Valve casting wax pattern shrinkage compensation method

Abstract

This application relates to the field of valve casting production technology, and particularly to a method for compensating the shrinkage rate of valve casting wax patterns. The method includes: determining the structural feature boundary points of the wax pattern based on three-dimensional structural data, and dividing the pattern into thick-walled shrinkage regions, thin-walled shrinkage regions, and edge shrinkage regions based on these boundary points; calculating a thick-walled shrinkage compensation benchmark value based on the thick-walled shrinkage ratio coefficient of the thick-walled shrinkage region and the corresponding first region dimension parameters; calculating a thin-walled shrinkage compensation benchmark value based on the thin-walled shrinkage ratio coefficient of the thin-walled shrinkage region and the corresponding second region dimension parameters; calculating an edge shrinkage compensation benchmark value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region dimension parameters; and obtaining the shrinkage compensation value corresponding to each shrinkage region based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value. This method significantly shortens the mold debugging cycle and reduces trial molding costs while improving compensation accuracy.

Description

Technical Field

[0001] This application belongs to the field of valve casting production technology, and in particular relates to a method for compensating for the shrinkage rate of valve casting wax molds. Background Technology

[0002] Valve castings are mostly manufactured using investment casting. The wax model, as the pre-casting substrate, directly determines the quality and assembly performance of the final casting. During preparation, cooling, and demolding, the wax model undergoes varying degrees of shrinkage deformation. Due to its structure, the wall thickness varies significantly between the thick-walled main body of the valve body, the thin-walled inner cavity, the flange, and the contour edges. The shrinkage rate and amount exhibit distinct characteristics in different areas. Without targeted compensation, this can easily lead to localized dimensional deviations and contour deformation of the wax model, ultimately resulting in the scrapping of the casting and increased production costs.

[0003] Currently, the shrinkage compensation of valve casting wax patterns generally adopts a global uniform compensation method based on empirical formulas. The global uniform compensation method treats areas with different wall thicknesses equally and uses the same compensation coefficient for proportional scaling. This inevitably leads to insufficient compensation in thick-walled areas, resulting in smaller final casting dimensions, and overcompensation in thin-walled areas, resulting in larger final casting dimensions. To address the shortcomings of global uniform compensation, the industry has tried an improved approach of zonal compensation. This involves artificially dividing the wax pattern into several areas according to differences in wall thickness and setting different shrinkage compensation coefficients for different areas. The zonal compensation scheme in the current scheme relies entirely on the visual judgment and experience of process engineers to define the zonal boundaries. After zoning, the shrinkage compensation value of each area usually depends on iterative cycles of multiple trial moldings, measurements, and mold repairs. This results in long trial molding cycles, high costs, and difficulty in adapting to the production needs of multiple varieties and variable batches. Summary of the Invention

[0004] This application provides a method for compensating the shrinkage rate of valve casting wax molds, which can solve the problems of long trial molding cycles, high costs, and difficulty in adapting to the production needs of multiple varieties and variable batches in the existing partition compensation scheme.

[0005] In a first aspect, embodiments of this application provide a method for compensating for the shrinkage rate of a valve casting wax pattern, including: Obtain the three-dimensional structural data of the wax model of the target valve casting; wherein, the three-dimensional structural data is used to indicate the wall thickness, contour dimensions and structural node information of each part of the wax model; Based on the three-dimensional structural data, the structural feature boundary points of the wax model are determined, and based on the structural feature boundary points, the thick-walled shrinkage region, the thin-walled shrinkage region, and the edge shrinkage region are divided. The thick-walled shrinkage compensation reference value is calculated based on the thick-walled shrinkage ratio coefficient of the thick-walled shrinkage region and the corresponding first region size parameter; the thin-walled shrinkage compensation reference value is calculated based on the thin-walled shrinkage ratio coefficient of the thin-walled shrinkage region and the corresponding second region size parameter; the edge shrinkage compensation reference value is calculated based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region size parameter. Based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value, shrinkage fitting calculation parameters suitable for the current wax mold are determined, and a shrinkage rate adaptive fitting algorithm is used to obtain the shrinkage compensation value corresponding to each shrinkage domain.

[0006] The technical solutions described in this application embodiment have at least the following technical effects: The valve casting wax model shrinkage compensation method provided in this application obtains three-dimensional structural data of the target valve casting wax model, characterizing the wall thickness, contour dimensions, and structural node information of various parts of the wax model. Then, based on the three-dimensional structural data, it determines the structural feature boundary points and divides the wax model into thick-walled shrinkage regions, thin-walled shrinkage regions, and edge shrinkage regions. It calculates the thick-walled shrinkage compensation benchmark value based on the thick-walled shrinkage ratio coefficient of the thick-walled shrinkage region and the dimensional parameters of the first region; it calculates the thin-walled shrinkage compensation benchmark value based on the thin-walled shrinkage ratio coefficient of the thin-walled shrinkage region and the dimensional parameters of the second region; and it calculates the edge shrinkage compensation benchmark value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the dimensional parameters of the third region. This method incorporates the differences in shrinkage characteristics between different wall thickness regions into the compensation using quantified independent coefficients. The calculation fundamentally overcomes the accuracy problems caused by insufficient compensation for thick-walled wax molds with uneven wall thickness distribution due to undercompensation and overcompensation for thin-walled wax molds caused by global uniform compensation. Based on three shrinkage compensation benchmark values, the shrinkage fitting calculation parameters adapted to the current wax mold are determined, and the shrinkage compensation value corresponding to each shrinkage domain is obtained by using an adaptive shrinkage rate fitting algorithm. This allows the determination of compensation values ​​to be automatically calculated based on the current wax mold structural characteristics, eliminating the need for repeated trial molding iterations for each new specification product. By automatically partitioning based on three-dimensional structural data and using an adaptive fitting algorithm, the shrinkage compensation value of each region is automatically determined, eliminating the dependence on manual experience in partitioning and compensation value determination, reducing the number of trial molding iterations, improving compensation accuracy, significantly shortening the mold debugging cycle, and reducing trial molding costs.

[0007] Secondly, embodiments of this application provide a valve casting wax mold shrinkage compensation system, comprising: The acquisition module is used to acquire the three-dimensional structural data of the wax model of the target valve casting; wherein, the three-dimensional structural data is used to indicate the wall thickness, contour dimensions and structural node information of each part of the wax model; The first determining module is used to determine the structural feature boundary points of the wax model based on the three-dimensional structural data, and to divide the thick-walled shrinkage region, thin-walled shrinkage region and edge shrinkage region based on the structural feature boundary points. The calculation module is used to calculate the thick-wall shrinkage compensation reference value based on the thick-wall shrinkage ratio coefficient of the thick-wall shrinkage region and the corresponding first region size parameter; calculate the thin-wall shrinkage compensation reference value based on the thin-wall shrinkage ratio coefficient of the thin-wall shrinkage region and the corresponding second region size parameter; and calculate the edge shrinkage compensation reference value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region size parameter. The second determining module is used to determine the shrinkage fitting calculation parameters adapted to the current wax mold based on the thick-wall shrinkage compensation benchmark value, the thin-wall shrinkage compensation benchmark value and the edge shrinkage compensation benchmark value, and to obtain the shrinkage compensation value corresponding to each shrinkage domain by using the shrinkage rate adaptive fitting algorithm.

[0008] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.

[0009] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect above.

[0010] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to execute the valve casting wax mold shrinkage compensation method described in the first aspect above.

[0011] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart illustrating the valve casting wax mold shrinkage compensation method provided in the embodiments of this application; Figure 2 This is a schematic diagram illustrating the effect of the wall thickness variation curve provided in the embodiments of this application; Figure 3 This is a schematic diagram of the valve casting wax mold shrinkage compensation system provided in the embodiments of this application; Figure 4This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0014] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0015] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0016] Currently, the shrinkage compensation of valve casting wax patterns generally adopts a global uniform compensation method based on empirical formulas. The global uniform compensation method treats areas with different wall thicknesses equally and uses the same compensation coefficient for proportional scaling. This inevitably leads to insufficient compensation in thick-walled areas, resulting in smaller final casting dimensions, and overcompensation in thin-walled areas, resulting in larger final casting dimensions. To address the shortcomings of global uniform compensation, the industry has tried an improved approach of zonal compensation. This involves artificially dividing the wax pattern into several areas according to differences in wall thickness and setting different shrinkage compensation coefficients for different areas. The zonal compensation scheme in the current scheme relies entirely on the visual judgment and experience of process engineers to define the zonal boundaries. After zoning, the shrinkage compensation value of each area usually depends on iterative cycles of multiple trial moldings, measurements, and mold repairs. This results in long trial molding cycles, high costs, and difficulty in adapting to the production needs of multiple varieties and variable batches.

[0017] To address the aforementioned problems, this application provides a method for compensating the shrinkage rate of a valve casting wax model. This method involves acquiring three-dimensional structural data of the target valve casting wax model, characterizing the wall thickness, contour dimensions, and structural node information of various parts of the wax model; then, based on the three-dimensional structural data, determining the structural feature boundary points and dividing the model into thick-walled shrinkage regions, thin-walled shrinkage regions, and edge shrinkage regions; calculating the thick-walled shrinkage compensation benchmark value based on the thick-walled shrinkage ratio coefficient of the thick-walled shrinkage region and the size parameters of the first region; calculating the thin-walled shrinkage compensation benchmark value based on the thin-walled shrinkage ratio coefficient of the thin-walled shrinkage region and the size parameters of the second region; and calculating the edge shrinkage compensation benchmark value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the size parameters of the third region. By introducing the differences in shrinkage characteristics of different wall thickness regions into the compensation calculation with quantified independent coefficients, the problem of insufficient accuracy caused by global uniform compensation for thick-walled wax molds with uneven wall thickness distribution and overcompensation for thin-walled wax molds is fundamentally overcome. Based on three shrinkage compensation benchmark values, the shrinkage fitting calculation parameters adapted to the current wax mold are determined, and the shrinkage compensation value corresponding to each shrinkage domain is obtained by using the shrinkage rate adaptive fitting algorithm. This allows the determination of the compensation value to be automatically calculated based on the current wax mold structural characteristics, eliminating the need for repeated trial molding and iteration for each new specification product. This significantly shortens the mold debugging cycle and reduces trial molding costs while improving compensation accuracy.

[0018] The valve casting wax mold shrinkage compensation method provided in this application embodiment can be applied to electronic devices. In this case, the electronic device is the executing subject of the valve casting wax mold shrinkage compensation method provided in this application embodiment. This application embodiment does not impose any restrictions on the specific type of electronic device.

[0019] For example, electronic devices can be mobile phones, tablets, wearable devices, augmented reality (AR) / virtual reality (VR) devices, laptops, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), desktop computers, smart screens, smart TVs, and other terminal devices; handheld devices with wireless communication capabilities; computing devices or other processing devices connected to a wireless modem; Internet of Things (IoT) terminals; computers; laptops; handheld communication devices; handheld computing devices; satellite wireless devices; wireless modem cards; set-top boxes (STBs); customer premises equipment (CPEs); and / or other devices used for communication over wireless systems, as well as next-generation communication systems, such as mobile terminals in 5G networks or mobile terminals in future evolved Public Land Mobile Networks (PLMNs).

