Pneumatic failure probability analysis method for front edge chipping blade
By parameterizing and generating different blade types, establishing the changes in aerodynamic characteristics and safety boundaries, and building a probability model, solving the aerodynamic failure problem caused by blocking the leading edge of the aerodynamic blade, achieving rapid assessment of the probability of aerodynamic failure, and improving the wartime situation Evaluate efficiency.
Patent Information
- Application Number
- CN202410941014.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-12
AI Technical Summary
In the service environment of aviation gas turbine engines, the aerodynamic performance caused by the loss of the leading edge of the blade is damaged, and it is difficult for the existing technology to quickly and efficiently evaluate the probability of aerodynamic failure, which affects aircraft performance and wartime emergency repair efficiency.
A method for aerodynamic failure probability analysis of the leading edge block-fall blade is proposed. By parameterizing the blade geometric characteristics, different blade types are generated, the aerodynamic characteristics change laws and safety boundaries are established, the probability function distribution model is constructed, and the aerodynamic failure probability is calculated.
It achieves rapid assessment of the probability of aerodynamic failure of the blade, improves the efficiency and accuracy of wartime damage assessment, and supports rapid emergency repair decisions.
Smart Images

Figure CN118940427A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for quickly evaluating aerodynamic failure of an aviation gas turbine engine blade under a service environment, and in particular to a method for analyzing the probability of aerodynamic failure of a blade with a leading edge falling off. Background Art
[0002] Blade leading edge fragmentation is a common form of compressor blade failure. Aviation gas turbine engines, especially military aviation gas turbine engines, will inevitably inhale hard objects such as sand, gravel, and birds in the service environment, which will then damage the high-speed rotating compressor blades. Hard objects damaging the compressor blades will not only affect the structural strength of the blades, but also affect the aerodynamic characteristics of the compressor, causing the performance of the aircraft engine to decline. Damage assessment is an important basic data source required for aircraft combat damage repair strategies and methods. On-site assessment is also called post-damage assessment. As an integral part of the combat damage repair process, time is a key limiting condition for on-site assessment. In real combat scenarios, aircraft encounter various types of threats, varying degrees of damage, and many types of combat damage. The time for combat damage assessment depends on the number of combat damage. In order to ensure that the on-site assessment meets the time constraint, the assessment should be fast and efficient.
[0003] The cost of compressor aerodynamic characteristics experiments is high, and numerical simulation analysis has the problems of long calculation cycle and large amount of calculation, which makes it difficult to directly apply it to the analysis of the impact of leading edge block loss on compressor aerodynamic characteristics. Therefore, the development of a method for analyzing the probability of aerodynamic failure of leading edge block loss blades is of great significance for quickly assessing whether aerodynamic failure occurs in aeroengine blades under service environment. Summary of the invention
[0004] In view of the problem that the aerodynamic performance of the compressor is impaired due to the falling of the leading edge of the blade, the present invention proposes a method for analyzing the aerodynamic failure probability of the blade due to the falling of the leading edge, which specifically includes the following steps:
[0005] Step 1: Parametrically describe the geometric features of the blade with leading edge block loss;
[0006] The details are as follows:
[0007] Step 1: Blade profile generation under compressor stage environment;
[0008] A spatial rectangular coordinate system is established, with the z-axis being the rotation axis of the compressor, and a rotation surface S2 being generated with the z-axis as the center; the cross-sectional curve formed by the intersection of the rotation surface S2 and the compressor blade is the compressor blade profile, and the compressor blade profiles at different blade height positions can be obtained by changing the values of the blade inlet radius r1 and the blade outlet radius r2;
[0009] Step 2: Parameterization of the geometric features of the falling blade;
[0010] For the leading edge of the blade tip that is damaged by a hard object, the scale of the falling block along the blade height direction Δr is first determined, and the compressor blade profiles at different blade heights are extracted within the range of Δr. For each compressor blade profile, there is a leading edge falling block area. The starting point coordinates x along the blade axis are used to calculate the starting point coordinates x of the blade suction surface and pressure surface. s and x p The angle between the block section and the mid-arc line is used to parameterize the block area at the leading edge of the blade in the coordinate system. The axial direction is the x-axis. For a normal compressor blade, the starting points of the suction and pressure surfaces of the blade along the x-axis are both at the leading edge, and the corresponding starting point coordinates are x 0 , x 0 represents the position of the leading edge of the blade in the coordinate system. The end points of the suction and pressure surfaces of the blade along the x-axis are both at the trailing edge, and the corresponding end point coordinates are x e , from x 0 to x e The distance along the x-axis is recorded as the total axial chord length Ca; when the leading edge falls off, the coordinates of the starting point of the suction surface along the x-axis become x s , the starting point coordinate of the pressure surface along the x-axis becomes x p , x 0 to x s The distance along the x-axis divided by Ca is recorded as the ratio of the axial dimension of the damaged part of the suction surface to the total axial chord length of the blade Sx, x 0 to x p The distance along the x-axis direction divided by Ca is recorded as the ratio Px of the axial dimension of the damaged part of the pressure surface to the total axial chord length of the blade. The geometric characteristics of the leading edge block area are characterized by parameters Sx and Px.
