CN104881548A - Discrete-domain force field based acquisition method of geometric evolution of adjacent process models - Google Patents

Abstract

The invention discloses a discrete-domain force field based acquisition method of geometric evolution of adjacent process models and solves the technical problem that the existing acquisition method of geometric evolution of process models is poor in practicality. According to the technical scheme, the method includes: subjecting the surface of a three-dimensional model to triangular meshing, and representing a current shape of the model surface according to a distribution status of stress borne by triangular facets of the surface of the three-dimensional model. The method is applicable to any type of three-dimensional models. effective matching facets of two adjacent process models are screened out by comparing model surface force field angles of view and model surface force field distances, evolutionary types of the facets of the process models are judged by measuring similarities of surface characteristic patterns of the effective matching facets, and the efficiency is higher. The model surface characteristic patterns are subjected to overall similarity measurement by an EMD (earth mover distance) distance comparison algorithm, the model surface characteristic patterns are subjected to regional comparison by a Frechet distanced-based similarity matching algorithm, and resolving power and precision are improved.

Description

Acquisition Method of Geometric Evolution of Adjacent Process Model Based on Discrete Domain Force Field

technical field

The invention relates to a method for acquiring the geometric evolution of a process model, in particular to a method for acquiring the geometric evolution of an adjacent process model based on a discrete domain force field.

Background technique

The document "Similarity Measurement of Geometric Evolution Sequence of 3D Process Model, Journal of Computer-Aided Design and Graphics, 2014, Vol26(7), p1176-1182" discloses a method for obtaining geometric evolution of process model. This method is aimed at the model obtained by the NURBS modeling method. According to the calculation of the control vertices of the NURBS model, a series of vectors with affine invariance are obtained. By comparing the series of vectors, the similarity of two adjacent process models is obtained. voxel, same voxel, identical voxel. On this basis, the attribute adjacency graph of voxels undergoing geometric changes is constructed, and the geometric evolution of two adjacent process models is expressed by the attribute adjacency graph. The method described in the literature is limited to the acquisition of the geometric evolution of the process model established by the NURBS method, and cannot be used for other types of process models; in addition, the method has a large amount of calculation when calculating a series of vectors with affine invariance, and the efficiency is not good. high.

Contents of the invention

In order to overcome the disadvantage of poor practicability of the existing method for obtaining the geometric evolution of the process model, the present invention provides a method for obtaining the geometric evolution of the adjacent process model based on the discrete domain force field. This method first divides two adjacent process models into triangular patches. It is required that the distribution of the divided triangular patches is basically uniform and the grid size is basically the same. Further, the two process models are solved on each surface of the two process models in the discrete domain force field. The force of the triangular surface, the force field angle, force field distance and surface feature map of each surface of the two process models are obtained, and the model is developed by comparing the force field angle, force field distance and surface feature map of each surface of the two models. The judgment of the surface geometric change type is to judge the concavo-convexity of the adjacent edge of the newly added surface and combine the type of the newly added surface to obtain the geometric evolution of the two adjacent process models. The geometric evolution of the obtained model is expressed as an extended attribute adjacency graph, and finally the extended attribute adjacency graph is expressed as an extended attribute adjacency matrix, so as to facilitate reading, storage and subsequent process knowledge retrieval in the computer. Based on the discrete domain force field to solve the geometric evolution of the adjacent process model, there is no requirement for the type of 3D model, that is, it is applicable to any type of 3D model; in the process of judging the type of geometric change of the model surface, through the perspective of the model surface force field , The comparison of the model surface force field distance screens out the effectively paired surfaces in two adjacent process models, and then only performs similarity measurement on the surface feature maps of the effective paired surfaces to judge the evolution type of the surfaces between the two process models, which improves the The efficiency of the method; the EMD distance comparison algorithm is used to measure the overall similarity of the model surface feature map, and the Fréchet distance similarity comparison algorithm is used to compare the area of the model surface feature map, so this method has high resolution and precision.

