Multi-element linear regression-based large-petal valve body cast steel part scale-down prediction method and device

Abstract

The present application relates to the technical field of casting process, in particular to a large-petal valve body cast steel piece scale reduction prediction method and device based on multiple linear regression. The method comprises the following steps: obtaining historical production data of different positions of a plurality of produced large-petal valve body cast steel pieces, determining actual scale reductions corresponding to each position according to the historical production data, and determining influence parameters including diameter, effective wall thickness, sand strength and position coefficient; performing multiple linear regression fitting according to the actual scale reductions corresponding to each position and the influence parameters to obtain a scale reduction prediction model; obtaining influence parameters corresponding to different positions of a to-be-produced large-petal valve body cast steel piece, inputting the influence parameters into the scale reduction prediction model to obtain predicted scale reductions corresponding to different positions, and determining a process scale of the to-be-produced large-petal valve body cast steel piece according to the predicted scale reductions. The present application can determine corresponding process scales for different positions, improve the adaptability of process scale setting to actual cast structure and production conditions, and reduce the dependence on artificial experience and complex numerical simulation.

Description

Technical Field

[0001] This invention relates to the field of casting technology, specifically to a method and apparatus for predicting the size reduction of large-valve body cast steel parts based on multiple linear regression. Background Technology

[0002] The large-disc valve body is a large annular cast steel component widely used in hydropower and other fields. During the casting process of the large-disc valve body, the casting undergoes volume shrinkage as it solidifies from a liquid state and cools to room temperature. Therefore, in the casting process design, it is usually necessary to pre-set a process scaling factor, that is, to enlarge the corresponding dimensions according to a certain ratio to compensate for the dimensional shrinkage that occurs during the solidification and cooling process, so that the dimensions of the casting after cooling meet the design requirements. The setting of the process scaling factor directly affects the dimensional accuracy of the casting. If the process scaling factor is not set properly, it is easy to cause the actual dimensions of the casting to deviate from the design dimensions, thereby increasing the workload of subsequent processing or repair, and in severe cases, it may even lead to the scrapping of the casting.

[0003] For large annular cast steel parts such as valve bodies with large flaps, their overall dimensions are quite large, with diameters reaching several meters, and they exhibit significant differences in wall thickness along the height direction. For example, the wall thickness is greater at the large and small ends, while the wall thickness in the middle section is relatively smaller. Due to the differences in geometry and wall thickness at different locations, and the fact that casting shrinkage is also affected by sand mold constraints and interactions between different locations, the actual degree of shrinkage exhibited by different locations of the valve body during solidification and cooling varies, making it difficult to rationally determine the process scaling.

[0004] Current methods for scaling down large cast steel parts typically rely on factors such as casting material, dimensions, and wall thickness. Fixed scaling down is determined using empirical data tables or industry manuals, or it is modified based on single factors like wall thickness and diameter. While these methods are relatively simple, they depend heavily on empirical data or single factors, making it difficult to comprehensively reflect the combined effects of different geometric dimensions, wall thicknesses, and locations on the actual shrinkage of the casting. Some existing methods can obtain the displacement field during the casting solidification process through numerical simulation and perform inverse deformation iteration accordingly. However, for large cast steel parts, the calculation process is complex and time-consuming, hindering rapid on-site process design.

[0005] In addition, there are methods that use machine learning models to predict casting dimensional deviations or geometric envelope density to correct process shrinkage. However, machine learning models typically require a large amount of training data and struggle to intuitively reflect the quantitative relationship between different influencing factors and shrinkage. Methods like geometric envelope density primarily analyze the geometric characteristics of the casting, failing to adequately consider sand mold strength and shrinkage differences at different locations within the casting. In particular, the sand mold constrains casting shrinkage during solidification and cooling, and existing methods for determining process shrinkage typically do not quantitatively consider actual sand mold strength as an independent influencing factor.

[0006] Therefore, existing technologies still struggle to comprehensively consider the combined effects of casting size, wall thickness, sand mold strength, and different locations on actual shrinkage when determining the process scaling of large annular cast steel parts such as large valve bodies. This makes it difficult to reasonably determine the corresponding process scaling for different locations of the casting. Summary of the Invention

[0007] This invention aims to address the technical problem that in existing technologies, the process reduction of large annular cast steel parts mainly relies on experience or single factors for determination, making it difficult to comprehensively reflect the combined influence of casting size, wall thickness, sand mold strength, and different positions on actual shrinkage, resulting in insufficient accuracy in setting process reduction at different positions. This invention provides a method and apparatus for predicting the reduction of large-valve body cast steel parts based on multiple linear regression. It can analyze historical production data of already produced large-valve body cast steel parts, establish quantitative relationships between diameter, effective wall thickness, sand mold strength, position coefficient, and actual reduction, and obtain a reduction prediction model. During the process design of large-valve body cast steel parts to be produced, reduction prediction can be performed based on the influence parameters corresponding to different positions. The predicted reduction at each position is obtained through the multiple linear regression model, and the corresponding process reduction is determined accordingly. This achieves a comprehensive characterization of the shrinkage characteristics of castings under different structures and production conditions, thereby improving the accuracy of process reduction setting and casting size control, and reducing reliance on manual experience and complex numerical simulations.

[0008] In a first aspect, embodiments of the present invention provide a method for predicting the size reduction of cast steel large-valve body parts based on multiple linear regression, the method comprising:

[0009] S1. Obtain historical production data for different positions of multiple produced large-valve body cast steel parts, and determine the actual scaling and influence parameters corresponding to each position based on the historical production data; the influence parameters include diameter, effective wall thickness, sand mold strength and position coefficient characterizing the influence of different positions on shrinkage behavior, and the effective wall thickness is determined according to the design wall thickness of different positions of the large-valve body cast steel parts. S2. Perform multiple linear regression fitting based on the actual scale and influencing parameters corresponding to each location to obtain the scale prediction model; S3. Obtain the influence parameters corresponding to different positions of the cast steel large-petal valve body to be produced, input the influence parameters into the scaling prediction model, and obtain the predicted scaling corresponding to different positions. S4. Determine the process scale of the cast steel large-valve body to be produced based on the predicted scale.

