A quick prediction method for infrared signature reduction of engine high-temperature component cooling
By constructing a surrogate model using full factorial design and response surface methodology, screening significant factors and performing regression analysis, the problem of rapidly predicting the infrared reduction effect of cooling schemes for high-temperature engine components was solved. This achieved efficient and accurate prediction of infrared indicators, and is applicable to infrared radiation assessment and stealth design of multiple components.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- AVIC GUIYANG ENGINE DESIGN & RES INST
- Filing Date
- 2026-03-10
- Publication Date
- 2026-07-14
AI Technical Summary
Existing technologies for evaluating the infrared reduction effect of cooling schemes for high-temperature engine components are cumbersome and time-consuming, making it difficult to quickly respond to multi-objective optimization design and stealth scheme selection. Furthermore, the model accuracy is insufficient, failing to meet the requirements of efficient and high-precision engineering design.
A full factorial experimental design method was adopted, and a surrogate model was constructed by DOE experimental design and response surface methodology. Significant influencing factors were screened, and linear or complete quadratic function regression analysis was used to establish a rapid prediction model, which simplifies the process of flow field simulation and infrared radiation simulation.
It significantly shortens the prediction cycle, improves prediction efficiency, ensures model fit R2>0.9, and provides high-precision infrared index prediction results, which are applicable to the infrared stealth design of various aero-engines.
Smart Images
Figure CN122389280A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of stealth technology for aero-engines, specifically relating to a rapid prediction method for the reduction of infrared indicators caused by the cooling of high-temperature engine components. Background Technology
[0002] Infrared stealth performance of aero-engines is one of the core indicators of modern aviation equipment. Infrared radiation from high-temperature components (such as turbine blades, internal cones, and nozzle walls) is the main source of infrared signals from the engine's exhaust. To reduce infrared radiation intensity, cooling designs are required for high-temperature components, and the effectiveness of the cooling scheme in reducing infrared performance directly determines the rationality of the stealth design.
[0003] In existing technologies, evaluating the infrared reduction effect of cooling schemes requires sequential modeling, flow field simulation, infrared radiation simulation, and data comparison for each component temperature combination. This process is cumbersome, time-consuming, and difficult to respond quickly to the needs of multi-objective optimization design and stealth scheme selection. Although some studies have used experimental design and approximate modeling methods to improve prediction efficiency, they lack a systematic design for the infrared radiation characteristics of high-temperature engine components. This results in problems such as inaccurate factor selection, insufficient model accuracy, and limited applicability, failing to meet the requirements of efficient and high-precision engineering design. Summary of the Invention
[0004] The purpose of this invention is to provide a rapid prediction method for the reduction of infrared indicators caused by the cooling of high-temperature engine components, addressing the technical problems described in the background art.
[0005] The technical solution of this invention: A rapid prediction method for the reduction of infrared parameters due to cooling of high-temperature engine components includes the following steps: S1: Identify optimization factors, based on demand identification, identify the prediction models to be carried out for the optimization design objectives, and identify the typical high-temperature component influencing factors for the target types of the optimization design objectives that need to be predicted, and identify the components that have a greater impact on the tail infrared radiation intensity as optimization factors. S2: Obtain the full factorial experimental point combination, and use the full factorial experimental design method to carry out the full factorial experimental point design with 2 levels and 4 factors for the optimized factor to obtain the full factorial experimental point combination to be carried out; S3: Obtain the optimized target value. For the combination of all factors, the flow field simulation and infrared test model are simplified and processed. Based on the flow field simulation analysis or empirical analysis, the optimized target value of the factor under different temperature combinations in each test point is obtained. S4: Screen the principal factors by using the Pareto plot of standardized effects, the normal plot of standardized effects, and factor significance effect analysis based on analysis of variance to obtain the principal factors and interaction factors that have a significant impact on the prediction target. S5: Obtain test points based on response surface methodology, design test points using Box-Behnken designs in response surface methodology, establish a mathematical model for predicting the objective function value by significant influencing factors, set factors, and generate sample space points; S6: Obtain the average value of infrared integrated radiation intensity, perform infrared radiation simulation calculations on the sample spatial points, and obtain the average value of infrared integrated radiation intensity within a range of ±45° for each test point; S7: Fit the prediction function model. Based on the corresponding surface design method of least squares, use linear, linear + square, linear + interaction or complete quadratic function forms for regression analysis, obtain the coefficients of each term, and fit the target prediction function model. S8: Prediction accuracy verification, using R 2 The method for testing the model fit is as follows: when R... 2 When R > 0.9, the fitted prediction function model is deemed acceptable, and the prediction accuracy meets the requirements. 2 When the value is ≤0.9, the fitted prediction function model is deemed unqualified and its accuracy does not meet the requirements. The range of factor variables is adjusted, and steps S5-S7 are repeated. S9: Infrared index prediction. The temperature combination of the high-temperature component to be predicted is input into a qualified function model, and the corresponding infrared index reduction effect is output as the infrared index prediction result.