[0011] Step 2: Generate geometric data of the blade with leading edge block loss;
[0012] For the parameter x in step 1 s and x p , based on the data of compressor blades being damaged by hard objects during the service life of aviation gas turbine engines, the maximum value x is estimated sm and x pm ; When the leading edge falls off, the blade geometry changes are unpredictable and highly random. Therefore, at (0, x sm ) and (0, x pm ) extract N values from the two intervals, that is, x si and x pi , i = 1 ~ N; x si and x pi The combination can obtain N sets of front edge block geometric features; for any set of x si and x pi First, the original compressor blade profile is generated by using the compressor blade geometry aerodynamic design method; then,si and x pi Draw a straight line between the two points on the corresponding pressure surface and suction surface, and use the straight line to truncate the original blade profile. The second half of the blade profile after truncation and the straight line together form a leading edge block blade profile. A total of N leading edge block blade profiles can be generated based on the original blade profile, where the second half of the blade profile refers to the blade profile with x greater than x si and x pi The front half is the broken part; for the broken blade, the revolving surface S1 at different blade heights is made within the range of Δr where the leading edge broken occurs to intersect with the broken blade to obtain different blade geometry. Assuming that a total of m compressor blade profiles are extracted, each blade profile can generate N leading edge broken blade profiles using the above method;
[0013] Step 3: Establish the correlation law and safety margin between the geometric characteristics of the leading edge falling blade and the change of the aerodynamic characteristics of the compressor blade profile;
[0014] The details are as follows:
[0015] Step 1: Establish the correlation between the geometric characteristics of the leading edge falling blade and the changes in the aerodynamic characteristics of the compressor blade profile;
[0016] For the kth compressor blade profile among the m compressor blade profiles in step 2, the aerodynamic parameters of the original blade profile and N leading edge chipped blade profiles are obtained by computational fluid dynamics or experimental testing, and the change of the total pressure loss of the compressor blade profile under the influence of the leading edge chipping is obtained; for a given kth compressor blade profile, N compressor blade profiles with leading edge chipping are generated by step 2, and the total pressure loss of the blade profile at the same incoming flow Mach number and inlet angle of attack is calculated for the original blade profile and the leading edge chipped blade profile respectively; for the original blade profile of the kth compressor blade profile, its corresponding total pressure loss is ωk; for the i-th leading edge chipped blade profile of the kth compressor blade profile, its corresponding total pressure loss is ωk i (k = 1 ~ m, i = 1 ~ N); by calculating the ωk of N leading edge block blades i , the correlation cloud diagram between the geometric characteristics of the leading edge falling blade and the changes in the aerodynamic characteristics of the compressor blade profile can be obtained;
[0017] Step 2: Define the security boundary;
[0018] Using the above correlation cloud diagram, the safety boundary of the leading edge block blade availability is determined from the perspective of the increase in total pressure loss. Referring to the definition of the compressor blade incident angle range, when ωk i When it is greater than n times of ωk, the blade is defined as aerodynamically failed, and the corresponding contour line is found in the associated cloud map. The leading edge block-dropping blade type corresponding to the geometric characteristic parameters inside the contour line is the aerodynamically failed blade type;
[0019] Step 4: Construct a probability function distribution model and calculate the failure probability;
[0020] The details are as follows:
[0021] Step 1: Construct a probability function distribution model;
[0022] The probability distribution function of the size of the pieces falling off the suction side and the pressure side is established; the distribution of Sx and Px should meet the following requirements: (1) For the same compressor blade, Sx and Px should have a certain correlation, rather than being independent of each other; (2) The values of Sx and Px can only take positive values and vary in the range of (0, 1); (3) Although theoretically, the values of Sx and Px can take 1, that is, the whole blade is broken due to a hard object, but this probability should be extremely small; According to the above three criteria, the weighted binary exponential distribution is selected to describe the probability distribution of Sx and Px:
[0023]
[0024] Among them, λ 1 , 2 and λ 3 are adjustable constants, mainly used to modify the characteristics of probability distribution functions, and λ = λ 1 +λ 2 +λ 3 , where λ is the value that depends on λ 1 , 2 and λ 3 constant; for the joint probability density function, the expectation, variance and correlation coefficient of Sx and Px are derived:
[0025]
[0026] Where, E(Sx) and E(Px) are the mathematical expectations of Sx and Px, respectively, and V(Sx) and V(Sx) are the variances of Sx and Px, respectively;
