CN118274740A - High-precision measuring method for blade profile of industrial cooling fan - Google Patents
High-precision measuring method for blade profile of industrial cooling fan Download PDFInfo
- Publication number
- CN118274740A CN118274740A CN202410675307.5A CN202410675307A CN118274740A CN 118274740 A CN118274740 A CN 118274740A CN 202410675307 A CN202410675307 A CN 202410675307A CN 118274740 A CN118274740 A CN 118274740A
- Authority
- CN
- China
- Prior art keywords
- value
- data points
- hierarchy
- acquiring
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001816 cooling Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims description 65
- 238000000605 extraction Methods 0.000 claims abstract description 101
- 239000013598 vector Substances 0.000 claims abstract description 74
- 230000000694 effects Effects 0.000 claims abstract description 47
- 238000004220 aggregation Methods 0.000 claims description 89
- 230000002776 aggregation Effects 0.000 claims description 89
- 238000009826 distribution Methods 0.000 claims description 74
- 238000005259 measurement Methods 0.000 abstract description 13
- 238000000691 measurement method Methods 0.000 abstract description 2
- 230000001419 dependent effect Effects 0.000 description 12
- 230000007423 decrease Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 5
- 230000003247 decreasing effect Effects 0.000 description 3
- 230000004931 aggregating effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000005452 bending Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012994 industrial processing Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000005057 refrigeration Methods 0.000 description 1
- 238000009423 ventilation Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of blade profile measurement, in particular to a high-precision blade profile measurement method of an industrial cooling fan. Firstly, acquiring all node sets corresponding to each level; in all node sets corresponding to the hierarchy, acquiring a classification effect value of the hierarchy according to the degree of difference between the feature vectors corresponding to the data points of all the node sets, the similarity degree of the feature vectors corresponding to all the data points in each node set and the fluctuation and the size of the information extraction feature values of all the node sets; further obtaining an optimal level; acquiring the data weight of each data point according to the characteristic vector of the data point in the node set corresponding to the optimal level; and finally, measuring the profile of the fan blade. According to the invention, the data weight capable of accurately reflecting the detail expression capability of the fan blade structural area where the data point is located is constructed, so that the accuracy of fan blade modeling for reflecting the fan blade profile is improved, and the effect of fan blade profile measurement is improved.
Description
Technical Field
The invention relates to the technical field of blade profile measurement, in particular to a high-precision blade profile measurement method of an industrial cooling fan.
Background
Industrial cooling fans are an important industrial refrigeration device. The industrial cooling fan effectively reduces the temperature by cooling down and convection ventilation. The industrial cooling fan plays an important role in industrial production and processing, can effectively prevent equipment from overheating, ensures normal operation of machines and improves production efficiency. The fan blade is a core component of the industrial cooling fan, and the reliability and performance of the fan blade structure directly influence the performance of the industrial cooling fan. The high-precision measurement of the profile of the fan blade is beneficial to ensuring that the fan blade meets the quality requirement and reducing the influence on the performance of the industrial cooling fan caused by unqualified fan blade.
Because the blade of the industrial cooling fan is large, the profile of the blade is difficult to directly measure, and the prior art point cloud data modeling method models the blade according to the point cloud data of the blade, so that the high-precision measurement of the blade is performed. However, in the modeling process of the fan blade by the point cloud data modeling method in the prior art, because the fan blade of the industrial cooling fan has a relatively complex shape and structure, data points at different positions in the point cloud data of the fan blade may have different detail expression capacities, the weight of the data points constructed by the traditional point cloud data modeling method is difficult to reflect the detail expression capacities of the fan blade structural area where the data points are located, so that the fan blade modeling is difficult to accurately reflect the actual fan blade profile, and the accuracy of the fan blade profile measurement is affected.
Disclosure of Invention
In order to solve the technical problems that in the modeling process of a fan blade by using a point cloud data modeling method in the prior art, as the fan blade of an industrial cooling fan has a relatively complex shape and structure, data points of different fan blade structure areas in the fan blade point cloud data possibly have different detail expression capacities, so that the fan blade modeling is difficult to accurately reflect a real fan blade profile and the accuracy of the fan blade profile measurement is affected, the invention aims to provide a high-precision measuring method for the fan blade profile of the industrial cooling fan, and the adopted technical scheme is as follows:
A high-precision measuring method for the blade profile of an industrial cooling fan comprises the following steps:
acquiring a point cloud data set of the fan blade; each data point in the point cloud data set has a corresponding feature vector;
In the point cloud data set, clustering all data points according to the difference of feature vectors and the space distribution interval among the data points to obtain all node sets corresponding to all levels; acquiring a distribution aggregation characteristic value of the data points according to the distribution characteristics of the node set to which the data points belong; acquiring information extraction characteristic values of the node set according to the size and fluctuation of the distribution aggregation characteristic values of the data points in the node set; in all node sets corresponding to the hierarchy, acquiring a classification effect value of the hierarchy according to the degree of difference between the feature vectors corresponding to the data points of all the node sets, the similarity degree of the feature vectors corresponding to all the data points in each node set and the fluctuation and the size of the information extraction feature values of all the node sets; obtaining an optimal level according to the classification effect value of the level;
Acquiring the data weight of each data point according to the characteristic vector of the data point in the node set corresponding to the optimal level; and measuring the profile of the fan blade according to the data weights of all the data points in the point cloud data set of the fan blade.
Further, the method for acquiring all node sets corresponding to the hierarchy comprises the following steps:
Based on hierarchical clustering method, determining the feature distance between two data points according to the difference of feature vectors between the two data points and the space distribution interval; and clustering all the data points according to the characteristic distance between every two data points to obtain all the node sets corresponding to each level.
Further, the method for acquiring the distribution aggregation characteristic value comprises the following steps:
Taking the data point as the data point to be analyzed; taking each data point in the node set to which the data point to be analyzed belongs as each reference data point of the data point to be analyzed; according to the distribution of the reference data points corresponding to the data points to be analyzed, a preset reference neighborhood range of the data points to be analyzed is constructed; and acquiring a distribution aggregation characteristic value of the data point to be analyzed according to the size of the preset reference neighborhood range of the data point to be analyzed and the duty ratio of the reference data point in the preset reference neighborhood range of the data point to be analyzed.