[0006] In step S1, the high-temperature components include turbine blades, turbine support, inner cone, and nozzle wall.
[0007] In step S1, the target type includes afterburning / non-afterburning turbojet, turbofan and turboshaft engines, and the optimization factors include turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
[0008] In step S2, level 2 refers to the minimum temperature TL and the maximum temperature TH, and factor 4 refers to the turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
[0009] In step S3, the simplification process retains the main component surface features, omits multi-edged edges and porous microstructures, and optimizes the target values including the mean, maximum, and median infrared radiation intensity of a specific angular domain.
[0010] In step S4, the principal factors are three factors that have a significant impact on the mean value of infrared integrated radiation intensity, including turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The interaction factors are factors that have no significant impact on the mean value of infrared integrated radiation intensity, including turbine support temperature B.
[0011] In step S5, the test point design adopts two levels of minimum temperature TL and maximum temperature TH, and three factors of turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The factors are set as minimum temperature TLA, TLC, TLD and maximum temperature THA, THC, THD of turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D.
[0012] The sample space points are generated based on the minimum temperature TLA, TLC, TLD and the maximum temperature THA, THC, THD, resulting in a total of 15 sample space points with 3 repeating central test points.
[0013] In step S7, regression fitting is performed using a complete quadratic function. The expression for the function model is: + + , where y is the predicted value of the infrared index. Significant factors is the regression coefficient.
[0014] The specific angular range is within a range of ±45° in the tail direction.
[0015] The beneficial effects of this invention are: In this invention, a surrogate model is constructed using Design of Experiments (DOE) and response surface methodology, eliminating the need for complete simulations of every temperature combination. This reduces the prediction cycle by more than 80% compared to traditional methods, significantly improving efficiency. Furthermore, the use of significant factor screening and quadratic regression fitting techniques results in a high model fit R0. 2 A value >0.9 ensures the reliability of the prediction results and guarantees the required accuracy. This invention is applicable to the infrared stealth design of various aero-engines and can be extended to the infrared radiation prediction of multiple components with different emissivity, or the infrared performance evaluation of similar multi-target combined structures, making it widely applicable. This invention can be directly applied to the preliminary screening of stealth schemes, the allocation of infrared indicators, and the decomposition of cooling requirements, providing an efficient tool for the overall infrared stealth design of engines. Specifically, this invention does not require the traditional processes of modeling, flow field simulation, infrared radiation simulation, data comparison and screening for each stealth combination scheme. Instead, it obtains a rapid prediction model containing significant influencing factors through DOE experimental design and response surface methodology, which has significant advantages such as high accuracy, high efficiency, and good repeatability. Attached Figure Description
[0016] Figure 1 This is a flowchart of the process steps of the present invention. Detailed Implementation
[0017] refer to Figure 1 A rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components includes the following steps: S1: Define the optimization factors. Based on demand identification, define the prediction model to be carried out for the optimization design target. For the target type of the optimization design target that needs to be predicted, identify the influence factors of typical high-temperature components. Identify the components that have a greater impact on the back infrared radiation intensity as optimization factors. Use the operating temperature of each component as the initial factor variable to ensure that the factor variables are relatively independent.