The technical solution adopted by the present invention to solve the technical problem is: a method for obtaining the geometric evolution of the adjacent process model based on the discrete domain force field, which is characterized in that the following steps are adopted:

(a) Divide the surfaces of two adjacent process models into triangular patches. It is required that the divided triangular patches are evenly distributed, the size of the patches is the same, and the consistency of the patch divisions on the surfaces of different process models is required. The surface of the process model is approximated by a set of planar triangle faces, and the change of the model surface is measured by the change of a limited number of small faces. In this way, any surface of each process model is represented as a particle set Q.

Q＝{q ₁ ,q ₂ ,…,q _n } ⑴

In the formula, q ₁ , q ₂ ,...,q _n are the centers of gravity of all triangles on any surface of the model, n is the number of triangles contained in the surface, and the coordinates of the center of gravity q _i are determined by the three vertices of the triangles calculated.

(b) Based on the discrete domain force field model, the surface characteristic map, the force field angle of view of the surface and the force field distance of the surface are solved for each surface of two adjacent process models. The distribution of gravitational force on different surfaces on the model is described by the surface feature map, the force field angle of view of the surface and the force field distance of the surface.

Define a discrete domain force field model:

Assuming that there is a mass point O at the center of gravity of the process model, and the coordinates of the mass point O do not change with the number of the process model in the same three-dimensional process specification, according to Newton’s third law, the elements q _i in the particle set Q on the surface of the object are uniform Under the gravitational action of the particle O, according to the formula of universal gravitation, we get:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}{{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In the formula, m _o and m _qi are the masses of particle O and q _i respectively, r _qi is the distance from particle O to q _i , and G is the gravitational constant.

Let m _qi =1, that is, element q _i is a unit mass point, then the model equation based on the discrete domain force field is as follows:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}}{{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}}{{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

In the formula, (x _o , y _o , z _o ) is the center of gravity of the process model, that is, the coordinates of the mass point O, and (x _qi , _yqi , z _qi ) is the coordinates of the element q _i .

Define a model surface feature map:

Based on the discrete domain force field model, the force of the center of gravity of all triangular patches on a certain surface of the model is solved, and the force is divided into several intervals according to the force of the surface, and the force interval is used as the abscissa. The number of the center of gravity of the force triangular patch is used as the ordinate to generate a frequency distribution histogram. The curve obtained by conformal interpolation after the scale normalization of the histogram is called the model surface feature map, denoted by the symbol T.

Define the force field perspective on the surface of the model:

force field view as vector and the model surface normal vector The formed angle is used to identify the orientation of different surfaces on the model in the force field, and is represented by the symbol θ.

Among them, the vector The line connecting the origin o of the coordinate system and the point O of the gravitational source points to O. It is stipulated here that the model surface normal vector is perpendicular to the model surface and the direction points to the outside of the model.

Define the force field distance on the surface of the model:

The force field distance is the distance from the particle O of the gravitational source to the surface of the model, which is used to identify the distance from the surface of the model to the gravitational source in the force field, represented by the symbol L.

(c) Changes in the surface gravitational force distribution of adjacent process models reflect the geometric evolution between the models. Through the comparison of the model surface force field angle and the model surface force field distance, the effectively paired surfaces in the two adjacent process models are selected, and then The similarity measurement is carried out on the surface feature map of the effective paired surface to judge the evolution type of the surface between the two process models. The similarity measurement of the surface feature map of the two surfaces adopts the EMD distance similarity comparison algorithm for overall comparison, and uses the Fréchet distance The similarity comparison algorithm performs regional comparison.

The evolution types of the surfaces of the two process models are divided into four types: modified surfaces, newly added surfaces, unchanged surfaces and disappearing surfaces. The modified surface and the unchanged surface do not participate in the construction of processing features, and the disappearing surface only exists in the previous process model of the two adjacent process models. The key to identifying the geometric evolution of the two adjacent process models is to find the new process model in the latter process model. Increase surface.

(d) Find the newly added surfaces in the process of geometric evolution of the two process models, judge the concavity and convexity of the adjacent edges of the newly added faces, obtain the topological relationship between the newly added faces, and combine the types of the newly added faces to obtain the two adjacent process models Geometric evolution.

(e) Express the geometric evolution of the acquired two adjacent process models as an extended attribute adjacency graph, and further express the extended attribute adjacency graph as an extended attribute adjacency matrix, which is convenient for computer reading, storage and subsequent process knowledge retrieval.

