Method and device for predicting shrinkage rate of turbine blade during casting and medium

Abstract

The invention provides a method and device for predicting the shrinkage rate of a turbine blade during casting and a medium, and belongs to the technical field of turbine blade manufacturing, and the method comprises the following steps: obtaining a sub-data set, obtaining a data set, constructing a shrinkage rate model, and importing the data set into the shrinkage rate model; carrying out training iterative optimization on the shrinkage rate model until the data and the process parameters of the geometric attributes of the turbine blade casting which are input and tested can calculate that the comparison between the test shrinkage rate of the turbine blade mold-casting and the actual shrinkage rate is smaller than a set error; and inputting the geometric attribute data and the process parameters of the turbine blade casting to be developed into the shrinkage rate model, and outputting the shrinkage rate of the turbine blade mold-casting by the shrinkage rate model. And meanwhile, a rapid forming wax mold does not need to be manufactured, so that the cost is saved.

Description

Technical Field

[0001] The invention belongs to the technical field of turbine blade manufacturing, and in particular relates to a method, a device and a medium for predicting shrinkage rate of turbine blades during casting. Background Art

[0002] Aerospace engines have been developing towards high precision and high thrust-to-weight ratio. On this basis, for turbine blades (reference Figure 1 ) have increasingly higher dimensional requirements, with general tolerances requiring CT (Circularity Tolerance) to be above level 5, and blade profiles to be below ±0.1mm. Due to the different requirements of different aerospace engines, turbine blades for different aerospace engines will also differ in size and shape; because turbine blades are huge and expensive, the size qualification rate of cast turbine blades has become an important factor plaguing the development of the industry. In the turbine blade investment casting process, the factor that most affects blade dimensional accuracy is the design of the mold-casting shrinkage rate. The current process is from mold to wax mold to casting, where there is a shrinkage rate from mold to wax mold, and there is also a shrinkage rate from wax mold to casting. The two shrinkage rates need to be added together to get the shrinkage rate from mold to casting.

[0003] However, there are two existing methods for obtaining the shrinkage rate of the mold-casting: the first method is to find blades with similar structures among similar blades that have been made, and give the shrinkage rate based on experience, but the cast blades often have large dimensional errors from the original design; the second method is to use rapid prototyping wax molds to conduct experiments and analyze the actual shrinkage rate from the wax mold to the casting; on the one hand, due to the low dimensional accuracy of the rapid prototyping wax mold, a large error will be introduced during the analysis, and on the other hand, this method can only analyze the shrinkage rate of the wax mold-casting, while the mold-wax mold shrinkage rate still needs to be estimated based on experience, and the required shrinkage rate is the mold-casting (the current algorithm is to calculate the wax mold-casting, then estimate the mold-wax mold, and then combine the above two shrinkage rates to calculate the shrinkage rate of the mold-casting); therefore, the dimensional error between the cast blade and the original design is also relatively large.

[0004] Therefore, a new method for predicting the mold-casting shrinkage during turbine blade casting is needed. Summary of the invention

[0005] In view of the above problems, the present invention proposes a method for predicting the shrinkage rate of turbine blades during casting, comprising the following steps: Get shrinkage data; Acquire a sub-dataset, wherein the sub-dataset includes data on geometric properties of a turbine blade casting, data on geometric properties of an inner cavity of a mold corresponding to the turbine blade casting, process parameters, and shrinkage rate data; Obtain a data set, where the data set includes N sub-data sets, and the geometric attributes of the turbine blade castings, the geometric attributes of the mold cavity, and the process parameters in the N sub-data sets are all different; Construct a shrinkage rate model and import the above data set into the shrinkage rate model; Perform training iteration optimization on the shrinkage rate model until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data of the geometric attributes of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate; Input the data of the geometric attributes of the turbine blade casting to be developed and the process parameters into the shrinkage rate model, and the shrinkage rate model outputs the shrinkage rate of the turbine blade mold-casting.

[0006] Further, the shrinkage rate data includes the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange.

