Gas compressor performance uncertainty evaluation method considering blade profile geometric deviation
The compressor performance model is corrected through linear interpolation method and low-dimensional flow surface calculation method, and the problems of large workload and unstable results in the existing technology are solved, and efficient and accurate evaluation of compressor performance dispersion is achieved, which is suitable for processing Changes in leaf geometry.
Patent Information
- Application Number
- CN202411101579.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-08
AI Technical Summary
When evaluating compressor performance, the numerical simulation workload is huge and the results are affected by the unconverged numerical simulation solutions. It has poor applicability to the nonlinear change law, so it is unable to effectively deal with the large changes in leaf geometry.
Linear interpolation method is used to obtain the probability density distribution of leaf type performance, correct the design point loss model, backward angle model and positive and negative angle of attack range model in the positive problem of S2 flow surface, and combine the low-dimensional S1 flow surface and S2 flow surface calculation method instead of The three-dimensional numerical simulation process with low efficiency is sampled through the Latin supercube and full-factor experimental design method to improve calculation efficiency and accuracy.
It realizes an efficient and stable evaluation of compressor performance dispersion, which can truly reflect the dispersion of blade type performance during processing and manufacturing, improves calculation accuracy and stability, and is suitable for considering the impact of real processing geometric deviations on compressor performance.
Smart Images

Figure CN118912018A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of compressor performance evaluation, and in particular relates to a compressor performance uncertainty evaluation method taking into account blade geometry deviation. Background Art
[0002] There is always a certain geometric deviation between the actual processed blade shape and the theoretical blade shape, which will inevitably lead to a certain dispersion of blade shape performance. In the face of the uncertainty of geometric deviation, it is necessary to quickly obtain the dispersion of multi-stage compressor performance and master the performance base of the compressor under a certain process level. The existing technology mainly uses the Monte Carlo method and Latin hypercube for random sampling, combines the chaotic polynomial method with compressor numerical simulation (CFD), and uses the accelerated Monte Carlo method to quantify performance uncertainty, establish the relationship between geometric deviation and compressor performance, and obtain the dispersion of compressor performance.
[0003] Existing uncertainty assessment methods based on polynomial chaos and acceleration require completing three-dimensional numerical simulation analysis of blade shape and compressor characteristics based on the generation of random blade geometry deviations. The workload of numerical simulation for multi-stage compressor performance evaluation is huge, and the results are affected by the non-convergent numerical simulation solutions.
[0004] Although the polynomial mixed method uses a sparse pseudo-spectral approximation method, which can reduce the amount of calculation samples to a certain extent according to the sensitivity direction, the above defects cannot be completely avoided.
[0005] The uncertainty assessment method based on the adjoint method is effective for small changes in the compressor geometry, but not for large changes in the shape. At the same time, the adjoint method solution is applicable to linear or approximately linear, but has poor applicability to nonlinear change laws. Summary of the invention
[0006] The purpose of the present invention is to provide a compressor performance uncertainty evaluation method taking into account blade geometry deviations, so as to solve the problem that the compressor numerical simulation evaluation method has a huge simulation workload and the results are affected by the unconverged numerical simulation solution.
[0007] The present invention adopts the following technical solution: a compressor performance uncertainty evaluation method considering blade geometry deviation, comprising:
[0008] Step 1: According to the characteristic curves of the designed blade profile and the measured blade profile, a performance probability density distribution of the measured blade profile relative to the designed blade profile is obtained by using a linear interpolation method, wherein the performance probability density distribution is a probability density distribution of a total pressure loss coefficient deviation, a trailing angle deviation, and a positive and negative angle of attack range deviation of the measured blade profile relative to the designed blade profile;
[0009] Step 2: sampling the total pressure loss coefficient deviation, the lagging angle deviation and the positive and negative angle of attack range deviation of multiple predetermined blade height positions of a given blade according to the performance probability density distribution of the measured blade profile, and obtaining the total pressure loss coefficient change, the lagging angle change and the positive and negative angle of attack range change at different blade height positions of the full blade height;
[0010] Step 3: According to the changes in the total pressure loss coefficient, the lag angle, and the positive and negative angle of attack range, the design point loss model, lag angle model, and positive and negative angle of attack range model based on the S2 flow surface positive problem are modified, and the dispersion of the compressor performance is evaluated.