In order to express the geometric feature information of the geometric evolution of the model more completely, the attributes of the nodes and arcs or edges of the attribute adjacency graph are extended, and the types of faces and edges are added to obtain the extended attribute adjacency graph. Avoid the ambiguity expressed by the adjacency matrix by increasing the number of columns of the adjacency matrix and the number of digits of the element a _ij value, and add a column a[i,n+1] to the adjacency matrix a[n,n], the n+1th column Corresponds to the type attribute of the face f _i .

In the adjacency matrix a[n,n], the number of digits of a _ij is increased from two digits to three digits, the first two digits of a _ij still represent the concavity and convexity of the adjacent sides of faces f _i and f _j , and the newly added third digit The type of the adjacent edge of f _i and f _j .

The beneficial effects of the present invention are: the method of the present invention divides the triangular mesh on the surface of the three-dimensional model, and further, uses the distribution state of the gravitational force on the triangular surface of the model surface in the discrete domain force field to characterize the current model surface shape. The change of the gravitational force distribution on the surface of the adjacent process model represents the geometric evolution between the models, and there is no requirement for the type of the 3D model, that is, it is applicable to any type of 3D model. By comparing the angle of view of the model surface force field and the distance of the model surface force field, the effectively paired surfaces in two adjacent process models are screened out, and then the similarity measurement is performed on the surface feature maps of the effective paired surfaces to determine the surface between the two process models The evolution type of , which improves the efficiency of the method. The EMD distance comparison algorithm is used to measure the overall similarity of the model surface feature map, and the Fréchet distance similarity comparison algorithm is used to compare the area of the model surface feature map, so this method has high resolution and precision.

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

Description of drawings

Figure 1 is a flow chart of the method of the present invention.

Figure 2 is two process models a and b of adjacent processes.

Fig. 3 is a schematic diagram of models a and b after triangular mesh division.

Fig. 4 is a characteristic diagram of each surface of model a.

Fig. 5 is a characteristic diagram of each surface of model b.

Fig. 6 is an extended attribute adjacency graph representation of the geometric evolution acquisition results of the process models a and b.

Detailed ways

Refer to Figure 1-6. The steps of the method for obtaining the geometric evolution of the adjacent process model based on the discrete domain force field in the present invention are as follows:

The surfaces of two adjacent process model a and model b are expressed as:

F _a ＝{f _a ¹ ,f _a ² ,f _a ³ ,f _a ⁴ ,f _a ⁵ ,f _a ⁶ ,f _a ⁷ ,f _a ⁸ }

F _b ＝{f _b ¹ ,f _b ² ,f _b ³ ,f _b ⁴ ,f _b ⁵ ,f _b ⁶ ,f _b ⁷ ,f _b ⁸ ,f _b ⁹ ,f _b ¹⁰ }

In ANSYS software, the models a and b are divided into Delaunay triangular patches.

The division of Delaunay triangular patches ensures that the distribution of the divided triangular patches is basically even, the size of the patches is basically the same, and the consistency of patch division on the surface of different process models is ensured.

Calculate the stress on the triangular faces of each surface of model a and b, the force field angle and force field distance of each face. Obtain the barycentric coordinates of the triangular faces on each surface of the model a and b from the vertices of the triangular faces on the surfaces of the process model a and b, and then import the barycenter q _k coordinates of the triangular faces and the barycentric coordinates of the models a and b into the Matlab software Here, according to the discrete domain force field model, the triangular surface forces on each surface of the two models are calculated, and the force field viewing angle and force field distance of each surface of model a and b are obtained at the same time.

Table 1 Force field viewing angles and force field distances of each surface of model a

Table 2 Force field viewing angles and force field distances of each surface of model b

Construct the surface feature maps of models a and b. Construct the force distribution histograms of the triangular patches on the surfaces of models a and b based on the stress of the triangular patches on each surface of the model, and perform normalization processing, and use the shape-preserving interpolation function of Matlab software to perform shape-preserving interpolation on the histograms to obtain Surface feature maps for each surface of the two models.