[0020] To better understand the valve casting wax pattern shrinkage compensation method provided in this application embodiment, the specific implementation process of the valve casting wax pattern shrinkage compensation method provided in this application embodiment will be described by way of example below.

[0021] Figure 1 A schematic flowchart of a valve casting wax pattern shrinkage compensation method provided in an embodiment of this application is shown. The valve casting wax pattern shrinkage compensation method includes: S100, acquire the three-dimensional structural data of the wax model of the target valve casting; wherein, the three-dimensional structural data is used to indicate the wall thickness, contour dimensions and structural node information of each part of the wax model.

[0022] It can be understood that a target valve casting wax model refers to a pre-made model of paraffin or resin-based metal casting, prepared according to the precision casting process requirements for a specific valve product specification. Three-dimensional structural data refers to a set of data representing the complete geometric shape of the wax model in digital form, obtained through non-contact or contact measurement methods. This includes the spatial coordinates of each discrete sampling point on the wax model surface, as well as wall thickness distribution, contour dimensions, and structural node information derived from point cloud data. Wall thickness refers to the material thickness of the wax model from the outer surface to the inner cavity surface or the relative outer surface along its surface normal direction. Wall thickness distribution in valve casting wax models is usually uneven; the difference in wall thickness at flange connections, valve body flow channel transition areas, and edge fillets can be several to tens of times. Contour dimensions refer to the maximum external dimensions of the wax model in three orthogonal directions, including total length, total width, and total height. Structural node information refers to node data extracted from the 3D scanned point cloud that reflects abrupt changes in the geometric features of the wax model, including points of abrupt changes in contour curvature, points of abrupt changes in wall thickness, and geometric corner vertices. By acquiring the three-dimensional structural data of the wax model of the target valve casting, comprehensive basic data support is provided for subsequent analysis of the wall thickness distribution characteristics of various parts of the wax model, determination of structural feature boundary points, and differentiated shrinkage rate compensation for different wall thickness areas.

[0023] For example, the acquisition of 3D structural data can be achieved in the following manner. The data acquisition equipment can be a blue light structured light 3D scanner or a laser line scanning 3D measurement system. The scanning accuracy of the blue light structured light 3D scanner is not less than 0.02 mm, the single-frame scanning range is not less than 200 mm x 150 mm, and the scanning point spacing is not less than 0.04 mm. Before scanning, the wax model needs to undergo surface pretreatment. A layer of titanium dioxide developer with a thickness not exceeding 5 micrometers is uniformly sprayed onto the surface of the wax model to eliminate the interference of the translucency of the paraffin material and surface reflection on the scanning accuracy. The wax model is placed on a rotatable scanning turntable. The turntable rotates in 30-degree increments. After each increment, the 3D scanner acquires one frame of point cloud data. A total of 12 frames of point cloud data are acquired after one rotation. After scanning, the point cloud data frames are registered and stitched together using a stitching algorithm to form a complete 3D point cloud of the wax model. Point cloud post-processing includes: using statistical filtering algorithms to remove outliers exceeding three standard deviations from the neighborhood mean; using moving least squares to smooth the point cloud to eliminate microscopic ripple noise generated during scanning; and using a Poisson surface reconstruction algorithm to convert the smoothed point cloud data into a continuous watertight triangular mesh model, with the side length of each triangular facet not exceeding 0.1 mm. On the triangular mesh model, a wall thickness distribution map is obtained by calculating the local thickness of the wax model at the center of each triangular facet: for each facet, a ray is emitted along the facet's normal direction, and the first intersection point between the ray and the other side of the triangular mesh model is found; the Euclidean distance from the facet's center to this intersection point is the local wall thickness at that location. The contour dimensions are obtained by establishing a minimum bounding box on the triangular mesh model, calculating the maximum and minimum values ​​of the point cloud along the X, Y, and Z coordinate axes; the differences are the contour dimensions in the three orthogonal directions. Structural node information is obtained by detecting curvature extrema and boundary feature points of the triangular mesh model. Specifically, a node extraction algorithm based on principal curvature analysis is used to calculate the average curvature and Gaussian curvature of each vertex. Vertices with an absolute average curvature value greater than a preset curvature threshold are marked as structural nodes, and the 3D coordinates and local wall thickness of each structural node are recorded. All the above 3D structural data are stored in the wax model structure data table in the host computer database. Each record contains fields such as node number, 3D coordinates, local wall thickness, node type label, and timestamp, providing structured input data for subsequent wall thickness mutation analysis, shrinkage domain division, and compensation value calculation.

[0024] It should be noted that in existing technologies, shrinkage compensation for valve casting wax models typically employs a global uniform compensation method based on empirical formulas. This involves proportionally enlarging the overall size of the wax model by multiplying it by a fixed shrinkage compensation coefficient. However, the wall thickness distribution of valve casting wax models is extremely uneven. The cooling rate and shrinkage of thick-walled areas are far greater than those of thin-walled areas. The difference in shrinkage between flange areas with an average wall thickness exceeding 20 mm and valve body thin-walled areas with an average wall thickness of less than 5 mm can be 2 to 3 times. The global uniform compensation method treats areas of different wall thicknesses equally, inevitably leading to undercompensation in thick-walled areas, resulting in smaller final casting dimensions, and overcompensation in thin-walled areas, resulting in larger final casting dimensions. This fails to meet the accuracy requirements of valve castings for the dimensional tolerances of mating surfaces (typically within ±0.5%). Some improved solutions adopt a zoned compensation approach in mold design, but the zone boundaries rely entirely on the visual judgment and experience of process engineers, lacking quantitative basis based on 3D scanning data. This results in poor consistency and repeatability of the zones, and significant fluctuations in compensation effects after personnel or production line changes. This method collects three-dimensional structural data from wax molds, quantitatively representing wall thickness distribution, contour dimensions, and structural node information in the form of digital point clouds and triangular meshes. This ensures that subsequent wall thickness mutation analysis, determination of structural feature boundary points, and shrinkage domain division are all based on objective measurement data. It fundamentally overcomes the defects of unstable compensation effects and non-repeatability caused by relying on manual experience, providing a reliable data foundation for precision shrinkage compensation of valve casting wax molds.

[0025] S200 determines the structural feature boundary points of the wax model based on three-dimensional structural data, and divides the thick-walled shrinkage region, thin-walled shrinkage region and edge shrinkage region based on the structural feature boundary points.

[0026] It can be understood that structural feature demarcation points refer to the geometric locations in the three-dimensional structural data of wax models where significant changes occur in wall thickness distribution, serving as the basis for dividing different shrinkage zones. Thick-walled shrinkage zones refer to areas in the wax model with a large average wall thickness and correspondingly large shrinkage during cooling and solidification. Typical examples include flange connection surfaces, valve body reinforcing ribs, and valve seat mounting bases—areas with concentrated wall thickness. These areas have high heat capacity, slow cooling rates, and long solidification shrinkage times, requiring significant shrinkage compensation. Thin-walled shrinkage zones refer to areas in the wax model with a small average wall thickness and correspondingly smaller shrinkage. Typical examples include thin-walled cylindrical sections of valve body flow channels. These areas have low heat capacity, fast cooling rates, and relatively small shrinkage, requiring relatively small shrinkage compensation. The edge shrinkage region refers to the geometric transition zone between thick-walled and thin-walled areas in a wax model. It includes areas of gradual wall thickness change, rounded corner transition areas, and contour edge areas. The wall thickness in this region exhibits a continuous change from thick to thin or from thin to thick. Its shrinkage behavior combines some characteristics of thick-walled and thin-walled shrinkage, but is not entirely identical to either, requiring an independent compensation strategy. By determining the structural feature boundary points and dividing the wax model into thick-walled, thin-walled, and edge shrinkage regions, the wax model is discretized into three independent regions with different shrinkage characteristics. This provides a basis for calculating the shrinkage compensation benchmark and compensation values ​​for each region, forming the foundation for differentiated shrinkage compensation by region.

[0027] It should be noted that the core problem with existing zoning compensation technologies lies in the fact that zoning boundaries rely solely on manual visual judgment, lacking quantifiable and reproducible objective demarcation criteria. Since the wall thickness transition of valve wax models is often not an abrupt step but a gradual slope, manual visual inspection struggles to accurately define the precise boundary between "thick" and "thin" sections, leading to inconsistent zoning results for the same valve model across different batches or among different operators. This method automatically extracts structural feature boundary points from three-dimensional structural data, using the location of abrupt wall thickness change nodes as objective, quantitative demarcation criteria. The location of these boundary points is automatically determined by data, unaffected by subjective operator judgment, thus improving the consistency and repeatability of zoning results and laying a reliable foundation for subsequent precise compensation.

[0028] In one possible implementation, step S200, determining the structural feature boundary points of the wax model based on three-dimensional structural data, includes: S210, based on the three-dimensional structural data, determine the location of the wall thickness abrupt change node, and based on the location of the wall thickness abrupt change node, determine the structural feature boundary point of the wax model.

[0029] It can be understood that the location of abrupt changes in wall thickness refers to the spatial location where a significant step change in the local wall thickness occurs when traversing the contour surface of the wax model in a 3D triangular mesh model. In valve casting wax models, abrupt changes in wall thickness typically occur at geometrically discontinuous locations such as the junction of the flange root and the valve body wall, the intersection of the end of the reinforcing rib and the valve body surface, and the transition between the edge of the valve seat boss and the flow channel wall. Determining the structural feature boundary point based on the location of abrupt changes in wall thickness uses wall thickness, a physical quantity directly related to shrinkage behavior, as the basis for boundary judgment. This gives the location of the boundary point a clear physical meaning; the difference in wall thickness on both sides indicates that the cooling and shrinkage behaviors on both sides are fundamentally different, thus providing a physical basis for dividing the two sides into different shrinkage domains.

[0030] Dividing the wax pattern into thick-walled, thin-walled, and edge shrinkage zones based on structural feature boundary points is a process of allocating the entire geometric surface of the wax pattern to three different shrinkage zones according to wall thickness characteristics. Thick-walled shrinkage zones correspond to areas with larger wall thickness amplitudes, thin-walled shrinkage zones to areas with smaller wall thickness amplitudes, and edge shrinkage zones to the wall thickness transition areas immediately adjacent to the structural feature boundary points. The edge shrinkage zone serves as a buffer zone between the thick-walled and thin-walled zones. Its necessity lies in the fact that the structural feature boundary point is a single node position determined by the minimum point on the wall thickness amplitude variation curve. The actual wall thickness change at the boundary point is not an infinitely steep zero-width step, but rather has a certain spatial transition range. If the areas on both sides of the boundary point are directly assigned to the thick-walled and thin-walled zones respectively without setting an edge shrinkage zone, the nodes in the transition area near the boundary line may have their shrinkage zone assignment repeatedly changed due to small fluctuations in wall thickness, leading to discontinuities in the compensation results near the boundary line, thus affecting the smoothness of the final casting surface.