[0011] Furthermore, step 1 specifically includes:
[0012] Step 101: obtaining a designed minimum total pressure loss coefficient, a designed backward angle, a designed positive angle of attack range, and a designed negative angle of attack range of a designed blade profile;
[0013] Step 102: obtaining a measured minimum total pressure loss coefficient, a measured backward angle, a measured positive angle of attack range, and a measured negative angle of attack range of the measured blade profile;
[0014] Step 103: According to the difference between the designed minimum total pressure loss coefficient and the measured minimum total pressure loss coefficient, the difference between the designed positive angle of attack range and the measured positive angle of attack range, the difference between the designed negative angle of attack range and the measured negative angle of attack range, and the difference between the designed trailing angle and the measured trailing angle, the probability density distribution of the deviation of the total pressure loss coefficient, trailing angle and positive and negative angle of attack range of the measured blade profile relative to the designed blade profile is calculated using kernel density analysis or Gaussian distribution statistical method.
[0015] Furthermore, step 101 is specifically as follows:
[0016] A design characteristic curve of the correlation between the angle of attack of the design blade and the total pressure loss coefficient is established, and the design minimum total pressure loss coefficient and the design angle of attack and design lag angle corresponding to the design minimum total pressure loss coefficient are obtained; then the linear interpolation method is used to interpolate the design characteristic curve within a range of 2 times the design minimum total pressure loss coefficient, and the design positive angle of attack and the design negative angle of attack are obtained. The absolute values of the differences between the design positive angle of attack and the design negative angle of attack and the design minimum total pressure loss coefficient are the design positive angle of attack range and the design negative angle of attack range, respectively.
[0017] Furthermore, step 102 is specifically as follows:
[0018] A measured characteristic curve of the correlation between the angle of attack of the measured blade profile and the total pressure loss coefficient is established, and the measured minimum total pressure loss coefficient and the measured angle of attack and the measured lagging angle corresponding to the measured minimum total pressure loss coefficient are obtained; then the measured characteristic curve is interpolated using the linear interpolation method within a range of 2 times the measured minimum total pressure loss coefficient, and the measured positive angle of attack and the measured negative angle of attack are obtained. The absolute values of the differences between the measured positive angle of attack and the measured negative angle of attack and the angle of attack corresponding to the measured minimum total pressure loss coefficient are the measured positive angle of attack range and the measured negative angle of attack range, respectively.
[0019] Furthermore, the modified formula for the design point loss model based on the S2 flow surface positive problem is:
[0020] ω t =ω p +ω s +ω sec +ω e +Δω cas
[0021] In the formula, ω t is the total pressure loss coefficient, ω p is the blade loss, ω s is the shock wave loss, ω sec is the secondary flow loss, ω e is the end wall loss, Δω cas is the change in total pressure loss coefficient.
[0022] Furthermore, the correction formula for the backward angle model based on the S2 flow surface positive problem is:
[0023]
[0024] Where δ is the backward angle, δ 2D is the backward angle of the plane cascade, is the backward angle of the plane blade cascade in the minimum loss state; i is the angle of attack; i * is the angle of attack corresponding to the minimum total pressure loss coefficient, δ v is the density ratio; δ 3D is the correction of the backward angle by three-dimensional flow; Δδ cas is the variation of the backward angle.
[0025] Furthermore, the correction formula for the positive and negative angle of attack range model based on the S2 flow surface positive problem is:
[0026] Among them, the positive attack angle range is:
[0027] In the formula, width + is the positive attack angle range; Δi + is the positive attack angle range of the designed blade; is the change in the positive attack angle range;
[0028] Among them, the negative angle of attack range is:
[0029] In the formula, width - is the negative angle of attack range; Δi - is the negative angle of attack range of the designed blade; is the change in the negative angle of attack range.
[0030] Furthermore, when sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples at 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is greater than or equal to 100, use the Latin hypercube sampling method to sample and combine the samples at 3-5 leaf height positions.
[0031] Furthermore, when sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples of 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is less than 100, use the full factorial experimental design method to sample and combine the samples of 3-5 leaf height positions.
[0032] The beneficial effects of the present invention are:
[0033] The present invention combines the probability distribution of the performance dispersion of the S1 flow surface of the primitive blade profile with the two-dimensional direct problem characteristic calculation method of the compressor, and uses the low-dimensional S1 flow surface and S2 flow surface calculation method to replace the less efficient three-dimensional numerical simulation process and performance uncertainty calculation to obtain the performance dispersion of the compressor;
[0034] The basic data required by the present invention is directly derived from the manufacturing process, which more truly reflects the dispersion of blade performance; the verified compressor performance S2 direct problem calculation method has high calculation accuracy and stability, and high calculation efficiency, and can be widely used to consider the impact of real processing geometry deviation on compressor performance;
[0035] The present invention requires correction of secondary flow loss, end wall loss and radial mixing loss for compressors with full three-dimensional flow to improve the accuracy of the evaluation method, and the correction amount of the above model can be extracted from the three-dimensional simulation analysis results of a single-stage compressor with superimposed processing geometric deviations. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 is a relationship curve between the angle of attack and the total pressure loss coefficient of the present invention;
[0037] Figure 2 is a relationship curve between the attack angle and the backward angle of the present invention;
[0038] Figure 3Schematic diagram of probability density distribution of blade performance deviation of the present invention;
[0039] Figure 4 It is a statistical curve of the minimum loss of the measured blade profile of the present invention relative to the designed blade profile. DETAILED DESCRIPTION
[0040] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.
[0041] The present invention discloses a compressor performance uncertainty evaluation method considering blade geometry deviation, comprising:
[0042] Step 1: Based on the characteristic curves of the designed blade profile and the measured blade profile, the performance probability density distribution of the measured blade profile relative to the designed blade profile is obtained by using the linear interpolation method. The performance probability density distribution is the probability density distribution of the total pressure loss coefficient deviation, the trailing angle deviation, and the positive and negative angle of attack range deviation of the measured blade profile relative to the designed blade profile.
[0043] The blade profile characteristics of the S1 flow surface are obtained through blade cascade tests or numerical simulations. The performance parameters characterizing the blade profile mainly include the total pressure loss coefficient ω, the lagging angle δ, the available angle of attack range ΔI, the positive angle of attack range and the negative angle of attack range. The present invention uses numerical simulation software such as Mises program, MAPS1 or NUMECA to calculate the blade profile characteristics of the S1 flow surface.
[0044] Step 101: Obtain the designed minimum total pressure loss coefficient, designed backward angle, designed positive angle of attack range, and designed negative angle of attack range of the designed blade profile
[0045] Step 101 is specifically as follows: establishing a design characteristic curve of the correlation between the angle of attack of the designed blade and the total pressure loss coefficient, and obtaining the designed minimum total pressure loss coefficient and the designed angle of attack and the designed lag angle corresponding to the designed minimum total pressure loss coefficient; then using a linear interpolation method to interpolate the design characteristic curve within a range of 2 times the designed minimum total pressure loss coefficient, and obtaining a designed positive angle of attack and a designed negative angle of attack, then the absolute values of the differences between the designed positive angle of attack and the designed negative angle of attack and the angle of attack corresponding to the designed minimum total pressure loss coefficient are the designed positive angle of attack range and the designed negative angle of attack range, respectively, as Figure 1 and 2 shown.
[0046] Step 102: Obtain the measured minimum total pressure loss coefficient, measured backward angle, measured positive angle of attack range, and measured negative angle of attack range of the measured blade profile
[0047] Step 102 is specifically as follows: establishing a measured characteristic curve of the correlation between the angle of attack of the measured blade profile and the total pressure loss coefficient, and obtaining the measured minimum total pressure loss coefficient and the measured angle of attack and the measured lagging angle corresponding to the measured minimum total pressure loss coefficient; then using a linear interpolation method to interpolate the measured characteristic curve within a range of 2 times the measured minimum total pressure loss coefficient, and obtaining the measured positive angle of attack and the measured negative angle of attack, then the absolute values of the differences between the measured positive angle of attack and the measured negative angle of attack and the angle of attack corresponding to the measured minimum total pressure loss coefficient are the measured positive angle of attack range and the measured negative angle of attack range, respectively.