Judging the change of the surface of the two-process model. By comparing the force field angles, force field distances and surface feature maps of the surfaces of models a and b, the changes of the model surfaces in the two processes are obtained. It is known from the results that f _b ⁹ and f _b ¹⁰ are newly added surfaces, and it is judged that the adjacent sides of f _b ⁹ and f _b ¹⁰ are circular and concave, and the geometry of the two adjacent process models can be obtained by combining the types of the two surfaces Evolved to blind hole manufacturing feature.

Table 3 Obtaining results of evolution types of each face of model b

Express the geometric evolution of the acquired two adjacent process models as an extended attribute adjacency graph, 1 and 2 in Fig. 6 represent faces f _b ⁹ and f _b ¹⁰ are nodes 1 and 2 of the extended attribute adjacency graph respectively; in 30 3 indicates that the surface f _b ⁹ is a cylindrical surface, 0 in 30 indicates that the inner surface of the surface f _b ⁹ constitutes a solid surface; 2 in 20 indicates that the surface f _b ¹⁰ is a plane, and the meaning of 0 is the same as that of 0 in 30; in 102 10 means that the faces f _b ⁹ and f _b ¹⁰ are adjacent, and the adjacent edges are concave edges, and 2 means that the type of adjacent edges is a circle, and then express the extended attribute adjacency graph as an extended adjacency matrix: $\left[\begin{array}{ccc}0& 102& 30\\ 102& 0& 20\end{array}\right].$

(a) Divide the surfaces of two adjacent process models into triangular patches. It is required that the distribution of the divided triangular patches is basically uniform, the size of the patches is basically the same, and the consistency of the patch divisions on the surfaces of different process models is required. According to the idea of differentiation, the surface of the process model is approximated by a set of plane triangle faces, and the change of the model surface is measured by the change of a limited number of small faces. In this way, any surface of each process model is represented as a particle set Q.

Q＝{q ₁ ,q ₂ ,…,q _n } ⑴

In the formula, q ₁ ,q ₂ ,…,q _n are the centers of gravity of all triangles on any surface of the model, n is the number of triangles contained in the surface, and the coordinates of the center of gravity q _i are calculated from the three vertices of the triangles get.

(b) Based on the discrete domain force field model, the surface feature map, the force field angle of the surface, and the force field distance of the surface are solved for each surface of two adjacent process models. The distribution of gravitational force on different surfaces on the model is described by the surface feature map, the force field angle of view of the surface and the force field distance of the surface.

Define a discrete domain force field model:

Assuming that there is a mass point O at the center of gravity of the process model, and the coordinates of the mass point O do not change with the number of the process model in the same three-dimensional process specification, according to Newton’s third law, the elements q _i in the particle set Q on the surface of the object are uniform Under the gravitational action of the particle O, according to the formula of universal gravitation, we get:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}{{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In the formula, m _o and m _qi are the masses of particle O and q _i respectively, r _qi is the distance from particle O to q _i , and G is the gravitational constant.

Let m _qi =1, that is, element q _i is a unit mass point, then the model equation based on the discrete domain force field is as follows:

$\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}}{{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; G</span>}\frac{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}}{{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; qi</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; z</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

In the formula (x _o , y _o , z _o ) is the center of gravity of the process model, that is, the coordinates of the mass point O, and (x _qi , y _qi , z _qi ) is the coordinate of the element q _i .

Define a model surface feature map:

Based on the discrete domain force field model, the force of the center of gravity of all triangular patches on a certain surface of the model is solved, and the force is divided into several intervals according to the force of the surface, and the force interval is used as the abscissa. The number of the center of gravity of the force triangular patch is used as the ordinate to generate a frequency distribution histogram. The curve obtained by conformal interpolation after the scale normalization of the histogram is called the model surface feature map, denoted by the symbol T.

Define the force field perspective on the surface of the model:

force field view as vector and the model surface normal vector The formed angle is used to identify the orientation of different surfaces on the model in the force field, and is represented by the symbol θ.

where the vector The line connecting the origin o of the coordinate system and the point O of the gravitational source points to O. It is stipulated here that the model surface normal vector is perpendicular to the model surface and the direction points to the outside of the model.

