A method for detecting consistency of compressor impeller efficiency

Abstract

The application discloses a kind of compressor impeller efficiency consistency detection methods, comprising: step 1, sample pretreatment and point cloud scanning;Step 2: color space conversion, RGB color space is converted into HSV color space and utilizes the characteristics of chroma H dimension;Step 3, deviation parameter calculation, by calculating average deviation quantifies the overall level of leaf shape deviation, calculates maximum deviation quantifies the extreme degree of leaf shape deviation, calculates large deviation ratio quantifies the sensitive area distribution density of leaf shape deviation;Step 4, efficiency deviation prediction, by establishing the quantitative correlation model of leaf shape deviation and efficiency deviation, converts geometric deviation data into intuitive performance deviation result;Step 5, performance consistency determination, according to efficiency deviation result determination. By establishing the quantitative correlation model of leaf shape deviation and efficiency deviation, combined with color space conversion realizes deviation accurate extraction, finally realizes the performance consistency detection of high efficiency, high precision.

Description

Technical Field

[0001] This invention belongs to the field of testing technology for internal combustion engine turbocharger components, and particularly relates to a method for testing the performance consistency of compressor impellers, which is suitable for fast-paced and high-precision testing of compressor impeller performance consistency in mass production scenarios. Background Technology

[0002] The compressor impeller is the core moving component of an internal combustion engine turbocharger. Its blade profile accuracy directly determines key performance indicators such as post-compression pressure and efficiency, thereby affecting engine power, economy, and emissions compliance. Currently, customers have stringent requirements for turbocharger performance consistency, such as controlling post-compression pressure variation within ΔP=±3kPa and strictly limiting compressor efficiency deviation. Simultaneously, emissions regulations are placing increasing demands on the sensitivity of turbocharger performance, requiring that performance fluctuations of each impeller be within a controllable range.

[0003] Current methods for testing the performance consistency of compressor impellers mainly rely on three types of methods, but all of them have significant drawbacks:

[0004] 1. Experimental testing method: Pressure ratio and efficiency deviation are obtained through bench testing of the booster. However, the testing accuracy is limited by the bench equipment (e.g., the test deviation of the Kaze test bench is ≥ ±0.55%), which cannot meet the theoretical efficiency deviation requirement of ±0.5%. Moreover, bench testing is costly and time-consuming (testing a single unit takes ≥ 2 hours), making it difficult to adapt to the needs of mass production.

[0005] 2. Simulation calculation method: The performance is predicted by using the process of "point cloud scanning-reverse modeling-CFD calculation". However, the reverse modeling process will generate deviation amplification. The accumulation of reverse deviation will lead to the deviation between the CFD calculation results and the actual performance, resulting in low simulation credibility.

[0006] 3. Blade shape detection method: Based on geometric deviation standards (such as the maximum blade shape deviation of cast impeller ±0.2mm), but this standard only focuses on the maximum blade shape deviation value and does not consider the influence of deviation location (such as deviations at the leading edge and trailing edge, which will increase boundary layer separation loss) and deviation distribution on performance. There are cases of "geometric compliance but performance exceeding the tolerance" and "geometric non-compliance but performance not exceeding the tolerance", so the standard is not reasonable enough.

[0007] In summary, existing testing methods cannot simultaneously meet the testing requirements of "high precision, high efficiency, and low cost." There is an urgent need to develop a new testing technology to establish a direct correlation between blade shape deviation and performance deviation, thereby achieving accurate and rapid testing of compressor impeller performance consistency. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide a method for detecting the efficiency consistency of compressor impellers in order to address the above-mentioned shortcomings. By establishing a quantitative correlation model between blade shape deviation and efficiency deviation, and combining color space conversion, the deviation can be accurately extracted, and finally, efficient and high-precision performance consistency detection can be achieved.

[0009] To solve the above technical problems, the present invention adopts the following technical solution:

[0010] A method for detecting the consistency of compressor impeller efficiency includes the following steps:

[0011] Step 1, Sample preprocessing and point cloud scanning;

[0012] Step 2: Color space conversion. Convert the RGB color space to the HSV color space and utilize the characteristics of the chromaticity H dimension to achieve accurate quantitative extraction of leaf shape deviation, while removing background interference.

[0013] Step 3: Deviation parameter calculation. The average deviation is calculated to quantify the overall level of the leaf shape deviation, the maximum deviation is calculated to quantify the extreme degree of the leaf shape deviation, and the large deviation ratio is calculated to quantify the distribution density of the sensitive area of ​​the leaf shape deviation, thus constructing a comprehensive characterization system for the leaf shape deviation.

[0014] Step 4, Efficiency Deviation Prediction: By establishing a quantitative correlation model between leaf shape deviation and efficiency deviation, geometric deviation data is transformed into intuitive performance deviation results.

[0015] Step 5, performance consistency determination, based on efficiency deviation results.