Acquiring the space distance of a preset reference neighborhood range of a data point to be analyzed;
Calculating the ratio between the total number of all the reference data points and the total number of all the data points in a preset reference neighborhood range of the data points to be analyzed to obtain a reference density characteristic value of the data points to be analyzed;
Acquiring a distribution aggregation characteristic value of the data points to be analyzed according to the space distance and the reference density characteristic value; the space distance and the distribution aggregation characteristic value are in a negative correlation relationship; the reference density characteristic value and the distribution aggregation characteristic value are in positive correlation.
Further, the method for acquiring the information extraction characteristic value comprises the following steps:
Acquiring a quantile aggregation value of the node set according to the size of the distribution aggregation characteristic value of the data points in the node set;
Acquiring an aggregation non-uniformity degree value of the node set according to the fluctuation of the distribution aggregation characteristic values of all data points in the node set;
Acquiring information extraction characteristic values of the node sets according to quantile aggregation values of the node sets and aggregation unevenness values of the node sets; the quantile aggregation value and the information extraction characteristic value are in positive correlation; the aggregation non-uniformity value and the information extraction characteristic value are in a negative correlation relationship.
Further, the method for acquiring the classification effect value comprises the following steps:
Acquiring a hierarchy extraction information value according to the size of the information extraction characteristic value of all node sets corresponding to the hierarchy and the fluctuation of the information extraction characteristic value of all node sets corresponding to the hierarchy;
Acquiring a level detail region value according to the degree of difference between the data point feature vectors of all node sets corresponding to the level;
acquiring a hierarchy detail unstable value according to fluctuation degrees of feature vectors corresponding to all data points in each node set corresponding to the hierarchy;
Acquiring a characteristic distinguishing degree value of the hierarchy according to the hierarchy detail distinguishing value and the hierarchy detail unstable value; the level detail distinguishing value and the characteristic distinguishing degree value are in positive correlation; the unstable level detail value and the characteristic distinguishing degree value are in a negative correlation relationship;
Extracting an information value and a characteristic distinguishing degree value according to the hierarchy to obtain a classification effect value of the hierarchy; the classification effect value and the hierarchy extraction information value are in positive correlation; the classification effect value and the characteristic distinguishing degree value are in positive correlation.
Further, the method for acquiring the hierarchical extraction information value comprises the following steps:
acquiring a hierarchy extraction characteristic value according to the information extraction characteristic value of all node sets corresponding to the hierarchy;
Acquiring a hierarchy extraction unstable value according to the fluctuation of the information extraction characteristic values of all node sets corresponding to the hierarchy;
acquiring a hierarchy extraction information value according to the hierarchy extraction characteristic value and the hierarchy extraction unstable value; the level extraction characteristic value and the level extraction information value are in positive correlation; the level extraction unstable value and the level extraction information value are in negative correlation.
Further, the method for acquiring the data weight comprises the following steps:
In all node sets corresponding to the optimal level, acquiring node vectors of the node sets according to the feature vectors of all data points in the node sets;
And acquiring the data weight of the data point according to the node vector of the node set to which the data point belongs and the node vectors of all the node sets.
Further, the method for obtaining the optimal hierarchy comprises the following steps:
and taking the hierarchy corresponding to the maximum classification effect value as the optimal hierarchy.
Further, the method for acquiring the preset reference neighborhood range comprises the following steps:
and taking the data points to be analyzed as the sphere centers, and constructing a preset reference neighborhood range, wherein the preset reference neighborhood range is the smallest sphere comprising the reference data points with preset reference quantity.
The invention has the following beneficial effects:
In order to cluster data points corresponding to different fan blade structure areas, all data points are clustered to obtain all node sets corresponding to all levels. The accuracy of different fan blade structure areas is reflected by different node sets corresponding to the analysis level. The worse the aggregation characteristics of the data points in the same node set is considered, the incomplete extraction of the fan blade structure area is reflected, and the distribution aggregation characteristic values of the data points are obtained; the larger the distribution aggregation characteristic value of the data points, the more similar data points around the data points are represented, and the data points are in the similar areas which are aggregated. Considering that the larger the distribution aggregation characteristic value of the data points in the node set is, the larger the stability of the distribution aggregation characteristic value of the data points is, the better the extraction capability of the node set to the same fan blade structure area is represented, the more the data points in the node set can reflect the same fan blade structure area, and the larger the information extraction characteristic value of the node set is.
Considering that the extraction capability of the node set to the same fan blade structure area is not comprehensive only according to the distribution condition of the data points in the same node set, and considering that the distribution aggregation characteristic values of the node sets corresponding to the levels are more similar and the distribution aggregation characteristic values are larger, the representative levels have better characteristic extraction capability to the fan blade structure area; considering that the feature vectors reflect the detail capability of the represented data points, the feature vectors of the same structural region are relatively close, the feature vectors of different structural regions have relatively large differences, and the feature vectors of different node sets corresponding to the layers are combined to be distinguished, so that the feature vectors of different node sets corresponding to the layers are relatively large in distinction, the feature vectors of the same node set are relatively small in distinction, and the clustering result corresponding to the layers has relatively good distinguishing capability on different fan blade structural regions. And obtaining the classification effect value of the hierarchy, wherein the larger the classification effect value is, the better the effect of the corresponding cluster of the representative hierarchy is, so that the best hierarchy with the best effect of the representative cluster is obtained. The data points of the fan blade structure area with larger influence degree on the modeling result have larger data weight; the optimal level can more accurately represent different fan blade structure areas corresponding to different node sets, data points in the same fan blade structure area have similar detail expression capability, and as the data points in different structure areas possibly have different detail expression capability and have different importance degrees on modeling, the data weight of each data point is obtained, so that the data point with larger influence degree on a modeling result is focused in subsequent analysis, and the accuracy of high-precision measurement of the fan blade profile is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for measuring the blade profile of an industrial cooling fan with high accuracy according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a three-dimensional coordinate system according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a hierarchical clustering relationship tree according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of a specific implementation, structure, characteristics and effects of a method for measuring the profile of a fan blade of an industrial cooling fan with high precision according to the invention in combination with the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a specific scheme of a high-precision measuring method for the blade profile of an industrial cooling fan, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for measuring a blade profile of an industrial cooling fan with high accuracy according to an embodiment of the invention is shown, and the method includes the following steps:
step S1, acquiring a point cloud data set of a fan blade; each data point in the point cloud data set has a corresponding feature vector.