Define the force field distance on the surface of the model:

The force field distance is the distance from the particle O of the gravitational source to the surface of the model, which is used to identify the distance from the surface of the model to the gravitational source in the force field, represented by the symbol L.

(c) Changes in the surface gravitational force distribution of adjacent process models reflect the geometric evolution between the models. Through the comparison of the model surface force field angle and the model surface force field distance, the effectively paired surfaces in the two adjacent process models are selected, and then The similarity measurement is carried out on the surface feature map of the effective paired surface to judge the evolution type of the surface between the two process models. The similarity measurement of the surface feature map of the two surfaces adopts the EMD distance similarity comparison algorithm for overall comparison, and uses the Fréchet distance The similarity comparison algorithm performs regional comparison.

The evolution types of the surfaces of the two process models are divided into four types: modified surfaces, newly added surfaces, unchanged surfaces and disappearing surfaces. The modified surface and the unchanged surface do not participate in the construction of processing features, and the disappearing surface only exists in the previous process model of the two adjacent process models. Therefore, the key to identifying the geometric evolution of the two adjacent process models is to find the Add facets. Refer to Table 4 for the judgment basis and results of the model surface evolution type.

Table 4 Judgment basis and results of model surface evolution types

(d) Find the newly added surfaces in the process of geometric evolution of the two process models, judge the concavity and convexity of the adjacent edges of the newly added faces, obtain the topological relationship between the newly added faces, and combine the types of the newly added faces to obtain the two adjacent process models Geometric evolution.

(e) Express the geometric evolution of the acquired two adjacent process models as an extended attribute adjacency graph, and further express the extended attribute adjacency graph as an extended attribute adjacency matrix, which is convenient for computer reading, storage and subsequent process knowledge retrieval.

In order to more completely express the geometric feature information of the geometric evolution of the model, the attributes of the nodes and arcs (or edges) of the attribute adjacency graph are extended to a certain extent, and the types of faces and edges are added to obtain the extended attribute adjacency graph. Avoid the ambiguity expressed by the adjacency matrix by increasing the number of columns of the adjacency matrix and the number of digits of the element a _ij value, and add a column a[i,n+1] to the adjacency matrix a[n,n], the n+1th column Corresponding to the type attribute of surface f _i , the value definition refers to Table 5.

The definition of the attributes of the surface in Table 5 in the extended adjacency matrix

Table 6 Definition of the attributes of the adjacent edges of the face in the extended adjacency matrix

In the adjacency matrix a[n,n], increase the number of digits of a _ij from 2 to 3, the first two digits of a _ij still represent the concavity and convexity of the adjacent sides of faces f _i and f _j , and the newly added third digit Refer to Table 6 for the value definition of the adjacent edge type of f _i and f _j .

Claims (1)

1., based on the acquisition methods that the adjacent operation model geometric in the discrete domain field of force develops, it is characterized in that comprising the following steps:

A tri patch division is carried out on () surface to adjacent two operation models, require that the tri patch divided is evenly distributed, the consistance that patch dimensions identical and different operation model surface upper panel divides; Operation model surface approaches expression by the set of plane triangle dough sheet, and the change of model surface is measured with the change of limited fettucelle; Like this, arbitrary surface of each operation model is just expressed as a particle set Q;

Q＝{q ₁,q ₂,…,q _n} ⑴

In formula, q ₁, q ₂..., q _nfor the center of gravity of all tri patchs in the arbitrary surface of model, n is the number of the tri patch that this surface comprises, center of gravity q _icoordinate calculated by three summits of triangular facet;

B () solves the surface characteristics figure on adjacent two each surfaces of operation model, the visual angle, the field of force on surface and the field of force distance on surface based on discrete domain force field model; By surface characteristics figure, the visual angle, the field of force on surface and the field of force on the surface distribution apart from gravitation suffered by different surfaces on common descriptive model, the shape on each surface of distribution characterization model of model surface gravitation;

Definition discrete domain force field model:

There is a particle O in the center of gravity place being located at operation model, and the coordinate of particle O does not change with the change of operation pattern number in same three-dimensional process code, according to Newton third law, and the element q in body surface set of particles Q _iall be subject to the universal gravitation effect of this particle O, according to Formula of Universal Gravitation:

$F\left(q_i\right)=G\frac{m_o{m}_{\mathrm{qi}}}{{r_{\mathrm{qi}}}^2}---\left(2\right)$

In formula, m _oand m _qibe respectively particle O and q _iquality, r _qifor particle O to q _idistance, G is gravitational constant;

Make m _qi=1, even element q _ifor unit particle, then as follows based on discrete domain force field model equation:

$F\left(q_i\right)=G\frac{m_o}{{r_{\mathrm{qi}}}^2}=G\frac{m_o}{{(x_{\mathrm{qi}}-x_o)}^2+{(y_{\mathrm{qi}}-y_o)}^2+{(z_{\mathrm{qi}}-z_o)}^2}---\left(3\right)$

In formula, (x _o, y _o, z _o) be the center of gravity of operation model and the coordinate of particle O, (x _qi, y _qi, z _qi) be element q _icoordinate;

Definition Model surface characteristics figure:

Based on the stressed size of all tri patch centers of gravity on a certain surface of discrete domain force field model solving model, several intervals are divided into according to the stressed size of large young pathbreaker that dough sheet is stressed, using the interval of power as horizontal ordinate, in interval, the number of stressed tri patch center of gravity is as ordinate, generates frequency distribution histogram; Being called model surface characteristic pattern to carrying out the curve that conforming interpolation obtains after histogrammic dimension normalization process, representing with symbol T;

The visual angle, the field of force on Definition Model surface:

Visual angle, the field of force is vector with model surface normal vector the angle formed, in order to the orientation of different surfaces on identification model in the field of force, represents by symbol theta;

Wherein, vector for the line of coordinate origin o and gravitation source particle O points to O; Here specify, model surface normal vector is perpendicular to model surface and outside the direction model of direction;

The field of force distance on Definition Model surface:

Field of force distance is the distance of gravitation source particle O to model surface, represents to the distance in gravitation source with symbol L in order to be identified at model surface in the field of force;

C geometry that the change of () adjacent operation model surface gravitation distribution embodies between model develops, by filtering out the face of effectively pairing in two adjacent operation models to the comparison at visual angle, the model surface field of force and model surface field of force distance, then similarity measurement is carried out to judge the differentiation type in face between two operation models to the surface characteristics figure in effective pairing face, the similarity measurement of the surface characteristics figure in two faces adopts EMD distance similarity comparison algorithm to carry out overall comparison, adopts Fr é chet distance similarity alignment algorithm to carry out regional correlation;

The differentiation type of two operation model surfaces is divided into four kinds: amendment face, newly-increased face, invariable plane and extinction face; Amendment face and invariable plane do not participate in the structure of machining feature, and extinction face only exists in the previous operation model of two adjacent operation models, identify that the key that two adjacent operation model geometrics develop finds the newly-increased face in a rear operation model;

D () finds the newly-increased face in two procedures model geometric evolution process, judge the concavity and convexity of newly-increased adjacent side, face, obtain the topological relation between newly-increased face, and the geometry that the type in conjunction with newly-increased face obtains adjacent two procedures model develops;

E the geometry of the adjacent two procedures model obtained is developed the attribute adjacent map being expressed as expansion by (), further extended attribute adjacent map is expressed as the attribute adjacency matrix expanded, and facilitates computing machine to read, stores and the retrieval of follow-up process knowledge;

In order to the geometric properties information of expression model geometry differentiation that can be more complete, the node of attribute adjacent map and the attribute on arc or limit are expanded, addition of the type in face, the type on limit, the attribute adjacent map be expanded; By increasing columns and the element a of adjacency matrix _ijthe ambiguity that the figure place of value avoids adjacency matrix to express, adds row a [i, n+1], the (n+1)th row corresponding surface f in adjacency matrix a [n, n] _itype attribute;

In adjacency matrix a [n, n], by a _ijfigure place be increased to three by two, a _ijfront two is presentation surface f still _iand f _jthe concavity and convexity of adjacent side, newly-increased the 3rd for face f _iand f _jthe type of adjacent side.