[0048] Step 103: According to the difference between the designed minimum total pressure loss coefficient and the measured minimum total pressure loss coefficient, the difference between the designed positive angle of attack range and the measured positive angle of attack range, the difference between the designed negative angle of attack range and the measured negative angle of attack range, and the difference between the designed trailing angle and the measured trailing angle, a probability density distribution of the deviation of the total pressure loss coefficient, trailing angle, and positive and negative angle of attack range of the measured blade profile relative to the designed blade profile is calculated using a kernel density analysis or a Gaussian distribution statistical method, such as Figure 3 shown.
[0049] Among them, the designed backward angle and the measured backward angle are obtained based on the curve of the correlation between the angle of attack and the backward angle.
[0050] Step 2: According to the performance probability density distribution of the measured blade type, the total pressure loss coefficient deviation, the lagging angle deviation and the positive and negative angle of attack range deviation of multiple predetermined blade height positions of a given blade are sampled to obtain the change in the total pressure loss coefficient, the lagging angle change and the positive and negative angle of attack range at different blade height positions of the full blade height.
[0051] According to the performance probability density distribution of the measured blade profile, the deviation of the total pressure loss coefficient at 3-5 blade height positions of a given blade is sampled according to the super Latin cube or Gaussian distribution. Among them, the 3-5 blade height positions are selected to include at least 10% of the blade height positions corresponding to the blade root, 50% of the blade height positions corresponding to the blade middle, and 90% of the blade height positions corresponding to the blade tip. The calculation samples of a given blade in each cross section are not less than 30 and can represent the dispersion of the manufacturing geometric deviation of the measured blade profile.
[0052] For the calculation samples of no less than 30 sections at different blade height positions mentioned above, super Latin cube design or full factorial experimental design is adopted to form a combination of the change of total pressure loss coefficient, the change of lagging angle and the change of positive and negative attack angle range as the input correction amount of the S2 flow surface forward problem model in step 3.
[0053] When sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples at 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is greater than or equal to 100, use the Latin hypercube sampling method to sample and combine the samples at 3-5 leaf height positions.
[0054] When sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples at 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is less than 100, use the full factorial experimental design method to sample and combine the samples at 3-5 leaf height positions.
[0055] Step 3: According to the changes in the total pressure loss coefficient, the lag angle, and the positive and negative angle of attack range, the design point loss model, lag angle model, and positive and negative angle of attack range model based on the S2 flow surface positive problem are modified, and the dispersion of the compressor performance is evaluated.
[0056] The S2 flow surface positive problem mainly includes the following calculation models:
[0057] Minimum loss angle of attack and angle of retreat model
[0058] The calculation method of the minimum loss angle of attack is mainly the method in NASAP-36. The minimum loss angle of attack is calculated according to different incoming flow Mach number levels. The model is:
[0059]
[0060] i * =i m2 Ma≥1.2
[0061] i * =i m2 +(i m1 -i m2 )f(Ma) 1.0<Ma<1.2
[0062] Where i* is the minimum loss angle of attack, i m1 and i m2 are the minimum loss angle of attack calculated by different methods; K i is the angle of attack model coefficient related to blade type and thickness; (i0) 10 is the reference angle of attack of the NACA-65 blade with zero camber of 10% relative thickness; n is the slope of the minimum loss angle of attack with the camber angle; θ' is the blade camber angle, is the relative position of maximum deflection; i Ma is the correction of Mach number to the minimum loss angle of attack, and f(Ma) is the Mach number correction formula.
[0063] The lagging angle model is:
[0064]
[0065] Where δ is the backward angle, δ 2D is the backward angle of the plane cascade, is the backward angle of the plane blade cascade with minimum loss, i is the angle of attack; i * is the angle of attack corresponding to the minimum total pressure loss coefficient, δ v is the density ratio; δ 3D Correction for the backward angle for three-dimensional flow.
[0066] The design point loss model is:
[0067] ω t =ω p +ω s +ω sec +ω e
[0068] In the formula, ω t is the total pressure loss coefficient, ω p is the blade loss, ω s is the shock wave loss, ω sec is the secondary flow loss, ω e is the end wall loss.