Specifically, in order to measure the fan blade of the industrial cooling fan, laser radar equipment is utilized to collect initial point cloud data of the fan blade of the industrial cooling fan. Because the environment of fan blade application is often comparatively abominable, lead to the initial point cloud data of industrial cooling fan's fan blade to have noise, noise can lead to industrial cooling fan's fan blade measurement inaccuracy. Therefore, denoising processing is required to be carried out on the point cloud data of the fan blade of the industrial cooling fan, and a point cloud data set of the fan blade is obtained. Based on a least square method, curvature and flatness corresponding to each data point in the point cloud data set are obtained. Constructing a feature vector of each data point in the point cloud data set according to the curvature and the flatness corresponding to each data point; the smaller the modular length of the feature vector of the data point indicates that the data point is more likely to be in a flatter region, the larger the modular length of the feature vector of the data point indicates that the data point is more likely to be in a detail region, and the feature vector reflects the detail capability and information richness indicating the data point.
It should be noted that, the point cloud data set is set in a three-dimensional space coordinate system, the X-Y plane of the three-dimensional space coordinate is parallel to the ground, and the Z axis is perpendicular to the ground. Fig. 2 is a schematic diagram of a three-dimensional space coordinate system according to an embodiment of the present invention, wherein three axes of the three-dimensional space coordinate system correspond to an X-axis, a Y-axis and a Z-axis, and fig. 2 is only a schematic diagram for helping understanding the three-dimensional space coordinate system.
It should be noted that, in order to facilitate the operation, all index data involved in the operation in the embodiment of the present invention is subjected to data preprocessing, so as to cancel the dimension effect. The specific means for removing the dimension influence is a technical means well known to those skilled in the art, and is not limited herein.
Step S2, clustering all data points in the point cloud data set according to the difference of feature vectors and the space distribution interval among the data points to obtain all node sets corresponding to all levels; acquiring a distribution aggregation characteristic value of the data points according to the distribution characteristics of the node set to which the data points belong; acquiring information extraction characteristic values of the node set according to the size and fluctuation of the distribution aggregation characteristic values of the data points in the node set; in all node sets corresponding to the hierarchy, acquiring a classification effect value of the hierarchy according to the degree of difference between the feature vectors corresponding to the data points of all the node sets, the similarity degree of the feature vectors corresponding to all the data points in each node set and the fluctuation and the size of the information extraction feature values of all the node sets; and obtaining the optimal level according to the classification effect value of the level.
In order to cluster data points corresponding to different fan blade structure areas, all data points are clustered to obtain all node sets corresponding to each level, so that different node sets corresponding to analysis levels can reflect the accuracy of different fan blade structure areas. The worse the aggregation characteristics of the data points in the same node set is considered, the incomplete extraction of the fan blade structure area is reflected, and the distribution aggregation characteristic values of the data points are obtained; the larger the distribution aggregation characteristic value of the data points, the more similar data points around the data points are represented, and the data points are in the similar areas which are aggregated. Considering that the larger the distribution aggregation characteristic value of the data points in the node set is, the larger the stability of the distribution aggregation characteristic value of the data points is, the better the extraction capability of the node set to the same fan blade structure area is represented, the more the data points in the node set can reflect the same fan blade structure area, and the larger the information extraction characteristic value of the node set is.
Considering that the extraction capability of the node set to the same fan blade structure area is not comprehensive only according to the distribution condition of the data points in the same node set, and considering that the distribution aggregation characteristic values of the node sets corresponding to the levels are more similar and the distribution aggregation characteristic values are larger, the representative levels have better characteristic extraction capability to the fan blade structure area; considering that the feature vectors reflect the detail capability of the represented data points, the feature vectors of the same structural region are relatively close, the feature vectors of different structural regions have relatively large differences, and the feature vectors of different node sets corresponding to the layers are combined to be distinguished, so that the feature vectors of different node sets corresponding to the layers are relatively large in distinction, the feature vectors of the same node set are relatively small in distinction, and the clustering result corresponding to the layers has relatively good distinguishing capability on different fan blade structural regions. And obtaining the classification effect value of the hierarchy, wherein the larger the classification effect value is, the better the effect of the corresponding cluster of the representative hierarchy is, so that the best hierarchy with the best effect of the representative cluster is obtained.
Preferably, the method comprises the steps of firstly clustering all data points to obtain all node sets corresponding to each level. In one embodiment of the present invention, the method for acquiring all node sets corresponding to a hierarchy includes:
Determining a characteristic distance between two data points according to the difference of the characteristic vectors between the two data points and the space distribution interval; the feature distance can reflect the possibility that two data points do not belong to the same fan blade structural area through the difference of the two data points in space position and the difference degree of the two data points in detail capability.
Obtaining a characteristic distance according to a characteristic distance formula, wherein the characteristic distance formula comprises:
; wherein, To be in the point cloud data set, the firstData point and the firstFeature distance between data points; To be in the point cloud data set, the first Feature vectors of data points; To be in the point cloud data set, the first Feature vectors of data points; To be in the point cloud data set, the first Data point and the firstEuclidean distance between the individual data points; Taking the modulus sign.
It should be noted that, the hierarchical clustering method is a technical means well known to those skilled in the art, and only the steps of obtaining all node sets corresponding to the hierarchy are briefly described herein:
(1) Taking each data point as a node set, wherein all node sets correspond to the bottom layer level, namely the 1 st level; (2) obtaining the characteristic distance of each two node sets. Typically by computing the average of the feature distances between all data points in the two node sets as the feature distance for the two node sets. And selecting two node sets with the minimum feature distance to be combined to form an updated node set. Taking the updated node set as the node set of the previous level, and updating the level; (3) Repeating (2) until an iteration termination condition is satisfied, the iteration termination condition typically being a preset maximum level; (4) outputting all node sets corresponding to each hierarchy. It should be noted that, all data points in all node sets corresponding to the hierarchy are all data points of the point cloud data set of the fan blade.