[0016] Furthermore, the specific implementation process of step 1 is as follows:

[0017] Clean the surface of the compressor impeller to be tested by wiping it with anhydrous ethanol to remove oil stains and sanding the blade roots and edges to remove burrs. Fix the impeller on the testing fixture and ensure that the parallelism between the impeller reference surface and the fixture surface is ≤0.01mm to avoid installation deviations affecting scanning accuracy.

[0018] A high-precision 3D scanner was used to perform a full-surface scan on the pre-processed impeller: the scanning accuracy was set to be higher than 0.005 mm, the scanning point density was not less than 100 points / mm², and the scanning range covered the main blades and the shunt blades of the impeller; after the scan was completed, the 3D scanner automatically generated 3D point cloud data of the blade shape deviation; the color image representing the blade shape deviation was adjusted so that the leading edge tip and trailing edge tip of the blade were placed as far left and far right as possible in the color area, and then the RGB image was exported.

[0019] Furthermore, the specific implementation process of step 2 includes the following steps:

[0020] Step 21, data preprocessing, normalization of true color data;

[0021] ;

[0022] Step 22, Calculation of key parameters;

[0023] Calculate the maximum, minimum, and difference values ​​in the RGB values;

[0024] Maximum value: ;

[0025] Minimum value: ;

[0026] Difference: ;

[0027] Calculate brightness: ;

[0028] Calculate saturation: ;

[0029] Calculate chromaticity: ;

[0030] Color correction value: , ;

[0031] Step 23, Background recognition and labeling: Separate the background and remove background interference;

[0032] The white background of the image is adjusted in chroma. H All dimensions are marked as " NaN By defining parameters after color space conversion, background interference is eliminated, and only leaf shape deviation data of the effective area of ​​the leaf is retained:

[0033] .

[0034] Furthermore, the specific implementation process of step 2 also includes the following steps:

[0035] Step 24, Color Correction Value Leaf shape deviation Mapping:

[0036] ;

[0037] in S d This represents the maximum downward deviation of the blade. S m The formula defines the maximum upper deviation of the leaf blade, and the if branch removes invalid background data and determines the chromaticity of the current pixel. Is it " NaN ",like for NaN This indicates that the pixel is a white background area, which has no valid leaf shape deviation data and is not included in subsequent calculations. The else branch converts the color of the valid leaf area into the specific leaf shape deviation. no NaN This pixel belongs to the effective area of ​​the leaf;

[0038] When inspecting the main blade, the main blade shape deviation ,in S Md The maximum downward deviation of the main blade. S Mu The maximum upper deviation of the main blade;

[0039] When inspecting the splitter blades, the blade shape deviation... ,in S Sd The maximum downward deviation of the main blade. S Su The maximum upper deviation of the main blade;

[0040] Therefore, the matrix for the blade shape deviation of the main blade and the spur blade is:

[0041] Leaf shape deviation and color of main leaf Mapping The chromaticity values ​​of the main blade region (0°-360°) are proportionally mapped to the deviation range of the main blade. S Md arrive S Mu The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... Corrected chromaticity values ​​for the main leaf area;

[0042] Split vane shape deviation and color Mapping The chromaticity values ​​of the splitter blade region (0°-360°) are proportionally mapped to the deviation range of the splitter blades. S Sd arrive S Su The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... This is the corrected chromaticity value for the splitter blade region.

[0043] Furthermore, the specific implementation process of step 3 also includes the following steps:

[0044] Step 31: After the leaf shape deviation mapping is completed, determine the following parameters:

[0045] Impeller diameter: Enter the impeller diameter according to the impeller model to be tested. d ;

[0046] The blade consists of a leading edge region and a trailing edge region. The leading edge region is the area 10% of the chord length from the blade inlet end; the trailing edge region is the area 10% of the chord length from the blade outlet end.

[0047] Large Deviation Identification Criteria: ;

[0048] Step 32, Eliminate the influence of impeller size on average blade profile deviation: Statistically calculate the deviation values ​​of all effective scanning points of the main blade and splitter blades, and then calculate the arithmetic mean. ;

[0049] Step 33, Maximum Deviation Calculation: Extract the maximum deviation value from the effective scanning points of the main blade and the splitter blade, eliminate the influence of impeller size, and quantify the degree of the most serious blade shape deviation in the compressor impeller;

[0050] ;

[0051] Step 34: In order to quantify the distribution density of severe blade shape deviations in the sensitive areas within the leading and trailing edge regions, and thus reflect the potential impact of blade shape deviations on aerodynamic performance, it is necessary to calculate the maximum blade deviation ratio. That is, to calculate the ratio of the number of pixels with large deviations exceeding a set threshold in the region to the total number of effective pixels in the region;

[0052] Step 35, Output Results: Output the key parameters that characterize the consistency deviation of the blade shape performance obtained from the calculation.

