CN118350237A

CN118350237A - Rolling die optimization method, blade rolling device, blade rolling equipment and medium

Info

Publication number: CN118350237A
Application number: CN202410364611.8A
Authority: CN
Inventors: 王晨霖; 张鸿富; 段易呈; 毛羽; 陈祉亦; 张世杰; 孔祥伟
Original assignee: 东北大学
Filing date: 2024-03-28
Publication date: 2024-07-16

Abstract

The invention relates to the technical field of blade rolling, and discloses a rolling die optimization method, a blade rolling method, a device, equipment and a medium; the method comprises the following steps: constructing an initial rolling die corresponding to the target blade, and extracting initial die profile data corresponding to the initial rolling die; simulating the rolling process according to the initial mold surface data; extracting a simulated blade profile matrix of the rolled blade; obtaining an error function value according to a theoretical blade profile matrix, a simulated blade profile matrix and a preset error function of the target blade; if the error function value is greater than or equal to the preset error value, updating the initial mold surface data by using a preset optimization algorithm; returning to simulate the rolling process according to the initial mold profile data; and extracting a simulated blade profile matrix of the rolled blade until the error function value is smaller than a preset error value. The method combines the optimization theory and the rolling simulation, and can greatly improve the performance of the rolling die and the quality of the rolled blade.

Description

Rolling die optimization method, blade rolling device, blade rolling equipment and medium

Technical Field

The present disclosure relates to the field of blade rolling technology, and in particular, to a rolling die optimization method, a blade rolling method, a device, an electronic apparatus, and a computer readable storage medium.

Background

As a key manufacturing process, the precise cold rolling of the blade has the advantages of high-precision forming, high material utilization rate, excellent surface quality, high production efficiency, low cost, material strength maintenance and the like, and has important roles in the fields of aeroengines, automobiles, energy and the like. Blade rolling dies are important tools for rolling blades, and their shape and parameters affect the accuracy of forming the blade to various degrees. However, in the actual rolling process, due to the influence of factors such as additional stress, forward sliding and backward sliding, elasticity of materials, temperature change and the like in the blade, a certain error exists between the rolled blade profile and the theoretically calculated target blade profile, so that the rolling die is required to be optimized and adjusted to reduce the deviation between the actual blade profile and the theoretical profile, and the accuracy of rolling the blade is improved.

In the actual rolling process, the rolling die usually requires a craftsman to manually repair the die for a plurality of times, and the method has high technical and experience requirements on the craftsman, long time consumption and more required materials; in addition, errors may occur in the trimming of the mold, and there may be differences in the trimming manner of the different molds, which may result in a decrease in production efficiency. In other methods, the deviation between the corresponding positions of the actual blade and the theoretical blade can be used for carrying out linear superposition on the contour line of the blade back of the basin of the die so as to realize the optimization of the rolling die. The method is simple to operate, and takes errors caused by the environment into consideration, but the method still has the problems of insufficient precision and possibly excessive iteration times because the contour of the die and the contour of the blade are not in a simple linear relation in precise rolling.

Disclosure of Invention

In view of the foregoing, embodiments of the present disclosure provide a rolling die optimization method, a blade rolling method, an apparatus, an electronic device, and a computer-readable storage medium, which aim to solve or at least partially solve the foregoing problems.

In a first aspect, embodiments of the present disclosure provide a roll die optimization method, the method comprising:

Constructing an initial rolling die corresponding to a target blade by using finite element simulation software, and extracting initial die profile data corresponding to the initial rolling die, wherein the initial die profile data comprises an initial leaf basin profile matrix and an initial leaf back profile matrix;

simulating a rolling process according to the initial mold profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade;

Obtaining an error function value according to the theoretical blade profile matrix corresponding to the target blade, the simulated blade profile matrix and a preset error function; the preset error function is generated according to a preset function and the theoretical blade profile matrix, and the preset function is used for representing a nonlinear mapping relation between the die profile data and the blade profile matrix;

And under the condition that the error function value is greater than or equal to a preset error value, updating the initial mold surface data by using a preset optimization algorithm, taking the obtained updated initial mold surface data as new initial mold surface data, and returning to the step: and simulating the rolling process according to the initial die profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade and the subsequent steps until the error function value is smaller than the preset error value.

In a second aspect, embodiments of the present disclosure also provide a blade rolling method, the method comprising:

obtaining a blade to be rolled and a blade rolling die, wherein the blade rolling die is obtained according to the method of the first aspect;

and rolling the blade to be rolled by using the blade rolling die to obtain the rolled blade.