Different hierarchical divisions of point cloud data are achieved through hierarchical clustering algorithms, at lower levels, data points may be divided into many small sets of nodes, each reflecting microscopic details of the data. With the promotion of the hierarchy, the number of node sets gradually decreases, and the scale of the node sets gradually increases, reflecting the macroscopic shape of the data and the higher-level structure. The different levels reflect the clustering results of different clustering scales of the data points, and the macroscopic shape and microscopic details of the fan blade can be reflected in hierarchical clustering, so that data reference is provided for subsequent analysis. Fig. 3 is a schematic diagram of a hierarchical clustering relationship tree provided in an embodiment of the present invention, in which the horizontal axis represents the number of data points and the vertical axis represents the feature distance, and the schematic diagram of the hierarchical clustering relationship tree shows the hierarchical structure of the point cloud data set, and fig. 3 is only a schematic diagram for helping to understand the hierarchical structure of the fan blade point cloud data set.
Preferably, in order to analyze different node sets corresponding to a hierarchy to reflect the accuracy of different fan blade structure areas, first, the aggregation characteristics of data points are analyzed, and in one embodiment of the present invention, the method for obtaining the distribution aggregation characteristic value includes:
Taking the data point as the data point to be analyzed; taking each data point in the node set to which the data point to be analyzed belongs as each reference data point of the data point to be analyzed; according to the distribution of the reference data points corresponding to the data points to be analyzed, a preset reference neighborhood range of the data points to be analyzed is constructed; the preset reference neighborhood range reflects the range in which similar data points around the data point are located.
And acquiring a distribution aggregation characteristic value of the data point to be analyzed according to the size of the preset reference neighborhood range of the data point to be analyzed and the duty ratio of the reference data point in the preset reference neighborhood range of the data point to be analyzed.
It should be noted that, in one embodiment of the present invention, the method for acquiring the preset reference neighborhood range includes: and taking the data points to be analyzed as the sphere centers, and constructing a preset reference neighborhood range, wherein the preset reference neighborhood range is the smallest sphere comprising the reference data points with preset reference quantity. In one embodiment of the present invention, the preset reference number is 9, and the practitioner can set the reference number according to the implementation scenario. It should be noted that, when the total number of all the reference data points of the data points to be analyzed is smaller than 9, the node set to which the data points to be analyzed belong is too small, the node set is inaccurate in characterizing the fan blade structure area, the preset reference neighborhood range of the data points to be analyzed is not constructed, and the information extraction characteristic value of the node set to which the data points to be analyzed belong is 0.
Preferably, in one embodiment of the present invention, the method for acquiring the distribution aggregation feature value includes:
The worse the aggregation degree of the data points in the same node set is, the incomplete extraction of the representative fan blade structure area is performed, and the space distance of a preset reference neighborhood range of the data points to be analyzed is obtained; the spatial distance reflects the degree of dispersion of data points in the same node set around the data point to be analyzed, and the larger the spatial distance is, the larger the degree of dispersion of similar data point distribution around the data point is, and the smaller the distribution aggregation characteristic value is.
In one embodiment of the invention, the radius of the preset reference neighborhood range is used as the space distance of the preset reference neighborhood range; in other embodiments of the present invention, the maximum distance in the preset reference neighborhood range may also be used as the spatial distance of the preset reference neighborhood range, which is not limited herein.
Calculating the ratio between the total number of all the reference data points and the total number of all the data points in a preset reference neighborhood range of the data points to be analyzed to obtain a reference density characteristic value of the data points to be analyzed; the reference density characteristic value reflects the duty ratio of the data points in the same node set around the data point to be analyzed, and the larger the duty ratio is, the less interference information in the preset reference neighborhood range of the data point to be analyzed is represented, and the more similar data points around the data point are represented.
Acquiring a distribution aggregation characteristic value of the data points to be analyzed according to the space distance and the reference density characteristic value; the space distance and the distribution aggregation characteristic value are in a negative correlation relationship; the reference density characteristic value and the distribution aggregation characteristic value are in positive correlation. The larger the distribution aggregation characteristic value of the data points, the more similar data points around the data points are represented, and the data points are in the similar areas which are aggregated.
The positive correlation relationship indicates that the dependent variable increases along with the increase of the independent variable, the dependent variable decreases along with the decrease of the independent variable, and the specific relationship can be multiplication relationship, addition relationship, idempotent of an exponential function and is determined by practical application; the negative correlation indicates that the dependent variable decreases with increasing independent variable, and the dependent variable increases with decreasing independent variable, which may be a subtraction relationship, a division relationship, or the like, and is determined by the actual application.
In one embodiment of the present invention, the distribution aggregation feature value formula includes:
; wherein, Aggregating eigenvalues for the distribution of data points to be analyzed; the spatial distance of a preset reference neighborhood range for the data point to be analyzed; the total number of all data points in a preset reference neighborhood range of the data points to be analyzed; the total number of all reference data points in a preset reference neighborhood range of the data points to be analyzed; is the reference density characteristic value of the data point to be analyzed.
Preferably, in order to analyze the extraction capability of the node set to the same fan blade structure region, in one embodiment of the present invention, the method for obtaining the information extraction feature value includes:
In order to reflect the distribution aggregation characteristic values of the whole node set, acquiring quantile aggregation values of the node set according to the distribution aggregation characteristic values of the data points in the node set; the larger the quantile aggregation value is, the more data points in the node set can reflect the same fan blade structure area, and the larger the information extraction characteristic value of the node set is.
In one embodiment of the present invention, the method for obtaining the quantile aggregate value includes:
Acquiring corresponding quantiles of each data point under the dimension of the distribution aggregation characteristic value; taking 25% quantiles as quantile aggregation values of the node set; the quantile aggregation value reflects the aggregation degree of smaller data in the node set; when the quantile aggregation value is smaller, the fact that partial data points exist in the node set is in a more scattered similar area is indicated, and the more difficult the data points in the node set reflect the same fan blade structure area.