[0027] The distribution characteristics of Sx and Px are determined by three adjustable parameters; the correlation coefficient between the two changes with the parameter λ 3 decreases monotonically with the decrease of ;
[0028] When Sx=Px=x0, the probability density function reaches an extreme value, at this time:
[0029]
[0030] Where x0 is the value of Sx and Px when the probability density function reaches an extreme value;
[0031] Step 2: Calculate the failure probability;
[0032] After fitting the safety boundary to the fitted safety boundary using the fitting function, the ray intersection method is used to determine whether the sample point is inside the safety boundary. When the point is inside the safety boundary, it is defined as a failure point, and the number of failed sample points m is obtained by statistics; given the total number of samples n, the failure probability P = (m / n) × 100% is calculated; assuming that the failure probability is q%, when ((1 / q) × 100) leading edge blades appear, it is very likely that one of the leading edge blades has failed aerodynamically.
[0033] In a specific embodiment of the present invention, in step 2, x sm and x pm The ratio to the axial chord length of the blade does not exceed 12%, that is, Sx and Px do not exceed 0.12.
[0034] In one embodiment of the present invention, in step 2, the compressor blade geometry aerodynamic design method adopts the mid-arc line modeling method.
[0035] In another embodiment of the present invention, in step 2, N is not less than 30, and m is not less than 3.
[0036] In another specific embodiment of the present invention, in step 4, Step 2, an elliptic function is used to fit the safety boundary.
[0037] The present invention specifically relates to a parameterized characterization method for the geometric characteristics of blades with leading edge debris and a probability analysis method for aerodynamic failure of blades with leading edge debris. By correlating the aerodynamic characteristics of a compressor with the geometric parameter characteristics of blades with leading edge debris, the correlation law between the geometric characteristics of blades with leading edge debris and changes in the aerodynamic characteristics of the compressor is obtained, a probability function distribution model is constructed and a safety boundary is established, so as to achieve a rapid assessment of whether aerodynamic failure occurs in blades with leading edge debris. The present invention has broad application prospects in the post-damage assessment link of emergency repairs during wartime. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is the geometric characteristics and parameterized characterization method of the compressor leading edge block-off blade;
[0039] Figure 2 It is the correlation cloud diagram between the change of compressor aerodynamic characteristics and the geometric characteristics of the compressor leading edge blade loss;
[0040] Figure 3 It is the probability density function distribution graph. DETAILED DESCRIPTION
[0041] The present invention will be further described below in conjunction with the accompanying drawings.
[0042] The present invention provides a method for analyzing the probability of aerodynamic failure of a blade with a leading edge falling off, which specifically comprises the following steps.
[0043] Step 1: Parametrically describe the geometric features of the blade with leading edge block loss;
[0044] According to the position of the leading edge debris in the blade height direction, the compressor blade geometry at the corresponding position is extracted, and the geometric characteristics of the blade with the leading edge debris (size and shape) are associated with the aerodynamic design parameters of the compressor blade (chord length, thickness, bend angle, etc.). The changes in the aerodynamic design parameters of the compressor blade are used to characterize the geometric characteristics of the blade damaged by hard objects.
[0045] The specific steps include:
[0046] Step 1: Blade profile generation under compressor stage environment;
[0047] like Figure 1 As shown, a spatial rectangular coordinate system is established, the z-axis is the rotation axis of the compressor, and the rotation surface S2 is generated with the z-axis as the center. According to the actual geometry of the compressor blades, the distance between the rotation surface S2 and the z-axis may vary, that is, Figure 1 The values of r1 and r2 in may be different, where r1 and r2 represent the radius of the blade inlet and outlet, i.e., the absolute position of the blade. Due to the curved, twisted and swept shape of the compressor blade, the blade height at the inlet and outlet may not be the same, so the absolute positions r1 and r2 corresponding to the same relative blade height at the inlet and outlet may not be the same. Therefore, the above-mentioned "rotation surface S2" is to select the corresponding r1 and r2 after determining the same relative blade height at the inlet and outlet, such as Figure 2 As shown, Figure 2 The ends of r1 and r2 are far from the axis (i.e. Figure 2 The method of generating the surface of revolution S2 is well known in the art and will not be described again. The cross-sectional curve formed by the intersection of the surface of revolution S2 and the compressor blade is the compressor blade profile. By changing the values of r1 and r2, the compressor blade profiles at different blade height positions can be obtained.