[0053] Furthermore, the specific process of implementing step 32 is as follows:

[0054] Count the number of valid pixels in the corresponding leaf-shaped deviation image: All leaf shape deviations in the image S ij Not for NaN The number of pixels is the total number of effective pixels. Num Where i and j are the row and column indices of the image pixels, i being the row number and j the column number, used to locate a specific pixel in the image. m , n These are the total number of rows and columns of the image, representing the pixel scale of the entire scanned image. S ij : Leaf shape deviation corresponding to the pixel in row i and column j;

[0055] Calculate the sum of leaf shape deviations in the image: The sum of leaf shape deviations of all pixels in the effective area of ​​the main leaf in the image is calculated.

[0056] The average blade profile deviation, after eliminating the influence of impeller size by combining the main blades and the splitter blades, is as follows:

[0057] , where Δ S sum_m Total leaf shape deviation of the main leaf, Δ S sum_S This represents the total blade profile deviation of the splitter blades. Num m The effective number of pixels for the main blade. Num s This represents the effective number of pixels for the splitter blades.

[0058] Furthermore, the specific process of implementing step 34 is as follows:

[0059] Get The matrix, consisting of column indices corresponding to the leaf shape deviation elements, forms a vector matrix of the effective pixels of the leaf:

[0060] ,in J k It is the first in the matrix k One element;

[0061] statistics J Number of elements in the vector: ;

[0062] Identify the row coordinate vector matrix corresponding to the pixels at the leading and trailing edges of the blade:

[0063] The row coordinate vector matrix of the leading edge region, starting with the effective region matrix of the blade. J In the middle, the pixels in the leading edge region with the largest absolute value of deviation are selected, and a matrix is ​​used. J top10% Mark the locations of these extreme deviations:

[0064] ,in J r It is a matrix J top10% The first in r One element, k It is a matrix J Element index, l It is the number of pixels with the maximum deviation in the leading edge region. r It is a matrix J top10% The element index, ranging from 1 to... l ;

[0065] The row coordinate vector matrix of the trailing edge region, starting with the effective area matrix of the blade. JIn the process, the pixels in the trailing edge region with the smallest absolute value of deviation are selected; a matrix is ​​then used. J bottom10% Mark the locations of these optimal regions, focusing on the areas where the leaf shape most closely resembles the standard model:

[0066] ,in J s It is a matrix J bottom10% The first in s One element, s It is a matrix J bottom10% The element index, the range of values ​​is N - l +1 to N ;

[0067] Column indexes for the leading and trailing regions: ;

[0068] Modulus of column index: ;

[0069] A matrix is ​​constructed by extracting all deviation values ​​that satisfy the large deviation threshold q from the pixel deviations of the leading and trailing edge regions of the blade. S locat_large :

[0070] ,in S ij_large It is the leaf shape deviation value corresponding to the pixel in the i-th row and j-th column within the leaf's sensitive area, and this deviation value is ≥ q. The first one, the second one, the third one... t Leaf shape deviation value corresponding to each pixel in the sensitive area;

[0071] Recognition Matrix S locat_large Elements that satisfy the large deviation condition for leaf shape are marked as 1, and elements that do not satisfy the large deviation condition are marked as 0.

[0072] ;

[0073] Statistical matrix S locat_large The number of pixels that satisfy the large deviation condition:

[0074] ;

[0075] Calculate the percentage of pixels in the leading and trailing edge regions of the blade that satisfy the large deviation condition relative to the total number of pixels in the corresponding leaf shape deviation image.

[0076] .

[0077] Furthermore, the specific implementation process of step 4 is as follows:

[0078] Leaf shape efficiency deviation is the result of the superposition of average leaf shape deviation and large leaf shape deviation; therefore, the formula for calculating deviation efficiency can be written as follows:

[0079] ;

[0080] in Represents the average leaf shape deviation Contribution to efficiency impact This represents the maximum leaf shape deviation. and the proportion of large deviations The overall contribution to efficiency.

[0081] Furthermore, the specific implementation process of step 4 is as follows:

[0082] The fitted correlation for the efficiency deviation is:

[0083] ;

[0084] Among them, the fitting coefficient A =0.2537, basic deviation B =-0.0018;

[0085] by When fitting to the form of independent variables, during adjustment a and b After the gain coefficient, it can be compared with... To maintain a good linear relationship, therefore The fitting correlation is:

[0086] ;

[0087] Where the fitting coefficient C =0.0112, additional loss base value D =-0.00002, the gain coefficient of the maximum deviation minus the average deviation. a =0.25, the gain coefficient for the large deviation proportion b =0.15.

[0088] The overall blade efficiency prediction bias is:

[0089] ;

[0090] .

[0091] Furthermore, the specific implementation process of step 5 is as follows:

[0092] If: the impeller performance is determined to be consistent and meets the requirements;

[0093] If the impeller performance is determined to be inconsistent, it is considered a defective product and must be returned to the production stage for reprocessing or scrapped.

[0094] The present invention adopts the above technical solution and has the following technical effects compared with the prior art:

[0095] 1. High detection accuracy: The leaf shape deviation is accurately quantified through RGB-HSV color space conversion. Combined with the deviation correlation model, the performance deviation prediction accuracy is high, meeting the efficiency deviation detection requirement of ±0.5%.

