Optimization method for height-speed characteristic test point of engine

Abstract

The invention belongs to the technical field of aero-engine performance, and particularly relates to an optimization method for height-speed characteristic test points of an engine. The method comprises the steps that engine height and speed characteristics are obtained based on engine overall performance simulation calculation; preprocessing the height and speed characteristic data; obtaining a height speed characteristic three-dimensional curved surface through interpolation; determining a nearest point of the test point; fitting the local plane to obtain a local plane normal vector; calculating a local plane normal vector to obtain cosine similarity; performing validity judgment on the cosine similarity to obtain optimized test point data; obtaining a test point interpolation curved surface change diagram before and after optimization; obtaining a relation graph of the change degree of the curved surface along with cosine similarity change; and obtaining an optimized test point concentrated area.

Description

Technical Field

[0001] This invention relates to, but is not limited to, the field of aero-engine performance technology, and specifically to a method for optimizing test points for engine altitude-speed characteristics. Background Technology

[0002] In section 5.2.2 of GJB 243A-2004, "Flight Test Requirements for Aviation Gas Turbine Power Plants," regarding the evaluation test of engine altitude-speed characteristics, it states that "select the engine operating state for measuring the engine's altitude-speed characteristics, and select a series of sufficient flight altitude and speed points within the flight envelope of the aircraft or helicopter." The term "sufficient" is not quantified. Using a "carpet-coverage" method for flight testing would be too costly; therefore, flight testing under relatively limited conditions is essential. The current main approach to engine performance flight testing abroad is to establish an engine performance calculation model based on flight test data, input the verification test results into the performance calculation model, replace flight tests, and determine the flight test state points and engine thrust. Currently, the selection method for altitude-velocity characteristic test points typically involves qualitatively selecting 2-3 altitudes based on experience from similar engine flight test projects, and qualitatively selecting 3-5 velocity points at each altitude to form a test point matrix. However, due to the differences in altitude-velocity characteristics among different types of engines, test points selected based on experience cannot determine their representativeness or suitability to specific engine characteristics, and cannot ensure that the number of selected test points is minimized. Current methods for optimizing altitude-velocity characteristic test points in a two-dimensional plane have drawbacks such as poor spatial adaptability and the inability to reflect the effects of altitude changes. Summary of the Invention

[0003] The purpose of this invention is to propose an optimization method for engine altitude-velocity characteristic test points, in order to solve the problems that existing test point selection mainly relies on flight test project experience, which makes it impossible to judge whether the selection results are representative and suitable for specific engine laws, and to achieve the goal of mastering the variation law information of altitude-velocity characteristics within the engine envelope. It also addresses the problems of poor spatiality and inability to reflect the influence of altitude changes when optimizing test points in a two-dimensional plane.

[0004] The technical solution of this invention is as follows: An optimization method for engine altitude-speed characteristic test points, comprising: Step 1: Based on the overall engine performance simulation, calculate the corresponding thrust according to the height-velocity characteristics within the engine envelope, and combine the height-velocity-thrust data into test point data. Step 2: Interpolate and process the test point data obtained in Step 1 to obtain a three-dimensional surface; Step 3: Find the nearest neighbor point of the test point among the interpolation points obtained in Step 2, and obtain the nearest neighbor point data; Step 4: Establish a plane equation for each test point and its nearest neighbor data, and solve for the local fitting plane equation and local plane normal vector for each test point based on the nearest neighbor data; Step 5: Calculate the similarity of the local plane normal vectors from Step 4 to obtain cosine similarity data; Step 6: Filter the cosine similarity data calculated in Step 5 according to the set cosine similarity threshold to obtain new test point data after kicking the point. Step 7: Repeat steps 4, 5, and 6 to obtain the optimized test point data; Step 8: Based on the optimized test point data obtained in Step 7, obtain the optimized 3D surface through interpolation, calculate the projection coordinates of the test point data before optimization onto the optimized 3D surface, calculate the Euclidean distance between the projection coordinates and the test point coordinates before optimization, and obtain the mean distance. Step 9: Change the set cosine similarity threshold, repeat steps 7 and 8, and obtain the relationship between the degree of surface change and the set cosine similarity threshold. The degree of surface change is represented by the mean distance. Step 10: Based on the relationship curve of the degree of surface change with the set cosine similarity threshold, determine the abrupt change point in the relationship curve where the mean distance is less than the first set value and the cosine similarity threshold is greater than the second set value, and then obtain the optimized test point data corresponding to the cosine similarity threshold at the abrupt change point as the final optimization result, and then determine the concentrated area of ​​the optimized test points.