[0048] Step 2: Parameterization of the geometric features of the falling blade;
[0049] like Figure 1 As shown in the figure, for the tip leading edge chipping caused by hard object damage, the scale Δr of the chipping along the blade height direction is first determined, and the compressor blade profiles at different blade heights are extracted within the range of Δr. For each compressor blade profile, there is a leading edge chipping area. The present invention uses the starting point coordinates of the blade suction surface and pressure surface profile along the blade axial direction (x axis) (as shown in the figure x s and x p ) and the angle between the block section and the mid-arc line are used to parameterize the block area at the leading edge of the blade in the coordinate system. For a normal compressor blade, the starting points of the suction and pressure surfaces along the x-axis are both at the leading edge, and the corresponding starting point coordinates are x0 (x 0 represents the position of the leading edge point of the blade in the coordinate system. The leading edge point has no necessary connection with the origin of the coordinate system. The end points of the suction and pressure surfaces of the blade along the x-axis are both at the trailing edge point, and the corresponding end point coordinates are x e , from x 0 to x e The distance along the x-axis is recorded as the total axial chord length Ca; when the leading edge falls off, the coordinates of the starting point of the suction surface along the x-axis become x s , the starting point coordinate of the pressure surface along the x-axis becomes x p , x 0 to x s The distance along the x-axis divided by Ca is recorded as the ratio of the axial dimension of the damaged part of the suction surface to the total axial chord length of the blade Sx, x 0 to x p The distance along the x-axis divided by Ca is recorded as the ratio Px of the axial scale of the damaged part of the pressure surface to the total axial chord length of the blade. The geometric characteristics of the leading edge chip area can be characterized by parameters Sx and Px.
[0050] Step 2: Generate geometric data of the blade with leading edge block loss;
[0051] According to the changes in aerodynamic design parameters of blades with different leading edge block dropouts, the compressor blade profile is redesigned and shaped to generate data for blades with different leading edge block dropouts.
[0052] For the parameter x in step 1 s and x p The maximum value x can be estimated based on the data of compressor blades being damaged by hard objects during the service of aviation gas turbine engines. sm and x pm , according to existing experience, x sm and x pm The ratio of the blade axial chord length will not exceed 12%, that is, Sx and Px will not exceed 0.12. When the leading edge block occurs, the blade geometry change is unpredictable and has a strong randomness. Therefore, at (0, x sm ) and (0, x pm ) extract N values from the two intervals, that is, x si and x pi , i = 1 to N. si and x pi We can obtain N sets of front edge block geometric features by combining them. si and x pi First, the original compressor blade profile is generated by using a compressor blade geometry aerodynamic design method, such as a mid-arc line modeling method (this method is well known to those skilled in the art); then, si and x piDraw a straight line between the two points on the corresponding pressure surface and suction surface, and use the straight line to cut off the original blade shape. The latter half of the blade shape after cutting off (the latter half of the blade shape refers to the point where x is greater than x si and x pi The front half of the blade is the broken part) and the straight line together form a leading edge broken blade profile, and a total of N leading edge broken blade profiles can be generated on the basis of the original blade profile. For the broken blade, a revolving surface S1 at different blade heights is made within the range of Δr where the leading edge broken occurs to intersect with the broken blade (the blade height within the range of Δr can be taken according to the method in step 1), and different blade geometry can be obtained. Assuming that a total of m compressor blade profiles are extracted, each blade profile can generate N leading edge broken blade profiles using the above method. Here N is generally not less than 30, and m is generally not less than 3.
[0053] Step 3: Establish the correlation law and safety margin between the geometric characteristics of the leading edge falling blade and the change of the aerodynamic characteristics of the compressor blade profile;
[0054] For the compressor blade profile affected by leading edge chipping, computational fluid dynamics or experimental tests are used to obtain the blade profile loss before and after the leading edge chipping occurs. The changes in blade profile loss before and after the leading edge chipping occur are compared, and the correlation law between the geometric characteristics of the blade with leading edge chipping and the changes in the aerodynamic characteristics of the compressor blade profile is established. The change law of the aerodynamic characteristics of the compressor blade profile with the geometric characteristics of the blade with leading edge chipping is obtained.