[0096] 2. High testing efficiency: The total testing time for a single impeller is ≤15 minutes (including scanning, calculation, and judgment), which is 8 times more efficient than bench testing (≥2 hours), making it suitable for mass production needs;

[0097] 3. Low testing cost: No need to build an expensive turbocharger test bench (test bench cost ≥ 8 million yuan), only a conventional 3D scanner and computer are needed, reducing equipment investment costs by more than 90%;

[0098] 4. Strong standard rationality: Establish a direct correlation between blade shape deviation (average deviation, maximum deviation, deviation ratio) and performance deviation, avoiding the problems of "geometric compliance but performance exceeding tolerance" and "geometric non-compliance but performance within tolerance", making the testing standard more in line with actual performance requirements. Attached Figure Description

[0099] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0100] Figure 1 This is a flowchart of the detection method in this invention;

[0101] Figure 2 This is a schematic diagram illustrating the conversion relationship between the RGB color space and the HSV color space in this invention;

[0102] Figure 3 This is a fitting formula for the efficiency deviation caused by the average leaf shape deviation in this invention.

[0103] Figure 4 This is a fitting formula for the efficiency deviation caused by the maximum leaf shape deviation and the proportion of large deviations in this invention. Detailed Implementation

[0104] Examples, such as Figure 1 As shown, a method for detecting the consistency of compressor impeller efficiency includes the following steps:

[0105] Step 1, Sample preprocessing and point cloud scanning;

[0106] Clean the surface of the compressor impeller to be tested by wiping it with anhydrous ethanol to remove oil stains, and polish it with sandpaper (grit ≥ 800 mesh) to remove burrs from the blade root and blade edge. Fix the impeller on the testing fixture and ensure that the parallelism between the impeller reference surface (such as the hub end face) and the plane of the fixture is ≤ 0.01mm to avoid installation deviations affecting scanning accuracy.

[0107] A high-precision 3D scanner was used to perform a full-surface scan of the pre-processed impeller: the scanning accuracy was set to be higher than 0.005 mm, the scanning point density to be no less than 100 points / mm², and the scanning range covered the main blades and branch blades of the impeller; after scanning, the 3D scanner automatically generated 3D point cloud data of blade shape deviation; the color image representing the blade shape deviation was adjusted so that the leading edge and trailing edge tips of the blades were placed as far left and right as possible in the color area, and then the RGB image was exported (image resolution). Different colors in the image represent blade shape deviation in different areas, and the maximum lower deviation of the main blade is recorded as . S Md Maximum upper deviation of the main blade S Mu Maximum downward deviation of the splitter blade S Sd Maximum upper deviation of the splitter blade S Su .

[0108] Step 2: Color space conversion. The RGB color space is converted to the HSV color space, and the characteristics of the chromaticity H dimension are used to achieve accurate quantitative extraction of leaf shape deviation. At the same time, background interference is removed to provide clean and effective data support for subsequent calculation of average deviation, maximum deviation and deviation ratio.

[0109] like Figure 2 As shown, since the point cloud data obtained by the 3D scanner is a true-color (RGB) image that reflects leaf shape deviation through color, the RGB color space is a three-dimensional matrix (red, green, and blue channels), meaning that each pixel has three corresponding values. Its mathematical expression is: ;

[0110] This means that it is impossible to use the RGB color space to statistically analyze one-dimensional leaf shape deviation data, and therefore impossible to obtain statistical data on average deviation, maximum deviation, and deviation ratio from true color images.

[0111] This invention converts the RGB color space to the HSV color space (chromaticity). Hsaturation S ,brightness V ), due to chromaticity H Dimensions possess full color attributes, therefore they can be utilized in the transformation color space. H Dimensional data is statistically analyzed to correspond to the statistics of leaf shape deviation.

[0112] The HSV color space expression is as follows: ;

[0113] Among them, chromaticity H : Indicates the type of color (such as red, yellow, blue, green, cyan, etc.). It is represented by an angle (0°-360°), forming a color wheel. Within the 360° range, each degree represents a color.

[0114] brightness V : Indicates the brightness or darkness of a color. That is, the degree to which a color is diluted by white light. The lower the brightness, the closer the color is to black (0% is black, 100% is the brightest state of the color).

[0115] Saturation S : Indicates the purity or vividness of a color.

[0116] The color space conversion process is as follows:

[0117] Step 21, data preprocessing, normalization of true color data;

[0118] ;

[0119] Step 22, Calculation of key parameters:

[0120] Calculate the maximum, minimum, and difference values ​​in the RGB values;

[0121] Maximum value: ;

[0122] Minimum value: ;

[0123] Difference: ;

[0124] Calculate brightness: ;

[0125] Calculate saturation: ;

[0126] Calculate chromaticity: ;

[0127] Color correction value: , .