[0031] For example, the partitioning process is as follows: For each triangular facet on the wax model's triangular mesh, calculate the wall thickness at the facet's center point. If the wall thickness is greater than a preset thick-wall threshold (1.3 times the average wall thickness of the entire wax model), the facet is initially assigned to the thick-wall candidate region; if the wall thickness is less than a preset thin-wall threshold (0.7 times the average wall thickness of the entire wax model), the facet is initially assigned to the thin-wall candidate region; otherwise, the facet is assigned to the intermediate transition candidate region. Then, for each boundary trajectory formed by connecting structural feature boundary points, extend a preset edge width (3% to 8% of the wax model's outline size) to both sides along the normal direction, centered on the boundary trajectory. All faces within the extended range are forcibly assigned to the edge contraction region, regardless of whether the wall thickness at the facet's center point falls within the thick-wall or thin-wall candidate region. Finally, the faces in the thick-wall candidate region not covered by the edge contraction region are ultimately determined as the thick-wall contraction region, and the faces in the thin-wall candidate region not covered by the edge contraction region are ultimately determined as the thin-wall contraction region. This two-step partitioning strategy, which involves "first coarsely dividing according to wall thickness and then forcibly delineating the edge domains centered on the boundary trajectory," ensures that the three contraction domains have significant differences in wall thickness statistically, while also guaranteeing the smoothness and structural rationality of the domain boundaries. This provides spatially continuous regional partitioning results with reasonable transition zones for subsequent zoning compensation.

[0032] For example, the determination of the location of the wall thickness abrupt change node can be achieved through the following process. After loading the wax model triangular mesh model into the 3D data analysis software configured in the host computer, multiple wall thickness sampling path lines are generated along the principal curvature direction of the wax model contour surface. At least 100 wall thickness sampling points are arranged at equal intervals on each path line. For each wall thickness sampling point, the local wall thickness value at the sampling point is obtained according to the local wall thickness calculation method described in S100. The wall thickness difference between two adjacent sampling points is scanned point by point along each path line. When the absolute value of the wall thickness difference between two adjacent points exceeds the preset wall thickness abrupt change threshold (which can be 3 times the standard deviation of the wall thickness of all sampling points on the path line), the midpoint position between the two sampling points is recorded as a candidate node for wall thickness abrupt change. Spatial clustering analysis is performed on all candidate nodes for wall thickness abrupt change on all path lines. The density-based DBSCAN clustering algorithm is used to group candidate nodes whose spatial distance is less than the preset clustering radius (which can be 5% of the wax model contour size) into the same cluster. The geometric center position of each cluster is determined as the final location of the wall thickness abrupt change node. The method for determining the structural feature boundary points based on the location of abrupt wall thickness change nodes can be as follows: First, determine the abrupt transition contour segment based on the location of the abrupt wall thickness change node. Then, generate the wall thickness amplitude variation curve based on the abrupt transition contour segment and determine the structural feature boundary points of the wax model based on the wall thickness amplitude variation curve. Alternatively, the locations of each abrupt wall thickness change node can be connected into boundary lines according to their connectivity on the wax model contour surface to form closed or nearly closed boundary contour lines. The nodes on the boundary contour lines are the structural feature boundary points.

[0033] In one possible implementation, step S210, determining the structural feature boundary point of the wax model based on the location of the wall thickness abrupt change node, includes: S211, Based on the location of the wall thickness abrupt change node, determine the abrupt change transition profile segment.

[0034] For example, a preset gradient interval can be extended along the normal direction of the wax model contour, based on the location of the wall thickness abrupt change node. Then, the wall thickness gradient normalization calculation is performed on all nodes within the extended interval, and continuous node segments with a wall thickness change rate greater than a preset gradient threshold are selected and identified as abrupt change transition contour segments. Alternatively, a fixed length contour segment can be cut off on both sides along the direction of the wax model contour, with the wall thickness abrupt change node as the center. The wall thickness fluctuation variance is calculated on the contour segment, and the segment with a wall thickness fluctuation variance greater than a preset variance threshold is identified as abrupt change transition contour segment, and so on, but not limited to these.

[0035] In one possible implementation, step S211 involves determining the abrupt transition profile segment based on the location of the wall thickness abrupt change node, including: S2111, based on the location of the wall thickness change node, extends the preset gradient interval along the normal direction of the wax model contour.

[0036] It can be understood that the preset gradient interval refers to the analysis window formed by extending a certain distance to both sides of the wax pattern contour surface, centered on the location of the wall thickness abrupt change node. The normal direction refers to the direction perpendicular to the tangent plane of the wax pattern contour surface at that node. Extending along the normal direction ensures that the wall thickness change within the interval can truly reflect the gradual trend of wall thickness transition from the abrupt change node to both sides. The length of the preset gradient interval needs to balance two contradictory requirements: if the interval is too short, it will contain too few wall thickness data points, and the statistical results will lack representativeness; if the interval is too long, it may cross the wall thickness transition zone and enter a region where the wall thickness tends to be stable, causing the abrupt change characteristics to be smoothed and weakened. The length of the preset gradient interval can be taken as 30% to 50% of the maximum difference in wall thickness of the wax pattern as an empirical range. In practical applications, the optimal value can be determined through a small number of trials based on the specific structural characteristics of the valve casting.

[0037] For example, the 3D data analysis software configured in the host computer calculates the normal vector direction of the surface where each wall thickness abrupt change node is located on the triangular mesh model for each node's location determined in S210. Specifically, this can be obtained by weighted averaging of the normal vectors of the set of triangular facets in the node's one-ring neighborhood. Then, taking the wall thickness abrupt change node location as the origin, the mesh model is expanded outwards facet by facet along the normal vector direction and its opposite direction, according to the topological adjacency relationship of the mesh model. The expansion distance is accumulated for each triangular facet expanded. When the accumulated expansion distance reaches the preset gradient interval length, the expansion stops, and the vertices and wall thickness data of all triangular facets passed during the expansion process are included in the current gradient interval. This equidistant expansion method along the normal direction ensures that the geometric distance between the sampling point in the gradient interval and the wall thickness abrupt change node location is equal on the normal projection, avoiding measurement deviations in the wall thickness change rate in the normal and non-orthogonal directions caused by the curvature of the contour surface.

[0038] S2112, perform wall thickness gradient normalization calculation on all nodes in the extended interval, screen out continuous node segments with wall thickness change rate greater than preset gradient threshold, and determine the continuous node segments as abrupt transition contour segments.

[0039] It can be understood that wall thickness gradient normalization calculation refers to the process of normalizing the wall thickness value of each node within the extended interval before calculating the wall thickness change rate between adjacent nodes. The purpose of normalization is to eliminate the influence of the difference in absolute wall thickness at different abrupt wall thickness change node locations, so that the comparison of wall thickness gradients can be carried out on the same scale. The wall thickness change rate is the ratio of the wall thickness difference between two adjacent nodes to the node spacing, reflecting the degree of drastic change in wall thickness along the normal direction of the contour surface. The preset gradient threshold is the criterion used to distinguish whether the wall thickness gradient is abrupt. A continuous node segment refers to a set of spatially adjacent nodes whose wall thickness change rate is continuously greater than the preset gradient threshold. This set corresponds to the transition region on the wax model contour surface where the wall thickness changes drastically from thick to thin or from thin to thick. Defining the continuous node segment as the abrupt transition contour segment provides an accurate input data range for subsequently generating the wall thickness amplitude change curve and accurately locating the structural feature boundary point.

[0040] For example, the normalization calculation process for the wall thickness gradient is as follows: For all nodes within each extended interval, firstly, the maximum and minimum values ​​of the wall thickness values ​​of the nodes within the interval are calculated. For the i-th node within the interval, the minimum value is subtracted from the wall thickness value of node i, and then divided by the difference between the maximum and minimum values ​​to obtain the normalized wall thickness value of node i. The normalized wall thickness value ranges from 0 to 1. Then, according to the topological adjacency order of the nodes on the contour surface, the wall thickness gradient between two adjacent nodes is calculated sequentially: the wall thickness gradient equals (the wall thickness value of node i+1 minus the wall thickness value of node i) divided by (the Euclidean distance from node i+1 to node i). The preset gradient threshold is taken as the 80th percentile of the wall thickness gradients of all adjacent nodes within the interval. Then, all adjacent node pairs within the interval are traversed, and adjacent node pairs with wall thickness gradients greater than the preset gradient threshold are marked as "mutation pairs". All spatially connected "mutation pairs" are merged into continuous node segments. For each continuous node segment, check the total number of nodes and the total length: if the total number of nodes is less than 5 or the total length is less than 2% of the wax model outline size, then the continuous node segment is regarded as a noise segment and is removed to avoid false abrupt transition outline segments caused by local measurement noise or surface micro-defects interfering with subsequent analysis.

[0041] S212, based on the abrupt transition profile segment, generates a curve showing the change in wall thickness amplitude.

[0042] It can be understood that the wall thickness amplitude variation curve is a two-dimensional curve plotted with the normalized position of each node along the normal direction on the abrupt transition contour segment as the abscissa and the local wall thickness value corresponding to each node as the ordinate. The wall thickness amplitude variation curve intuitively shows the continuous change process of the wax model wall thickness from the thick-walled region through the transition segment to the thin-walled region (or vice versa). The shape of the curve, including the absolute value of the curve slope, the concavity and convexity of the curve, and the position of the inflection point of the curve, contains complete quantitative information on the wall thickness transition behavior, and is the core analysis object for subsequent precise location of structural feature boundary points.

[0043] For example, the 3D data analysis software configured in the host computer extracts the spatial coordinates and corresponding local wall thickness values ​​of all nodes within each abrupt transition contour segment. The first node of the abrupt transition contour segment is set as the zero point. Following the order of nodes in the topological connection relationship, the cumulative arc length distance from each node to the first node along the contour surface is calculated sequentially, serving as the node's position coordinates. A scatter plot is drawn point by point in a Cartesian coordinate system, with the position coordinates of each node as the horizontal axis and the local wall thickness value of each node as the vertical axis. Then, a cubic spline interpolation algorithm is used to connect the scatter points into a smooth wall thickness amplitude variation curve. The boundary conditions for the cubic spline interpolation algorithm are natural boundary conditions, meaning the second derivative of the curve at the beginning and end is zero, to ensure the smoothness of the curve at the ends. After the wall thickness amplitude variation curve is drawn, it is stored in the data analysis module of the host computer, simultaneously recording metadata such as the abrupt transition contour segment number corresponding to the curve, the total number of nodes contained in the curve, and the total length of the curve along the normal direction.

[0044] S213, determine the structural feature boundary point of the wax model based on the wall thickness amplitude variation curve.

[0045] It can be understood that the wall thickness variation curve contains information about the location where the wall thickness change rate is the maximum or the wall thickness amplitude reaches a steady state transition. These locations correspond to the substantial boundary between thick and thin or thin and thick wall thickness in the wax model. The physical logic of extracting structural feature boundary points from the wall thickness variation curve is as follows: In the transition section between thick-walled and thin-walled regions, there is a turning point in the wall thickness variation characteristics. On one side of this point, the wall thickness value is generally at a higher level and the change tends to be gradual (thick-walled characteristic); on the other side, the wall thickness value is generally at a lower level and the change tends to be gradual (thin-walled characteristic). This turning point itself is the location where the wall thickness change rate is the maximum or the midpoint of the wall thickness transition zone, making it most suitable as the boundary point between two different contraction regions.

[0046] For example, the wall thickness amplitude variation curve can be processed by sliding smoothing filtering to locate the contour node corresponding to the minimum amplitude point, and then the contour node corresponding to the minimum amplitude point can be determined as the structural feature boundary point of the wax model; or the first derivative of the wall thickness amplitude variation curve can be obtained to get the wall thickness variation rate curve, and the contour node corresponding to the extreme point with the largest absolute value of the rate in the curve can be determined as the structural feature boundary point of the wax model, and so on, but not limited to these.