[0069] The non-design point loss model is:
[0070] ω=ω t [5-4(ξ+2)] ξ<-2
[0071] ω=ω t (1+ξ 2 ) -2≤ξ≤1
[0072] ω=2ω t ξ ξ>1
[0073] Where ξ=(ii * ) / width ± ,width ± is the range of positive and negative angles of attack.
[0074] The dispersion of blade performance caused by the geometric uncertainty of the measured blade profile is mainly manifested in the changes of the minimum total pressure loss, the backward angle and the positive and negative angle of attack range, such as Figure 4 As shown in the figure, the total pressure loss has a strong correlation with the trailing angle, that is, the larger the trailing angle, the greater the total pressure loss. According to the probability distribution of the measured blade performance dispersion, the minimum total pressure loss, trailing angle and the correction value of the positive and negative available angle of attack range are obtained by sampling.
[0075] According to the total pressure loss coefficient change Δω cas , the change of the lagging angle Δδ cas The specific correction method for the change in positive and negative attack angles is:
[0076] Design point loss model correction: ω t =ω p +ωs +ω sec +ω e +Δω cas ;
[0077] Lag angle model correction:
[0078] Positive and negative angle of attack range model correction:
[0079] Among them, the positive attack angle range is:
[0080] In the formula, width + is the positive attack angle range; Δi + is the positive attack angle range of the designed blade; is the change in the positive attack angle range;
[0081] Among them, the negative angle of attack range is:
[0082] In the formula, width - is the negative angle of attack range; Δi - is the negative angle of attack range of the designed blade; is the change in the negative angle of attack range.
[0083] Finally, the compressor performance dispersion is evaluated, specifically:
[0084] Linear interpolation is used to calculate the intersection of the flow-pressure ratio characteristics of the compressor at different speeds and the common working line to determine the flow and pressure ratio of the working point. The efficiency of the working point is then interpolated through the efficiency-pressure ratio characteristic curve. Finally, the surge margin of the compressor is calculated based on the flow, pressure ratio of the working point and the flow and pressure ratio of the surge point. The mean, standard deviation, maximum and minimum values of the flow, efficiency and surge margin of the working point at the same speed are calculated to complete the dispersion calculation of the compressor performance.
[0085] Among them, the comprehensive surge margin
[0086] Among them, π s 、m s is the pressure ratio and flow rate at the burst point, π w and m w is the pressure ratio and flow rate at the operating point.
[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A compressor performance uncertainty assessment method considering blade geometry deviation, characterized in that: include: Step 1: According to the characteristic curves of the designed blade profile and the measured blade profile, a performance probability density distribution of the measured blade profile relative to the designed blade profile is obtained by using a linear interpolation method, wherein the performance probability density distribution is a probability density distribution of a total pressure loss coefficient deviation, a trailing angle deviation, and a positive and negative angle of attack range deviation of the measured blade profile relative to the designed blade profile; Step 2: sampling the total pressure loss coefficient deviation, the lagging angle deviation and the positive and negative angle of attack range deviation of multiple predetermined blade height positions of a given blade according to the performance probability density distribution of the measured blade profile, and obtaining the total pressure loss coefficient change, the lagging angle change and the positive and negative angle of attack range change at different blade height positions of the full blade height; Step 3: According to the changes in the total pressure loss coefficient, the lag angle, and the positive and negative angle of attack range, the design point loss model, lag angle model, and positive and negative angle of attack range model based on the S2 flow surface positive problem are modified, and the dispersion of the compressor performance is evaluated.
2. A compressor performance uncertainty assessment method considering blade profile geometry deviation according to claim 1, characterized in that: Step 1 specifically includes: Step 101: obtaining a designed minimum total pressure loss coefficient, a designed backward angle, a designed positive angle of attack range, and a designed negative angle of attack range of a designed blade profile; Step 102: obtaining a measured minimum total pressure loss coefficient, a measured backward angle, a measured positive angle of attack range, and a measured negative angle of attack range of the measured blade profile; Step 103: According to the difference between the designed minimum total pressure loss coefficient and the measured minimum total pressure loss coefficient, the difference between the designed positive angle of attack range and the measured positive angle of attack range, the difference between the designed negative angle of attack range and the measured negative angle of attack range, and the difference between the designed trailing angle and the measured trailing angle, the probability density distribution of the deviation of the total pressure loss coefficient, trailing angle and positive and negative angle of attack range of the measured blade profile relative to the designed blade profile is calculated using kernel density analysis or Gaussian distribution statistical method.