[0007] Further, the obtaining of the shrinkage rate data includes the following steps: Take the mold cavity contour model as the reference object and the casting outer contour model as the test object, perform least squares fitting to obtain a comparison model; Divide cross-sections along the transverse and / or longitudinal directions on the comparison model to obtain cross-sectional views; Obtain the points of the casting outer contour model and the points of the mold cavity contour model at the same position on the cross-sectional view, and obtain the shrinkage rate through the points of the casting outer contour model and the points of the mold cavity contour model; Repeat the steps of obtaining the points of the casting outer contour model and the points of the mold cavity contour model at the same position on the cross-sectional view, and obtaining the shrinkage rate through the points of the casting outer contour model and the points of the mold cavity contour model, and sequentially obtain the shrinkage rates at different positions on the cross-sectional view until the shrinkage rates at all positions on the cross-sectional view are obtained; Repeat the step of dividing cross-sections along the transverse and / or longitudinal directions on the comparison model to obtain cross-sectional views, and sequentially obtain different cross-sections until all cross-sections are obtained; Obtain the required shrinkage rate by screening the above shrinkage rates.

[0008] Further, the casting outer contour model is a casting outer contour point cloud model obtained by blue light scanning the casting outer contour; The mold cavity contour model is a mold cavity point cloud model obtained by blue light scanning the mold cavity contour or the three-dimensional digital model of the mold design.

[0009] Further, the data of the geometric properties of the turbine blade casting and the geometric properties of the mold cavity both include the overall length, the blade length, the shroud length, the shroud width, the shroud height, the maximum chord length of the blade, the minimum chord length of the blade, and the maximum thickness of the blade cross-section.

[0010] Further, the shrinkage rate model is trained with iterative optimization until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data of the geometric properties of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate, including the following steps: The shrinkage rate model uses the support vector machine algorithm and / or the random forest algorithm and / or the deep neural network algorithm for training with iterative optimization; Continuously input the test data of the geometric properties of the turbine blade casting and the process parameters into the shrinkage rate model multiple times, and the shrinkage rate model calculates and outputs the test shrinkage rate of the turbine blade mold-casting.

[0011] By comparing the error between the corresponding test shrinkage rate and the actual shrinkage rate, If the errors are all within the set error, then save the above model.

[0012] If any of the errors is greater than the set error, then repeat the step of training and iteratively optimizing the shrinkage rate model using the support vector machine algorithm and / or the random forest algorithm and / or the deep neural network algorithm.

[0013] Further, the process parameters include the wax injection temperature, the injection pressure, the injection flow rate, the holding pressure time, the shell mold system, the number of shell mold layers, the shell mold preheating temperature, the pouring temperature, and the alloy type.

[0014] A device for predicting the shrinkage rate during the casting of a turbine blade, comprising: A first collection unit for obtaining shrinkage rate data; the sub-dataset includes the data of the geometric properties of the turbine blade casting, the data of the geometric properties of the mold cavity corresponding to the turbine blade casting, the process parameters, and the shrinkage rate data; A second collection unit for obtaining a dataset, the dataset including N sub-datasets, and the data of the geometric properties of the turbine blade casting, the data of the geometric properties of the mold cavity, and the process parameters in the N sub-datasets are all different; A construction unit for constructing a shrinkage rate model and importing the above dataset into the shrinkage rate model; A processing unit for training and iteratively optimizing the shrinkage rate model until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data of the geometric properties of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate; A prediction unit for inputting the data of the geometric properties of the turbine blade casting to be developed and the process parameters into the shrinkage rate model, and the shrinkage rate model outputs the shrinkage rate of the turbine blade mold-casting.

[0015] A device for predicting the shrinkage rate during the casting of turbine blades, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0016] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0017] Advantages of the present invention: In the present invention, existing data is imported into the shrinkage rate model, and the shrinkage rate model is trained, iteratively optimized until the test shrinkage rate of the turbine blade mold-casting calculated from the input data of the geometric properties and process parameters of the turbine blade casting for testing is less than the set error compared with the actual shrinkage rate; then the data and process parameters of the blade to be developed are input into the shrinkage rate model, and the shrinkage rate model can output the shrinkage rate of the turbine blade mold-casting, improving the acquisition speed and accuracy of the shrinkage rate. At the same time, since there is no need to make a rapid prototyping wax mold, the cost is saved.

[0018] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification or be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures pointed out in the specification and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 Shows a schematic structural diagram of a turbine blade in the prior art.

[0021] Figure 2 Shows a schematic flowchart of the method for predicting the shrinkage rate during the casting of turbine blades in an embodiment of the present invention.