[0047] This setup, by first locating the transition contour segment based on the location of the wall thickness abrupt change node, then generating a dedicated wall thickness amplitude variation curve based on this contour segment, and finally accurately determining the structural feature boundary point from the curve, can effectively eliminate noise interference and contour burr effects in the three-dimensional structural data of the wax model. It achieves automated, high-precision, and adaptive division of thick-walled shrinkage domains, thin-walled shrinkage domains, and edge shrinkage domains. Compared with the traditional method of dividing regions based on manual experience, it effectively improves the stability and consistency of boundary point identification, avoids distortion in subsequent shrinkage compensation calculations due to boundary positioning deviations, and thus improves the accuracy and reliability of overall shrinkage compensation for valve casting wax models.

[0048] In one possible implementation, step S213, determining the structural feature boundary point of the wax model based on the wall thickness amplitude variation curve, includes: S2131, the wall thickness amplitude variation curve is subjected to sliding smoothing filtering to locate the contour node corresponding to the minimum amplitude point.

[0049] As can be understood, sliding smoothing filtering refers to using a data window of fixed width to slide along the wall thickness amplitude variation curve with a preset step size. Each time, the average wall thickness of all data points within the window is calculated as the filtered wall thickness value at the window's center point. The purpose of sliding smoothing filtering is to eliminate high-frequency noise spikes caused by small fluctuations in local node wall thickness measurements during cubic spline interpolation, making the overall trend of the wall thickness amplitude variation curve clearer and avoiding misidentification of local small fluctuations caused by noise as amplitude minima. An amplitude minima refers to the inflection point on the filtered wall thickness amplitude variation curve where the function value changes from decreasing to increasing, i.e., the point where the first derivative of the curve changes from negative to positive at a zero-crossing point, such as... Figure 2 As shown, the minimum amplitude point represents the thinnest position in the transition section. In the "concave" shape of wall thickness change from thick to thin to thick, this position is the geometric boundary between two abrupt changes in wall thickness, and it is also the midpoint of the transition from one side of thick wall features to the other side of thick wall features (passing through a thin wall region in between), making it suitable as a boundary point for structural features.

[0050] For example, the specific parameters of the sliding smoothing filter are as follows: the sliding window width is 10% of the total number of nodes on the wall thickness amplitude variation curve, but not less than 5 nodes, and the sliding step size is 1 node. For each node on the curve, the wall thickness values ​​of all nodes within half a window width before and after that node are taken, and the arithmetic mean of these wall thickness values ​​is calculated as the filtered wall thickness value of that node. For nodes at the beginning and end that are less than half a window width, the data points within the window are supplemented by mirror extension to avoid the edge effect of the filter. After filtering, the sign change of the first derivative is calculated point by point on the filtering curve: for the j-th node, the difference between the wall thickness value of the (j+1)-th node and the wall thickness value of the j-th node is calculated. If the sign of the difference changes from negative to positive (i.e., the difference changes from less than zero to greater than zero), then the j-th node is the amplitude minimum point. The amplitude minimum point is mapped back to the corresponding contour node on the wax model triangular mesh model, and the three-dimensional coordinates and node number of the contour node are recorded.

[0051] S2132, the contour node corresponding to the minimum amplitude point is determined as the structural feature boundary point of the wax model.

[0052] It can be understood that the contour node corresponding to the minimum amplitude point is the geometric position where the wall thickness is thinnest in the wall thickness transition section. The wall thickness values ​​on both sides of this position increase towards the thick wall direction and the thin wall direction, respectively. This is the most reasonable dividing point for dividing the wall thickness transition section into the thick wall side and the thin wall side. Determining the contour node corresponding to the minimum amplitude point as the structural feature dividing point means that this node geometrically marks the transition of the wax model wall thickness from being dominated by thick wall features to being dominated by thin wall features. Dividing the thick wall contraction region and the thin wall contraction region based on this boundary can make the wall thickness distribution within the two contraction regions as uniform as possible, while maximizing the wall thickness difference between regions, which conforms to the principle of region division that is as homogeneous as possible within the region and as heterogeneous as possible between the regions.

[0053] For example, the 3D data analysis software configured in the host computer stores the contour node numbers corresponding to the minimum amplitude points located in S2131 into the structural feature boundary point list. Each structural feature boundary point in the list contains the following attributes: node number, 3D spatial coordinates, topological patch number on the wax model triangular mesh model, corresponding wall thickness amplitude variation curve number, and normalized wall thickness value of the minimum amplitude point. For all structural feature boundary points detected on the 3D structure of the wax model, the trajectory of the boundary points is tracked according to spatial connectivity. Specifically, a breadth-first search algorithm based on the topological adjacency relationship of the triangular mesh model is adopted: taking any structural feature boundary point as the starting seed point, searching for nodes in the one-ring neighborhood of the seed point that are also marked as structural feature boundary points, adding them to the current boundary trajectory, and using the newly added nodes as new seed points to continue expanding the search until there are no more unvisited structural feature boundary points in the one-ring neighborhood. After the search corresponding to a complete closed boundary trajectory is completed, the boundary trajectory topologically divides the triangular mesh model of the wax model into several disconnected regions. All boundary trajectories are traversed and verified to ensure that each boundary trajectory is either a closed loop or an open curve terminating at both ends of the natural boundary of the wax model. For open boundary trajectories that are not closed, they are automatically extended and completed along the outline boundary of the wax model to the model boundary to form a complete dividing line.

[0054] This setup first applies a sliding smoothing filter to the wall thickness amplitude variation curve, effectively filtering out burrs and random data noise generated during the 3D scanning and contour extraction process, ensuring that the curve shape truly reflects the wall thickness variation law of the wax model; then, by locating the minimum amplitude point, the structural feature boundary point is determined, and the boundary point is accurately located based on the inherent morphological characteristics of the wall thickness transition area, further improving the anti-interference ability and positioning accuracy of the boundary identification, and providing an accurate basis for subsequent division of each shrinkage domain and shrinkage compensation calculation.

[0055] S300, calculate the thick-wall shrinkage compensation reference value based on the thick-wall shrinkage ratio coefficient of the thick-wall shrinkage region and the corresponding first region size parameter; calculate the thin-wall shrinkage compensation reference value based on the thin-wall shrinkage ratio coefficient of the thin-wall shrinkage region and the corresponding second region size parameter; calculate the edge shrinkage compensation reference value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region size parameter.

[0056] It can be understood that the thick-walled shrinkage ratio coefficient is a proportional constant pre-calibrated based on the statistical characteristics of the wall thickness of the thick-walled shrinkage zone and the material properties of the wax used, used to characterize the linear shrinkage per unit wall thickness during the cooling and solidification process of this region. The first region dimensional parameter refers to the characteristic dimensional value extracted from the thick-walled shrinkage zone that represents the overall dimensional level of the region, such as the arithmetic mean of the projected dimensions of the thick-walled shrinkage zone in the three orthogonal directions of the wax model, or the average distance from all facets in the thick-walled shrinkage zone to the geometric center of the region. The thick-walled shrinkage compensation reference value is a preliminary reference value obtained by multiplying the thick-walled shrinkage ratio coefficient by the first region dimensional parameter, used to reflect the basic compensation amount required for the thick-walled shrinkage zone under the current wax model structure. Similarly, the thin-walled shrinkage ratio coefficient and the edge shrinkage ratio coefficient correspond to the shrinkage behavior characteristics of the thin-walled shrinkage zone and the edge shrinkage zone, respectively; the second region dimensional parameter and the third region dimensional parameter correspond to the characteristic dimensional values ​​of the thin-walled shrinkage zone and the edge shrinkage zone, respectively; and the thin-walled shrinkage compensation reference value and the edge shrinkage compensation reference value are preliminary reference values ​​for the basic compensation amount required for the corresponding shrinkage zone. The physical significance of the three shrinkage compensation benchmark values ​​is as follows: due to the different wall thicknesses, the cooling rate and solidification shrinkage amount are different. The compensation required per unit size of the thick-walled shrinkage domain is greater than that of the thin-walled shrinkage domain. The edge shrinkage domain is in between due to the continuous change of wall thickness. Therefore, it is necessary to use different shrinkage ratio coefficients to calculate the compensation benchmark value of each shrinkage domain separately. This provides an initial calculation starting point for determining the final shrinkage compensation value by domain, avoiding the problems of undercompensation for thick walls and overcompensation for thin walls caused by the global uniform compensation coefficient.

[0057] For example, the calibration of the thick-wall shrinkage ratio coefficient, thin-wall shrinkage ratio coefficient, and edge shrinkage ratio coefficient can be completed in advance through the following offline experimental procedure. Select paraffin or resin-based wax of the same grade as the wax model of the target valve casting, and prepare three types of standard test blocks in a constant temperature and humidity environment (temperature controlled at 23±1℃, relative humidity controlled at 50%±5%): a thick-walled block with a thickness of 15 mm and dimensions of 100 mm × 40 mm × 15 mm; a thin-walled block with a thickness of 4 mm and dimensions of 100 mm × 40 mm × 4 mm; and a gradually thickening wall block with dimensions of 100 mm × 40 mm, where the wall thickness linearly transitions from 15 mm at one end to 4 mm at the other, used to simulate the wall thickness transition characteristics of the edge shrinkage region. At least 10 of each type of standard test block should be prepared to ensure statistical reliability. The actual dimensions of each standard test sample were measured using a high-precision 3D scanner after 24 hours of full cooling and stabilization following casting. The actual dimensions after cooling were subtracted from the standard dimensions of the casting mold, and then divided by the standard dimensions of the casting mold to obtain the measured linear shrinkage rate of each sample. The arithmetic mean of the measured linear shrinkage rates of all 10 thick-walled samples was taken as the thick-wall shrinkage ratio coefficient; the arithmetic mean of the measured linear shrinkage rates of 10 thin-walled samples was taken as the thin-wall shrinkage ratio coefficient; and the arithmetic mean of the measured linear shrinkage rates of 10 samples with gradually changing wall thickness was taken as the edge shrinkage ratio coefficient. These three ratio coefficients are stored in a process parameter database in the non-volatile memory of the host computer and are updated periodically according to changes in wax batches.

[0058] For example, the calculation method for the first region size parameter is as follows: Statistical analysis is performed on the vertex coordinates of all triangular facets in the thick-walled shrinkage region, and the differences between the maximum and minimum values ​​of the X-coordinate, Y-coordinate, and Z-coordinate are calculated respectively. The arithmetic mean of the differences in the three directions is taken as the first region size parameter. The second and third region size parameters are calculated using the same method for the thin-walled shrinkage region and the edge shrinkage region, respectively. The calculation formula for the thick-walled shrinkage compensation reference value is: the thick-walled shrinkage compensation reference value is equal to the thick-walled shrinkage ratio coefficient multiplied by the first region size parameter. The thin-walled shrinkage compensation reference value and the edge shrinkage compensation reference value are calculated in the same way. The calculation results of the three shrinkage compensation reference values ​​are stored in the controller's memory in parameter form for subsequent shrinkage fitting calculation parameter determination steps.