3. A compressor performance uncertainty assessment method considering blade profile geometry deviation according to claim 2, characterized in that: Step 101 is specifically as follows: Establish a design characteristic curve of the correlation between the angle of attack of the designed blade and the total pressure loss coefficient, and obtain the design minimum total pressure loss coefficient and the design angle of attack and design lag angle corresponding to the design minimum total pressure loss coefficient; Then, the linear interpolation method is used to interpolate the design characteristic curve within the range of 2 times the design minimum total pressure loss coefficient, and the design positive angle of attack and the design negative angle of attack are obtained. The absolute values of the differences between the design positive angle of attack and the design negative angle of attack and the angle of attack corresponding to the design minimum total pressure loss coefficient are the design positive angle of attack range and the design negative angle of attack range, respectively.
4. A compressor performance uncertainty assessment method considering blade profile geometry deviation according to claim 3, characterized in that: Step 102 is specifically as follows: A measured characteristic curve of the correlation between the angle of attack of the measured blade profile and the total pressure loss coefficient is established, and the measured minimum total pressure loss coefficient and the measured angle of attack and the measured lagging angle corresponding to the measured minimum total pressure loss coefficient are obtained; then the measured characteristic curve is interpolated using the linear interpolation method within a range of 2 times the measured minimum total pressure loss coefficient, and the measured positive angle of attack and the measured negative angle of attack are obtained. The absolute values of the differences between the measured positive angle of attack and the measured negative angle of attack and the angle of attack corresponding to the measured minimum total pressure loss coefficient are the measured positive angle of attack range and the measured negative angle of attack range, respectively.
5. The compressor performance uncertainty assessment method considering blade profile geometry deviation according to claim 1, characterized in that: The modified formula for the design point loss model based on the S2 flow surface positive problem is: oh t =ω p +oh s +oh sec +oh e +See cas In the formula, ω t is the total pressure loss coefficient, ω p is the blade loss, ω s is the shock wave loss, ω sec is the secondary flow loss, ω e is the end wall loss, Δω cas is the change in total pressure loss coefficient.
6. The compressor performance uncertainty assessment method considering blade geometry deviation according to claim 1, characterized in that: The modified formula for the backward angle model based on the S2 flow surface positive problem is: Where δ is the backward angle, δ 2D is the backward angle of the plane cascade, is the backward angle of the plane blade cascade in the minimum loss state; i is the angle of attack; i * is the angle of attack corresponding to the minimum total pressure loss coefficient, δ v is the density ratio; δ 3D is the correction of the backward angle by three-dimensional flow; Δδ cas is the variation of the backward angle.
7. The compressor performance uncertainty assessment method considering blade geometry deviation according to claim 1, characterized in that: The correction formula for the positive and negative angle of attack range model based on the S2 flow surface positive problem is: Among them, the positive attack angle range is: In the formula, width + is the positive attack angle range; Δi + is the positive attack angle range of the designed blade; is the change in the positive attack angle range; Among them, the negative angle of attack range is: In the formula, width - is the negative angle of attack range; Δi - is the negative angle of attack range of the designed blade; is the change in the negative angle of attack range.
8. A compressor performance uncertainty assessment method considering blade profile geometry deviation according to any one of claims 1 to 7, characterized in that: When sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples at 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is greater than or equal to 100, use the Latin hypercube sampling method to sample and combine the samples at 3-5 leaf height positions.
9. A compressor performance uncertainty assessment method considering blade profile geometry deviation according to any one of claims 1 to 7, characterized in that: When sampling in step 2, first sample 3-5 leaf height positions of the entire leaf height according to the probability density distribution to obtain samples of 3-5 leaf height positions. When the sample size of a single section at a certain leaf height position is less than 100, use the full factorial experimental design method to sample and combine the samples of 3-5 leaf height positions.
Citation Information
Cited By
Compressor impeller efficiency consistency detection method
CN121345809A