It should be noted that, the method for obtaining the quantiles is a technical means well known to those skilled in the art, and is not described herein, the corresponding quantiles of the data points in the dimension of the distribution and aggregation feature value reflect the distribution situation of the data points in the distribution and aggregation feature value, for example, the corresponding 25% quantiles of the data points in the dimension of the distribution and aggregation feature value means that the value of the distribution and aggregation feature value of 25% of the data points in the dimension of the distribution and aggregation feature value is less than or equal to 25% quantiles, and the value of the distribution and aggregation feature value of 75% of the data points is greater than 25% quantiles.
In other embodiments of the present invention, the method for obtaining the quantile aggregate value includes:
Calculating the average value of the distribution aggregation characteristic values of all the data points in the node set to obtain a quantile aggregation value of the node set; the quantile aggregation value reflects the possibility that the integral node set is an aggregated similar area, and the greater the quantile aggregation value is, the greater the possibility that the integral node set is the aggregated similar area is represented, which means that the more data points in the node set can reflect the same fan blade structure area, and the greater the information extraction characteristic value of the node set is.
Acquiring an aggregation non-uniformity degree value of the node set according to the fluctuation of the distribution aggregation characteristic values of all data points in the node set; the larger the aggregation non-uniformity value is, the worse the stability of the distribution aggregation characteristic value of the data points in the node set is, and the smaller the information extraction characteristic value of the node set is.
In one embodiment of the invention, the variance of the distribution aggregation characteristic value of all data points in the node set is calculated to obtain the aggregation non-uniformity degree value of the node set. In other embodiments of the present invention, standard deviations of distribution aggregation feature values of all data points in the node set may also be calculated to obtain an aggregation non-uniformity value of the node set, which is not described herein.
Acquiring information extraction characteristic values of the node sets according to quantile aggregation values of the node sets and aggregation unevenness values of the node sets; the quantile aggregation value and the information extraction characteristic value are in positive correlation; the aggregation non-uniformity value and the information extraction characteristic value are in a negative correlation relationship.
The positive correlation relationship indicates that the dependent variable increases along with the increase of the independent variable, the dependent variable decreases along with the decrease of the independent variable, and the specific relationship can be multiplication relationship, addition relationship, idempotent of an exponential function and is determined by practical application; the negative correlation indicates that the dependent variable decreases with increasing independent variable, and the dependent variable increases with decreasing independent variable, which may be a subtraction relationship, a division relationship, or the like, and is determined by the actual application.
Obtaining an information extraction feature value according to an information extraction feature value formula, wherein the information extraction feature value formula comprises:
; wherein, Extracting characteristic values for the information of the node set; Aggregating values for quantiles of a node set; aggregate non-uniformity values for a set of nodes.
Preferably, in order to analyze the effect of the hierarchy corresponding cluster, in one embodiment of the present invention, the method for obtaining the classification effect value includes:
Acquiring a hierarchy extraction characteristic value according to the information extraction characteristic value of all node sets corresponding to the hierarchy; the larger the level extraction feature value is, the better the feature extraction capability of the level on the fan blade structure area is represented. In one embodiment of the invention, the minimum information extraction characteristic value is used as a level extraction characteristic value, and the larger the level extraction characteristic value is, the better the extraction capability of each node set corresponding to the level on the same fan blade structure area is represented. In other embodiments of the present invention, the average value of the information extraction feature values of all the node sets corresponding to the hierarchy is used as the hierarchy extraction feature value, and the larger the hierarchy extraction feature value is, the better the extraction capability of the whole node set corresponding to the hierarchy to the fan blade structure region is represented.
Acquiring a hierarchy extraction unstable value according to the fluctuation of the information extraction characteristic values of all node sets corresponding to the hierarchy; the larger the extraction instability value of the level is, the more dissimilar the extraction capability of the fan blade structure region corresponding to all the node sets corresponding to the level is represented, and the worse the feature extraction capability of the level to the fan blade structure region is represented. In one embodiment of the invention, the variance of the information extraction feature values of all node sets corresponding to the hierarchy is used as the hierarchy extraction unstable value. In other embodiments of the present invention, standard deviation of feature values extracted from information of all node sets corresponding to a hierarchy is used as an unstable value extracted from the hierarchy.
Acquiring a hierarchy extraction information value according to the hierarchy extraction characteristic value and the hierarchy extraction unstable value; the level extraction characteristic value and the level extraction information value are in positive correlation; the hierarchy extraction unstable value and the hierarchy extraction information value are in negative correlation; the larger the level extraction information value is, the better the feature extraction capability of the level on the fan blade structure area is represented.
Acquiring a level detail region value according to the degree of difference between the data point feature vectors of all node sets corresponding to the level; in one embodiment of the invention, the average value of the feature vectors of all data points in the node set is calculated and used as the feature average value vector of the node set; taking the modulus value of the difference value between the characteristic mean value vectors of the two node sets as the characteristic difference value of the two node sets; and calculating the average value of the characteristic difference values of all the two node sets corresponding to the hierarchy to obtain the detail difference value of the hierarchy. The larger the detail difference value of the level is, the larger the detail difference between different node sets corresponding to the representative level is, and the clustering result corresponding to the reflecting level has better distinguishing capability on different fan blade structure areas.
Acquiring a hierarchy detail unstable value according to fluctuation degrees of feature vectors corresponding to all data points in each node set corresponding to the hierarchy; in one embodiment of the invention, the variance of the eigenvectors of all data points in the node set is calculated and used as the eigenvalue of the node set; calculating the average value of the characteristic fluctuation values of all node sets corresponding to the hierarchy to obtain a hierarchy detail unstable value; the smaller the detail unstable value of the level is, the smaller the feature vector difference between the same node set in the representative level is, and the clustering result corresponding to the reflection level has better distinguishing capability on different fan blade structure areas.