[0059] It should be noted that existing technologies using global uniform shrinkage rate compensation only scale up all areas proportionally based on a single nominal shrinkage rate provided by the wax supplier, completely ignoring the differences in actual shrinkage caused by varying cooling rates in different wall thickness areas of the wax model. Taking a typical valve casting wax model as an example, the flange area has a wall thickness of 20 mm, while the thin-walled valve body area has a wall thickness of 4 mm. Based on the specific heat capacity and thermal conductivity of paraffin wax, the cooling time of the thick-walled area is approximately 5 to 10 times that of the thin-walled area, and the final total shrinkage of the thick-walled area can be 2 to 3 times that of the thin-walled area. When using a single nominal shrinkage rate for global compensation, it is only barely usable when this nominal shrinkage rate is exactly equal to a certain compromise between the actual shrinkage rates of the thick and thin walls. However, this compromise cannot precisely match the actual needs of any given area. This method calibrates independent shrinkage ratio coefficients for thick-walled shrinkage regions, thin-walled shrinkage regions, and edge shrinkage regions, and introduces the differences in shrinkage characteristics of wax molds with different wall thicknesses into the compensation calculation in the form of quantified coefficients. This fundamentally overcomes the defect of insufficient compensation accuracy of global uniform compensation for wax molds with uneven wall thickness distribution.

[0060] S400 determines the shrinkage fitting calculation parameters to fit the current wax mold based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value, and uses the shrinkage rate adaptive fitting algorithm to obtain the shrinkage compensation value corresponding to each shrinkage domain.

[0061] It can be understood that shrinkage fitting calculation parameters refer to the set of calculation parameters that integrate three independent shrinkage compensation benchmark values ​​into a unified set that reflects the overall shrinkage characteristics of the current wax model. This set of parameters is used to drive the subsequent shrinkage rate adaptive fitting algorithm. Adapting to the current wax model means that the values ​​of the shrinkage fitting calculation parameters are not fixed preset values, but customized parameters calculated in real time based on the current wax model's three-dimensional structural data, the size proportions of the three shrinkage zones, and the actual shrinkage compensation benchmark values ​​of each shrinkage zone. The shrinkage rate adaptive fitting algorithm is a calculation method that uses the wax model's structural data and shrinkage fitting calculation parameters as input, and automatically determines the optimal shrinkage compensation value for each shrinkage zone through adaptive fitting. The shrinkage compensation value refers to the specific compensation value that is finally output after calculation by the adaptive fitting algorithm, used to guide the enlargement of the wax model mold cavity size. The shrinkage compensation value of each shrinkage zone is calculated independently and applied to the mold size correction of the corresponding shrinkage zone.

[0062] It should be noted that some existing technologies set fixed shrinkage compensation values ​​for each region based on zonal compensation. These fixed compensation values ​​are typically derived from empirical corrections based on dimensional measurements of several trial molds. However, valve casting wax molds with different valve diameters, pressure ratings, and casting alloy grades exhibit significant differences in structural dimensions and wall thickness distribution. A single fixed compensation value cannot cover the entire product series. Each new specification requires multiple iterative cycles of trial molding, measurement, and mold modification to obtain a suitable compensation value, resulting in long trial molding cycles and high costs. This method employs an adaptive shrinkage rate fitting algorithm. Using the three-dimensional structural data of the wax mold and shrinkage fitting calculation parameters as input, it automatically calculates the shrinkage compensation value for each shrinkage region based on the current structural characteristics of the wax mold. This eliminates the need for repeated trial molding iterations for each new specification, achieving automated determination of shrinkage compensation values. This significantly shortens the mold debugging cycle in new product development and reduces trial molding costs.

[0063] In one possible implementation, in step S400, based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value, shrinkage fitting calculation parameters adapted to the current wax mold are determined, including: S410 calculates the weighted sum of the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value to obtain the comprehensive shrinkage benchmark value of the wax mold.

[0064] The comprehensive shrinkage benchmark value of a wax pattern can be understood as a comprehensive numerical index representing the overall average shrinkage level of the wax pattern, obtained by weighting and averaging the shrinkage compensation benchmark values ​​of the three shrinkage domains according to the area proportion of each shrinkage domain in the overall wax pattern. The weighted summation method reasonably reflects the different contributions of each shrinkage domain to the overall shrinkage behavior of the wax pattern: areas with larger area proportions have greater weight in the weighted summation, while areas with smaller area proportions have correspondingly smaller weights, thus making the comprehensive shrinkage benchmark value closer to the average shrinkage performance of the wax pattern in the actual casting process. The comprehensive shrinkage benchmark value of the wax pattern serves as a bridge connecting the shrinkage compensation benchmark values ​​of the three independent domains with the subsequent adaptive fitting calculation parameters. It integrates the three dispersed benchmark values ​​into a unified comprehensive index, facilitating the subsequent determination of shrinkage fitting calculation parameters suitable for the current wax pattern structure based on this comprehensive index and the size proportion of each domain.

[0065] For example, the specific process of weighted summation calculation is as follows: First, calculate the total area of ​​the triangular facets contained in the thick-walled shrinkage region, thin-walled shrinkage region, and edge shrinkage region. Let the area of ​​the thick-walled shrinkage region be A1, the area of ​​the thin-walled shrinkage region be A2, and the area of ​​the edge shrinkage region be A3, with the total area being A1 + A2 + A3. The weighting coefficient for the thick-walled region is w1, which equals A1 divided by the total area; the weighting coefficient for the thin-walled region is w2, which equals A2 divided by the total area; and the weighting coefficient for the edge region is w3, which equals A3 divided by the total area. The comprehensive shrinkage benchmark value of the wax model is equal to w1 × thick-walled shrinkage compensation benchmark value + w2 × thin-walled shrinkage compensation benchmark value + w3 × edge shrinkage compensation benchmark value. The result of the weighted summation is a numerical value with the dimension of length, in millimeters, reflecting the average benchmark level of overall shrinkage compensation under the current three-shrinkage region area distribution pattern of the wax model.

[0066] S420, based on the comprehensive shrinkage benchmark value of the wax mold and the size ratio of each shrinkage zone, determines the shrinkage fitting calculation parameters that are suitable for the current wax mold structure.

[0067] It is understandable that the shrinkage fitting calculation parameters are the core set of control parameters used to configure the adaptive fitting algorithm for shrinkage rate. These parameters determine the algorithm's behavioral characteristics, such as fitting accuracy, convergence speed, and fitting range, when calculating shrinkage compensation values. The size proportion of each shrinkage region refers to the proportion of the size parameters of the thick-walled shrinkage region, the thin-walled shrinkage region, and the edge shrinkage region in the sum of the size parameters of the three regions. The shrinkage fitting calculation parameters are determined based on the comprehensive shrinkage benchmark value of the wax mold and the size proportion of each shrinkage region. This ensures that the shrinkage fitting calculation parameters can simultaneously reflect the overall shrinkage level of the wax mold (reflected by the comprehensive shrinkage benchmark value) and the structural distribution characteristics of the wax mold (reflected by the size proportion of each region), guaranteeing that the adaptive fitting algorithm has sufficient adaptability to the current wax mold in subsequent calculations.

[0068] For example, the shrinkage fitting calculation parameters include the following four component parameters: the fitting step size parameter is taken as 1% of the comprehensive shrinkage benchmark value of the wax model, but not less than 0.005 mm, used to control the search step size of the adaptive fitting algorithm during the iterative optimization process; the fitting tolerance parameter is taken as 5% of the comprehensive shrinkage benchmark value of the wax model, but not less than 0.02 mm, used to control the convergence criterion of the fitting algorithm; the upper limit parameter of the fitting range is taken as 1.3 times the thick-walled shrinkage compensation benchmark value, and the lower limit parameter of the fitting range is taken as 0.7 times the thin-walled shrinkage compensation benchmark value. These two parameters together define the legal value range when the adaptive fitting algorithm searches for the optimal compensation value. The above four component parameters constitute the shrinkage fitting calculation parameters, which are passed to the shrinkage rate adaptive fitting algorithm module in the form of a parameter set as the initialization configuration input of the algorithm.

[0069] In one possible implementation, before step S400, where a shrinkage rate adaptive fitting algorithm is used to obtain the shrinkage compensation value corresponding to each shrinkage domain, the method further includes: S401, calculate the first average shrinkage amplitude of the thick-walled shrinkage region and the second average shrinkage amplitude of the thin-walled shrinkage region respectively.

[0070] It can be understood that the first average shrinkage amplitude refers to the arithmetic mean of the wall thickness values ​​of all facet nodes in the thick-walled shrinkage domain after being converted into shrinkage amounts using the thick-walled shrinkage ratio coefficient. This reflects the average shrinkage amplitude of the thick-walled shrinkage domain under the current 3D structural data. The second average shrinkage amplitude refers to the arithmetic mean of the wall thickness values ​​of all facet nodes in the thin-walled shrinkage domain after being converted into shrinkage amounts using the thin-walled shrinkage ratio coefficient. This reflects the average shrinkage amplitude of the thin-walled shrinkage domain.

[0071] For example, the calculation process for the first average shrinkage amplitude is as follows: traverse each triangular facet within the thick-walled shrinkage region, read the wall thickness value d at the center point of the facet, multiply d by the thick-walled shrinkage ratio coefficient to obtain the theoretical shrinkage amount of the facet, sum the theoretical shrinkage amounts of all facets within the thick-walled shrinkage region, and divide by the total number of facets in the thick-walled shrinkage region. The quotient is the first average shrinkage amplitude, in millimeters. The second average shrinkage amplitude is calculated for all facets within the thin-walled shrinkage region using the same method. Since the wall thickness of the edge shrinkage region changes continuously and it is in a transitional state, the average shrinkage amplitude is not calculated separately. The structural data of the edge shrinkage region is processed in the amplitude correction stage with reference to the weighted value of the first and second amplitude correction coefficients.

[0072] S402, based on the first average shrinkage amplitude and the second average shrinkage amplitude, determine the first amplitude correction coefficient of the thick-walled shrinkage domain and the second amplitude correction coefficient of the thin-walled shrinkage domain respectively, and obtain the third amplitude correction coefficient corresponding to the molding interference factor.

[0073] It can be understood that the first amplitude correction coefficient refers to the adjustment coefficient used to correct the amplitude of the structural data of the thick-walled shrinkage region, determined based on the first average shrinkage amplitude. The second and third amplitude correction coefficients are the amplitude correction coefficients corresponding to the thin-walled shrinkage region and the forming interference factors, respectively. Forming interference factors refer to non-structural factors that, in addition to wall thickness, will cause changes in the shrinkage of the wax pattern during the actual casting process, such as fluctuations in ambient temperature, viscosity differences between wax batches, and uneven mold preheating temperature.

[0074] For example, the first amplitude correction coefficient is 1 divided by the first average shrinkage amplitude, meaning the first amplitude correction coefficient is equal to the reciprocal of the first average shrinkage amplitude. The second amplitude correction coefficient is 1 divided by the second average shrinkage amplitude. The third amplitude correction coefficient is obtained through regression analysis based on the changes in wax mold shrinkage at different ambient temperatures measured in offline experiments. Specifically, five standard wax molds of the same mold are poured under different ambient temperatures (e.g., 20℃, 23℃, 26℃, 29℃), and the actual shrinkage after cooling is measured. Using the shrinkage at 23℃ as the baseline value of 1.0, the ratio of the shrinkage under other temperature conditions to the baseline value is calculated. This ratio is then linearly regressed with the ambient temperature to obtain a linear function relationship between the third amplitude correction coefficient and the ambient temperature deviation. During the online detection phase, the current ambient temperature is read by a temperature sensor, and the current third amplitude correction coefficient is calculated by substituting it into the function relationship.