[0055] Step 1: Establish the correlation between the geometric characteristics of the leading edge falling blade and the changes in the aerodynamic characteristics of the compressor blade profile;
[0056] For the kth compressor blade profile among the m compressor blade profiles in step 2, the aerodynamic parameters of the original blade profile and N blade profiles with leading edge block loss can be obtained by computational fluid dynamics or experimental testing (this method is known to those skilled in the art and will not be described again), and the change in the total pressure loss of the compressor blade profile under the influence of leading edge block loss (i.e., the change before and after the block loss) is obtained. The definition of the total pressure loss of the blade profile is well known in the art, and the change in the total pressure loss of the compressor blade profile under the influence of leading edge block loss obtained based on the aerodynamic parameters is known to those skilled in the art. For a given kth compressor blade profile, N compressor blade profiles with leading edge block loss are generated by step 2, and the total pressure loss of the blade profile at the same incoming flow Mach number and inlet angle of attack is calculated for the original blade profile and the blade profile with leading edge block loss (this method is known to those skilled in the art and will not be described again), and the definition and calculation of the incoming flow Mach number and inlet angle of attack are well known in the art. For the original blade of the k-th compressor blade, the corresponding total pressure loss is ωk; for the i-th leading edge block blade of the k-th compressor blade, the corresponding total pressure loss is ωk i (k = 1 ~ m, i = 1 ~ N). By calculating the ωk of N leading edge block loss blades i , we can get Figure 2The cloud diagram shown is the correlation between the geometric characteristics of the leading edge block-dropping blade and the changes in the aerodynamic characteristics of the compressor blade profile.
[0057] Step 2: Define the security boundary;
[0058] use Figure 2 , from the perspective of the increase in total pressure loss, the critical boundary of the availability of the leading edge block blade, that is, the safety boundary, is determined, referring to the definition of the compressor blade incidence angle range (Goodhand, MN, Miller, RJ, Lung, HW, 2015, "The Impact of Geometric Variation on Compressor Two-Dimensional Incidence Range," Journal of Turbomachinery, 137(2), p.021007.), when ωk i When it is greater than n times of ωk, the blade is defined as aerodynamically failed. In this paper, taking n = 1.5 as an example, the value of the safety margin is ω = 1.5 × ωk. Find the corresponding contour line in the cloud map ( Figure 2 The double-dotted line in the middle), the leading edge block-dropping blade shape corresponding to the geometric characteristic parameters inside the contour line is the aerodynamic failure blade shape.
[0059] Step 4: Construct a probability function distribution model and calculate the failure probability;
[0060] The probability distribution of the geometric characteristics of the leading edge chipped blade profile follows a certain rule. By constructing a probability function distribution model to describe the distribution of the geometric characteristics of the leading edge chipped blade profile, combined with the safety margin given in step three, the aerodynamic failure probability of the leading edge chipped blade is calculated.
[0061] Step 1: Construct a probability function distribution model;
[0062] In order to analyze the probability of blade aerodynamic failure caused by a hard object hitting the leading edge and causing a piece to fall off, the probability distribution function of the size of the piece falling off the suction side and the pressure side is first established. The distribution of Sx and Px should meet the following requirements: (1) For the same compressor blade, Sx and Px should have a certain correlation, rather than being independent of each other; (2) The values of Sx and Px can only take positive values and vary in the range of (0, 1); (3) Although theoretically, the values of Sx and Px can be 1, that is, the whole blade is broken due to a hard object hitting, but this probability should be extremely small. According to the above three criteria, the weighted binary exponential distribution is selected here to describe the probability distribution satisfied by Sx and Px:
[0063]
[0064] Among them, λ 1 ,2 and λ 3 are adjustable constants, mainly used to modify the characteristics of probability distribution functions, and λ = λ 1 +λ 2 +λ 3 , where λ is the value that depends on λ 1 , 2 and λ 3 For this joint probability density function, the expectation, variance and correlation coefficient of Sx and Px can be derived:
[0065]
[0066] Where E(Sx) and E(Px) are the mathematical expectations of Sx and Px, respectively, and V(Sx) and V(Sx) are the variances of Sx and Px, respectively.
[0067] It can be found that the distribution characteristics of Sx and Px are determined by three adjustable parameters. The correlation coefficient between the two increases with the parameter λ 3 decreases monotonically with the decrease of .
[0068] For the above joint probability density function, we also want to know when Sx and Px have the highest probability. It can be found by derivation that when Sx=Px=x0, the probability density function will reach an extreme value. At this time:
[0069]
[0070] Where x0 is the value of Sx and Px when the probability density function reaches its extreme value.
[0071] Combined with the statistical data of blade leading edge fragments caused by hard objects hitting the blade, λ 1 , 2 and λ 3 The values of and the corresponding probability distribution characteristic parameters are shown in the following table:
[0072] λ1 λ2 λ3 x0 E(Sx) E(Px) <![CDATA[ρ sp ]]> 50 50 1 0.01 0.0299 0.0299 0.1968
[0073] According to the above table, λ 1 ,λ 2 The values of are the same, so the conditional probability distributions of Sx and Px are symmetrical, which is manifested by the same expected values. The correlation coefficient between the two is 0.1968. It should be pointed out that according to the probability density function, when the conditional probability distributions of Sx and Px are symmetrical, in λ 3 When it approaches 0, the maximum correlation coefficient is 0.2. x0 = 0.01 means that when a hard object is hit, the size of the blade leading edge fragments is the maximum probability of 1% of the blade axial chord length. Based on the parameters in the above table, the probability density distribution diagram is as follows Figure 3 shown.
[0074] Step 2: Calculate the failure probability;
[0075] According to the probability density distribution function in Step 1, samples are drawn within the range of Sx, Px∈(0,1). The safety boundary in Step 3 and the drawn samples are combined to calculate the blade failure probability, as described below. First, the ray method is used to determine whether the sample point is inside the safety boundary (Chen Ruiqing, Zhou Jian, Yu Lie. A fast algorithm for determining the relationship between points and polygons [J]. Journal of Xi'an Jiaotong University, 2007, 41(1):5.). The principle of this method is to draw a ray through the judgment point P and calculate the number of intersections between the ray and the safety boundary. If the number of intersections is an even number, the point is outside the safety boundary; if it is an odd number, the point is inside the safety boundary. At the same time, the solution to the situation where point P does not intersect with the boundary line of the safety boundary and point P falls on the boundary line is as follows: if point P does not intersect with the boundary line of the safety boundary, point P is outside the safety boundary; when the point just falls on the edge line of the safety boundary, it is no longer calculated whether it intersects, and it is directly treated as a point on the edge line.
[0076] Use a suitable function to fit the safety margin. Figure 2 Take the safety boundary of as an example, use the elliptic function to fit, get the fitted safety boundary, and use the ray intersection method to determine whether the sample point is inside the safety boundary. When the point is inside the safety boundary, it is defined as a failure point, and the number of failed sample points m is obtained by statistics. Given the total number of samples n, the failure probability P = (m / n) × 100% is calculated. Assuming the failure probability is q%, when there are ((1 / q) × 100) leading edge blades falling off, it is very likely that one of the leading edge blades has already failed aerodynamically.
[0077] Specific implementation cases
[0078] Step 1: Parametrically describe the geometric features of the blade with leading edge block loss;
[0079] A 1.5-stage compressor was selected, and the blade profile and main aerodynamic design parameters of its stator blade at 50% blade height were obtained, as shown in Table 1.
[0080] Table 1 Geometry and aerodynamic parameters
[0081]
[0082]
[0083] Step 2: Generate geometric data of the blade with leading edge block loss;
[0084] Referring to the actual falling of compressor blades under service environment, the proportion of the flow dimension of the damaged parts of the suction side and the pressure side to the total flow chord length of the blade generally does not exceed 12%. Therefore, Sx and Px are set to increase by 2% every time from 0%, reaching a maximum of 12%, with a total of 49 blade profiles.
[0085] Step 3: Establish the correlation law and safety margin between the geometric characteristics of the leading edge falling blade and the change of the aerodynamic characteristics of the compressor blade profile;
[0086] Calculation of the loss coefficient ωk of 49 leading edge block loss blade profiles i (i=1~49), we get Figure 2 The correlation cloud diagram of the geometric characteristics of the leading edge block blade and the change of the aerodynamic characteristics of the compressor blade profile is shown. The loss coefficient of the calculated reference blade profile is ωk=0.0241. When ωk i When it is greater than n times of ωk, the blade is defined as aerodynamically failed. Taking n = 1.5 as an example, the value of the safety margin is ω = 1.5 × 0.0241 = 0.0362, and the corresponding contour is obtained as follows: Figure 2 Indicated by double-dotted lines.