[0010] Optionally, the historical production data includes process design data, actual production process measurement data, and actual finished product size measurement data; The process design data includes the process dimensions, process scale, diameter and design wall thickness corresponding to each position; the measured data of the production process includes the sand mold strength; and the measured data of the finished product size includes the actual dimensions corresponding to each position obtained through three-dimensional detection. The determination of the actual scale includes: The process dimensions are enlarged according to the process reduction scale to obtain the enlarged dimensions; The difference between the enlarged size and the actual size at the corresponding position is determined, and the ratio of the difference to the process size is determined as the actual scale-down at the corresponding position.

[0011] Optionally, the different positions of the cast steel part of the large valve body include the large end, the middle section and the small end; The determination of the effective wall thickness includes: The design wall thickness of the large end is determined as the effective wall thickness of the large end. The design wall thickness of the small end is determined as the effective wall thickness of the small end. The effective wall thickness of the middle section is obtained by weighted averaging the designed wall thickness of the large end and the small end.

[0012] Optionally, the design wall thickness of the large end and the design wall thickness of the small end have the same weight in the weighted average process, and the arithmetic mean of the design wall thickness of the large end and the design wall thickness of the small end is determined as the effective wall thickness of the middle section.

[0013] Optionally, the position coefficient is a quantitative parameter characterizing the influence of different height positions of the large valve body cast steel part on the shrinkage behavior. The large end, the middle section and the small end correspond to different position coefficients, and the position coefficient is calibrated by regression fitting.

[0014] Optionally, the large end, the middle section, and the small end correspond to the first position coefficient, the second position coefficient, and the third position coefficient, respectively. The first position coefficient, the second position coefficient, and the third position coefficient are calibrated by regression fitting based on the contraction characteristics of different positions.

[0015] Optionally, the multiple linear regression fitting adopts the least squares method, and determines the intercept term and the regression coefficients corresponding to the diameter, effective wall thickness, sand mold strength and position coefficient respectively based on the actual scale corresponding to each position and the diameter, effective wall thickness, sand mold strength and position coefficient. The historical production data used for the multiple linear regression fitting includes no fewer than three produced large-valve body cast steel parts projects, and each of the produced large-valve body cast steel parts projects includes sample data corresponding to the large end, middle section and small end.

[0016] Optionally, the scaling prediction model satisfies: S=β0+β1D+β2T+β3SS+β4P; Where S is the predicted scale (in %), D is the diameter (in mm), T is the effective wall thickness (in mm), SS is the sand mold strength (in MPa), P is the position coefficient, β0 is the intercept term, β1 is the diameter regression coefficient, β2 is the effective wall thickness regression coefficient, β3 is the sand mold strength regression coefficient, and β4 is the position regression coefficient.

[0017] Secondly, embodiments of the present invention provide a device for predicting the scale reduction of cast steel large-valve body parts based on multiple linear regression, the device comprising: The data acquisition module is used to acquire historical production data from different locations of multiple manufactured large-valve body cast steel parts; The actual scale determination module is used to determine the actual scale corresponding to each position based on the historical production data. The influence parameter determination module is used to determine the influence parameters corresponding to each position based on the historical production data. The influence parameters include diameter, effective wall thickness, sand mold strength, and position coefficient characterizing the influence of different positions on shrinkage behavior. The effective wall thickness is determined according to the design wall thickness of different positions of the large valve body cast steel part. The regression model building module is used to perform multiple linear regression fitting based on the actual scaling and influencing parameters corresponding to each location to obtain the scaling prediction model. The scaling prediction module is used to obtain the influence parameters corresponding to different positions of the large-petal valve body cast steel part to be produced, and input the influence parameters into the scaling prediction model to obtain the predicted scaling corresponding to different positions. The process scaling setting module is used to determine the process scaling of the large valve body cast steel part to be produced based on the predicted scaling.

[0018] Optionally, the different positions of the cast steel part of the large valve body include the large end, the middle section and the small end; The influence parameter determination module is used to determine the design wall thickness of the large end as the effective wall thickness of the large end, determine the design wall thickness of the small end as the effective wall thickness of the small end, and perform a weighted average processing on the design wall thickness of the large end and the design wall thickness of the small end to obtain the effective wall thickness of the middle section. The regression model construction module is used to determine the intercept term and the regression coefficients corresponding to the diameter, effective wall thickness, sand mold strength, and position coefficient respectively, based on the actual scale corresponding to each position, as well as the diameter, effective wall thickness, sand mold strength, and position coefficient.

[0019] The method and apparatus for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to the present invention have at least the following beneficial effects: By acquiring historical production data of different positions of the already produced large-valve body cast steel parts, the actual scale reduction corresponding to each position is determined. The diameter, effective wall thickness, sand mold strength, and position coefficients characterizing shrinkage behavior at different positions are used as influencing parameters for multiple linear regression fitting. This establishes a quantitative relationship between the casting geometry, wall thickness characteristics, sand mold constraints, positional differences, and actual scale reduction. Furthermore, by inputting the influencing parameters corresponding to different positions of the large-valve body cast steel parts to be produced into the scale reduction prediction model, the predicted scale reduction for different positions can be obtained. Based on this, the corresponding process scale reduction can be determined, enabling the process scale reduction to adapt to the actual structure and production conditions of different positions of the large-valve body, thereby improving the accuracy of casting size control.

[0020] Compared to methods that primarily rely on empirical data or single dimensions and wall thickness factors to determine process reduction, this invention comprehensively considers factors such as sand mold strength, effective wall thickness, and positional differences, thus more fully reflecting the shrinkage characteristics of large annular cast steel parts at different locations. Furthermore, compared to numerical simulation methods that require complex solidification process calculations, this invention can predict the reduction of the casting to be produced using a multiple linear regression model. The calculation process is relatively simple, reducing reliance on complex calculation tools and manual experience, and helping to shorten the process design cycle. The device implements the above methods through modules such as data acquisition, actual reduction determination, influencing parameter determination, regression model construction, reduction prediction, and process reduction setting, thereby achieving modular processing for the prediction and determination of reduction of large-valve body cast steel parts.

[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. Attached Figure Description

[0022] Figure 1 This is a flowchart of the steps in Example 1 for predicting the size reduction of a large-valve body cast steel part based on multiple linear regression. Figure 2This is a schematic diagram showing the position of the large valve body in Example 2; Figure 3 This is a schematic diagram of the wall thickness of the large valve body in Example 2; Figure 4 This is a schematic diagram of the diameter of the large valve body in Example 2; Figure 5 This is a structural block diagram of the large-valve body cast steel part scaling prediction device based on multiple linear regression in Example 3. Detailed Implementation

[0023] The present invention will be further described in detail below with reference to experimental examples and specific embodiments. However, this should not be construed as limiting the scope of the above-mentioned subject matter of the present invention to the following embodiments. All technologies implemented based on the content of the present invention fall within the scope of protection of the present invention.

[0024] Unless otherwise specified, the terms "upper," "lower," "left," "right," "center," "inner," "outer," and "side" used in the description of specific embodiments of the present invention to indicate orientation or positional relationships are based on the orientation or positional relationships shown in the accompanying drawings, or the orientation or positional relationship in which the product / equipment / device is usually placed during use. These terms are merely for the purpose of facilitating the description of the present invention or simplifying the description in specific embodiments, and for enabling those skilled in the art to quickly understand the solution, and do not indicate or imply that a particular device / component / element must have a specific orientation, or be constructed and operated in a specific positional relationship. Therefore, they should not be construed as limitations on the present invention.

[0025] In the description of the embodiments of this invention, technical terms such as "first" and "second" only distinguish one entity or operation from another, and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary or secondary relationship of the indicated technical features. In the description of the embodiments of this invention, "multiple" means two or more, unless otherwise explicitly defined.

[0026] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0027] Example 1 This embodiment provides a method for predicting the shrinkage of large-petal valve body cast steel parts based on multiple linear regression, applicable to determining the process shrinkage of large annular cast steel parts such as large-petal valve bodies. During the solidification and cooling process of the large-petal valve body, the actual shrinkage at different locations is influenced by factors such as the casting diameter, wall thickness, sand mold strength, and location. Therefore, this embodiment utilizes historical production data from completed large-petal valve body cast steel parts to establish a quantitative relationship between the aforementioned influencing factors and the actual shrinkage. The established relationship is then used to predict the shrinkage at different locations of the large-petal valve body cast steel parts to be produced, providing a basis for setting the process shrinkage. Specifically, the method includes the following steps.

[0028] like Figure 1 As shown, the method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression in this embodiment includes the following steps: S1. Obtain historical production data for different positions of multiple produced large-valve body cast steel parts, and determine the actual scale and influencing parameters corresponding to each position based on the historical production data.

[0029] The historical production data comes from the large-petal valve body project that has been completed, including the corresponding project's process design data, actual production measurement data, and actual measurement data of the finished product dimensions after casting production. The process design data can at least reflect the process dimensions, process reduction, diameter, and design wall thickness at different locations of the casting. The actual production measurement data includes the actual measured sand mold strength, and the actual measurement data of the finished product dimensions reflects the actual dimensions of the casting after solidification and cooling.

[0030] For each historical project, based on the dimensional data used in the process design phase and the actual dimensions of the finished casting, the actual scaling at the corresponding location is determined. Specifically, the process dimensions can be enlarged according to the corresponding process scaling, and the enlarged dimensions are compared with the actual dimensions of the finished product to determine the actual scaling at the corresponding location. The obtained actual scaling reflects the degree of dimensional shrinkage actually produced by the casting under the corresponding geometric structure and production conditions, and serves as the target data for subsequent multiple linear regression fitting.

[0031] Simultaneously, influencing parameters corresponding to each location are determined based on historical production data. These parameters include diameter, effective wall thickness, sand mold strength, and position coefficient. Specifically, the diameter characterizes the radial dimension of the casting at the corresponding location; the effective wall thickness characterizes the influence of wall thickness conditions on shrinkage behavior at the corresponding location, and is determined based on the design wall thickness at different locations of the large valve body; the sand mold strength uses actual sand mold strength data obtained during production to reflect the constraints imposed by the sand mold on the solidification shrinkage of the casting; and the position coefficient transforms the influence of different height positions of the casting on shrinkage behavior into quantitative parameters that can participate in subsequent regression fitting. Therefore, each historical sample has a set of diameter, effective wall thickness, sand mold strength, and position coefficients corresponding to the actual shrinkage, providing input data for establishing a shrinkage prediction model.

[0032] S2, based on the actual scale and influencing parameters corresponding to each position, performs multiple linear regression fitting to obtain the scale prediction model.

[0033] The actual scales at different locations in multiple historical projects obtained in step S1 are correlated with the corresponding diameters, effective wall thicknesses, sand mold strengths, and position coefficients. Multiple linear regression is then performed based on this data. In the regression fitting process, the actual scale is used as the result to be fitted, while the diameter, effective wall thickness, sand mold strength, and position coefficient are used as variables affecting the actual scale. The regression coefficients and intercept terms for each influencing parameter are determined using historical samples, thereby establishing a quantitative relationship between each influencing parameter and the actual scale.

[0034] In one specific implementation, the least squares method can be used to complete the multiple linear regression fitting, that is, the corresponding regression coefficients are determined based on the deviation between the actual reduction size in historical samples and the model prediction results. As subsequent production projects and measured data increase, new production data can be used as new historical samples to participate in the regression fitting, updating the regression coefficients in the model, allowing the reduction size prediction model to be continuously improved based on actual production data. Through the above processing, the diameter, wall thickness, sand mold strength, and positional differences, which originally relied on manual experience for judgment, are transformed into model parameters with a quantitative relationship to the actual reduction size, resulting in a reduction size prediction model for subsequent casting reduction size prediction.

[0035] S3, obtain the influence parameters corresponding to different positions of the cast steel part of the large valve body to be produced, and input the influence parameters into the scale prediction model to obtain the predicted scale corresponding to different positions.