[0128] Step 23, Background recognition and labeling: Separate the background and remove background interference;

[0129] The white background of the image is adjusted in chroma. H All dimensions are marked as " NaN By defining parameters after color space conversion, background interference is eliminated, and only leaf shape deviation data of the effective area of ​​the leaf is retained:

[0130] ;

[0131] The saturation in the HSV space is set using the formula above. S ,brightness V The threshold is used to filter out all pixels that meet the white characteristic, and then the chromaticity of these pixels is... H Marked as NaN This achieves separation between the background and the effective area of ​​the blade.

[0132] Step 24, Color Correction Value Leaf shape deviation Mapping:

[0133] ;

[0134] Where S d This represents the maximum downward deviation of the blade. S m The formula defines the maximum upper deviation of the leaf blade, and the if branch removes invalid background data and determines the chromaticity of the current pixel. Is it " NaN ",like for NaN This indicates that the current pixel is in a white background area. White background areas have no valid leaf shape deviation data and are not included in subsequent calculations. The else branch converts the color of the valid leaf area into the specific leaf shape deviation. no NaN If so, the current pixel belongs to the valid area of ​​the leaf.

[0135] When inspecting the main blade, the main blade shape deviation S Md The maximum downward deviation of the main blade. S Mu The maximum upper deviation of the main blade;

[0136] When inspecting the splitter blades, the blade shape deviation... ,in S Sd The maximum downward deviation of the main blade. S Su The maximum upper deviation of the main blade;

[0137] Therefore, the matrix for the blade shape deviation of the main blade and the spur blade is:

[0138] Leaf shape deviation and color of main leaf Mapping The chromaticity values ​​of the main blade region (0°-360°) are proportionally mapped to the deviation range of the main blade. S Md arrive S Mu The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... Corrected chromaticity values ​​for the main leaf area;

[0139] Split vane shape deviation and color Mapping The chromaticity values ​​of the splitter blade region (0°-360°) are proportionally mapped to the deviation range of the splitter blades. S Sd arrive S Su The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... This is the corrected chromaticity value for the splitter blade region.

[0140] Step 3, Deviation Parameter Calculation: The average deviation is calculated to quantify the overall level of blade shape deviation, the maximum deviation is calculated to quantify the extreme degree of blade shape deviation, and the large deviation ratio is calculated to quantify the distribution density of the sensitive area of ​​blade shape deviation. A comprehensive characterization system of blade shape deviation is constructed to provide key data support for the subsequent establishment of a correlation model between blade shape deviation and efficiency deviation and accurate prediction of impeller efficiency, while avoiding the limitation of focusing only on the value of a single geometric deviation.

[0141] Step 31: After the leaf shape deviation mapping is completed, determine the following parameters:

[0142] Impeller diameter: Enter the impeller diameter according to the impeller model to be tested. d ;

[0143] Leading edge / trailing edge definition range: The blade includes a leading edge region and a trailing edge region. The leading edge region is the area 10% of the chord length from the blade inlet end. It is the initial action area where the airflow enters the blade. Blade shape deviations in this area (such as bulges and depressions) will directly change the initial flow direction of the airflow, easily causing airflow separation and turbulence loss. It is a highly sensitive area affecting compressor efficiency. The trailing edge region is the area 10% of the chord length from the blade outlet end. It is the final adjustment area where the airflow leaves the blade. Blade shape deviations in this area will affect the uniformity of the airflow, and thus affect the inlet conditions of the subsequent impeller stage.

[0144] Large Deviation Identification Criteria: .

[0145] Step 32, eliminate the influence of impeller size on average blade profile deviation. : Calculate the arithmetic mean of the deviation values ​​of all effective scanning points of the main blade and the splitter blade;

[0146] Count the number of valid pixels in the corresponding leaf-shaped deviation image: All leaf shape deviations in the image S ij Not for NaN The number of pixels is the total number of effective pixels. Num ,in i , j It is the row and column index of the image pixels. i It's a line number. j These are column numbers, used to locate a specific pixel in the image. m , n These are the total number of rows and columns of the image, representing the pixel scale of the entire scanned image. S ij : No. i line, number j Leaf shape deviation corresponding to column pixels.

[0147] Calculate the sum of leaf shape deviations in the image: The sum of leaf shape deviations of all pixels in the effective area of ​​the main leaf in the image is calculated.

[0148] The average blade profile deviation, after eliminating the influence of impeller size by combining the main blades and the splitter blades, is as follows:

[0149] , where Δ S sum_m Total leaf shape deviation of the main leaf, Δ S sum_S This represents the total blade profile deviation of the splitter blades. Num m The effective number of pixels for the main blade. Num s This represents the effective number of pixels for the splitter blades.

[0150] Step 33, Maximum Deviation Calculation: Extract the maximum deviation value from the effective scanning points of the main blade and the splitter blade, eliminate the influence of impeller size, and quantify the degree of the most serious blade shape deviation in the compressor impeller;

[0151] .

[0152] Step 34: In order to quantify the distribution density of severe blade shape deviations in the sensitive areas within the leading and trailing edge regions, and thus reflect the potential impact of blade shape deviations on aerodynamic performance, it is necessary to calculate the maximum blade deviation ratio. This involves calculating the ratio of the number of pixels in the region that have a large deviation exceeding a set threshold to the total number of valid pixels in the region. The specific calculation process is as follows:

[0153] Get The matrix, consisting of column indices corresponding to the leaf shape deviation elements, forms a vector matrix of the effective pixels of the leaf:

[0154] This process clearly defines the pixel location range of the effective area of ​​the blade, allowing subsequent deviation statistical calculations to accurately focus on the blade area and eliminate background interference. J k It is the first in the matrix k Each element.

[0155] statistics J Number of elements in the vector: .

[0156] Identify the row coordinate vector matrix corresponding to the pixels at the leading and trailing edges of the blade:

[0157] The row coordinate vector matrix of the leading edge region, starting with the effective region matrix of the blade. J In the process, the leading edge region pixels with the largest absolute deviation are selected; a matrix is ​​then used. J top10% By marking the locations of these extreme deviations and focusing on the areas with the most severe blade shape deviations, a precise pixel range is provided for analyzing the impact of local extreme deviations on impeller efficiency.

[0158] ,in J r It is a matrix J top10% The first in r One element, k It is a matrix J Element index, l It is the number of pixels with the maximum deviation in the leading edge region. r It is a matrix J top10% The element index, ranging from 1 to... l .

[0159] The row coordinate vector matrix of the trailing edge region, starting with the effective area matrix of the blade. J In the process, the pixels in the trailing edge region with the smallest absolute value of deviation are selected; a matrix is ​​then used. J bottom10% Mark the locations of these optimal regions, focusing on the areas where the leaf shape most closely resembles the standard model, to provide a precise pixel range for comparing the performance differences between optimal and extreme deviation regions:

[0160] ,inJ s It is a matrix J bottom10% The first in s One element, s It is a matrix J bottom10% The element index, the range of values ​​is N - l +1 to N .

[0161] Column indexes for the leading and trailing regions: ;

[0162] Modulus of column index: ;

[0163] Extract all pixels that meet the large deviation threshold from the pixel deviations of the leading and trailing edge regions of the blade. q The deviation values ​​form a matrix S locat_large :

[0164] ,in S ij_large It is the first in the leaf sensitive area i line, number j The leaf shape deviation value corresponding to the column pixel, and this deviation value is ≥ q. The first one, the second one, the third one... t The leaf shape deviation value corresponding to each pixel in the sensitive area.

[0165] Recognition Matrix S locat_large The element that satisfies the condition for large deviation in leaf shape is denoted as . 1 Elements that do not meet the large deviation condition are marked as 0:

[0166] ;

[0167] Statistical matrix S locat_large The number of pixels that satisfy the large deviation condition:

[0168] ;

[0169] Calculate the percentage of pixels in the leading and trailing edge regions of the blade that satisfy the large deviation condition relative to the total number of pixels in the corresponding leaf shape deviation image.

[0170] .

[0171] Step 35, Output Results: Calculate the key parameters characterizing the consistency deviation of blade shape performance: , , .

[0172] Step 4, Efficiency Deviation Prediction: By establishing a quantitative correlation model between leaf shape deviation and efficiency deviation, geometric deviation data is transformed into intuitive performance deviation results, providing accurate and quantifiable basis for subsequent performance consistency judgment.

[0173] Leaf shape efficiency deviation is the result of the superposition of average leaf shape deviation and large leaf shape deviation; therefore, the formula for calculating deviation efficiency can be written as follows:

[0174] ;

[0175] in Represents the average leaf shape deviation Contribution to efficiency impact This represents the maximum leaf shape deviation. and the proportion of large deviations The overall contribution to efficiency.

[0176] Fitting was performed based on a large amount of CFD simulation and experimental data. The impact on efficiency is almost linear, such as Figure 3 As shown, therefore The fitted correlation for the efficiency deviation is:

[0177] ;

[0178] Among them, the fitting coefficient A= 0.2537, basic deviation B =-0.0018;

[0179] After numerous attempts to fit the formula, it was found that... When fitting to the form of independent variables, during adjustment a and b After the gain coefficient, it can be compared with... Maintain a good linear relationship, such as Figure 4 As shown, therefore The fitting correlation is:

[0180] ;

[0181] Where the fitting coefficient C =0.0112, additional loss base value D =-0.00002, the gain coefficient of the maximum deviation minus the average deviation. a =0.25, the gain coefficient for the large deviation proportion b =0.15.

[0182] The overall blade efficiency prediction bias is:

[0183] ;

[0184] .

[0185] Step 5, performance consistency determination;

[0186] Based on the efficiency deviation results:

[0187] If: the impeller performance is determined to be consistent and meets the requirements;

[0188] If the impeller performance is determined to be inconsistent, it is considered a defective product and must be returned to the production stage for reprocessing or scrapped.