[0075] S403 uses the first amplitude correction coefficient, the second amplitude correction coefficient, and the third amplitude correction coefficient to perform amplitude correction processing on the structural data of each contraction domain to obtain standardized wax mold structural data.

[0076] It can be understood that amplitude correction processing refers to multiplying the wall thickness values ​​in the original three-dimensional structural data of each shrinkage region by the corresponding amplitude correction coefficient and the third amplitude correction coefficient, respectively, so that the corrected structural data are comparable between different shrinkage regions and between different production batches. Standardized wax mold structural data refers to unified structural data that has been corrected by three amplitude correction coefficients, eliminating the influence of shrinkage amplitude differences and molding interference factors. It is the direct input data for the shrinkage rate adaptive fitting algorithm.

[0077] For example, the amplitude correction process is as follows: For each patch node in the thick-walled shrinkage domain, the node wall thickness value is multiplied by a first amplitude correction coefficient and then by a third amplitude correction coefficient to obtain the standardized node wall thickness value; for each patch node in the thin-walled shrinkage domain, the node wall thickness value is multiplied by a second amplitude correction coefficient and then by a third amplitude correction coefficient to obtain the standardized node wall thickness value; for each patch node in the edge shrinkage domain, the node wall thickness value is multiplied by the geometric mean of the first and second amplitude correction coefficients and then by a third amplitude correction coefficient to obtain the standardized node wall thickness value. The standardized wall thickness values ​​of all nodes, the original three-dimensional coordinate values, and the topological adjacency relationships together constitute the standardized wax model structure data, which is stored in the standardized structure data table of the host computer and serves as the input to the shrinkage rate adaptive fitting algorithm.

[0078] It should be noted that the wall thickness values ​​in the original 3D structural data of each shrinkage region reflect the actual geometry of the wax model. However, due to the influence of the molding environment and batch differences in wax materials, the wall thickness values ​​of wax models from different batches may exhibit systematic deviations. If the original 3D structural data is directly input into the shrinkage rate adaptive fitting algorithm, the deviation in wall thickness measurement caused by molding interference factors will be misinterpreted by the algorithm as structural feature differences, resulting in the final calculated shrinkage compensation value deviating from the actual requirements of the wax model. This method introduces amplitude correction processing before adaptive fitting. By calculating the average shrinkage amplitude of each region and determining the amplitude correction coefficient accordingly, the systematic influence of molding interference factors on the structural data is eliminated, the input data of the adaptive fitting algorithm is standardized, and the accuracy and batch-to-batch consistency of the shrinkage compensation value calculation results are ensured.

[0079] In one possible implementation, in S400, a shrinkage rate adaptive fitting algorithm is used to obtain the shrinkage compensation value corresponding to each shrinkage domain, including: S430 generates a shrinkage compensation value-fit weight mapping curve based on standardized wax mold structural data and shrinkage fitting calculation parameters.

[0080] The shrinkage compensation value-fitting weight mapping curve can be understood as a continuous curve plotted in a Cartesian coordinate system, with the possible values ​​of the shrinkage compensation value on the x-axis and the fitting weights corresponding to each shrinkage compensation value on the y-axis. The fitting weight is a quantitative indicator that measures the goodness of fit of a specific shrinkage compensation value to the current structural data of the wax model; a higher fitting weight indicates a better fit of the shrinkage compensation value to the current wax model structural data. The process of generating the mapping curve involves calculating the fitting weight value of each candidate compensation value point-by-point within the legal range of shrinkage compensation values, using a preset fitting step size parameter as the interval, and connecting the discrete fitting weight values ​​into a continuous curve using spline interpolation. The legal range of values ​​is defined by the upper limit and lower limit parameters of the fitting range, and the fitting step size is determined by the fitting step size parameter. For each candidate compensation value, the fitting weight is calculated as follows: apply the candidate compensation value to all nodes of the standardized wax model structure data, calculate the deviation between the wall thickness value of each node after compensation and the wall thickness value of the wax model design, sum the absolute values ​​of all node deviations and take the reciprocal or convert them into fitting weight values ​​in the form of a negative exponent. The smaller the total deviation, the higher the fitting weight, indicating that the candidate compensation value has a better overall fit to the current wax model structure data.

[0081] For example, the specific calculation process of the fitting weight is as follows: Let the candidate compensation value be C. For each node i in the standardized wax model structure data, the original standardized wall thickness of the node is di. After applying the compensation value C, the compensated wall thickness of the node is di+C, and the designed wall thickness of the wax model is di_target. The wall thickness deviation of node i is ei=|di+C-di_target|. The total deviation E=Σei is obtained by summing the wall thickness deviations of all N nodes. The fitting weight W=exp(-λ×E), where λ is a scaling factor, which is 1 divided by (average wall thickness of wax model × N × 0.01) to ensure that the numerical range of the fitting weight is within a reasonable dynamic range. Starting from the lower limit parameter of the fitting range, and using the fitting step size parameter as the step size, the calculation is performed point by point until the upper limit parameter of the fitting range is reached. Each candidate compensation value C and the corresponding fitting weight W are recorded as a data point (C, W) in the data point list. After all candidate compensation values ​​are calculated, a cubic spline interpolation algorithm is used to connect all data points into a smooth shrinkage compensation value-fitting weight mapping curve.

[0082] S440, based on a preset initial weight threshold, divides the shrinkage compensation value-fit weight mapping curve into multiple continuous independent feature intervals, and selects feature intervals with fitting weights greater than the preset initial weight threshold.

[0083] It can be understood that the preset initial weight threshold is a critical value for judging whether the fitted weights at each point on the mapping curve are statistically significant. The purpose of extracting feature intervals from the mapping curve where the fitted weights are greater than the preset initial weight threshold is to eliminate noisy intervals with excessively low fitted weights and focus on the range of high-fit compensation value candidates with statistically significant fitted weights. A feature interval refers to the horizontal axis segment on the mapping curve where the fitted weights are continuously greater than the preset initial weight threshold. Each feature interval corresponds to an independent range of shrinkage compensation value candidates. The method for dividing continuous independent feature intervals is as follows: scan the mapping curve from the lower limit to the upper limit of the fitted range. When the curve moves from a region below the preset initial weight threshold to a region above the preset initial weight threshold, it is recorded as the starting point of a feature interval; when the curve falls back from a region above the threshold to a region below the threshold, it is recorded as the ending point of that feature interval. The complete horizontal axis segment between the starting and ending points constitutes a continuous independent feature interval. Selecting feature intervals where the fitted weights are greater than the preset initial weight threshold means retaining the segments on the mapping curve where the fitted weights are statistically significant, while discarding invalid segments with excessively low weights or dominated by noise.

[0084] For example, the preset initial weight threshold is set to 30% of the maximum fitted weight of all data points on the mapping curve. This value can distinguish between the true high-fit compensation region determined by the structural characteristics of the wax model and random spurious peaks caused by measurement noise or wall thickness calculation errors. The division process is as follows: starting from the lower limit of the fitting range of the mapping curve, using the fitting step size parameter as the scanning step size, the fitted weight value of the current point is compared with the preset initial weight threshold point by point along the increasing direction of the horizontal axis. Each crossing point from "below the threshold" to "above the threshold" is recorded as the starting point coordinate of the interval, and each crossing point from "above the threshold" to "below the threshold" is recorded as the ending point coordinate of the interval. Each pair of starting point-ending point coordinates constitutes a feature interval. All feature intervals are recorded in the feature interval list. Each feature interval contains information such as the interval starting point compensation value, the interval ending point compensation value, the maximum fitted weight value within the interval, and the compensation value coordinates corresponding to the maximum fitted weight value within the interval.

[0085] S450, based on multiple feature intervals, obtains the contraction compensation value corresponding to each contraction domain.

[0086] It can be understood that the shrinkage compensation value corresponding to each shrinkage region refers to the three independent shrinkage compensation values ​​output by the adaptive fitting algorithm, respectively corresponding to the thick-walled shrinkage region, the thin-walled shrinkage region, and the edge shrinkage region. These three shrinkage compensation values ​​are applied to the mold cavity size correction of their respective shrinkage regions. The logic of obtaining the shrinkage compensation value of each shrinkage region based on multiple feature intervals is that the multi-peak structure of the mapping curve contains physical information about different shrinkage regions adapting to different compensation values—the highest peak usually corresponds to the compensation value dominated by the shrinkage region with the largest area proportion, the second highest peak corresponds to the compensation value dominated by the second largest shrinkage region, and the third peak corresponds to the compensation value dominated by the edge shrinkage region. By performing integral weight calculation and sorting and filtering on the feature intervals, the optimal compensation value corresponding to the three shrinkage regions can be objectively separated from the multi-peak mapping curve.

[0087] It should be noted that existing methods for determining compensation values ​​typically output only a single optimal compensation value, completely ignoring the possibility of multiple local peaks in the mapping curve. In a multi-peak mapping curve, the compensation value corresponding to the highest peak may be the optimal choice under specific molding conditions, but when molding conditions change slightly (e.g., ambient temperature fluctuates by 1 to 2°C), this peak may rapidly decay and be replaced by another neighboring peak. Therefore, determining the compensation value based solely on a single global peak lacks robustness in the face of unavoidable process fluctuations in actual production. This method generates a complete shrinkage compensation value-fitting weight mapping curve and extracts multiple feature intervals, preserving the multi-peak structure information of the mapping curve. This ensures that the determination of the compensation value considers not only the global optimal value but also the second-best alternative value, providing an information basis for subsequent selection and ranking of the three target intervals with the largest weights based on feature intervals, thus enhancing the robustness of the compensation strategy in dealing with process fluctuations.

[0088] In one possible implementation, in S450, based on multiple feature intervals, the contraction compensation value corresponding to each contraction domain is obtained, including: S451, if the number of feature intervals is not less than 3, perform integral calculations on the region enclosed by the shrinkage compensation value-fitted weight mapping curve and the preset initial weight threshold in each feature interval to obtain the corresponding integral weight. Select the three target feature intervals with the largest integral weights from all feature intervals, and determine the compensation values ​​corresponding to the peak values ​​of the three target feature intervals in ascending order of shrinkage compensation values ​​as the shrinkage compensation values ​​of the thick-walled shrinkage region, the thin-walled shrinkage region, and the edge shrinkage region, respectively.