[0036] For the cast steel large-petal valve body parts to be produced that require process design, the diameter and design wall thickness corresponding to different positions are obtained according to the design data, and the corresponding effective wall thickness is determined according to the design wall thickness at different positions; the sand mold strength can be determined according to the molding process to be adopted, and the estimated sand mold strength can be obtained by using the sand mold strength data under the same sand mold process conditions in the past or the corresponding target control value; at the same time, the position coefficient corresponding to the position to be predicted is determined.

[0037] The diameter, effective wall thickness, estimated sand mold strength, and position coefficient of each location on the cast steel part of the large valve body to be produced are input into the shrinkage prediction model established in step S2. The shrinkage prediction model determines the predicted shrinkage corresponding to each location based on the influencing parameters and the corresponding regression coefficients obtained from the fitting. By processing different locations on the cast steel part of the large valve body to be produced in the same way, the predicted shrinkage corresponding to different locations can be obtained, thereby reflecting the shrinkage differences caused by geometric dimensions, wall thickness conditions, sand mold constraints, and positional differences at different locations of the same casting.

[0038] S4. Determine the process scale of the cast steel large-valve body to be produced based on the predicted scale.

[0039] The predicted scaling measures obtained in step S3 for different positions are used as the basis for setting the process scaling measures for the cast steel large-valve body to be produced. The corresponding process scaling measures are determined based on the prediction results for different positions, ensuring that the dimensional enlargement used in the mold design is compatible with the expected actual shrinkage at the corresponding position. Therefore, it is not necessary to uniformly set the process scaling measures for the entire large-valve body based solely on fixed empirical scaling measures. Instead, the process scaling measures for different positions can be specifically determined by considering the dimensions of the casting to be produced, wall thickness, sand mold strength, and shrinkage characteristics at different positions.

[0040] Through the above steps, this embodiment establishes a multivariate quantitative relationship between diameter, effective wall thickness, sand mold strength, position coefficient, and actual shrinkage using actual production data from completed production projects. This quantitative relationship is then used to predict the shrinkage of different positions in the large-valve body castings to be produced. This approach comprehensively considers the influence of multiple actual structural parameters and production process parameters on casting shrinkage, providing a quantitative basis for setting process shrinkage at different positions. This reduces the deviation caused by relying solely on empirical data or a single factor to determine process shrinkage, while avoiding complex solidification numerical simulations for each casting to be produced. This is beneficial for improving the accuracy of process shrinkage determination and the efficiency of process design.

[0041] Furthermore, this invention can also be extended to the process design of cast steel parts such as ball valve bodies that have large annular structures and exhibit differences in size, wall thickness, and shrinkage at different locations. By re-collecting historical production data for specific castings and performing multiple linear regression fitting using the diameter, effective wall thickness, sand mold strength, and position coefficient at the corresponding locations, a scaling prediction model adapted to the production conditions of this type of casting can be established. This provides a quantitative reference for setting process scaling at different locations, improving the adaptability of scaling settings to the specific casting structure and molding conditions.

[0042] Example 2 Based on the overall method and process described in Example 1, this embodiment specifically explains the acquisition of historical data for large-valve body cast steel parts, the determination of actual scale reduction, the construction of influencing parameters, the establishment of a multiple linear regression model, and the scale reduction prediction process for large-valve body cast steel parts to be produced. The prediction results are verified in conjunction with three large-valve body cast steel part projects.

[0043] like Figures 2-4 As shown, the large-disc valve body processed in this embodiment is a large annular cast steel part, wherein, Figure 2 The diagram illustrates the positional division of the large-end, middle section, and small-end of the large-disc valve body along its height. Figure 3 The wall thicknesses T1, T2, and T3 at different positions of the large valve body are shown. Figure 4 The diagram shows the diameters D1, D2, and D3 corresponding to different positions of the large valve body. Based on the differences in diameter, wall thickness, and contraction state at different positions of the large valve body, this embodiment constructs samples for the large end, middle section, and small end respectively and performs scaling prediction to ensure that the established scaling prediction model can reflect the actual contraction characteristics at different positions.

[0044] First, several completed large-valve body projects were selected, and process design data, actual production data, and finished product dimensional measurement data were collected for each project. Each project was divided into three inspection locations: the large end, the middle section, and the small end. The process design data included the process scale set for each project and the process dimensions in the casting design drawings. These process dimensions are the mold design dimensions, including the scaled-down dimensions. The inner diameters of the large end (D1), the middle section (D2), and the small end (D3), as well as the corresponding design wall thicknesses (T1, T2, and T3) for each location, were also recorded. The actual production data included sand mold strength, and the finished product dimensional measurement data were the actual dimensions obtained through 3D inspection after casting production. By correlating the design stage data, production process data, and final measurement results of the completed production projects, historical samples can be formed to characterize the actual shrinkage state of castings under different structural and production conditions, providing foundational data for subsequent regression model fitting.

[0045] For each inspection location, the corresponding actual scale is calculated based on the process dimensions, process scale, and the actual dimensions obtained from 3D inspection. This actual scale is then used as the dependent variable in a multiple linear regression model. The actual scale is determined according to the following formula: S_actual = [L_process × (1 + S_process) L_actual] / L_process Where S_actual represents the actual scale at the corresponding position, L_process represents the process dimension, S_process represents the corresponding process scale, and L_actual represents the actual dimension at the corresponding position after the casting is produced. That is, scale = (process dimension × (1 + process scale) - actual dimension) / process dimension. Through the above processing, the actual dimensional changes that occur after production are converted into scale data that can be used for regression fitting, so that the actual shrinkage state of different projects and different positions can be represented in a unified data form.

[0046] After obtaining the actual scale at each location, the parameters affecting the scale are defined based on the shrinkage characteristics of the large-valve body. This embodiment uses four types of influencing parameters: diameter D, effective wall thickness T, sand mold strength SS, and position coefficient P. Specifically, diameter D uses the diameter data of the corresponding detection location; sand mold strength SS uses the sand mold strength data obtained during production; effective wall thickness T characterizes the influence of different wall thicknesses on shrinkage behavior; and position coefficient P characterizes the influence of different height positions of the large-valve body on shrinkage behavior.

[0047] For the effective wall thickness T, the design wall thickness of each end (large and small) is used as the effective wall thickness. For the middle section, since its wall thickness is thinner, its shrinkage is also affected by the thicker wall regions at both ends. Therefore, the design wall thickness of the middle section is not directly used as the effective wall thickness; instead, a weighted average is applied based on the wall thicknesses of the large and small ends. In this embodiment, the arithmetic mean is preferred, i.e.: T_middle = (T_large + T_small) / 2 Where T_middle represents the effective wall thickness of the middle section, T_large represents the designed wall thickness of the large end, and T_small represents the designed wall thickness of the small end. Through the above effective wall thickness processing, the influence of the thick-walled regions of the large and small ends on the shrinkage behavior of the thin-walled region in the middle is incorporated into the model input, so that the wall thickness parameter not only reflects the local geometry, but also reflects the shrinkage driving relationship between different wall thickness regions of the large valve body.

[0048] This embodiment further quantifies the impact of different height positions of the large valve body on the contraction behavior using a position coefficient P. Based on the historical project data used in this embodiment, and after regression fitting calibration, the following position coefficients are used for the large end, middle section, and small end: Large end (near the riser): P = -300; Middle section: P = 300; Small end (farthest from the riser): P = 500.

[0049] The aforementioned position coefficients are used to reflect the shrinkage pattern: the shrinkage at the small end is the most significant, followed by the middle section, and the shrinkage at the large end is the least significant. It should be noted that the specific values ​​of the position coefficients are determined through regression fitting based on the historical production data used. When establishing a scaling prediction model for different historical production data, the corresponding position coefficients can be determined based on the regression fitting results.

[0050] After completing the above sample construction and determining the influencing parameters, a multiple linear regression equation was established based on the actual scale, diameter D, effective wall thickness T, sand mold strength SS, and position coefficient P corresponding to each position in several produced large-valve body castings. The general multiple linear regression relationship used in this embodiment is: S = β0+ β1D + β2T + β3SS + β4P + ξ Where S represents the shrinkage, D represents the diameter, T represents the effective wall thickness, SS represents the sand mold strength, P represents the position coefficient, β0 represents the intercept term, β1, β2, β3, and β4 represent the regression coefficients corresponding to the diameter, effective wall thickness, sand mold strength, and position coefficient, respectively, and ξ represents the error term in the regression relationship. The least squares method was used to fit the model and solve for the corresponding regression coefficients using data from multiple collected items and different positions. Through the above regression processing, the influence of diameter, effective wall thickness, sand mold strength, and different height positions on casting shrinkage was transformed into a quantitative relationship between each influencing parameter and the shrinkage.

[0051] This embodiment utilizes data from at least three completed large-valve body production projects for initial model building. Since each project includes three sample locations—the large end, the middle section, and the small end—at least three projects can generate at least nine sample points. As production projects and measured data accumulate, new data can be used for regression fitting to iteratively optimize the model. The regression coefficients can also determine the scaling change corresponding to unit changes in each influencing parameter, thus forming a quantitative reference that can be used for process adjustments.

[0052] It should be noted that, for a given set of historical production data for large-valve bodies, with diameter D, effective wall thickness T, sand mold strength SS, and position coefficient P as independent variables and actual scale reduction as the dependent variable, a set of regression coefficients corresponding to this set of historical production data can be determined by using the least squares method for multiple linear regression fitting. The determined regression coefficients are used to characterize the quantitative relationship between each influencing parameter and the actual scale reduction under the corresponding historical production data conditions. Therefore, given the historical production data and variable settings, the corresponding scale reduction prediction model can be determined based on the least squares fitting results. When new production projects are added or historical production data changes, multiple linear regression fitting can be performed again based on the updated historical production data to determine the corresponding regression coefficients and update the scale reduction prediction model.

[0053] Based on data from multiple previous projects, the regression coefficients obtained in this embodiment are: β0 = 1.48 β1 = -0.0001114 β2 = 0.00042 β3 = -0.328 β4 = 0.00067 Therefore, the scaling prediction model used in this embodiment is obtained as follows: S = 1.48 + 0.00042T 0.0001114D 0.328SS + 0.00067P Where S represents the predicted scale (%), T represents the effective wall thickness (mm), D represents the diameter (mm), SS represents the sand mold strength (MPa), and P represents the position coefficient. This specific model was used for scale prediction and verification in the following three large-disc valve body projects.

[0054] First set of project data: This group takes a large-disc valve body product from a certain project as an example. Its material is ZG20Mn, the total height H of the casting is 2460 mm, the whole is in the shape of a ring, and it is divided into three areas: the large end, the middle section and the small end. First, the process dimensions, design wall thickness, sand mold strength and position coefficient corresponding to the three positions of the project were collected. The basic data are shown in Table 1.

[0055] Table 1 Basic data of a certain large-valve body No. 1 cast steel component

[0056] Based on the above wall thickness data, the effective wall thickness corresponding to the three locations is determined. The large end and small end each use their own designed wall thickness; therefore: T1 = 365 mm T3 = 360 mm Considering the shrinkage effect of the thick walls at both ends in the middle section, the arithmetic mean of the design wall thicknesses of the large and small ends is used as the effective wall thickness. Therefore: T2= (365 + 360) / 2 = 362.5 mm Therefore, the diameter, effective wall thickness, sand mold strength, and position coefficient at the three locations are substituted into the scaling prediction model.

[0057] For the larger end, P = -300, we get: S1 = 1.48 + 0.00042 × 365 0.0001114×3500 0.328×0.66 + 0.00067×(-300) = 1.48 + 0.1533 0.3899 0.21648 0.201 = 0.82592% → 0.83% For the middle segment, P = 300, we get: S2 = 1.48 + 0.00042 × 362.5 0.0001114×3760 0.328×0.66 + 0.00067×300 = 1.48 + 0.15225 0.418864 0.21648 + 0.201 = 1.197906% → 1.20% For the little end, P = 500, we get: S3 = 1.48 + 0.00042 × 360 0.0001114×2520 0.328×0.66 + 0.00067×500 = 1.48 + 0.1512 0.280728 0.21648 + 0.335 = 1.468992% → 1.47% After the casting was completed, 3D inspection revealed actual shrinkages of 0.88%, 1.23%, and 1.43% at the large end, middle section, and small end, respectively. Comparing these with the predicted values ​​of 0.83%, 1.20%, and 1.47%, the differences between the predicted and actual values ​​were -0.05%, -0.03%, and 0.04%, respectively. This demonstrates that, under the given dimensions, wall thickness, and sand mold strength conditions of this project, the predicted shrinkages at the three locations obtained from the established model are quite close to the actual shrinkages after production. This indicates that the aforementioned multiple influencing parameters can be used to quantitatively characterize the shrinkage state at different locations.

[0058] Second set of project data: This group also uses a large-valve body product made of ZG20Mn material as an example, and uses the above-mentioned scaling prediction model to predict the scaling at different positions, and compares it with the actual scaling after production. First, the basic data corresponding to the three positions are obtained, as shown in Table 2.

[0059] Table 2 Basic Data of Cast Steel Part No. 2 of a Large Valve Body

[0060] Based on the method for determining the effective wall thickness, the effective wall thicknesses for the large end and the small end are respectively: T1 = 402 mm T3 = 409 mm The middle section uses the arithmetic mean of the wall thicknesses at both ends: T2= (402 + 409) / 2 = 405.5 mm Substitute the parameters corresponding to each position into the regression model.

[0061] For the larger end, P = -300: S1 = 1.48 + 0.00042 × 402 0.0001114×4290 0.328×0.606 + 0.00067×(-300) = 1.48 + 0.16884 0.477906 0.198768 0.201 = 0.771166% → 0.77% For the middle segment, P = 300: S2 = 1.48 + 0.00042 × 405.5 0.0001114×4420 0.328×0.606 + 0.00067×300 = 1.48 + 0.17031 0.492388 0.198768 + 0.201 = 1.160154% → 1.16% For the little end, P = 500: S3 = 1.48 + 0.00042 × 409 0.0001114×3172 0.328×0.606 + 0.00067×500 = 1.48 + 0.17178 0.353361 0.198768 + 0.335 = 1.4346512% → 1.43% After the casting was produced, 3D measurement revealed actual reductions of 0.81%, 1.22%, and 1.51% for the large end, middle section, and small end, respectively. The difference between the predicted value of 0.77% for the large end and the actual value of 0.81% was -0.04%, the difference between the predicted value of 1.16% for the middle section and the actual value of 1.22% was -0.06%, and the difference between the predicted value of 1.43% for the small end and the actual value of 1.51% was -0.08%. These results indicate that, under conditions different from the diameter, wall thickness, and sand mold strength of the first group of projects, the reduction prediction model established in this embodiment can still obtain prediction results that are relatively close to the actual reduction.

[0062] Third set of project data: This group continues to use a large-disc valve body product made of ZG20Mn material as an example to illustrate the application of the method described in this embodiment under conditions of larger diameter and lower sand mold strength. The basic data corresponding to the three positions are shown in Table 3.

[0063] Table 3. Basic data of a certain large-valve body No. 3 cast steel component.

[0064] Based on the method for determining the effective wall thickness, the effective wall thicknesses for the large end and the small end are as follows: T1 = 400 mm T3 = 395 mm The effective wall thickness of the intermediate section is: T2= (400 + 395) / 2 = 397.5 mm Input the above parameters into the scaling prediction model respectively.

[0065] For the larger end, P = -300: S1 = 1.48 + 0.00042 × 400 0.0001114×5070 0.328×0.366 + 0.00067×(-300) = 1.48 + 0.168 0.564798 0.120048 0.201 = 0.762154% → 0.76% For the middle segment, P = 300: S2 = 1.48 + 0.00042 × 397.5 0.0001114×5220 0.328×0.366 + 0.00067×300 = 1.48 + 0.16695 0.581508 0.120048 + 0.201 = 1.146394% → 1.15% For the little end, P = 500: S3 = 1.48 + 0.00042 × 395 0.0001114×3734 0.328×0.366 + 0.00067×500 = 1.48 + 0.1659 0.415968 0.120048 + 0.335 = 1.4448844% → 1.44% The conditions for this project group include a diameter exceeding 5000 mm and a relatively low sand mold strength of 0.366 MPa. Predicted results show a shrinkage of approximately 0.76% at the large end, 1.15% in the middle section, and 1.44% at the small end. Actual shrinkages after production were 0.82%, 1.20%, and 1.47%, respectively. The difference between the predicted and actual values ​​for the large end was -0.06%, for the middle section -0.05%, and for the small end -0.03%. Furthermore, the predicted results for the three locations show a relationship of larger shrinkage at the small end, followed by the middle section, and smaller shrinkage at the large end, corresponding to the shrinkage differences at different locations characterized by position coefficients in this embodiment.

[0066] For the large-petal valve body cast steel parts to be produced, the above model application process can be followed. First, obtain the diameter D and design wall thickness corresponding to different positions, and calculate the corresponding effective wall thickness T according to the aforementioned method for determining effective wall thickness. For the sand mold strength SS, the average sand mold strength or target control value under the same sand mold process conditions in the past can be used as the estimated sand mold strength. At the same time, the position coefficient P corresponding to each position is determined. Subsequently, the diameter D, effective wall thickness T, estimated sand mold strength SS, and position coefficient P are input into the scaling prediction model established based on historical production data. The predicted actual scaling S of each position is calculated respectively, and the process scaling setting value of the corresponding position of the new large-petal valve body is determined based on the obtained prediction results. Thus, the quantitative relationship of "structural parameters, production conditions - actual scaling" formed by historical projects can be further used for new casting process design, realizing the transfer from historical production data to the process scaling setting of the product to be produced.

[0067] As can be seen from the above three projects, this embodiment utilizes a multiple linear regression model to incorporate factors such as sand mold strength, casting diameter, effective wall thickness, and axial position into the scaling prediction process for large-valve body cast steel parts. Specifically, the effective wall thickness treatment considers the influence of the thick-walled regions at both ends on the shrinkage of the thin-walled region in the middle; the sand mold strength reflects the constraints of actual molding conditions on shrinkage; and the position coefficient distinguishes the shrinkage state at different height positions. Compared to methods that rely solely on empirical values ​​or single factors to correct process scaling, this embodiment can generate corresponding quantitative prediction results for specific structures and production conditions. Furthermore, its calculation process does not require time-consuming and complex solidification numerical simulations for each new product, and the obtained regression coefficients can reflect the quantitative relationship between different influencing parameters and scaling, thus facilitating its application in the process design of large annular cast steel parts such as large-valve bodies.

[0068] Example 3 This embodiment provides a device for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression. It is used to establish a scale reduction prediction model based on historical production data of already produced large-valve body cast steel parts, and to use this model to predict the scale reduction at different locations on the large-valve body cast steel parts to be produced, thus providing a basis for setting the process scale reduction at the corresponding locations. Figure 5 As shown, the device includes a data acquisition module, an actual scaling determination module, an influencing parameter determination module, a regression model construction module, a scaling prediction module, and a process scaling setting module. Each module interacts with the data in the following order: historical data processing, model building, input of parameters for the casting to be produced, scaling prediction, and process scaling determination.

[0069] The data acquisition module is used to acquire historical production data from different locations of multiple produced large-disc valve body cast steel parts. This historical production data includes process design data, actual production process measurement data, and actual finished product dimension measurement data. The process design data includes the process dimensions, process reduction, diameter, and design wall thickness corresponding to different locations. The actual production process measurement data includes sand mold strength, and the actual finished product dimension measurement data includes the actual dimensions obtained through 3D inspection after casting production. For a project involving a produced large-disc valve body cast steel part, data corresponding to the large end, middle section, and small end can be collected separately. The acquired historical production data is then provided to the actual reduction determination module and the influence parameter determination module, respectively, to serve as sample data for constructing the subsequent regression model.

[0070] The actual scale determination module is used to determine the actual scale corresponding to each position based on historical production data provided by the data acquisition module. Specifically, the actual scale is determined based on the process dimensions, process scale, and actual dimensions after production at the corresponding position. This can be achieved by enlarging the process dimensions according to the corresponding process scale, determining the size difference between the enlarged size and the actual size, and then obtaining the actual scale at the corresponding position based on the relationship between the size difference and the process dimensions. Thus, the dimensional shrinkage state of the large-valve body after actual production is converted into data that can participate in multiple linear regression fitting, and the obtained actual scale is provided to the regression model construction module.

[0071] The influencing parameter determination module is used to determine the diameter, effective wall thickness, sand mold strength, and position coefficient corresponding to each location based on historical production data. Specifically, the diameter uses the diameter data corresponding to the location, the sand mold strength uses the sand mold strength data obtained from actual measurements during production, and the position coefficient is used to quantitatively characterize the influence of different height positions of the large-valve body on the shrinkage behavior. For the effective wall thickness, the designed wall thickness of the large and small ends is used as the effective wall thickness; for the middle section, due to the shrinkage caused by the thick walls at both ends, the effective wall thickness of the middle section is obtained by weighted averaging of the designed wall thicknesses of the large and small ends, preferably using the arithmetic mean of the designed wall thicknesses at both ends. Through the above processing, a set of diameter, effective wall thickness, sand mold strength, and position coefficient corresponding to each actual scale can be formed, and these influencing parameters are provided to the regression model construction module.

[0072] The regression model construction module receives the actual scale output from the actual scale determination module and the diameter, effective wall thickness, sand mold strength, and position coefficient output from the influencing parameter determination module. Based on the above data from multiple already produced large-valve body castings, it performs multiple linear regression fitting to establish a quantitative relationship between each influencing parameter and the actual scale. In this embodiment, the least squares method can be used to determine the intercept term, diameter regression coefficient, effective wall thickness regression coefficient, sand mold strength regression coefficient, and position regression coefficient, thereby obtaining the scale prediction model. This scale prediction model is provided as the output of the model construction stage to the scale prediction module for processing the influencing parameters corresponding to the large-valve body castings to be produced. As subsequent production data accumulates, newly added production data can be used as historical samples to re-participate in the regression fitting to update the scale prediction model.

[0073] The scaling prediction module is used to obtain the influence parameters corresponding to different positions of the large-petal valve body casting steel part to be produced, and to determine the predicted scaling for each position based on the scaling prediction model obtained by the regression model construction module. Specifically, for the large-petal valve body casting steel part to be produced, the diameter and design wall thickness at the corresponding position are obtained, and the effective wall thickness is obtained according to the aforementioned effective wall thickness determination method; the sand mold strength can be estimated by using the average sand mold strength or the target control value corresponding to the same sand process in the past, and the position coefficients corresponding to different positions are determined. The diameter, effective wall thickness, estimated sand mold strength, and position coefficients are input into the scaling prediction model, and the predicted scaling for the corresponding position is determined according to each influence parameter and the corresponding regression coefficient. The above processing is performed on the large end, the middle section, and the small end respectively, so as to obtain the predicted scaling for different positions of the large-petal valve body casting steel part to be produced, and the predicted scaling is provided to the process scaling setting module.

[0074] The process scaling setting module receives the predicted scaling corresponding to different positions output by the scaling prediction module, and determines the process scaling setting value for the corresponding position of the large valve body casting to be produced based on the predicted scaling. Thus, during the operation of the device, historical samples are first formed using the process design data of the produced castings, sand mold strength, and actual measured data of the finished product dimensions. The actual scaling determination module and the influence parameter determination module obtain the result quantity and influence parameters required for model fitting, respectively. Then, the scaling prediction model is established by the regression model construction module. For the casting to be produced, the scaling prediction module obtains the predicted scaling for different positions based on its diameter, effective wall thickness, estimated sand mold strength, and position coefficient. Finally, the process scaling setting module generates the corresponding process scaling setting result, thereby completing the data processing flow from historical production data to the determination of the process scaling of the casting to be produced.

[0075] Through data interaction between the aforementioned modules, this embodiment can incorporate the diameter, effective wall thickness, sand mold strength, and shrinkage differences at different locations of the large-valve body cast steel part into the scaling prediction process, enabling the process scaling at different locations to be determined separately based on corresponding structural parameters and production conditions. Compared to methods that primarily rely on fixed empirical scaling or correction based on a single factor, this device can utilize historical actual production data to form quantitative predictive relationships. Simultaneously, through functional modules, it achieves continuous processing of historical data, model fitting, scaling prediction, and process scaling setting, providing a corresponding data processing foundation for determining the process scaling of large ring-shaped cast steel parts.