[0005] Step 2, specifically: Interpolate the test points obtained in step 1 at equal intervals within the data range to generate an equally spaced grid and obtain interpolation point data; standardize the original test points and interpolation data as a whole to obtain standardized test point data and interpolation data; draw a three-dimensional surface based on the standardized test point data and interpolation point data, requiring the three-dimensional surface to strictly pass through all standardized test point data and ensure the smoothness of the surface.

[0006] Step 3 specifically involves: Step 3: Find k nearest neighbors for each test point in the interpolation point data obtained in Step 2. The nearest neighbor data is obtained by gradually expanding the range to find k points with the closest spatial distance, where k is not less than 3.

[0007] Step 5 specifically involves: Calculate the cosine of the angle between two local plane normal vectors obtained in step 4 in sequence to obtain the directional similarity between two adjacent local plane normal vectors and obtain cosine similarity data. Step 6 specifically involves: Based on the cosine similarity data obtained in step 5, obtain the two normal vectors corresponding to the first cosine similarity data that is continuously greater than the cosine similarity threshold, remove the test point data corresponding to the second normal vector, and obtain the new test point data after the kick-out point.

[0008] In step 10, the first setting value is 0.7-0.75, and the second setting value is 0.97-0.98.

[0009] In step 4: The local plane and its equation are obtained by fitting the nearest neighbor data using the least squares method, and the local plane normal vector is obtained based on the local plane equation.

[0010] The beneficial effects of this invention are as follows: This invention provides an optimized design method for test points of engine altitude-velocity characteristics to solve the problem of selecting test points based on flight test project experience. At the same time, by optimizing test points in three-dimensional space, it is possible to grasp the variation law of altitude-velocity characteristics within the engine envelope and provide a range for verification flight tests. This can effectively reduce the number of flight tests, improve flight test efficiency, and improve the accuracy of test point selection. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating an optimization method for engine altitude-speed characteristic test points. Figure 2 This is a schematic diagram for calculating cosine similarity. Figure 3 This is a schematic diagram illustrating the repeated steps of cosine similarity. Figure 4 This is a diagram illustrating the nearest neighbor point. Figure 5 This is a graph showing the relationship between the degree of surface variation and optimization conditions. Detailed Implementation

[0012] To make the method of the present invention clearer, the following uses a certain type of engine as an example, based on its altitude-speed characteristic test point data, to clearly and completely describe the optimization design method of the present invention.

[0013] To address the problems in selecting test points for engine altitude-velocity characteristics, this invention proposes an optimization method for test points. By optimizing the design of test points for altitude-velocity characteristics within the engine envelope, the variation law of engine altitude-velocity characteristics within the envelope can be grasped, the test points in representative areas can be accurately obtained, the number of test points can be significantly reduced, flight test efficiency can be improved, and the main spatial surface characteristics are still preserved.

[0014] This invention proposes an optimization method for engine altitude-speed characteristic test points, such as... Figure 1 As shown, it includes: Step 1: Based on the component characteristics and the zero-dimensional aerodynamic parameters thermodynamic calculation method within the component, construct the error function and nonlinear equation system. By solving the nonlinear equation system, evaluate whether the error function meets the accuracy requirements, obtain the steady-state iterative solution, perform numerical simulation of the engine, calculate the corresponding thrust based on the height-velocity characteristics within the engine envelope, and combine the height-velocity-thrust data into test point data.

[0015] Step 2: Perform equidistant interpolation on the test points obtained in Step 1 within the data range to generate an equidistant grid and obtain interpolation point data; standardize the original test points and interpolation data as a whole to obtain standardized test point data and interpolation data; draw a three-dimensional surface based on the standardized test point data and interpolation point data, requiring the three-dimensional surface to strictly pass through all standardized test point data and ensure the smoothness of the surface.