[0018] The high-temperature components include turbine blades, turbine support, inner cone, and nozzle wall. The target types include afterburning / non-afterburning turbojet, turbofan, and turboshaft engines, and the optimization factors include turbine blade temperature (A), turbine support temperature (B), inner cone temperature (C), and nozzle wall temperature (D).
[0019] This step specifically involves: based on the target requirements, clarifying the prediction model to be developed for the optimization design objective. For example, in this embodiment, the average infrared integrated radiation intensity within a range of ±45° is used as the prediction target. For the target type to be predicted, such as afterburning, non-afterburning turbojet, turbofan, or turboshaft engines, identify components that significantly affect the exhaust infrared radiation intensity at the rear end as optimization factors. In this embodiment, the main infrared radiating components of the exhaust system at the rear end of a non-afterburning turbofan engine are used as the object, including key factors such as turbine blade temperature (A), turbine support temperature (B), inner cone temperature (C), and nozzle wall temperature (D).
[0020] S2: Obtain the full factorial experimental point combination, and use the full factorial experimental design method to carry out the full factorial experimental point design with 2 levels and 4 factors for the optimized factor to obtain the full factorial experimental point combination to be carried out; The two levels are the minimum temperature TL and the maximum temperature TH, and the four factors are the turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
[0021] This step is based on the DOE program in Minitab software. First, according to the structural form of the afterburning / non-afterburning turbofan engine, several relatively independent factor variables are initially screened. Then, a full factorial test model is established using Minitab software. Specifically, the full factorial test design method is adopted, and the full factorial test points are designed according to 2 levels (minimum temperature TL, maximum temperature TH) and 4 factors (factor turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D) to obtain the required test point combination.
[0022] S3: Obtain the optimized target value. For the combination of all factors, the flow field simulation and infrared test model are simplified and processed. Based on the flow field simulation analysis or empirical analysis, the optimized target value of the factor under different temperature combinations in each test point is obtained. The simplification process retains the main component surface features, omits multi-edged edges and porous microstructures, and optimizes the target values including the mean, maximum, and median infrared radiation intensity in a specific angular region. The specific angular range is within ±45° of the tail.
[0023] The specific operation is as follows: the flow field simulation and infrared test model to be carried out for the test object are reasonably simplified. In principle, the surface features of the main components are retained, and the micro-structures such as multi-edged edges and multi-pores are simplified. Based on the flow field simulation analysis or empirical analysis, the optimized target values of factors under different temperature combinations in each test point are obtained. Typical examples include the mean, maximum, and median infrared radiation intensity in the range of ±45° in the tail direction.
[0024] In this step, the simplified processing of the flow field simulation and infrared experimental model helps to improve simulation efficiency.
[0025] This step involves simulating the infrared radiation characteristics based on the data from each set in the experimental model to obtain the infrared radiation intensity. After all the necessary data for the experimental model has been collected, the model is fitted and simplified to screen out factors with significant influence, i.e., step S4. S4: Screen the principal factors by using the Pareto plot of standardized effects, the normal plot of standardized effects, and factor significance effect analysis based on analysis of variance to obtain the principal factors and interaction factors that have a significant impact on the prediction target. The principal factors are three factors that have a significant impact on the mean of the infrared integrated radiation intensity, including turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The interaction factors are factors that have no significant impact on the mean of the infrared integrated radiation intensity, including turbine support temperature B.
[0026] Specifically, this step involves using the Pareto plot of standardized effects, the normal plot of standardized effects, and factor significance effect analysis based on analysis of variance to identify the principal and interaction factors that significantly influence the prediction target. In this embodiment, three factors—turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D—are identified as having a significant impact on the mean of the backward infrared integrated radiation intensity within a ±45° range, and are therefore considered principal factors. Turbine support temperature is a weak factor and an interaction factor with no significant impact; therefore, only these three factors—turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D—are selected for constructing the backward prediction model.