Acquiring a characteristic distinguishing degree value of the hierarchy according to the hierarchy detail distinguishing value and the hierarchy detail unstable value; the level detail distinguishing value and the characteristic distinguishing degree value are in positive correlation; the unstable level detail value and the characteristic distinguishing degree value are in a negative correlation relationship; the larger the feature discrimination degree value is, the larger the feature vector distinction between different node sets corresponding to the representative level is, the smaller the feature vector distinction between the same node set is, and the clustering result corresponding to the reflection level has better discrimination capability on different fan blade structure areas.
Extracting an information value and a characteristic distinguishing degree value according to the hierarchy to obtain a classification effect value of the hierarchy; the classification effect value and the hierarchy extraction information value are in positive correlation; the classification effect value and the characteristic distinguishing degree value are in positive correlation. The larger the classification effect value is, the better the feature extraction capability of the level on the fan blade structure area is shown, the better the differentiation capability of the clustering result corresponding to the level on different fan blade structure areas is shown, and the better the clustering effect corresponding to the level is shown.
The positive correlation relationship indicates that the dependent variable increases along with the increase of the independent variable, the dependent variable decreases along with the decrease of the independent variable, and the specific relationship can be multiplication relationship, addition relationship, idempotent of an exponential function and is determined by practical application; the negative correlation indicates that the dependent variable decreases with increasing independent variable, and the dependent variable increases with decreasing independent variable, which may be a subtraction relationship, a division relationship, or the like, and is determined by the actual application.
Obtaining a classification effect value according to a classification effect value formula, wherein the classification effect value formula comprises:
; wherein, A classification effect value for a hierarchy; extracting characteristic values for information of all node sets corresponding to the hierarchy; Extracting the minimum value in the characteristic values for the information of all node sets corresponding to the hierarchy; Extracting variances of the eigenvalues for the information of all node sets corresponding to the hierarchy; To be in all node sets corresponding to the hierarchy The means of the feature vectors of all data points in the individual node sets; To be in all node sets corresponding to the hierarchy The means of the feature vectors of all data points in the individual node sets; The total number of all node sets corresponding to the hierarchy; To be in all node sets corresponding to the hierarchy Feature vectors of all data points in the node set; To at the first Feature vector for all data points in a set of individual nodesIs a variance of (2); taking a modular value symbol; Extracting information values for the hierarchy; a level detail region score; Is a level detail unstable value.
Preferably, considering that the larger the classification effect value is, the better the effect of the corresponding cluster of the representative hierarchy is, in order to analyze the best cluster effect corresponding hierarchy, in one embodiment of the present invention, the method for obtaining the best hierarchy includes:
and taking the hierarchy corresponding to the maximum classification effect value as the optimal hierarchy.
Step S3, according to the characteristic vector of the data points in the node set corresponding to the optimal level, acquiring the data weight of each data point; and measuring the profile of the fan blade according to the data weights of all the data points in the point cloud data set of the fan blade.
By the steps, the optimal level with the best clustering effect is determined, the optimal level corresponds to different node sets, different fan blade structure areas can be more accurately represented, data points in the same fan blade structure area have similar detail expression capability, and as the data points in different structure areas may have different detail expression capability, the importance degree of modeling is different. The data points are positioned in the flat areas of the fan blades, the detail expression capability is relatively weak, and the influence degree on the modeling result is small; the data points are located in the intersection areas of different structures of the fan blade or the concave-convex areas of the edges of the fan blade, the detail expressive power is relatively strong, and the influence degree of the modeling result is relatively high. Acquiring the data weight of each data point; the data points of the fan blade structure area with larger influence degree on the modeling result have larger data weight; therefore, data points with larger influence degree on modeling results are focused in subsequent analysis, and accuracy of high-precision measurement of the fan blade profile is improved.
Preferably, considering that data points with larger influence degree on a modeling result should be focused during modeling, the modeling result of the fan blade cannot reflect sufficient detail information of the fan blade due to detail loss in the modeling process, and high-precision measurement of the fan blade profile is influenced.
In all node sets corresponding to the optimal level, acquiring node vectors of the node sets according to the feature vectors of all data points in the node sets; in one embodiment of the invention, the average value of the feature vectors of all data points in the node set is used as the node vector of the node set.
And acquiring the data weight of the data point according to the node vector of the node set to which the data point belongs and the node vectors of all the node sets.
; Wherein,Data weights for data points; the average value of the feature vectors of all data points in the node set to which the data points belong; to correspond to node set in the best hierarchy The means of the feature vectors of all data points in the individual node sets; the total number of all node sets corresponds to the optimal hierarchy; Taking the modulus sign.
In the formula, because the optimal level corresponds to different node sets, different fan blade structure areas can be more accurately represented, data points in the same fan blade structure area have similar detail expression capability, in order to enable the data points in the fan blade structure area with larger influence degree on modeling results to have larger data weight,Reflects the detail expression degree of the fan blade structure area represented by the node set to which the data point belongs, the larger the value is, the more likely the fan blade structure area represented by the node set is the detail area,Reflecting the detail representation degree of the best level corresponding to all node sets byFor a pair ofAnd carrying out normalization, wherein the larger the data weight is, the larger the influence degree of the fan blade structural area to which the data point belongs on the modeling result is.
And measuring the profile of the fan blade according to the data weights of all the data points in the point cloud data set of the fan blade. Specifically, modeling is performed on point cloud data by using the data weights of all data points in a point cloud data set of the fan blade and adopting a grid generation algorithm with distance weighting, so as to obtain fan blade modeling. In this process, points with greater weight will have greater impact in the grid generation process. In one embodiment of the invention, the mesh generation algorithm with distance weighting can use a poisson surface reconstruction method.
And (3) carrying out fan blade profile measurement by using fan blade modeling, and calculating fan blade profile data such as curvature, length, bending degree and the like of the profile. And comparing the calculated fan blade profile data with the fan blade profile data of a standard ideal fan blade. During the comparison, error indicators, such as root mean square error, maximum error, etc., between the contours may be calculated to quantify the difference between the two. And analyzing the error index obtained by calculation to know the degree of difference and the distribution condition between the fan blade profile and the standard fan blade profile. The method is favorable for finding possible problems of the fan blade in the production or use process and making corresponding improvement measures or optimization schemes so as to improve the quality and performance of the fan blade.