[0016] Step 3: In the interpolation point data obtained in Step 2, find k nearest neighbors for each test point data and obtain the nearest neighbor data; the search method is: gradually expand the range to find k points with the closest spatial distance, requiring k to be no less than 3; Step 4: Establish a plane equation for each test point and its nearest neighbor data, and solve for the local fitting plane equation and local plane normal vector for each test point based on the nearest neighbor data; Step 5: Calculate the cosine of the angle between the two local plane normal vectors obtained in Step 4 in sequence, see... Figure 2 The example shown obtains the directional similarity between two adjacent local plane normal vectors and acquires cosine similarity data. Cosine similarity calculation formula:

[0017] Step 6: Based on the cosine similarity data obtained in Step 5, obtain the two normal vectors corresponding to the cosine similarity data that are greater than the set cosine similarity threshold, remove the test point data corresponding to either of the two normal vectors, and obtain the new test point data after removing the point. The optimization conditions include: when there are at least two consecutive cosine similarity data that satisfy the optimization conditions, intermediate test point data that are used to calculate the cosine similarity of these two data that satisfy the optimization conditions can be removed. Step 7, repeat steps 4, 5, and 6, see below for details. Figure 3 Obtain optimized test point data; Step 8: Based on the optimized test point data obtained in Step 7, obtain the concentrated area of ​​optimized test points, obtain the optimized three-dimensional surface through interpolation, calculate the projection coordinates of the test point data before optimization onto the optimized three-dimensional surface, and calculate the Euclidean distance between the projection coordinates and the test point coordinates before optimization. Step 9: Change the set cosine similarity threshold, repeat steps 7 and 8, and obtain the relationship between the degree of surface change and the optimization conditions. The degree of surface change is represented by Euclidean distance, and the change of optimization conditions is represented by the set cosine similarity threshold. Step 10: Based on the relationship curve of the degree of surface change with the set cosine similarity threshold, determine the abrupt change point in the relationship curve where the mean distance is less than the first set value and the cosine similarity threshold is greater than the second set value, and then obtain the optimized test point data corresponding to the cosine similarity threshold at the abrupt change point as the final optimization result.

[0018] The first setting is 0.7-0.75, and the second setting is 0.97-0.98.

[0019] This scheme optimizes the test points for engine height-speed characteristics, and the specific implementation method is as follows: Step 1: Based on the overall engine performance simulation, calculate the engine height-velocity characteristics and obtain test point data composed of height Hp, velocity Ma, and thrust Fn.

[0020] Step 2: Perform equidistant interpolation on the test points within the data range to generate an equidistant grid and obtain the interpolation point data; standardize the original test points and interpolation data as a whole to obtain the standardized test point data and interpolation data; draw a three-dimensional surface based on the standardized test point data and interpolation point data.

[0021] Step 3: Find the 6 nearest neighbors for each test point in the interpolation point data and obtain the nearest neighbor data. See [link to relevant documentation]. Figure 4 Example shown; Step 4: Establish a plane equation in the form ax + by + cz + d = 0 for each test point and its nearest neighbor data, construct the error function, find the derivative and establish a system of linear equations, solve a, b, c, d to obtain the local fitting plane equation and the local plane normal vector (a, b, c). Step 5: Calculate the cosine of the angle between two local plane normal vectors in sequence based on the local plane normal vectors to obtain the similarity in direction between two adjacent local plane normal vectors, i.e., obtain the cosine similarity data; Cosine similarity calculation formula:

[0022] Step 6: In the cosine similarity data, obtain the two normal vectors corresponding to the first cosine similarity data point that is consecutively greater than the cosine similarity threshold, and remove the test point data corresponding to the second normal vector. Figure 2 As shown, when the cosine similarity m and n are continuously greater than the cosine similarity threshold, the test points corresponding to the normal vector b are removed, and new test point data are obtained after the removal of the points. Step 7: Repeat steps 4, 5, and 6 to obtain the optimized test point data; Step 8: Interpolate the optimized test point data to obtain the optimized 3D surface, calculate the projection coordinates of the test point data before optimization onto the optimized 3D surface, and obtain the mean distance by calculating the Euclidean distance between the projection coordinates and the test point coordinates before optimization. Step 9: Change the set cosine similarity threshold, and repeat steps 7 and 8 to obtain the relationship between the degree of surface change and the set cosine similarity threshold. See [link to step 9]. Figure 5 In the example shown, the degree of surface variation is characterized by the mean distance; Step 10: Based on the relationship curve of the degree of surface change with the set cosine similarity threshold, determine the abrupt change points in the relationship curve where the mean distance is less than the first set value and the cosine similarity threshold is greater than the second set value. See [link to relevant documentation]. Figure 5 As shown in the example, the coordinates of the mutation point are (0.98, 0.73). The optimized test point data corresponding to the cosine similarity threshold at the mutation point is then obtained as the final optimization result, and the concentrated area of ​​the optimized test points is determined.