[0027] In this step, by screening out the main factors and interaction factors that have a significant impact on the prediction target and eliminating weak factors such as turbine support temperature, the complexity of the subsequent model is reduced and the prediction efficiency is improved.
[0028] S5: Obtain test points based on response surface methodology, design test points using Box-Behnken designs in response surface methodology, establish a mathematical model for predicting the objective function value by significant influencing factors, set factors, and generate sample space points; The test point design uses two levels: minimum temperature TL and maximum temperature TH, and three factors: turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The factors are set as the minimum temperature TLA, TLC, TLD and the maximum temperature THA, THC, THD of turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D.
[0029] The sample space points are generated based on the minimum temperature TLA, TLC, TLD and the maximum temperature THA, THC, THD, resulting in a total of 15 sample space points with 3 repeating central test points.
[0030] In this step, to further establish a mathematical model for predicting the objective function value by significantly influencing factors, Box-Behnken designs (BBD) experimental design from the response surface design method is used to design test points for 2 levels (minimum temperature TL, maximum temperature TH) and 3 factors (factor turbine blade temperature A, inner cone temperature C, nozzle wall temperature D). The minimum temperatures TLA, TLC, TLD and the maximum temperatures THA, THC, THD of factor turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D are set respectively, generating a total of 15 sample space points with 3 repeated central test points. In the actual test, the same central points can be removed.
[0031] In this step, duplicate center points can be removed to simplify the calculation.
[0032] S6: Obtain the average value of infrared integrated radiation intensity, perform infrared radiation simulation calculations on the sample spatial points, and obtain the average value of infrared integrated radiation intensity within a range of ±45° for each test point; Specifically, this step involves performing infrared radiation simulation calculations on the sample space points from step S5S to obtain the average value of the infrared integrated radiation intensity within a range of ±45° for each test point.
[0033] This step is a repetition of step S3, performing infrared radiation simulation calculations to ensure the accuracy and reliability of infrared index data for each sample point, providing high-quality data support for fitting the surrogate model.
[0034] S7: Fit the prediction function model. Based on the corresponding surface design method of least squares, use linear, linear + square, linear + interaction or complete quadratic function forms for regression analysis, obtain the coefficients of each term, and fit the target prediction function model. Regression fitting is performed using a complete quadratic function, and the expression of the function model is: + + , where y is the predicted value of the infrared index. Significant factors is the regression coefficient.
[0035] S8: Prediction accuracy verification, using R 2 The method for testing the model fit is as follows: when R... 2 When R > 0.9, the fitted prediction function model is deemed acceptable, and the prediction accuracy meets the requirements. 2 When the value is ≤0.9, the fitted prediction function model is deemed unqualified and its accuracy does not meet the requirements. The range of factor variables is adjusted, and steps S5-S7 are repeated. In this invention, R 2 The calculation formula is R 2 =1-(RSS / TSS), where RSS is the residual sum of squares and TSS is the total sum of squares. Residual sum of squares (RSS): Represents the sum of squared deviations between actual and predicted values, indicating the degree of unknown variation in a variable; Total Sum of Squares (TSS): Represents the sum of squares between the actual and expected values, indicating the total degree of variation of the variable. Generally speaking, the higher the proportion of the variable variation in a predictive model to the total variation, the more accurate the model. When RSS=0, it means the model can completely simulate the total variation of the variable. Therefore, R... 2 A value close to 1 indicates a good model fit. In practical applications, R is generally assumed to be... 2 A value greater than 0.8 indicates that the model is accurate.
[0036] S9: Infrared Indicator Prediction. This step inputs the combined temperatures of the high-temperature components to be predicted into a qualified function model, outputting the corresponding infrared indicator reduction effect as the infrared indicator prediction result. This step can quickly output the corresponding infrared indicator reduction effect prediction result without repeating complex simulation calculations, significantly improving design efficiency.
