 are assigned to the target sample, and the HOH structured grid is calculated and generated.

[0074] The above process can be programmed in general-purpose computer languages ​​such as Python and implemented on general-purpose computers. The time consumed by using the impeller machinery meshing database to quickly divide the blade mesh is 1 to 2 orders of magnitude lower than that of manually adjusting the mesh control parameters for meshing, which can significantly improve the simulation efficiency of the impeller machinery. In a preferred embodiment, it takes 1 minute to generate the mesh control parameters of 10 rows of blades using the impeller machinery meshing database, while it takes 100 minutes to manually adjust the mesh control parameters of 10 rows of blades.

[0075] The purpose of the above embodiments is to further explain the present invention in detail in conjunction with the accompanying drawings so that those skilled in the art can understand the technical concept of the present invention. Within the scope of the present invention, the optimization or equivalent replacement of the method steps involved, and the combination of the implementation methods in different embodiments without conflict of principle, all fall within the protection scope of the present invention.

Claims

1. A method for establishing a turbomachinery mesh database, characterized in that: The following steps are involved: Providing a plurality of blade profile samples, and providing blade profile characteristic parameters and grid control parameters of the blade profile samples; Normalizing the blade profile characteristic parameters to construct a distance function; Performing k-means clustering analysis according to the distance function, and dividing the leaf type database into k clusters given a k value; Setting an iteration condition, when the blade profile characteristic parameters and the grid control parameters in the same cluster do not satisfy the iteration condition, repeating clustering until the blade profile characteristic parameters and the grid control parameters in the same cluster satisfy the iteration condition; The impeller machinery meshing database including blade characteristic parameters, mesh control parameters, classification results and cluster centers of blade samples is obtained.

2. The method for establishing a turbomachinery mesh database according to claim 1, characterized in that: The blade profile characteristic parameters include multiple combinations of an inlet geometric angle, an outlet geometric angle, a blade rotation direction, an inlet twist angle, an outlet twist angle, a grid fillet size, and whether it is an end-bent blade.

3. The method for establishing a turbomachinery mesh database according to claim 2, characterized in that: The iteration conditions include: blade rotation direction is consistent, whether the end-bend blades are consistent, the inlet geometric angle difference is not greater than 12°, the outlet geometric angle difference is not greater than 12°, the inlet torsion angle difference is not greater than 25°, and the outlet torsion angle difference is not greater than 25°.

4. The method for establishing a turbomachinery mesh database according to claim 1 or 2, characterized in that: The iteration conditions include: the difference in the number of grid nodes in the grid control parameters does not exceed 10, and the difference in the inclination angle at the same position in the grid control parameters does not exceed 20°.

5. The method for establishing a turbomachinery mesh database according to claim 1 or 2, characterized in that: The normalized calculation method for extracting the characteristic parameters of the blade profile is: X inew =W i (X i -X imin ) / (X imax -X imin ), where X inew is the normalized data, i represents different characteristic parameters, X imin is the minimum value of the characteristic parameter i, X imax is the maximum value of the characteristic parameter i, W i is the given weight value.

6. The method for establishing a turbomachinery mesh database according to claim 5, characterized in that: The distance function D is calculated using the Euclidean distance: Where j and k represent any two leaf types.

7. The method for establishing a turbomachinery mesh database according to claim 1 or 2, characterized in that: The process of giving a k value comprises the following steps: Establishing the silhouette coefficient SC i =(b i -a i ) / max(b i ,a i ), where b i is the minimum distance between leaf type sample i and samples in other clusters, a i is the average distance from i to the samples in this cluster; Given an initial k value, clustering is performed and the silhouette coefficient SC of all the leaf profile samples is calculated. i The average value of SC; The k value is adjusted, clustering is performed again, and SC is calculated until a k value that satisfies |SC-1|=min|SC-1| is obtained.

8. A computing device comprising a memory and a processor, characterized in that: The memory stores a calculation program, and when the processor executes the calculation program, the method for establishing a turbomachinery meshing database as claimed in any one of claims 1 to 7 can be implemented.

9. A method for meshing an impeller machine, characterized in that: Providing a turbomachinery mesh database, wherein the turbomachinery mesh database is obtained by using the method for establishing turbomachinery mesh data according to any one of claims 1 to 7; Providing a target blade profile to be meshed, extracting blade profile characteristic parameters of the target blade profile and normalizing them, and calculating a distance function of the target blade profile to obtain a target distance function; Comparing the target distance function values ​​of the target blade profile relative to each cluster center in the turbomachinery grid partition database, and classifying the target blade profile into the cluster where the cluster center with the smallest target distance function value is located as the target cluster; The target distance function value between the target blade profile and each blade profile in the target cluster is calculated, and the grid control parameter of the blade profile sample corresponding to the minimum value of the target distance function is assigned to the target blade profile to complete grid division.

10. The application of the turbomachinery meshing database according to claim 9, characterized in that: The grid division is generated by HOH structured grid calculation.

Images