[0087] Step 4: Build a probability function distribution model and calculate the failure probability
[0088] Combining the probability distribution function and the statistical data of the blade leading edge falling off caused by hard objects hitting the blade, λ 1 , 2 and λ 3 The values of and the corresponding probability distribution characteristic parameters are shown in the following table:
[0089] λ1 λ2 λ3 x0 E(Sx) E(Px) <![CDATA[ρ sp ]]> 50 50 1 0.01 0.0299 0.0299 0.1968
[0090] Based on the parameters in the above table, the probability density distribution diagram is as follows Figure 3 shown.
[0091] Use a suitable function to fit the safety margin. Figure 2 Taking the safety boundary as an example, the elliptic function is used for fitting, and the expression of the fitting curve is:
[0092] 4.12S x 2 -6.02S x P x +1.55P x 2 +0.0049=0
[0093] According to the probability distribution function, the total number of sample points n = 108080 is extracted. According to the ray method, the number of failure sample points m = 390 is statistically obtained, and the failure probability is calculated.
[0094]
[0095] The failure probability is 0.361%. When (1 / 0.361)×100=277 blades with leading edge fragments appear, it is considered that one blade with leading edge fragments has aerodynamically failed.
Claims
1. A method for analyzing the probability of aerodynamic failure of a blade with a leading edge falling off, characterized in that: Specifically include the following steps: Step 1: Parametrically describe the geometric features of the blade with leading edge block loss; The details are as follows: Step 1: Blade profile generation under compressor stage environment; A spatial rectangular coordinate system is established, with the z-axis being the rotation axis of the compressor, and a rotation surface S2 being generated with the z-axis as the center; the cross-sectional curve formed by the intersection of the rotation surface S2 and the compressor blade is the compressor blade profile, and the compressor blade profiles at different blade height positions can be obtained by changing the values of the blade inlet radius r1 and the blade outlet radius r2; Step 2: Parameterization of the geometric features of the falling blade; For the leading edge of the blade tip that is damaged by a hard object, the scale of the falling block along the blade height direction Δr is first determined, and the compressor blade profiles at different blade heights are extracted within the range of Δr. For each compressor blade profile, there is a leading edge falling block area. The starting point coordinates x along the blade axis are used to calculate the starting point coordinates x of the blade suction surface and pressure surface. s and x p The angle between the block section and the mid-arc line is used to parameterize the block drop area at the leading edge of the blade in the coordinate system. The axial direction is the x-axis. For a normal compressor blade, the starting points of the suction surface and pressure surface of the blade along the x-axis are both at the leading edge point, and the corresponding starting point coordinates are x0, which represents the position of the leading edge point of the blade in the coordinate system. The ending points of the suction surface and pressure surface of the blade along the x-axis are both at the trailing edge point, and the corresponding ending point coordinates are x e , from x0 to x e The distance along the x-axis is recorded as the total axial chord length Ca; when the leading edge falls off, the coordinates of the starting point of the suction surface along the x-axis become x s , the starting point coordinate of the pressure surface along the x-axis becomes x p , x0 to x s The distance along the x-axis divided by Ca is recorded as the ratio of the axial dimension of the damaged part of the suction surface to the total axial chord length of the blade Sx, x0 to x p The distance along the x-axis direction divided by Ca is recorded as the ratio Px of the axial dimension of the damaged part of the pressure surface to the total axial chord length of the blade. The geometric characteristics of the leading edge block area are characterized by parameters Sx and Px. Step 2: Generate geometric data of the blade with leading edge block loss; For the parameter x in step 1 s and x p , based on the data of compressor blades being damaged by hard objects during the service life of aviation gas turbine engines, the maximum value x is estimated sm and x pm ; When the leading edge falls off, the blade geometry changes are unpredictable and highly random. Therefore, at (0, x sm ) and (0, x pm ) extract N values from the two intervals, that is, x si and x pi , i = 1 ~ N; x si and x pi The combination can obtain N sets of front edge block geometric features; Any set of x si and x pi First, the original compressor blade profile is generated by using the compressor blade geometry aerodynamic design method; then, si and x pi Draw a straight line between the two points on the corresponding pressure surface and suction surface, and use the straight line to truncate the original blade profile. The second half of the blade profile after truncation and the straight line together form a leading edge block blade profile. A total of N leading edge block blade profiles can be generated based on the original blade profile, where the second half of the blade profile refers to the blade profile with x greater than x si and x pi The front half is the broken part; for the broken blade, the revolving surface S1 at different blade heights is made within the range of Δr where the leading edge broken occurs to intersect with the broken blade to obtain different blade geometry. Assuming that a total of m compressor blade profiles are