[0089] It is understandable that the number of feature intervals reflects the richness of the highly fit regions on the mapping curve. When the number of feature intervals is not less than 3, it indicates that the mapping curve has at least three independent fitting weight peaks, sufficient to provide different candidate sources of compensation values ​​for the thick-walled contraction region, thin-walled contraction region, and edge contraction region, respectively. When the number of feature intervals is less than 3, it indicates that the multi-peak characteristics of the mapping curve are not obvious, possibly because the contraction characteristics of the three contraction regions of the wax model are relatively small, resulting in fewer peaks merging on the mapping curve. It is necessary to relax the screening conditions by lowering the weight threshold to restore the weaker peaks masked by the higher threshold. The integral weight refers to the result of numerically integrating the area of ​​the closed region enclosed by the curve segment of the mapping curve above the preset initial weight threshold and the preset initial weight threshold horizontal line within each feature interval. The integral weight comprehensively reflects the overall significance of the feature interval in two dimensions: the range of compensation values ​​and the height of fitting weights—the larger the area, the more consistently high the fitting weights of the interval remain within a wider range of compensation values, and the more robust and reliable the interval is statistically. The selection of the three target feature intervals with the largest integral weights from all feature intervals is based on the dual criteria of "optimal fit and statistical significance" to choose candidate sources for compensation values. The three target feature intervals are then sorted by compensation value from smallest to largest and assigned sequentially to the thin-walled contraction region, the edge contraction region, and the thick-walled contraction region. The physical basis for this is that the thin-walled contraction region has the smallest contraction amount corresponding to a smaller compensation value, the thick-walled contraction region has the largest contraction amount corresponding to a larger compensation value, and the edge contraction region is in the middle. After sorting the compensation values ​​from smallest to largest, the minimum value corresponds to the thin-walled region, the middle value to the edge region, and the maximum value to the thick-walled region, consistent with the order of contraction amounts in the three regions.

[0090] For example, the integral calculation adopts the Romberg numerical integration method. For each feature interval, the compensation value at the start of the interval is the lower limit of integration, and the compensation value at the end of the interval is the upper limit of integration. The integrand is the fitted weight function value of the mapping curve minus a preset initial weight threshold. The integral result is the integral weight of that feature interval. The integral weights of all feature intervals are sorted in descending order, and the top three are taken as the three target feature intervals. The three target feature intervals are sorted in ascending order according to the value of the compensation value corresponding to their peak values. The three compensation values ​​after sorting are as follows: the minimum compensation value is determined as the shrinkage compensation value of the thin-walled shrinkage region, the middle compensation value is determined as the shrinkage compensation value of the edge shrinkage region, and the maximum compensation value is determined as the shrinkage compensation value of the thick-walled shrinkage region.

[0091] S452, If the number of feature intervals is less than 3, gradually reduce the preset initial weight threshold and redivide the feature intervals until the number of feature intervals is not less than 3.

[0092] It is understandable that gradually reducing the preset initial weight threshold means decreasing the weight threshold step by step with a preset reduction step size. After each reduction, the interval division and filtering process is re-executed until the condition of "the number of feature intervals is not less than 3" is met or the weight threshold drops to the preset lower limit. The reason for reducing the threshold is that when the shrinkage characteristics of the three contraction regions of the wax model are small, the absolute fitting weights of multiple peaks on the mapping curve are not significantly different, but the weaker peaks may be truncated by the initially high threshold. By appropriately reducing the threshold, these truncated weaker peaks can be restored, allowing the multi-peak structure of the mapping curve to be fully presented. However, the threshold cannot be reduced indefinitely. When the threshold drops to the preset lower limit (which can be 10% of the maximum fitting weight value on the mapping curve) and still cannot obtain no less than 3 feature intervals, it indicates that the shrinkage difference of the three contraction regions of the current wax model is indeed insufficient to form a distinguishable independent peak in the fitting weight dimension. At this time, the compensation value corresponding to the number of feature intervals can be output and processed in the way of sharing adjacent regions.

[0093] For example, the operation process of gradually reducing the preset initial weight threshold is as follows: Let the current weight threshold be T, the initial value T0 be 30% of the maximum fitted weight value, and the reduction step size Δ be 5% of the maximum fitted weight value. When the number of feature intervals is less than 3 and the current weight threshold T is greater than the preset lower limit (10% of the maximum fitted weight value), the current weight threshold is updated to T=T-Δ, and the entire process is re-executed with the updated threshold. The above judgment-threshold reduction-re-division process is repeated until the number of feature intervals is not less than 3 or the current weight threshold is reduced to the preset lower limit. If the number of feature intervals is still 2 when the current weight threshold is reduced to the preset lower limit, the compensation values ​​corresponding to the peak values ​​of the two feature intervals are assigned to the thin-walled shrinkage region and the thick-walled shrinkage region respectively, and the shrinkage compensation value of the edge shrinkage region is the arithmetic mean of the two compensation values; if the number of feature intervals is only 1, the compensation value corresponding to the peak value of the only feature interval is assigned to all three shrinkage regions (degenerated into uniform compensation), and a prompt message is output to the host computer, prompting the operator that the difference between the three shrinkage regions of the current wax mold is not significant, and it is recommended to pay attention to the consistency of the molding process parameters. This setup, through integral calculation to obtain the integral weight of each feature interval, filters out the three target intervals with the highest signal effectiveness, effectively eliminating data noise and invalid interference intervals. Combining the inherent patterns of shrinkage in different regions of the wax model, the corresponding shrinkage domains are matched sequentially according to the size of the shrinkage compensation value, improving the reasonable accuracy of compensation value allocation. Simultaneously, an adaptive threshold adjustment mechanism is implemented, automatically re-dividing intervals when the number of effective intervals is insufficient, reducing feature loss due to improper threshold settings, improving the overall robustness of the algorithm, and ultimately obtaining compensation values ​​that closely match the actual shrinkage characteristics of each region.

[0094] In one possible implementation, after obtaining the shrinkage compensation value corresponding to each shrinkage domain using a shrinkage rate adaptive fitting algorithm in step S400, the method further includes: S500, based on the shrinkage compensation value corresponding to the thick-walled shrinkage domain and the first preset interval coefficient, determines the thick-walled compensation verification interval, and takes the compensation value with the largest fitting weight within the thick-walled compensation verification interval as the final shrinkage compensation value of the thick-walled structure.

[0095] It can be understood that the shrinkage compensation value corresponding to the thick-walled shrinkage domain is the compensation value obtained by the adaptive fitting algorithm of shrinkage rate from the multi-peak mapping curve, which corresponds to the peak position of the target feature interval with the largest integral weight. This preliminary compensation value is the optimal solution determined under the global fitting weight distribution. However, since the mapping curve is a continuous curve generated by spline interpolation of discrete sampling points, the sampling step size and interpolation algorithm may introduce small positional deviations, resulting in the preliminary compensation value not being a strictly local optimum. The first preset interval coefficient is a scaling factor used to define the local fine search range of the compensation value of the thick-walled shrinkage domain. The thick-walled compensation verification interval is a narrow interval formed by expanding to both sides with the preliminary compensation value of the thick-walled shrinkage domain as the center and the product of the first preset interval coefficient and the preliminary compensation value as the radius. The compensation value with the largest fitting weight within the thick-walled compensation verification interval is taken as the final shrinkage compensation value of the thick-walled shrinkage domain. Its core function is to perform local fine-tuning verification of the preliminary compensation value output by the adaptive fitting algorithm, and to re-search for the local extreme points of the fitting weight in the vicinity of the preliminary compensation value with finer granularity, so as to ensure that the final output compensation value has the best fit in the local range and eliminate the positioning error that may be introduced by the discrete sampling and interpolation process.

[0096] For example, the value of the first preset interval coefficient can be 0.5 times the thick-wall shrinkage ratio coefficient. This value controls the radius of the verification interval within half of the thick-wall shrinkage amount, which can cover the possible positioning deviation of the initial compensation value without expanding the search range to the peak areas of other adjacent mapping curves. The method for determining the thick-wall compensation verification interval is as follows: if the initial compensation value of the thick-wall shrinkage domain is C_thick and the first preset interval coefficient is α1, then the lower limit of the verification interval is C_thick×(1-α1) and the upper limit of the verification interval is C_thick×(1+α1). The data analysis module configured in the host computer extracts the curve segment between the lower limit and the upper limit of the verification interval on the shrinkage compensation value-fitting weight mapping curve generated by S430, and performs dense resampling of the curve segment with a preset verification step size (one-tenth of the fitting step size parameter), calculates the fitting weight value at each resampling point, and takes the compensation value corresponding to the resampling point with the largest fitting weight value as the final shrinkage compensation value of the thick-wall. S600, based on the shrinkage compensation value corresponding to the thin-wall shrinkage domain and the second preset interval coefficient, determines the thin-wall compensation verification interval, and takes the compensation value with the largest fitting weight within the thin-wall compensation verification interval as the final shrinkage compensation value of the thin-wall.

[0097] It is understandable that the shrinkage compensation value corresponding to the thin-walled shrinkage domain is usually the smallest among the three preliminary compensation values ​​output by the adaptive fitting algorithm. The shape of the peak region of its mapping curve may be sharp or flat, depending on the consistency of the wall thickness distribution of the thin-walled domain. The second preset interval coefficient is a scaling factor for the verification interval set for the shrinkage characteristics of the thin-walled shrinkage domain. The construction method and local refinement search logic of the verification interval for thin-walled compensation are the same as those for the thick-walled domain in S500. However, the value of the second preset interval coefficient needs to be set independently. The reason is that the absolute value of the shrinkage in the thin-walled domain is smaller than that in the thick-walled domain. The positioning deviation of the preliminary compensation value relative to the shrinkage in the thin-walled domain may be different. Directly using the interval coefficient of the thick-walled domain may result in the verification interval being too wide and covering the peak region of the adjacent shrinkage domain, or too narrow and failing to cover the true local optimum position.

[0098] For example, the value of the second preset interval coefficient can be 1.0 times the thin-walled shrinkage ratio coefficient. The thin-walled region has a smaller wall thickness, faster cooling rate, and is more significantly affected by environmental temperature fluctuations during the shrinkage process. This results in a wider flat peak region on the mapping curve or the presence of multiple adjacent small secondary peaks. Therefore, a larger interval coefficient than that used for the thick-walled region is employed to appropriately expand the local fine search range, enabling it to capture the true extreme values ​​of the fitted weights within this region. The calculation method for the thin-walled compensation verification interval is as follows: Let the initial compensation value for the thin-walled shrinkage region be C_thin, and the second preset interval coefficient be α2. Then, the lower limit of the verification interval is C_thin × (1-α2), and the upper limit is C_thin × (1+α2). This interval segment is intercepted on the mapping curve, and dense resampling is performed with a verification step size. The fitted weights are calculated point by point, and the compensation value corresponding to the maximum fitted weight is taken as the final thin-walled shrinkage compensation value.

[0099] S700, based on the shrinkage compensation value corresponding to the edge shrinkage domain and the third preset interval coefficient, determines the edge compensation verification interval, and takes the compensation value with the largest fitting weight in the edge compensation verification interval as the final edge shrinkage compensation value.

[0100] It is understandable that the edge contraction region, as a transitional area with gradually changing wall thickness between the thick-walled and thin-walled regions, exhibits some characteristics of both regions in its contraction behavior. The peak value corresponding to the contraction compensation value-fitting weight mapping curve is usually located between the peak value of the thick-walled region and the peak value of the thin-walled region, and the peak shape may be asymmetrical due to the superposition of mapping characteristics of the two regions. The third preset interval coefficient is a verification interval coefficient set for the transitional characteristics of the edge contraction region. Its value needs to consider that the initial compensation value of the edge region is located in the middle position in the ranking. The verification interval should avoid intruding upward into the feature interval of the thick-walled region and downward into the feature interval of the thin-walled region. Therefore, the third preset interval coefficient is usually set as a compromise value between the first and second preset interval coefficients to ensure that the verification interval captures the local optimum while maintaining the independence of the neighboring peak regions.