[0022] Figure 3 Shows a schematic diagram of comparing the outer contour point cloud of a casting with the outer contour point cloud of a mold cavity model in an embodiment of the present invention.

[0023] Figure 4 Shows Figure 3 a schematic diagram of the position of the cross-sectional view taken.

[0024] Figure 5 shows Figure 4 a schematic diagram of one of the cross-sectional views.

[0025] Figure 6 shows Figure 3 a schematic diagram of one of the longitudinal sectional views.

[0026] Figure 7 shows a schematic diagram of the support vector machine algorithm in an embodiment of the present invention.

[0027] Figure 8 shows a schematic diagram of the random forest algorithm in an embodiment of the present invention.

[0028] Figure 9 shows a schematic diagram of the deep neural network algorithm in an embodiment of the present invention.

[0029] Figure 10 shows a schematic flow diagram of the device for predicting the shrinkage rate during the casting of turbine blades in Embodiment 2 of the present invention.

[0030] Figure 11 shows a schematic flow diagram of the device for predicting the shrinkage rate during the casting of turbine blades in Embodiment 4 of the present invention. Specific embodiments

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0032] Refer to Figure 2 , a method for predicting the shrinkage rate during the casting of turbine blades, comprising the following steps: S1. Obtain shrinkage rate data; specifically, the shrinkage rate data includes the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange.

[0033] The method for obtaining the output result Y (the shrinkage rate of the mold-casting of the turbine blade, that is, the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange.) is as follows: S11. Through a computer, using the inner cavity contour model of the mold as the reference object and the outer contour model of the casting as the test object, perform least squares fitting to obtain a comparison model (refer to Figure 3 ); S12. Divide cross-sections along the horizontal and / or vertical directions on the comparison model to obtain cross-sectional views. Specifically, a plurality of cross-sections need to be divided equidistantly along the horizontal direction on the comparison model (refer to Figure 4 ) and a plurality of cross-sections along the vertical direction.

[0034] S13. Obtain the points of the casting outer contour model and the points of the mold inner cavity contour model at the same position parts on the cross-sectional view, and obtain the shrinkage rate through the points of the casting outer contour model and the points of the mold inner cavity contour model. Specifically, through the cross-section (refer to Figure 5 ) or the longitudinal section (refer to Figure 6 ), cut the casting outer contour model and the mold inner cavity contour model, and calculate the shrinkage rate at this point within this cross-section for the deviation analysis at different points.

[0035] Specifically, refer to Figure 6 , taking the height as an example, the calculation formula is as follows: Shrinkage rate = ΔZ / ; where ΔZ = - , where is the height of the casting outer contour model, is the height of the mold inner cavity contour model. When the deviation value is positive ( Figure 6 the red part in it is positive, and D in the figure is the comparison deviation value), the cast model shrinks. When the shrinkage rate is negative ( Figure 6 the blue part in it, and D in the figure is the comparison deviation value), the cast model expands.

[0036] S14. Repeat the steps of obtaining the points of the casting outer contour model and the points of the mold inner cavity contour model at the same position parts on the cross-sectional view, and obtaining the shrinkage rate through the points of the casting outer contour model and the points of the mold inner cavity contour model, and sequentially obtain the shrinkage rates at different positions on the cross-sectional view until the shrinkage rates at different positions in all directions on the cross-sectional view are obtained.

[0037] S15. Repeat the step of dividing cross-sections along the horizontal and / or vertical directions on the comparison model to obtain cross-sectional views, and sequentially obtain different cross-sections until all cross-sections are obtained; S16. Obtain the required shrinkage rate by screening the above shrinkage rates.

[0038] Furthermore, the casting outer contour model is a casting outer contour point cloud model obtained by blue light scanning the casting outer contour; the mold inner cavity contour model is a mold inner cavity point cloud model obtained by blue light scanning the mold inner cavity contour or the mold design 3D digital model.