[0037] In this invention, a surrogate model is constructed using Design of Experiments (DOE) and response surface methodology, eliminating the need for complete simulations of every temperature combination. This reduces the prediction cycle by more than 80% compared to traditional methods, significantly improving efficiency. Furthermore, the use of significant factor screening and quadratic regression fitting techniques results in a high model fit R0. 2A value >0.9 ensures the reliability of the prediction results and guarantees the required accuracy. This invention is applicable to the infrared stealth design of various aero-engines and can be extended to the infrared radiation prediction of multiple components with different emissivity, or the infrared performance evaluation of similar multi-target combined structures, making it widely applicable. This invention can be directly applied to the preliminary screening of stealth schemes, the allocation of infrared indicators, and the decomposition of cooling requirements, providing an efficient tool for the overall infrared stealth design of engines. Specifically, this invention does not require the traditional processes of modeling, flow field simulation, infrared radiation simulation, data comparison and screening for each stealth combination scheme. Instead, it obtains a rapid prediction model containing significant influencing factors through DOE experimental design and response surface methodology, which has significant advantages such as high accuracy, high efficiency, and good repeatability.
[0038] The following are examples for illustration: Taking a non-afterburning turbofan engine as an example, the reduction effect of the mean infrared integrated radiation intensity within a range of ±45° in the tail direction is predicted. The specific steps are as follows: (1) Define the optimization factors: The prediction object is the average value of infrared integrated radiation intensity within a range of ±45° in the tail direction of a non-afterburning turbofan engine. High-temperature components are identified, including turbine blades, turbine support, inner cone, and nozzle wall. The corresponding initial factor variables are turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
[0039] (2) Obtaining the full factorial experimental point combination: Based on the Minitab software DOE program, the minimum temperature TL and maximum temperature TH of each factor were set (A: 800K-1200K, B: 600K-1000K, C: 700K-1100K, D: 500K-900K), and 2 4 =A full factorial experimental model with 16 test sites.
[0040] (3) Obtain the optimization target value: simplify the engine flow field and infrared radiation model, omit the micropore structure, conduct simulations at 16 test points respectively, and collect the average value data of infrared integrated radiation intensity.
[0041] (4) Screening of principal factors: Through standardized effects Pareto plot, normal plot and variance analysis, turbine blade temperature A, inner cone temperature C and nozzle wall temperature D were identified as significant principal factors, while turbine support temperature B was a weak factor (interaction factor) and was removed.
[0042] (5) Obtain test points based on response surface methodology: Box-Behnken designs (BBD) were used for the three significant factors A, C, and D, and the temperature range was set (A: 850K-1150K, C: 750K-1050K, D: 550K-850K) to generate 15 sample points (including 3 replicate center test points).
[0043] (6) Obtain the mean value of infrared integrated radiation intensity: Conduct infrared radiation simulation on 15 sample points and obtain the corresponding mean value data of infrared integrated radiation intensity.
[0044] (7) Fitting the prediction function model: Regression analysis was performed using a complete quadratic function to obtain the surrogate model expression: y=a0+a1A+a2C+a3D+a4A 2 +a5C 2 +a6D 2 +a7AC+a8AD+a9CD, solve for each coefficient using the least squares method.
[0045] (8) Validation of prediction accuracy: Calculate the model R 2 =0.94, which meets the accuracy requirements, and the model is determined to be a qualified surrogate model.
[0046] (9) Infrared index prediction: Input the temperature combination to be predicted (A=950K, C=850K, D=650K), and the model will quickly output the prediction result of the mean value of infrared integrated radiation intensity, so as to achieve efficient evaluation of the infrared index reduction effect.
Claims
1. A rapid prediction method for the reduction of infrared indexes due to cooling of high-temperature engine components, characterized in that... Includes the following steps: S1: Identify optimization factors, based on demand identification, identify the prediction models to be carried out for the optimization design objectives, and identify the typical high-temperature component influencing factors for the target types of the optimization design objectives that need to be predicted, and identify the components that have a greater impact on the tail infrared radiation intensity as optimization factors. S2: Obtain the full factorial experimental point combination, and use the full factorial experimental design method to carry out the full factorial experimental point design with 2 levels and 4 factors for the optimized factor to obtain the full factorial experimental point combination to be carried out; S3: Obtain the optimized target value. For the combination of all factors, the flow field simulation and infrared test model are simplified and processed. Based on the flow field simulation analysis or empirical analysis, the optimized target value of the factor under different temperature combinations in each test point is obtained. S4: Screen the principal factors by using the Pareto plot of standardized effects, the normal plot of standardized effects, and factor significance effect analysis based on analysis of variance to obtain the principal factors and interaction factors that have a significant impact on the prediction target. S5: Obtain test points based on response surface methodology, design test points using Box-Behnken designs in response surface methodology, establish a mathematical model for predicting the objective function value by significant influencing factors, set factors, and generate sample space points; S6: Obtain the average value of infrared integrated radiation intensity, perform infrared radiation simulation calculations on the sample spatial points, and obtain the average value of infrared integrated radiation intensity within a range of ±45° for each test point; S7: Fit the prediction function model. Based on the corresponding surface design method of least squares, use linear, linear + square, linear + interaction or complete quadratic function forms for regression analysis, obtain the coefficients of each term, and fit the target prediction function model. S8: Prediction accuracy verification, using R 2 The method for testing the model fit is as follows: when R... 2 When R > 0.9, the fitted prediction function model is deemed acceptable, and the prediction accuracy meets the requirements. 2 When the value is ≤0.9, the fitted prediction function model is deemed unqualified and its accuracy does not meet the requirements. The range of factor variables is adjusted, and steps S5-S7 are repeated. S9: Infrared index prediction. The temperature combination of the high-temperature component to be predicted is input into a qualified function model, and the corresponding infrared index reduction effect is output as the infrared index prediction result.
2. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 1, characterized in that: In step S1, the high-temperature components include turbine blades, turbine support, inner cone, and nozzle wall.
3. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 1, characterized in that: In step S1, the target type includes afterburning / non-afterburning turbojet, turbofan and turboshaft engines, and the optimization factors include turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
4. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 3, characterized in that: In step S2, level 2 refers to the minimum temperature TL and the maximum temperature TH, and factor 4 refers to the turbine blade temperature A, turbine support temperature B, inner cone temperature C, and nozzle wall temperature D.
5. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 1, characterized in that: In step S3, the simplification process retains the main component surface features, omits multi-edged edges and porous microstructures, and optimizes the target values including the mean, maximum, and median infrared radiation intensity of a specific angular domain.
6. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 4, characterized in that: In step S4, the principal factors are three factors that have a significant impact on the mean value of infrared integrated radiation intensity, including turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The interaction factors are factors that have no significant impact on the mean value of infrared integrated radiation intensity, including turbine support temperature B.
7. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 6, characterized in that: In step S5, the test point design adopts two levels of minimum temperature TL and maximum temperature TH, and three factors of turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D. The factors are set as minimum temperature TLA, TLC, TLD and maximum temperature THA, THC, THD of turbine blade temperature A, inner cone temperature C, and nozzle wall temperature D.
8. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 7, characterized in that: The sample space points are generated based on the minimum temperature TLA, TLC, TLD and the maximum temperature THA, THC, THD, resulting in a total of 15 sample space points with 3 repeating central test points.
9. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 1, characterized in that: In step S7, regression fitting is performed using a complete quadratic function. The expression for the function model is: + + , where y is the predicted value of the infrared index. Significant factors is the regression coefficient.
10. The rapid prediction method for the reduction of infrared indicators due to cooling of high-temperature engine components according to claim 5, characterized in that: The specific angular range is within a range of ±45° in the tail direction.