[0076] It should be understood that the various modules of the large-valve body cast steel part scaling prediction device based on multiple linear regression provided in the above embodiments are only illustrated by the division of each functional module in the above description. In practical applications, the above functions can be assigned to different functional modules as needed. That is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0077] The functional modules in the above embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units 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 functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of the embodiments of the present invention.

[0078] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0079] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for predicting the scale reduction of cast steel large-valve body parts based on multiple linear regression, characterized in that, Includes the following steps: S1. Obtain historical production data for different positions of multiple produced large-valve body cast steel parts, and determine the actual scaling and influence parameters corresponding to each position based on the historical production data; the influence parameters include diameter, effective wall thickness, sand mold strength and position coefficient characterizing the influence of different positions on shrinkage behavior, and the effective wall thickness is determined according to the design wall thickness of different positions of the large-valve body cast steel parts. S2. Perform multiple linear regression fitting based on the actual scale and influencing parameters corresponding to each location to obtain the scale prediction model; S3. Obtain the influence parameters corresponding to different positions of the cast steel large-petal valve body to be produced, input the influence parameters into the scaling prediction model, and obtain the predicted scaling corresponding to different positions. S4. Determine the process scale of the cast steel large-valve body to be produced based on the predicted scale.

2. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 1, characterized in that, The historical production data includes process design data, actual production process data, and actual finished product size data. The process design data includes the process dimensions, process scale, diameter and design wall thickness corresponding to each position; the measured data of the production process includes the sand mold strength; and the measured data of the finished product size includes the actual dimensions corresponding to each position obtained through three-dimensional detection. The determination of the actual scale includes: The process dimensions are enlarged according to the process reduction scale to obtain the enlarged dimensions; The difference between the enlarged size and the actual size at the corresponding position is determined, and the ratio of the difference to the process size is determined as the actual scale-down at the corresponding position.

3. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 1, characterized in that, The large valve body cast steel component has different positions including the large end, the middle section and the small end; The determination of the effective wall thickness includes: The design wall thickness of the large end is determined as the effective wall thickness of the large end. The design wall thickness of the small end is determined as the effective wall thickness of the small end. The effective wall thickness of the middle section is obtained by weighted averaging the designed wall thickness of the large end and the small end.

4. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 3, characterized in that, The design wall thickness of the large end and the design wall thickness of the small end have the same weight in the weighted average process, and the arithmetic mean of the design wall thickness of the large end and the design wall thickness of the small end is determined as the effective wall thickness of the middle section.

5. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 3, characterized in that, The position coefficient is a quantitative parameter characterizing the influence of different height positions of the large valve body cast steel part on the shrinkage behavior. The large end, the middle section and the small end correspond to different position coefficients, and the position coefficient is calibrated by regression fitting.

6. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 5, characterized in that, The large end, middle section, and small end correspond to the first position coefficient, the second position coefficient, and the third position coefficient, respectively. The first position coefficient, the second position coefficient, and the third position coefficient are calibrated by regression fitting based on the contraction characteristics of different positions.

7. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 1, characterized in that, The multiple linear regression fitting adopts the least squares method. Based on the actual scale corresponding to each position, as well as the diameter, effective wall thickness, sand mold strength, and position coefficient, the intercept term and the regression coefficients corresponding to the diameter, effective wall thickness, sand mold strength, and position coefficient are determined respectively. The historical production data used for the multiple linear regression fitting includes no fewer than three produced large-valve body cast steel parts projects, and each of the produced large-valve body cast steel parts projects includes sample data corresponding to the large end, middle section and small end.

8. The method for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 7, characterized in that, The scaling prediction model satisfies: S=β0+β1D+β2T+β3SS+β4P; Where S is the predicted scale (in %), D is the diameter (in mm), T is the effective wall thickness (in mm), SS is the sand mold strength (in MPa), P is the position coefficient, β0 is the intercept term, β1 is the diameter regression coefficient, β2 is the effective wall thickness regression coefficient, β3 is the sand mold strength regression coefficient, and β4 is the position regression coefficient.

9. A device for predicting the scale reduction of cast steel large-valve body parts based on multiple linear regression, characterized in that, include: The data acquisition module is used to acquire historical production data from different locations of multiple manufactured large-valve body cast steel parts; The actual scale determination module is used to determine the actual scale corresponding to each position based on the historical production data. The influence parameter determination module is used to determine the influence parameters corresponding to each position based on the historical production data. The influence parameters include diameter, effective wall thickness, sand mold strength, and position coefficient characterizing the influence of different positions on shrinkage behavior. The effective wall thickness is determined according to the design wall thickness of different positions of the large valve body cast steel part. The regression model building module is used to perform multiple linear regression fitting based on the actual scaling and influencing parameters corresponding to each location to obtain the scaling prediction model. The scaling prediction module is used to obtain the influence parameters corresponding to different positions of the large-petal valve body cast steel part to be produced, and input the influence parameters into the scaling prediction model to obtain the predicted scaling corresponding to different positions. The process scaling setting module is used to determine the process scaling of the large valve body cast steel part to be produced based on the predicted scaling.

10. The device for predicting the scale reduction of large-valve body cast steel parts based on multiple linear regression according to claim 9, characterized in that, The large valve body cast steel component has different positions including the large end, the middle section and the small end; The influence parameter determination module is used to determine the design wall thickness of the large end as the effective wall thickness of the large end, determine the design wall thickness of the small end as the effective wall thickness of the small end, and perform a weighted average processing on the design wall thickness of the large end and the design wall thickness of the small end to obtain the effective wall thickness of the middle section. The regression model construction module is used to determine the intercept term and the regression coefficients corresponding to the diameter, effective wall thickness, sand mold strength, and position coefficient respectively, based on the actual scale corresponding to each position, as well as the diameter, effective wall thickness, sand mold strength, and position coefficient.

Images