In summary, the embodiment of the invention provides a high-precision measuring method for the profile of a fan blade of an industrial cooling fan, which comprises the steps of firstly obtaining all node sets corresponding to each level; in all node sets corresponding to the hierarchy, acquiring a classification effect value of the hierarchy according to the degree of difference between the feature vectors corresponding to the data points of all the node sets, the similarity degree of the feature vectors corresponding to all the data points in each node set and the fluctuation and the size of the information extraction feature values of all the node sets; further obtaining an optimal level; acquiring the data weight of each data point according to the characteristic vector of the data point in the node set corresponding to the optimal level; and finally, measuring the profile of the fan blade. According to the embodiment of the invention, the data weight capable of accurately reflecting the detail expression capability of the fan blade structural area where the data point is located is constructed, so that the accuracy of fan blade modeling for reflecting the fan blade profile is improved, and the effect of fan blade profile measurement is improved.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
1. The high-precision measuring method for the blade profile of the industrial cooling fan is characterized by comprising the following steps of:
Acquiring a point cloud data set of the fan blade; each data point in the point cloud data set has a corresponding feature vector;
in the point cloud data set, clustering all data points according to the difference of feature vectors and the space distribution interval among the data points to obtain all node sets corresponding to all levels; acquiring a distribution aggregation characteristic value of the data points according to the distribution characteristics of the node set to which the data points belong; acquiring information extraction characteristic values of the node set according to the size and fluctuation of the distribution aggregation characteristic values of the data points in the node set; in all node sets corresponding to the hierarchy, acquiring a classification effect value of the hierarchy according to the degree of difference between the feature vectors corresponding to the data points of all the node sets, the similarity degree of the feature vectors corresponding to all the data points in each node set and the fluctuation and the size of the information extraction feature values of all the node sets; obtaining an optimal level according to the classification effect value of the level;
Acquiring the data weight of each data point according to the characteristic vector of the data point in the node set corresponding to the optimal level; and measuring the profile of the fan blade according to the data weights of all the data points in the point cloud data set of the fan blade.
2. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 1, wherein the method for acquiring all node sets corresponding to the hierarchy comprises the following steps:
Based on hierarchical clustering method, determining the feature distance between two data points according to the difference of feature vectors between the two data points and the space distribution interval; and clustering all the data points according to the characteristic distance between every two data points to obtain all the node sets corresponding to each level.
3. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 1, wherein the method for obtaining the distribution aggregation characteristic value comprises the following steps:
Taking the data point as the data point to be analyzed; taking each data point in a node set to which the data point to be analyzed belongs as each reference data point of the data point to be analyzed; according to the distribution of the reference data points corresponding to the data points to be analyzed, a preset reference neighborhood range of the data points to be analyzed is constructed; and acquiring a distribution aggregation characteristic value of the data points to be analyzed according to the size of the preset reference neighborhood range of the data points to be analyzed and the duty ratio of the reference data points in the preset reference neighborhood range of the data points to be analyzed.
4. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 3, wherein the method for acquiring the distribution and aggregation characteristic value comprises the following steps:
acquiring the space distance of a preset reference neighborhood range of a data point to be analyzed;
Calculating the ratio between the total number of all the reference data points and the total number of all the data points in a preset reference neighborhood range of the data points to be analyzed to obtain a reference density characteristic value of the data points to be analyzed;
acquiring a distribution aggregation characteristic value of the data points to be analyzed according to the space distance and the reference density characteristic value; the space distance and the distribution aggregation characteristic value are in a negative correlation relationship; and the reference density characteristic value and the distribution aggregation characteristic value are in positive correlation.
5. The method for measuring the profile of a fan blade of an industrial cooling fan with high precision according to claim 1, wherein the method for acquiring the information extraction characteristic value comprises the following steps:
Acquiring a quantile aggregation value of the node set according to the size of the distribution aggregation characteristic value of the data points in the node set;
Acquiring an aggregation non-uniformity degree value of the node set according to the fluctuation of the distribution aggregation characteristic values of all the data points in the node set;
Acquiring an information extraction characteristic value of the node set according to the quantile aggregation value of the node set and the aggregation unevenness value of the node set; the quantile aggregation value and the information extraction characteristic value are in positive correlation; and the aggregation non-uniformity degree value and the information extraction characteristic value are in a negative correlation relationship.
6. The method for measuring the profile of a fan blade of an industrial cooling fan with high precision according to claim 1, wherein the method for obtaining the classification effect value comprises the following steps:
Acquiring a hierarchy extraction information value according to the size of the information extraction characteristic value of all node sets corresponding to the hierarchy and the fluctuation of the information extraction characteristic value of all node sets corresponding to the hierarchy;
Acquiring a level detail region value according to the degree of difference between the data point feature vectors of all node sets corresponding to the level;
Acquiring a hierarchy detail unstable value according to fluctuation degrees of the feature vectors corresponding to all data points in each node set corresponding to the hierarchy;
Acquiring a characteristic distinguishing degree value of the hierarchy according to the hierarchy detail distinguishing value and the hierarchy detail unstable value; the level detail distinguishing value and the characteristic distinguishing degree value are in positive correlation; the level detail unstable value and the characteristic distinguishing degree value are in a negative correlation relationship;
Extracting an information value and the characteristic distinguishing degree value according to the hierarchy to obtain a classification effect value of the hierarchy; the classification effect value and the hierarchy extraction information value are in positive correlation; and the classification effect value and the characteristic distinguishing degree value are in positive correlation.
7. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 6, wherein the method for acquiring the level extraction information value comprises the following steps:
acquiring a hierarchy extraction characteristic value according to the information extraction characteristic value of all node sets corresponding to the hierarchy;
Acquiring a hierarchy extraction unstable value according to the fluctuation of the information extraction characteristic values of all node sets corresponding to the hierarchy;
Acquiring a hierarchy extraction information value according to the hierarchy extraction characteristic value and the hierarchy extraction unstable value; the level extraction characteristic value and the level extraction information value are in positive correlation; the hierarchy extraction unstable value and the hierarchy extraction information value are in negative correlation.
8. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 1, wherein the method for acquiring the data weight comprises the following steps:
Acquiring node vectors of all the node sets in all the node sets corresponding to the optimal level according to the characteristic vectors of all the data points in the node sets;
And acquiring the data weight of the data point according to the node vector of the node set to which the data point belongs and the node vectors of all the node sets.
9. The method for measuring the profile of the fan blade of the industrial cooling fan with high precision according to claim 1, wherein the method for obtaining the optimal level comprises the following steps:
And taking the hierarchy corresponding to the maximum classification effect value as the optimal hierarchy.
10. The method for measuring the profile of a fan blade of an industrial cooling fan with high precision according to claim 3, wherein the method for obtaining the preset reference neighborhood range comprises the following steps:
And taking the data points to be analyzed as sphere centers, and constructing a preset reference neighborhood range, wherein the preset reference neighborhood range is the smallest sphere comprising the reference data points with preset reference quantity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410675307.5A CN118274740B (en) | 2024-05-29 | 2024-05-29 | High-precision measuring method for blade profile of industrial cooling fan |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410675307.5A CN118274740B (en) | 2024-05-29 | 2024-05-29 | High-precision measuring method for blade profile of industrial cooling fan |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118274740A true CN118274740A (en) | 2024-07-02 |
| CN118274740B CN118274740B (en) | 2024-09-10 |
Family
ID=91642266
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410675307.5A Active CN118274740B (en) | 2024-05-29 | 2024-05-29 | High-precision measuring method for blade profile of industrial cooling fan |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118274740B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050102301A1 (en) * | 2002-03-04 | 2005-05-12 | Adrian Flanagan | Mechanism for unsupervised clustering |
| CN112101393A (en) * | 2019-06-18 | 2020-12-18 | 上海电机学院 | Wind power plant fan clustering method and device |
| CN112284290A (en) * | 2020-10-20 | 2021-01-29 | 湖南大学 | Autonomous measurement method and system for aero-engine blade robot |
| CN114419085A (en) * | 2022-01-19 | 2022-04-29 | 中煤航测遥感集团有限公司 | Automatic building contour line extraction method and device, terminal device and storage medium |
| CN115439839A (en) * | 2022-08-04 | 2022-12-06 | 三峡大学 | Multi-type structural surface layered identification method based on point cloud data |
| CN116883606A (en) * | 2023-09-04 | 2023-10-13 | 安徽继远软件有限公司 | Three-dimensional model construction method and system based on laser point cloud |
-
2024
- 2024-05-29 CN CN202410675307.5A patent/CN118274740B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050102301A1 (en) * | 2002-03-04 | 2005-05-12 | Adrian Flanagan | Mechanism for unsupervised clustering |
| CN112101393A (en) * | 2019-06-18 | 2020-12-18 | 上海电机学院 | Wind power plant fan clustering method and device |
| CN112284290A (en) * | 2020-10-20 | 2021-01-29 | 湖南大学 | Autonomous measurement method and system for aero-engine blade robot |
| CN114419085A (en) * | 2022-01-19 | 2022-04-29 | 中煤航测遥感集团有限公司 | Automatic building contour line extraction method and device, terminal device and storage medium |
| CN115439839A (en) * | 2022-08-04 | 2022-12-06 | 三峡大学 | Multi-type structural surface layered identification method based on point cloud data |
| CN116883606A (en) * | 2023-09-04 | 2023-10-13 | 安徽继远软件有限公司 | Three-dimensional model construction method and system based on laser point cloud |
Non-Patent Citations (2)
| Title |
|---|
| 张辉: "基于曲面的点云直接分层算法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, no. 03, 15 March 2022 (2022-03-15) * |
| 解则晓;刘静晓;潘成成;张梦泽;: "一种散乱分层点云的有序化精简方法", 图学学报, no. 03, 15 June 2016 (2016-06-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118274740B (en) | 2024-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113344019A (en) | K-means algorithm for improving decision value selection initial clustering center | |
| Yang et al. | A Point Cloud Simplification Method Based on Modified Fuzzy C‐Means Clustering Algorithm with Feature Information Reserved | |
| CN112819299A (en) | Differential K-means load clustering method based on center optimization | |
| CN113221467B (en) | Turbine gas-thermal performance uncertainty visual analysis method and system | |
| CN115797418B (en) | Complex mechanical part measurement point cloud registration method and system based on improved ICP | |
| CN113436223B (en) | Point cloud data segmentation method and device, computer equipment and storage medium | |
| CN114676749A (en) | Power distribution network operation data abnormity judgment method based on data mining | |
| CN111291822A (en) | Equipment running state judgment method based on fuzzy clustering optimal k value selection algorithm | |
| CN111709454A (en) | Multi-wind-field output clustering evaluation method based on optimal copula model | |
| CN110310322B (en) | Method for detecting assembly surface of 10-micron-level high-precision device | |
| CN115081515A (en) | Energy efficiency evaluation model construction method and device, terminal and storage medium | |
| CN117252863A (en) | Quick detection and analysis method for geographic information abnormal data | |
| CN111353529A (en) | Mixed attribute data set clustering method for automatically determining clustering center | |
| CN118274740B (en) | High-precision measuring method for blade profile of industrial cooling fan | |
| JP5715445B2 (en) | Quality estimation apparatus, quality estimation method, and program for causing computer to execute quality estimation method | |
| CN117057201A (en) | Bayesian optimization-based turbine blade wax pattern data acquisition and prediction method | |
| CN117194872A (en) | Wind power plant wind speed correction method, system, computer equipment and storage medium | |
| Marx et al. | A machine learning based approach of performance estimation for high-pressure compressor airfoils | |
| CN108490912B (en) | A kind of multi-modal process modal identification method based on pivot similarity analysis | |
| CN113043274B (en) | Robot performance evaluation method and system | |
| CN111985826B (en) | Visual quality grading method and system for multi-index industrial products | |
| CN109101998B (en) | Clustering method and system based on spatial information of residential ground context | |
| CN118552528B (en) | Wafer warpage quick measurement method based on data analysis | |
| CN118626579B (en) | Intelligent geographic mapping data management method and system | |
| CN118196327B (en) | Intelligent green building construction information management method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |