CN103292760A - Thin-wall blade error analytical method - Google Patents

Abstract

The invention discloses an error analysis method for thin-walled blades, which uses a three-coordinate measuring machine to measure three sets of data points on the processed thin-walled blades using the Z-axis contour method, and aims at extracting three sets of equal heights between the design blades and the measured data points. Group data, accurately calculate the error value of thin-walled blades, and analyze and denoise the measured data points, and register and fit the simplified measured data points with the design data to establish the objective function; use particle swarm algorithm Optimize and solve the objective function; and quickly and accurately analyze the error between the measured data of the processed thin-walled blade and the theoretical data of the blade, and intuitively reflect the size of the error value, shorten the error analysis cycle, and improve the efficiency of error analysis . Realize the effective control in the finishing process of the thin-walled blade, and greatly improve the machining accuracy and efficiency.

Description

An error analysis method for thin-walled blades

technical field

The invention belongs to the field of manufacturing and finishing of aeroengine blades, and in particular relates to an error analysis method for thin-walled blades.

Background technique

Blade is the core component of aero-engine and gas turbine, and it is also a typical thin-walled part. Its manufacturing and processing level directly affects the aerodynamic performance of the engine. With the continuous development of aerodynamic design technology, structural technology and material technology, the blades of aero-engines and gas turbines have structural characteristics of bending, twisting, thinning, sweeping, and lightness. Not only the processing technology is complex, the precision requirements are high, and the processing quality is direct. affect the overall performance of the engine. In engineering practice, in order to eliminate the adverse effect on the processing accuracy of thin-walled blades, some technological measures or auxiliary measurement methods are often used to reduce the processing deformation and processing errors of the blades, but this will inevitably increase many additional processes , and these technological measures are mainly based on qualitative analysis and actual processing experience, lacking quantitative analysis and operating specifications, not only the accuracy and quality of blade parts are difficult to guarantee, but also the requirements for the measurement environment are high, and the detection accuracy of error values is low. Therefore, how to quickly and accurately analyze the machining error of the blade is an important link to ensure that the machining accuracy meets the design requirements.

Invention patent 201210255187.0 discloses a method and equipment for rapid full-scale detection of blades. The method is based on a three-dimensional optical measurement system and surface structured light projection profilometry to measure blades. In the invention patent 201210350382.1, a method of quickly realizing the detection of the blade profile by the inductance meter is proposed, which is to use the inductance meter to adjust the initial zero position, and then use the three-coordinate measuring machine to measure the blade data to obtain the error value. Its disadvantage is that it has high requirements on the measurement environment, and the detection accuracy of the error value is low.

Contents of the invention

In order to avoid the deficiencies of the existing technology and overcome the high requirements for the measurement environment and the low detection accuracy of the error value, the present invention proposes an error analysis method for thin-walled blades, which uses the measurement data point model and the design data point model Perform matching and fitting, quickly and accurately analyze the error between the measured data of the processed blade and the theoretical data of the blade, and intuitively reflect the size of the error value. In this error analysis method, three sets of data points are measured on the processed blade by a three-coordinate measuring machine using the Z-axis contour method, and three sets of data that are equal to the height of the measured data points are extracted for the designed blade, and the measured data points are analyzed and removed. Noise processing, registration of the simplified measurement data points and design data to establish the objective function; use particle swarm optimization algorithm to optimize and solve the objective function; obtain the blade error analysis results.

Thin-walled blade error analysis method of the present invention is characterized in that comprising the following steps:

Step 1. Use a three-coordinate measuring machine to measure the processed blade, and use the Z-axis contour method to obtain three sets of cross-sectional data points;

Step 2. Obtain three groups of blade cross-section data points corresponding to the same height as step 1 for the blade design model;

Step 3. The measurement data is analyzed and denoised preprocessing, and its specific steps are as follows:

(1) For n ordered data points P _i (i=0,1,2...n) of the measurement point set, P _i-1 , P _i , and P _i+1 are three adjacent points, forming a triangle;

与

的夹角，d=|P_i-1P_i|sinα表示弦高；(2) α means

and

The included angle, d=|P _i-1 P _i |sinα represents the chord height;

(3) For the fixed angle △α and the chord height △d limit, filter out the points within the two thresholds. △α takes 0～10 degrees, and △d is determined according to formula (1):

$\mathrm{<span\; class="notranslate">\; \&Delta;d</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&Delta;\&alpha;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, N _b represents the initial number of points, N _a represents the number of simplified points,

表示相邻点距离的均值；

Indicates the mean value of the distance between adjacent points;

Step 4. Registering the streamlined measurement data point set with the design data, the specific steps are as follows:

(1) Transform the two sets of data into the same coordinate system by rotating along the Z-axis and moving around the X-axis and Y-axis. The rotation transformation matrix is:

$\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ {\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}& {\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right)$

γ is the rotation angle of the measurement data point around the design data point in the processing coordinate system,

P _x , P _y are the displacements of the measured data points relative to the design data points;

(2) Establish the objective function to minimize the average distance value between the corresponding points of the two sets of data, and introduce the weight m _i to improve the registration accuracy. The objective function is determined by formula (2):

$f\left(x\right)=\frac{1}{N}\underset{i=0}{\overset{\mathrm{no}}{\mathrm{\&Sigma;}}}{m}_i{(Q_i-{\mathrm{AP}}_i)}^2$ (2)

${\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}}$

Among them, Q _i is the design data point, P _i is the measurement data point, l _i represents the thickness of the blade section at Q _i , and l _max represents the maximum value of the thickness on the same section;

(3) Use the particle swarm optimization algorithm to solve the optimal rotation transformation matrix that minimizes the objective function value;

Step 5. Calculate the error of the processed blade according to the result after registration, and the calculated error includes contour error, twist error, inclination error, and curvature error;

Step 6. Analyzing the obtained error results.

Beneficial effect

The thin-wall blade error analysis method of the present invention adopts the method of matching the measurement data point model with the design data point model to accurately calculate the error value of the blade, so as to ensure that the detection result is more accurate and reliable during the detection process of the blade profile, and at the same time improve improve the efficiency of error analysis. The analysis method can quickly and accurately analyze the error between the measured data of the processed blade and the theoretical data of the blade, and intuitively reflect the size of the error value; the error analysis cycle is shortened, and automatic analysis is realized.

Description of drawings

A method for analyzing errors of thin-walled blades according to the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

Fig. 1 is a flow chart of the error analysis method for thin-walled blades of the present invention.

Figure 2 is a schematic diagram of three sets of data points measured for a machined blade.

Fig. 3 is three sets of data points extracted from the blade design model that are equal in height to the measured data points.

Detailed ways:

This embodiment is an error analysis method for a thin-walled blade.

Referring to Fig. 1, Fig. 2 and Fig. 3, the present invention adopts the method of matching the measured data point model with the design data point model, quickly and accurately analyzes the error between the measured data of the processed blade and the theoretical data of the blade, and intuitively reflects The magnitude of the error value.

Next, the error analysis method is used to analyze the thin-walled blades processed by a certain type of aero-engine, and the specific steps are as follows:

The first step is to measure the processed blade with a three-coordinate measuring machine; and use the Z-axis contour method to obtain three sets of cross-sectional data points, as shown in Figure 2;

The second step is to obtain three sets of blade cross-section data points corresponding to the same height as step 1 for the blade design model, as shown in Figure 3;

The third step is to analyze and denoise the measurement data, and the specific steps are as follows:

(1) For n ordered data points P _i (i=0,1,2...n) of the measurement point set, P _i-1 , P _i , and P _i+1 are three adjacent points, forming a triangle;

的夹角，d=|P_i-1P_i|sinα表示弦高；(2) α means and

The included angle, d=|P _i-1 P _i |sinα represents the chord height;

(3) For the limit value of angle △α and chord height △d, filter out the points within the two thresholds, △α is 0 to 10 degrees, and △d is determined according to formula (1):

$\mathrm{<span\; class="notranslate">\; \&Delta;d</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&Delta;\&alpha;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, N _b represents the initial number of points, N _a represents the number of simplified points,

表示相邻点距离的均值；

Indicates the mean value of the distance between adjacent points;

The fourth step is to register the simplified measurement data point set with the design data, and the specific steps are as follows:

(1) Transform the two sets of data into the same coordinate system by rotating along the Z-axis and moving around the X-axis and Y-axis. The rotation transformation matrix is:

$\mathrm{<span\; class="notranslate">\; A</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{cccc}\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&gamma;</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ {\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}& {\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right)$

γ is the rotation angle of the measurement data point around the design data point in the processing coordinate system,

P _x , P _y are the displacements of the measured data points relative to the design data points;

(2) Establish the objective function to minimize the average distance value between the corresponding points of the two sets of data, and introduce the weight m _i to improve the registration accuracy. The objective function is determined by formula (2):

$f\left(x\right)=\frac{1}{N}\underset{i=0}{\overset{\mathrm{no}}{\mathrm{\&Sigma;}}}{m}_i{(Q_i-{\mathrm{AP}}_i)}^2$ (2)

${\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; l</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}}$

Among them, Q _i is the design data point, P _i is the measurement data point, l _i represents the thickness of the blade section at Q _i , and l _max represents the maximum value of the thickness on the same section;

(3) Use the particle swarm optimization algorithm to solve the optimal rotation transformation matrix that minimizes the objective function value;

The fifth step is to calculate the error of the processed blade according to the registration result, and the calculated error includes contour error, twist error, inclination error, and curvature error;

The sixth step is to get the error analysis results.

The error analysis method of the thin-walled blade of the present invention analyzes the processed thin-walled blade to find out the error of the blade, and obtains the error analysis result; realizes effective control in the finishing process of the thin-walled blade, and greatly improves the machining accuracy and efficiency. Through the error analysis of a certain type of engine blade, it is verified that the method of the present invention can fully guarantee the accuracy of blade error analysis and realize automatic analysis.

Claims (1)

1. thin wall vane error analysis method is characterized in that may further comprise the steps:

Blade after step 1. adopts three coordinate measuring machine to processing is measured, and adopts Z axle equal altitude method to obtain three groups of cross-section data points;

Step 2. pair blade designs a model and obtains the corresponding three group blade profile data points contour with step 1;

Step 3. is analyzed the denoising pre-service with measurement data, and its concrete steps are as follows:

(1) for n the ordered data point P that measures point set _i(i=0,1,2 ... n), P _I-1, P _i, P _I+1Be adjacent 3 points, form a triangle;

(2) α represents

With

Angle, d=|P _I-1P _i| sin α represents action;

(3) to deciding angle △ α and action △ d limit value, the point that will be in two threshold values filters out.△ α gets 0～10 degree, and △ d determines by formula (1):

$\mathrm{\&Delta;d}=\frac{N_a}{N_b}\mathrm{\&mu;}\sin \mathrm{\&Delta;\&alpha;}---\left(1\right)$

Wherein, N _bN is initially counted in expression _aCounting after expression is simplified,

The average of expression consecutive point distance;

Measurement data point set and design data after step 4. pair is simplified are carried out registration, and its concrete steps are as follows:

(1) two groups of data-switching are arrived under the identical coordinate system with the movement around X-axis, Y-axis by the rotation along the Z axle, the rotational transform matrix is:

$A=\left(\begin{array}{cccc}\cos \mathrm{\&gamma;}& \sin \mathrm{\&gamma;}& 0& 0\\ -\sin \mathrm{\&gamma;}& \cos \mathrm{\&gamma;}& 0& 0\\ 0& 0& 1& 0\\ p_x& p_y& 0& 1\end{array}\right)$

γ is the angle of measurement data points winding counting strong point rotation in machining coordinate system,

P _x, P _yBe the displacement of measurement data points with respect to design data point;

(2) set up objective function and make two groups of distance value mean values between the data corresponding point reach minimum, and introduce weights m _iImprove registration accuracy, objective function is determined by (2) formula:

$f\left(x\right)=\frac{1}{N}\underset{i=0}{\overset{n}{\mathrm{\&Sigma;}}}{m}_i{(Q_i-{\mathrm{AP}}_i)}^2$ （2）

$m_i=\frac{l_i}{l_{\max }}$

Wherein, Q _iBe design data point, P _iBe measurement data points, l _iBe illustrated in Q _iThe thickness of place's blade profile, l _MaxBe illustrated in the maximal value of thickness on the same cross section;

(3) find the solution with particle cluster algorithm, make target function value reach minimum optimum rotational transform matrix;

The error of the as a result calculating processing rear blade of step 5. after according to registration, the error of asking for comprises profile tolerance error, torsion resistance error, bank error, flexibility error;

The step 6. couple error analysis result who obtains.

Images