[0039] S2. Obtain a sub-dataset, where the sub-dataset includes data on the geometric attributes of the turbine blade casting, data on the geometric attributes of the mold cavity corresponding to the turbine blade casting, process parameters, and shrinkage rate data. Specifically, the influencing factor X affecting the dimensional shrinkage rate of the turbine blade includes a total of 17 factors, namely data on the geometric attributes of the turbine blade casting and process parameters. When inputting new data on the geometric attributes of the turbine blade casting and process parameters into the shrinkage rate model, the shrinkage rate model outputs the result Y (the shrinkage rate of the mold-casting of the turbine blade), and Y includes a total of 5 result values for multiple cross-sections in different directions.

[0040] S3. Obtain a dataset, where the dataset includes N sub-datasets, and the data on the geometric attributes of the turbine blade casting, the geometric attributes of the wax mold cavity, and the process parameters in the N sub-datasets are all different. Specifically, 20 groups of data are selected and imported into the shrinkage rate model. For the 20 groups of data, please refer to Table 1 and Table 2 below, where Table 2 is a continuation of Table 1. Table 1:

[0041] Table 2:

[0042] Furthermore, although not mentioned in the present invention, any dataset that can be used to construct the relationship between the geometric attributes or process parameters of the turbine blade and the blade shrinkage rate is within the scope of the present invention's rights protection.

[0043] S4. Construct a shrinkage rate model and import the above-mentioned dataset into the shrinkage rate model. When building the shrinkage rate model (i.e., a machine learning model), this project uses a multi-model fusion algorithm that includes a support vector machine algorithm (refer to Figure 7 ), and / or a random forest algorithm (refer to Figure 8 ), and / or a deep neural network algorithm (refer to Figure 9 ) to build the model. Specifically, the present invention can adopt different machine learning algorithms to build the machine learning model, but any model that can be used to train the prediction of the turbine blade shrinkage rate is within the scope of the present invention's rights protection.

[0044] Specifically, for the support vector machine algorithm (refer to Figure 7 , Figure 7 is a schematic diagram of the standard support vector machine algorithm. This algorithm is used to draw a separation band as wide as possible in a multi-dimensional space ( is the separation band), to separate the points with the determination result being true (black points) and the points with the determination result being false (black circles). The red and red circles are the points on the separation band; here, x is a two-dimensional matrix formed by the input of 17 factors in the present invention, and w and b are the parameter values continuously optimized during the iterative optimization process.

[0045] Random forest algorithm (reference Figure 8 , Figure 8 is a schematic diagram of the standard random forest algorithm. Each random tree selects several factors (x) for decision-making to give results, and then the results of several random trees are voted to form the final result output (for example, the random tree on the left sequentially selects the random trees on the right, gives the decision result, and votes this result with the results of other random trees to form the final result y); where x are the 17 factors input in the present invention, and y is the shrinkage rate finally required in the present invention (the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange).

[0046] Deep neural network algorithm (reference Figure 9 , Figure 9 is a schematic diagram of the standard deep neural network algorithm. According to the input factors (the left pink circles), it is passed layer by layer until the final result (the rightmost circle) is obtained (where layer-by-layer non-linear learning is adopted to extract features, for example, the factors in the first pink circle pass the features to five neural points in the next layer, and each neural point then passes the features to five neural points in the next layer, and so on for passing and learning), where the left circles represent the 17 factors input in the present invention, and the right circles represent the shrinkage rate finally required in the present invention (the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange).

[0047] In the above embodiments, optionally, another implementation manner is that the present invention can use existing intelligent models, such as deepseek, OpenAI, and Tongyi Qianwen QWQ-32B, etc., and obtain the shrinkage rate model by inputting data for training in the above models.

[0048] S5. The shrinkage rate model is trained and iteratively optimized until the test shrinkage rate of the turbine blade mold-casting calculated from the input data of the geometric properties of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate. Specifically, it includes the following steps: S51. The shrinkage rate model uses a multi-model fusion algorithm of the support vector machine algorithm (reference Figure 7 ) and / or the random forest algorithm (reference Figure 8 ) and / or the deep neural network algorithm (reference Figure 9 ) to build a model for training and iterative optimization; specifically, the optimization process can be understood as a black box, and the support vector machine algorithm and / or the random forest algorithm (reference Figure 8 ) and / or the deep neural network algorithm used are all existing algorithms, and their technologies are relatively mature.

[0049] S52. Continuously input the data of the geometric properties of the turbine blade casting (i.e., the data of the turbine blade to be obtained) and the process parameters into the shrinkage rate model multiple times. After calculation, the shrinkage rate model outputs the test shrinkage rate of the turbine blade die-casting.