[0189] The description of this invention is given for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A method for detecting uniformity of efficiency of a compressor impeller, characterized by: Includes the following steps: Step 1, Sample preprocessing and point cloud scanning; Step 2: Color space conversion. Convert the RGB color space to the HSV color space and utilize the characteristics of the chromaticity H dimension to achieve accurate quantitative extraction of leaf shape deviation, while removing background interference. Step 3: Deviation parameter calculation. The average deviation is calculated to quantify the overall level of the leaf shape deviation, the maximum deviation is calculated to quantify the extreme degree of the leaf shape deviation, and the large deviation ratio is calculated to quantify the distribution density of the sensitive area of ​​the leaf shape deviation, thus constructing a comprehensive characterization system for the leaf shape deviation. Step 4, Efficiency Deviation Prediction: By establishing a quantitative correlation model between leaf shape deviation and efficiency deviation, geometric deviation data is transformed into intuitive performance deviation results. Step 5, performance consistency determination, based on efficiency deviation results; The specific implementation process of step 3 also includes the following steps: Step 31: After the leaf shape deviation mapping is completed, determine the following parameters: Impeller wheel diameter: input impeller wheel diameter according to the type of impeller to be detected d ; The blade consists of a leading edge region and a trailing edge region. The leading edge region is the area 10% of the chord length from the blade inlet end; the trailing edge region is the area 10% of the chord length from the blade outlet end. Large Deviation Identification Criteria: ; Step 32, Eliminate the influence of impeller size on average blade profile deviation: Statistically calculate the deviation values ​​of all effective scanning points for the main blade and splitter blades. Calculate the arithmetic mean; Step 33, Maximum Deviation Calculation: Extract the maximum deviation value from the effective scanning points of the main blade and the splitter blade, eliminate the influence of impeller size, and quantify the degree of the most serious blade shape deviation in the compressor impeller; , To account for leaf shape deviation, the white background of the image is marked as "NaN" in the chromaticity H dimension; Step 34: In order to quantify the distribution density of severe blade shape deviations in the sensitive areas within the leading and trailing edge regions, and thus reflect the potential impact of blade shape deviations on aerodynamic performance, it is necessary to calculate the maximum blade deviation ratio. That is, the ratio of the number of pixels with large deviations exceeding a set threshold in the leading and trailing edge regions to the total number of effective pixels in the leading and trailing edge regions; Step 35, Output Results: Output the key parameters that characterize the consistency deviation of the blade shape performance obtained from the calculation. The specific process of step 32 is as follows: Count the number of valid pixels in the corresponding leaf-shaped deviation image: All leaf shape deviations in the image S ij Not for NaN The number of pixels is the total number of effective pixels. Num ,in i , j It is the row and column index of the image pixels. i It's a line number. j These are column numbers, used to locate a specific pixel in the image. m , n These are the total number of rows and columns of the image, representing the pixel scale of the entire scanned image. S ij : No. i line, number j Leaf shape deviation corresponding to column pixels; Calculate the sum of leaf shape deviations in the image: The sum of leaf shape deviations of all pixels in the effective area of ​​the main leaf in the image is calculated. The average blade profile deviation, after eliminating the influence of impeller size by combining the main blades and the splitter blades, is as follows: , where Δ S sum_m Total leaf shape deviation of the main leaf, Δ S sum_S This represents the total blade profile deviation of the splitter blades. Num m The effective number of pixels for the main blade. Num s The effective number of pixels for the diverter blades; The specific process of step 34 is as follows: Get The matrix, consisting of column indices corresponding to the leaf shape deviation elements, forms a vector matrix of the effective pixels of the leaf: ,in J k It is the first in the matrix k One element; statistics J Number of elements in the vector: ; Identify the row coordinate vector matrix corresponding to the pixels at the leading and trailing edges of the blade: The row coordinate vector matrix of the leading edge region, starting with the effective region matrix of the blade. J In the middle, the pixels in the leading edge region with the largest absolute value of deviation are selected, and a matrix is ​​used. J top10% Mark the locations of these extreme deviations: ,in J r It is a matrix J top10% The first in r One element, k It is a matrix J Element index, l It is the number of pixels with the maximum deviation in the leading edge region. r It is a matrix J top10% The element index, ranging from 1 to... l ; The row coordinate vector matrix of the trailing edge region, starting with the effective area matrix of the blade. J In the process, the pixels in the trailing edge region with the smallest absolute value of deviation are selected; a matrix is ​​then used. J bottom10% Mark the locations of these optimal regions, focusing on the areas where the leaf shape most closely resembles the standard model: ,in J s It is a matrix J bottom10% The first in s One element, s It is a matrix J bottom10% The element index, the range of values ​​is N - l +1 to N ; Column indexes for the leading and trailing regions: ; Modulus of column index: ; Extract all pixels that meet the large deviation threshold from the pixel deviations of the leading and trailing edge regions of the blade. q The deviation values ​​form a matrix S locat_large : ,in S ij_large It is the first in the leaf sensitive area i line, number j The leaf shape deviation value corresponding to the column pixel, and the leaf shape deviation value S ij_large ≥ q , The first one, the second one, the third one... t Leaf shape deviation value corresponding to each pixel in the sensitive area; Recognition Matrix S locat_large The element that satisfies the condition for large deviation in leaf shape is denoted as . 1 Elements that do not meet the large deviation condition are marked as 0: ； Statistical matrix S locat_large The number of pixels that satisfy the large deviation condition: ； Calculate the percentage of pixels in the leading and trailing edge regions of the blade that satisfy the large deviation condition relative to the total number of pixels in the corresponding leaf shape deviation image. ； The specific implementation process of step 4 is as follows: Leaf shape efficiency deviation is the result of the superposition of average leaf shape deviation and large leaf shape deviation; therefore, the formula for calculating deviation efficiency can be written as follows: ； in Represents the average leaf shape deviation Contribution to efficiency impact This represents the maximum leaf shape deviation. and the proportion of large deviations The overall contribution to efficiency; The specific implementation process of step 4 is as follows: The fitted correlation for the efficiency deviation is: ； Among them, the fitting coefficient A = 0.2537 and the basic deviation B = -0.0018; by When fitting to the form of independent variables, after adjusting the gain coefficients of a and b, it is compared with... To maintain a good linear relationship, therefore The fitting correlation is: ； The fitting coefficient C = 0.0112, the base value of the additional loss D = -0.00002, the gain coefficient a = 0.25 (maximum deviation minus average deviation), and the gain coefficient b = 0.15 (large deviation percentage). The overall blade efficiency prediction bias is: ； 。 2. The method for detecting the consistency of compressor impeller efficiency as described in claim 1, characterized in that: The specific implementation process of step 1 is as follows: Clean the surface of the compressor impeller to be tested by wiping it with anhydrous ethanol to remove oil stains and sanding the blade roots and edges to remove burrs. Fix the impeller on the testing fixture and ensure that the parallelism between the impeller reference surface and the fixture surface is ≤0.01mm to avoid installation deviations affecting scanning accuracy. A high-precision 3D scanner was used to perform a full-surface scan on the pre-processed impeller: the scanning accuracy was set to be higher than 0.005 mm, the scanning point density was not less than 100 points / mm², and the scanning range covered the main blades and the shunt blades of the impeller; after the scan was completed, the 3D scanner automatically generated 3D point cloud data of the blade shape deviation; the color image representing the blade shape deviation was adjusted so that the leading edge tip and trailing edge tip of the blade were placed as far left and far right as possible in the color area, and then the RGB image was exported.

3. The method for detecting the consistency of compressor impeller efficiency as described in claim 1, characterized in that: The specific implementation process of step 2 includes the following steps: Step 21, data preprocessing, normalization of true color data; ； ； ； Step 22, Calculation of key parameters; Calculate the maximum, minimum, and difference values ​​in the RGB values; Maximum value: ; Minimum value: ; Difference: ; Calculate brightness: ; Calculate saturation: ; Calculate chromaticity: ; Color correction value: , ; Step 23, Background recognition and labeling: Separate the background and remove background interference; By defining parameters after color space conversion, background interference is eliminated, and only leaf shape deviation data of the effective area of ​​the leaf is retained: 。 4. The method for detecting the consistency of compressor impeller efficiency as described in claim 3, characterized in that: The specific implementation process of step 2 also includes the following steps: Step 24, Color Correction Value Leaf shape deviation Mapping: ； in S d This represents the maximum downward deviation of the blade. S m The formula defines the maximum upper deviation of the leaf blade, and the if branch removes invalid background data, determining whether the chroma of the current pixel is "". NaN If it is NaN This indicates that the current pixel is a white background area. White background areas have no valid leaf shape deviation data and are not included in subsequent calculations. The else branch converts the color of the valid leaf area into a specific leaf shape deviation. The current pixel is not... NaN If so, the current pixel belongs to the valid area of ​​the leaf. When inspecting the main blade, the main blade shape deviation ,in S Md The maximum downward deviation of the main blade. S Mu The maximum upper deviation of the main blade; When inspecting the splitter blades, the blade shape deviation... ,in S Sd The maximum downward deviation of the main blade. S Su The maximum upper deviation of the main blade; Therefore, the matrix for the blade shape deviation of the main blade and the spur blade is: Main leaf shape deviation and color mapping The chromaticity values ​​of the main blade region (0°-360°) are proportionally mapped to the deviation range of the main blade. S Md arrive S Mu The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... Corrected chromaticity values ​​for the main leaf area; Split vane shape deviation and color mapping The chromaticity values ​​of the splitter blade region (0°-360°) are proportionally mapped to the deviation range of the splitter blades. S Sd arrive S Su The chromaticity values ​​of the leaf area are converted into specific physical deviation data, among which... This is the corrected chromaticity value for the splitter blade region.

5. The method for detecting the consistency of compressor impeller efficiency as described in claim 1, characterized in that: The specific implementation process of step 5 is as follows: If: the impeller performance is determined to be consistent and meets the requirements; If the impeller performance is determined to be inconsistent, it is considered a defective product and must be returned to the production stage for reprocessing or scrapped.

Images