extracted, each blade profile can generate N leading edge broken blade profiles using the above method; Step 3: Establish the correlation law and safety margin between the geometric characteristics of the leading edge falling blade and the change of the aerodynamic characteristics of the compressor blade profile; The details are as follows: Step 1: Establish the correlation between the geometric characteristics of the leading edge falling blade and the changes in the aerodynamic characteristics of the compressor blade profile; For the kth compressor blade profile among the m compressor blade profiles in step 2, the aerodynamic parameters of the original blade profile and N leading edge chipped blade profiles are obtained by computational fluid dynamics or experimental testing, and the change of the total pressure loss of the compressor blade profile under the influence of the leading edge chipping is obtained; for a given kth compressor blade profile, N compressor blade profiles with leading edge chipping are generated by step 2, and the total pressure loss of the blade profile at the same incoming flow Mach number and inlet angle of attack is calculated for the original blade profile and the leading edge chipped blade profile respectively; for the original blade profile of the kth compressor blade profile, its corresponding total pressure loss is ωk; for the i-th leading edge chipped blade profile of the kth compressor blade profile, its corresponding total pressure loss is ωk i (k = 1 ~ m, i = 1 ~ N); by calculating the ωk of N leading edge block blades i , the correlation cloud diagram between the geometric characteristics of the leading edge falling blade and the changes in the aerodynamic characteristics of the compressor blade profile can be obtained; Step 2: Define the security boundary; Using the above correlation cloud diagram, the safety boundary of the leading edge block blade availability is determined from the perspective of the increase in total pressure loss. Referring to the definition of the compressor blade incident angle range, when ωk i When it is greater than n times of ωk, the blade is defined as aerodynamically failed, and the corresponding contour line is found in the associated cloud map. The leading edge block-dropping blade type corresponding to the geometric characteristic parameters inside the contour line is the aerodynamically failed blade type; Step 4: Construct a probability function distribution model and calculate the failure probability; The details are as follows: Step 1: Construct a probability function distribution model; The probability distribution function of the size of the pieces falling off the suction side and the pressure side is established; the distribution of Sx and Px should meet the following requirements: (1) For the same compressor blade, Sx and Px should have a certain correlation, rather than being independent of each other; (2) The values of Sx and Px can only take positive values and vary in the range of (0, 1); (3) Although theoretically, the values of Sx and Px can take 1, that is, the whole blade is broken due to a hard object, but this probability should be extremely small; According to the above three criteria, the weighted binary exponential distribution is selected to describe the probability distribution of Sx and Px: Among them, λ1, λ2 and λ3 are all adjustable constants, which are mainly used to modify the characteristics of the probability distribution function, and λ = λ1 + λ2 + λ3, where λ is a constant whose value depends on λ1, λ2 and λ3; for the joint probability density function, the expectation and variance of Sx and Px and the correlation coefficient between the two are derived: Where, E(Sx) and E(Px) are the mathematical expectations of Sx and Px, respectively, and V(Sx) and V(Sx) are the variances of Sx and Px, respectively; The distribution characteristics of Sx and Px are determined by three adjustable parameters; the correlation coefficient between the two decreases monotonically as the parameter λ3 decreases; When Sx=Px=x0, the probability density function reaches an extreme value, at this time: Where x0 is the value of Sx and Px when the probability density function reaches an extreme value; Step 2: Calculate the failure probability; After fitting the safety boundary to the fitted safety boundary using the fitting function, the ray intersection method is used to determine whether the sample point is inside the safety boundary. When the point is inside the safety boundary, it is defined as a failure point, and the number of failure sample points m is obtained by counting; Given the total number of samples n, the failure probability is calculated to be P = (m / n) × 100%. Assuming the failure probability is q%, when ((1 / q) × 100) leading edge blades are lost, it is very likely that one of the leading edge blades has failed aerodynamically.
2. The aerodynamic failure probability analysis method for a leading edge chipped blade according to claim 1, characterized in that: In step 2, x sm and x pm The ratio to the axial chord length of the blade does not exceed 12%, that is, Sx and Px do not exceed 0.
12.
3. The aerodynamic failure probability analysis method for a leading edge chipped blade according to claim 1, characterized in that: In step 2, the aerodynamic design method of the compressor blade geometry surface adopts the mid-arc line modeling method.
4. The aerodynamic failure probability analysis method for a leading edge chipped blade according to claim 1, characterized in that: In step 2, N is not less than 30, and m is not less than 3.
5. The aerodynamic failure probability analysis method for a leading edge chipped blade according to claim 1, characterized in that: In step 4, Step 2, an elliptic function is used to fit the safety boundary.