[0050] S453. By comparing the error between the corresponding test shrinkage rate and the actual shrinkage rate, if the errors are all within the set error range (the set error is less than or equal to 1%), then save the above model.

[0051] S54. If any of the errors is greater than the set error, repeat the steps of S51.

[0052] Specifically, referring to Table 3 and Table 4, where Table 3 and Table 4 are test data, i.e., the data input in step S52, and Table 5 is the data calculated by the shrinkage rate model and the data obtained by blue light scanning (refer to step S1). The difference between the two data is within the allowable error range. Therefore, the shrinkage rate of the die-casting of the turbine blade can be predicted by the method of the present invention.

[0053] Table 3

[0054] Table 4:

[0055] Table 5:

[0056] S6. Input the geometric property data and process parameters of the turbine blade casting to be developed into the shrinkage rate model, and the shrinkage rate model can output the corresponding die-casting shrinkage rate value.

[0057] Furthermore, the data of the geometric properties of the turbine blade casting and the geometric properties of the die cavity include the overall length, blade length, flange length, flange width, flange height, maximum chord length of the blade, minimum chord length of the blade, and maximum thickness of the blade cross-section.

[0058] Furthermore, the process parameters include the wax injection temperature, injection pressure, injection flow rate, holding time, shell mold system, number of shell mold layers, shell mold preheating temperature, pouring temperature, and alloy type (the alloy type is the alloy grade in Table 2).

[0059] Refer to Figure 10 , a device for predicting the shrinkage rate during the casting of a turbine blade, comprising: A first collection unit, which acquires shrinkage rate data; acquires a sub-dataset, and the sub-dataset includes the data of the geometric properties of the turbine blade casting, the data of the geometric properties of the die cavity corresponding to the turbine blade casting, process parameters, and shrinkage rate data; A second collection unit that obtains a data set, where the data set includes N sub-data sets, and the geometric attributes of the turbine blade castings, the geometric attributes of the mold inner cavity, and the process parameters in the N sub-data sets are all different; A construction unit that constructs a shrinkage rate model and imports the above data set into the shrinkage rate model; A processing unit that performs training iteration optimization on the shrinkage rate model until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data of the geometric attributes of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate; A prediction unit that inputs the data of the geometric attributes of the turbine blade casting to be developed and the process parameters into the shrinkage rate model, and the shrinkage rate model outputs the shrinkage rate of the turbine blade mold-casting.

[0060] Example 3 Reference Figure 11 A device for predicting the shrinkage rate during turbine blade casting, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method described in Example 1. Specifically, the device for predicting the shrinkage rate during turbine blade casting can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The device for predicting the shrinkage rate during turbine blade casting may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the schematic diagram is only an example of the device for predicting the shrinkage rate during turbine blade casting, and does not constitute a limitation on the device for predicting the shrinkage rate during turbine blade casting. It may include more or fewer components than shown, or combine certain components or different components. For example, the device for predicting the shrinkage rate during turbine blade casting may also include a power supply component, an input / output interface, a network access device, a bus, etc.

[0061] Example 4 A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the method described in Example 1.

[0062] Specifically, the computer program can be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the method for predicting the shrinkage rate during turbine blade casting.

[0063] Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements 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.

Claims

1. A method for predicting the shrinkage rate during the casting of a turbine blade, characterized in that, Including the following steps: Obtain shrinkage rate data; Obtain a sub-dataset, where the sub-dataset includes data on the geometric properties of the turbine blade casting, data on the geometric properties of the mold cavity corresponding to the turbine blade casting, process parameters, and shrinkage rate data; Obtain a dataset, where the dataset includes N sub-datasets, and the data on the geometric properties of the turbine blade casting, the geometric properties of the mold cavity, and the process parameters in the N sub-datasets are different; Construct a shrinkage rate model and import the above dataset into the shrinkage rate model; The shrinkage rate model is trained and iteratively optimized until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data on the geometric properties of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate; Input the data on the geometric properties of the turbine blade casting to be developed and the process parameters into the shrinkage rate model, and the shrinkage rate model outputs the shrinkage rate of the turbine blade mold-casting.

2. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 1, characterized in that, The shrinkage rate data includes the Z-direction shrinkage rate of the intake edge, the Z-direction shrinkage rate of the middle part, the Z-direction shrinkage rate of the exhaust edge, the X-direction shrinkage rate of the flange, and the Y-direction shrinkage rate of the flange.

3. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 1 or 2, characterized in that, The step of obtaining the shrinkage rate data includes the following steps: Take the mold cavity contour model as the reference object and the casting outer contour model as the test object, perform least squares fitting to obtain a comparison model; Divide cross-sections along the horizontal and / or vertical directions on the comparison model to obtain cross-sectional views; Obtain the points of the casting outer contour model and the points of the mold cavity contour model at the same position on the cross-sectional view, and obtain the shrinkage rate through the points of the casting outer contour model and the points of the mold cavity contour model; Repeat the step of obtaining the points of the casting outer contour model and the points of the mold cavity contour model at the same position on the cross-sectional view, and obtain the shrinkage rate through the points of the casting outer contour model and the points of the mold cavity contour model, and sequentially obtain the shrinkage rates at different positions on the cross-sectional view until the shrinkage rates at all positions on the cross-sectional view are obtained; Repeat the step of dividing cross-sections along the horizontal and / or vertical directions on the comparison model to obtain cross-sectional views, and sequentially obtain different cross-sections until all cross-sections are obtained; Obtain the required shrinkage rate by screening the above shrinkage rates.

4. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 3, characterized in that, The casting outer contour model is a casting outer contour point cloud model obtained by blue light scanning the casting outer contour; The mold cavity contour model is a mold cavity point cloud model obtained by blue light scanning the mold cavity contour or the three-dimensional digital model of the mold design.

5. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 1, characterized in that, The data on the geometric properties of the turbine blade casting and the geometric properties of the mold cavity both include the overall length, the blade length, the flange length, the flange width, the flange height, the maximum chord length of the blade, the minimum chord length of the blade, and the maximum thickness of the blade cross-section.

6. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 1, characterized in that, The step that the shrinkage rate model is trained and iteratively optimized until the test shrinkage rate of the turbine blade mold-casting calculated from the input test data on the geometric properties of the turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate includes the following steps: The shrinkage rate model uses the support vector machine algorithm and / or the random forest algorithm and / or the deep neural network algorithm for training and iterative optimization; Continuously input the data of the geometric properties of the turbine blade casting and the process parameters into the shrinkage rate model for multiple times, and the shrinkage rate model calculates and outputs the test shrinkage rate of the turbine blade mold-casting; By comparing the error between the corresponding test shrinkage rate and the actual shrinkage rate; If the errors are all within the set error, save the above model; If any of the errors is greater than the set error, repeat the steps of training and iterative optimization of the shrinkage rate model using the support vector machine algorithm and / or the random forest algorithm and / or the deep neural network algorithm.

7. The method for predicting the shrinkage rate during the casting of a turbine blade according to claim 1, characterized in that, The process parameters include the wax injection temperature, injection pressure, injection flow rate, holding pressure time, shell system, number of shell layers, shell preheating temperature, pouring temperature, and alloy type.

8. A device for predicting the shrinkage rate during the casting of a turbine blade, characterized in that, Including: The first collection unit acquires the shrinkage rate data; Obtain a sub-dataset, where the sub-dataset includes the data of the geometric properties of the turbine blade casting, the data of the geometric properties of the mold cavity corresponding to the turbine blade casting, the process parameters, and the shrinkage rate data; The second collection unit acquires a dataset, where the dataset includes N sub-datasets, and the data of the geometric properties of the turbine blade casting, the data of the geometric properties of the mold cavity, and the process parameters in the N sub-datasets are all different; The construction unit constructs a shrinkage rate model and imports the above dataset into the shrinkage rate model; The processing unit performs training and iterative optimization on the shrinkage rate model until the test shrinkage rate of the turbine blade mold-casting calculated from the input data of the geometric properties of the test turbine blade casting and the process parameters is less than the set error compared with the actual shrinkage rate; The prediction unit inputs the data of the geometric properties of the turbine blade casting to be developed and the process parameters into the shrinkage rate model, and the shrinkage rate model outputs the shrinkage rate of the turbine blade mold-casting.

9. A device for predicting the shrinkage rate during the casting of a turbine blade, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1-7.

Images