[0023] The first setting is 0.7-0.75, and the second setting is 0.97-0.98.

[0024] The above detailed embodiments are a description of the present invention. It should not be considered that the specific embodiments of the present invention are limited to these descriptions. For those skilled in the art, several simple deductions and substitutions can be made without departing from the concept of the present invention, and all of these should be considered to fall within the protection scope of the present invention.

Claims

1. A method for optimizing engine altitude-speed characteristic test points, characterized in that, include: Step 1: Based on the overall engine performance simulation, calculate the corresponding thrust according to the height-velocity characteristics within the engine envelope, and combine the height-velocity-thrust data into test point data. Step 2: Interpolate and process the test point data obtained in Step 1 to obtain a three-dimensional surface; Step 3: Find the nearest neighbor point of the test point among the interpolation points obtained in Step 2, and obtain the nearest neighbor point data; Step 4: Establish a plane equation for each test point and its nearest neighbor data, and solve for the local fitting plane equation and local plane normal vector for each test point based on the nearest neighbor data; Step 5: Calculate the similarity of the local plane normal vectors from Step 4 to obtain cosine similarity data; Step 6: Filter the cosine similarity data calculated in Step 5 according to the set cosine similarity threshold to obtain new test point data after kicking the point. Step 7: Repeat steps 4, 5, and 6 to obtain the optimized test point data; Step 8: Based on the optimized test point data obtained in Step 7, obtain the optimized 3D surface through interpolation, calculate the projection coordinates of the test point data before optimization onto the optimized 3D surface, calculate the Euclidean distance between the projection coordinates and the test point coordinates before optimization, and obtain the mean distance. Step 9: Change the set cosine similarity threshold, repeat steps 7 and 8, and obtain the relationship between the degree of surface change and the set cosine similarity threshold. The degree of surface change is represented by the mean distance. Step 10: Based on the relationship curve of the degree of surface change with the set cosine similarity threshold, determine the abrupt change point in the relationship curve where the mean distance is less than the first set value and the cosine similarity threshold is greater than the second set value, and then obtain the optimized test point data corresponding to the cosine similarity threshold at the abrupt change point as the final optimization result, and then determine the concentrated area of ​​the optimized test points.

2. The method for optimizing engine altitude-speed characteristic test points according to claim 1, characterized in that, Step 2, specifically: Interpolate the test points obtained in step 1 at equal intervals within the data range to generate an equally spaced grid and obtain interpolation point data; standardize the original test points and interpolation data as a whole to obtain standardized test point data and interpolation data; draw a three-dimensional surface based on the standardized test point data and interpolation point data, requiring the three-dimensional surface to strictly pass through all standardized test point data and ensure the smoothness of the surface.

3. The method for optimizing engine altitude-speed characteristic test points according to claim 2, characterized in that, Step 3 specifically involves: Step 3: Find k nearest neighbors for each test point in the interpolation point data obtained in Step 2. The nearest neighbor data is obtained by gradually expanding the range to find k points with the closest spatial distance, where k is not less than 3.

4. The method for optimizing engine altitude-speed characteristic test points according to claim 3, characterized in that, Step 5 specifically involves: Calculate the cosine of the angle between two local plane normal vectors obtained in step 4 in sequence to obtain the directional similarity between two adjacent local plane normal vectors and obtain cosine similarity data.

5. The method for optimizing engine altitude-speed characteristic test points according to claim 4, characterized in that, Step 6 specifically involves: Based on the cosine similarity data obtained in step 5, obtain the two normal vectors corresponding to the first cosine similarity data that is continuously greater than the cosine similarity threshold, remove the test point data corresponding to the second normal vector, and obtain the new test point data after the kick-out point.

6. The method for optimizing engine altitude-speed characteristic test points according to claim 1, characterized in that, In step 10, the first setting value is 0.7-0.75, and the second setting value is 0.97-0.

98.

7. The method for optimizing engine altitude-speed characteristic test points according to claim 5, characterized in that, In step 4: The local plane and its equation are obtained by fitting the nearest neighbor data using the least squares method, and the local plane normal vector is obtained based on the local plane equation.

8. A storage medium, characterized in that, It stores a computer program that performs the method as described in any one of claims 1-7.

Images