[0101] For example, the third preset interval coefficient can be 0.7 times the edge shrinkage ratio coefficient. The initial compensation value of the edge domain is set as C_edge, the third preset interval coefficient is α3, the lower limit of the approved interval is C_edge×(1-α3), and the upper limit of the approved interval is C_edge×(1+α3). The interval segment on the mapping curve is densely resampled with the approved step size, and the fitting weight of each point is calculated. The compensation value corresponding to the point with the largest fitting weight is taken as the final edge shrinkage compensation value.

[0102] This setup uses the initial shrinkage compensation value obtained from each shrinkage region as a benchmark, and combines it with the corresponding preset interval coefficient to define a dedicated compensation verification interval. Within the interval, the value with the largest fitting weight is selected as the final compensation value. This allows for a secondary fine calibration of the initial compensation result, further mitigating the impact of algorithm calculations and data deviations, correcting minor errors in the initial value selection, and significantly improving the accuracy of the shrinkage compensation value in each region.

[0103] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0104] Corresponding to the valve casting wax mold shrinkage compensation method described in the above embodiments, this application also provides a valve casting wax mold shrinkage compensation system, the various modules of which can realize the various steps of the valve casting wax mold shrinkage compensation method. Figure 3 The diagram shows a structural block diagram of a valve casting wax mold shrinkage compensation system provided in an embodiment of this application. For ease of explanation, only the parts related to the embodiment of this application are shown.

[0105] Reference Figure 3 The valve casting wax mold shrinkage compensation system includes: The acquisition module is used to acquire the three-dimensional structural data of the wax model of the target valve casting; the three-dimensional structural data is used to indicate the wall thickness, contour dimensions and structural node information of each part of the wax model.

[0106] The first determining module is used to determine the structural feature boundary points of the wax model based on the three-dimensional structural data, and to divide the thick-walled shrinkage region, thin-walled shrinkage region and edge shrinkage region based on the structural feature boundary points.

[0107] The calculation module is used to calculate the thick-wall shrinkage compensation benchmark value based on the thick-wall shrinkage ratio coefficient of the thick-wall shrinkage region and the corresponding first region size parameter; to calculate the thin-wall shrinkage compensation benchmark value based on the thin-wall shrinkage ratio coefficient of the thin-wall shrinkage region and the corresponding second region size parameter; and to calculate the edge shrinkage compensation benchmark value based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region size parameter.

[0108] The second determining module is used to determine the shrinkage fitting calculation parameters that are suitable for the current wax mold based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value and the edge shrinkage compensation benchmark value, and to obtain the shrinkage compensation value corresponding to each shrinkage domain by using the shrinkage rate adaptive fitting algorithm.

[0109] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0110] Those skilled in the art will understand that, for the sake of convenience and brevity, the above-described module division is merely an example. In practical applications, the above functions can be assigned to different modules as needed, that is, the internal structure of the system can be divided into different modules to complete all or part of the functions described above. The modules in the embodiments can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0111] This application also provides an electronic device 6. Figure 4 This is a schematic diagram of the structure of an electronic device 6 provided in an embodiment of this application. Figure 4 As shown, the electronic device 6 of this embodiment includes: at least one processor 60 ( Figure 4 Only one is shown in the image), at least one memory 61 ( Figure 4(Only one is shown in the image) and a computer program 62 stored in the at least one memory 61 and executable on the at least one processor 60, wherein when the processor 60 executes the computer program 62, it causes the electronic device 6 to perform the steps in any of the above embodiments of the valve casting wax mold shrinkage compensation method, or causes the electronic device 6 to perform the functions of each module in the above system embodiments.

[0112] For example, the computer program 62 may be divided into one or more modules / units, which are stored in the memory 61 and executed by the processor 60 to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 62 in the electronic device 6.

[0113] The electronic device 6 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. This electronic device may include, but is not limited to, a processor 60 and a memory 61. Those skilled in the art will understand that... Figure 4 This is merely an example of electronic device 6 and does not constitute a limitation on electronic device 6. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0114] The processor 60 can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0115] In some embodiments, the memory 61 may be an internal storage unit of the electronic device 6, such as a hard disk or memory of the electronic device 6. In other embodiments, the memory 61 may be an external storage device of the electronic device 6, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 6. Furthermore, the memory 61 may include both internal and external storage units of the electronic device 6. The memory 61 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 61 can also be used to temporarily store data that has been output or will be output.

[0116] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in any of the above method embodiments.

[0117] This application provides a computer program product that, when run on an electronic device, causes the electronic device to perform the steps in any of the above method embodiments.

[0118] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0119] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0120] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0121] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for compensating for the shrinkage rate of wax patterns in valve castings, characterized in that, include: Obtain the three-dimensional structural data of the wax model of the target valve casting; wherein, the three-dimensional structural data is used to indicate the wall thickness, contour dimensions and structural node information of each part of the wax model; Based on the three-dimensional structural data, the structural feature boundary points of the wax model are determined, and based on the structural feature boundary points, the thick-walled shrinkage region, the thin-walled shrinkage region, and the edge shrinkage region are divided. The thick-walled shrinkage compensation reference value is calculated based on the thick-walled shrinkage ratio coefficient of the thick-walled shrinkage region and the corresponding first region size parameter; the thin-walled shrinkage compensation reference value is calculated based on the thin-walled shrinkage ratio coefficient of the thin-walled shrinkage region and the corresponding second region size parameter; the edge shrinkage compensation reference value is calculated based on the edge shrinkage ratio coefficient of the edge shrinkage region and the corresponding third region size parameter. Based on the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value, shrinkage fitting calculation parameters suitable for the current wax mold are determined, and a shrinkage rate adaptive fitting algorithm is used to obtain the shrinkage compensation value corresponding to each shrinkage domain.

2. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 1, characterized in that, The determination of the structural feature boundary points of the wax model based on the three-dimensional structural data includes: Based on the three-dimensional structural data, the locations of abrupt changes in wall thickness are determined, and the structural feature boundary points of the wax model are determined based on the locations of these abrupt changes in wall thickness.

3. The valve casting wax mold shrinkage compensation method as described in claim 2, characterized in that, The determination of the structural feature boundary point of the wax model based on the location of the wall thickness abrupt change node includes: Based on the location of the wall thickness abrupt change node, the abrupt change transition profile segment is determined; Based on the abrupt transition profile segment, a wall thickness amplitude variation curve is generated; The structural feature boundary point of the wax model is determined based on the wall thickness amplitude variation curve.

4. The valve casting wax pattern shrinkage compensation method as described in claim 3, characterized in that, The step of determining the abrupt transition profile segment based on the location of the wall thickness abrupt change node includes: Based on the location of the wall thickness abrupt change node, a preset gradient interval is extended along the normal direction of the wax model contour; For all nodes within the extended interval, perform wall thickness gradient normalization calculation, screen out continuous node segments with wall thickness change rate greater than a preset gradient threshold, and determine the continuous node segments as abrupt transition contour segments.

5. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 3, characterized in that, The determination of the structural feature boundary point of the wax model based on the wall thickness amplitude variation curve includes: The wall thickness amplitude variation curve is subjected to sliding smoothing filtering to locate the contour node corresponding to the minimum amplitude point; The contour node corresponding to the minimum amplitude point is determined as the structural feature boundary point of the wax model.

6. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 1, characterized in that, The step of determining shrinkage fitting calculation parameters adapted to the current wax mold based on the thick-wall shrinkage compensation benchmark value, the thin-wall shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value includes: The weighted summation of the thick-walled shrinkage compensation benchmark value, the thin-walled shrinkage compensation benchmark value, and the edge shrinkage compensation benchmark value yields the comprehensive shrinkage benchmark value of the wax mold. Based on the comprehensive shrinkage benchmark value of the wax mold and the size ratio of each shrinkage zone, the shrinkage fitting calculation parameters adapted to the current wax mold structure are determined.

7. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 1, characterized in that, Before employing the shrinkage rate adaptive fitting algorithm to obtain the shrinkage compensation value corresponding to each shrinkage region, the method further includes: Calculate the first average shrinkage amplitude of the thick-walled shrinkage region and the second average shrinkage amplitude of the thin-walled shrinkage region, respectively; Based on the first average shrinkage amplitude and the second average shrinkage amplitude, the first amplitude correction coefficient of the thick-walled shrinkage domain and the second amplitude correction coefficient of the thin-walled shrinkage domain are determined respectively, and the third amplitude correction coefficient corresponding to the molding interference factor is obtained. The first amplitude correction coefficient, the second amplitude correction coefficient, and the third amplitude correction coefficient are used to perform amplitude correction processing on the structural data of each contraction region to obtain standardized wax mold structural data.

8. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 7, characterized in that, The method employs an adaptive shrinkage rate fitting algorithm to obtain shrinkage compensation values ​​corresponding to each shrinkage domain, including: Based on the standardized wax mold structure data and the shrinkage fitting calculation parameters, a shrinkage compensation value-fit weight mapping curve is generated. Based on a preset initial weight threshold, the shrinkage compensation value-fit weight mapping curve is divided into multiple continuous independent feature intervals, and the feature intervals with fitting weights greater than the preset initial weight threshold are selected. Based on multiple feature intervals, the contraction compensation value corresponding to each contraction domain is obtained.

9. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 8, characterized in that, The step of obtaining the contraction compensation value corresponding to each contraction region based on multiple feature intervals includes: If the number of feature intervals is not less than 3, the region enclosed by the shrinkage compensation value-fit weight mapping curve and the preset initial weight threshold in each feature interval is integrated to obtain the corresponding integral weight; the three target feature intervals with the largest integral weights are selected from all the feature intervals, and the compensation values ​​corresponding to the peak values ​​of the three target feature intervals are determined in ascending order of shrinkage compensation value as the shrinkage compensation values ​​of the thick-walled shrinkage region, the thin-walled shrinkage region and the edge shrinkage region, respectively; If the number of feature intervals is less than 3, the preset initial weight threshold is gradually reduced and the feature intervals are re-divided until the number of feature intervals is not less than 3.

10. The method for compensating for shrinkage rate of valve casting wax patterns as described in claim 1, characterized in that, After obtaining the shrinkage compensation value corresponding to each shrinkage region using the shrinkage rate adaptive fitting algorithm, the method further includes: Based on the shrinkage compensation value corresponding to the thick-walled shrinkage domain and the first preset interval coefficient, the thick-walled compensation verification interval is determined, and the compensation value with the largest fitting weight in the thick-walled compensation verification interval is taken as the final shrinkage compensation value of the thick-walled shrinkage. Based on the shrinkage compensation value corresponding to the thin-wall shrinkage domain and the second preset interval coefficient, the thin-wall compensation verification interval is determined, and the compensation value with the largest fitting weight in the thin-wall compensation verification interval is taken as the final shrinkage compensation value of the thin-wall. Based on the shrinkage compensation value corresponding to the edge shrinkage domain and the third preset interval coefficient, the edge compensation verification interval is determined, and the compensation value with the largest fitting weight within the edge compensation verification interval is taken as the final edge shrinkage compensation value.

Images