In a third aspect, embodiments of the present disclosure also provide a roll die optimizing apparatus, the apparatus comprising:

The die construction module is used for constructing an initial rolling die corresponding to the target blade by utilizing finite element simulation software, extracting initial die profile data corresponding to the initial rolling die, wherein the initial die profile data comprises an initial leaf basin profile matrix and an initial leaf back profile matrix;

The simulation module is used for simulating the rolling process according to the initial die profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade;

The error generation module is used for obtaining an error function value according to the theoretical blade profile matrix corresponding to the target blade, the simulated blade profile matrix and a preset error function; the preset error function is generated according to a preset function and the theoretical blade profile matrix, and the preset function is used for representing a nonlinear mapping relation between the die profile data and the blade profile matrix;

The optimizing module is used for updating the initial mold surface data by utilizing a preset optimizing algorithm under the condition that the error function value is larger than or equal to a preset error value, taking the obtained updated initial mold surface data as new initial mold surface data, and returning to the step: and simulating the rolling process according to the initial die profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade and the subsequent steps until the error function value is smaller than the preset error value.

In a fourth aspect, embodiments of the present disclosure also provide a blade rolling apparatus, the apparatus comprising:

an acquisition module for acquiring a blade to be rolled and a blade rolling die, wherein the blade rolling die is obtained according to the method of the first aspect;

and the rolling module is used for rolling the blade to be rolled by utilizing the blade rolling die to obtain the rolled blade.

In a fifth aspect, embodiments of the present disclosure further provide an electronic device, including: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to perform the steps of the roll die optimization method/blade rolling method described above.

In a sixth aspect, the disclosed embodiments also provide a computer-readable storage medium storing one or more programs that, when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the steps of the rolling die optimization method/blade rolling method described above.

By means of the technical scheme, compared with the prior art, the rolling die optimizing method, the blade rolling method, the device, the electronic equipment and the computer readable storage medium do not treat the outline of the die and the outline of the blade as a simple linear relation, but utilize a preset function for representing the nonlinear mapping relation between die profile data and blade profile matrix, obtain a preset error function based on the preset function, find an optimal solution of the preset error function by using an optimizing method, enable the simulated blade profile matrix to converge along the trend of error reduction, ensure the effectiveness of optimization, reduce the iteration times and avoid negative optimization caused by excessive adjustment; meanwhile, the optimization method can be flexibly applied to different types of objective functions, the nonlinear optimization of the mold surface is well adapted, and the precision of the rolling mold can be greatly improved by the optimized mathematical model and the calculation method, so that the molding quality of the rolled blade is improved. In addition, the embodiment of the disclosure utilizes the computer simulation technology to model and optimize the blade rolling die, and the numerical calculation and analysis method avoids the defects of long experimental period, higher requirements on workers, high resource consumption and the like, reduces the production cost and improves the production efficiency.

The foregoing description is merely an overview of the technical solutions of the present disclosure, and may be implemented according to the content of the specification in order to make the technical means of the present disclosure more clearly understood, and in order to make the above and other objects, features and advantages of the present disclosure more clearly understood, the following specific embodiments of the present disclosure are specifically described.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate and explain the present disclosure, and together with the description serve to explain the present disclosure. In the drawings:

FIG. 1 shows a flow diagram of a roll die optimization method provided by an embodiment of the present disclosure;

FIG. 2 illustrates a flow diagram of a blade rolling method provided by another embodiment of the present disclosure;

FIG. 3 shows a schematic view of the structure of a roll die optimizing apparatus provided by an embodiment of the present disclosure;

FIG. 4 shows a schematic view of a blade rolling apparatus provided by an embodiment of the present disclosure;

Fig. 5 shows a schematic structural diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

For the purposes of promoting an understanding of the principles and advantages of the disclosure, reference will now be made to the drawings and specific examples thereof, together with the following description. It will be apparent that the described embodiments are only some, but not all, of the embodiments of the present disclosure. Based on the embodiments in this disclosure, all other embodiments that a person of ordinary skill in the art would obtain without making any inventive effort are within the scope of protection of this disclosure.

It should be noted that: like reference numerals and letters denote like items in the following figures, and thus once an item is defined in one figure, no further definition or explanation thereof is necessary in the following figures.

It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the foregoing figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that such uses may be interchanged where appropriate such that embodiments of the disclosure described herein may be implemented in sequences other than those illustrated or described herein. Furthermore, the terms "include" and variations thereof are to be interpreted as open-ended terms that mean "include, but are not limited to.

According to research, aiming at the problem of optimization of a rolling die, the prior art realizes the optimization of the rolling die by utilizing the deviation between the corresponding positions of an actual blade and a theoretical blade to linearly superimpose the contour line of the blade basin and the blade back of the die. The method is simple and convenient to operate, and takes errors caused by the environment into consideration, but the method still has the problems of insufficient precision of the optimized die and possibly excessive iteration times because the contour of the die and the contour of the blade are not in a simple linear relation in precision rolling. Based on this, the present invention proposes a rolling die optimizing method, a blade rolling method, a device, an electronic apparatus, and a computer-readable storage medium, and the present disclosure is described in detail below by way of specific embodiments.

For the sake of understanding the present embodiment, first, a detailed description will be given of a rolling die optimizing method disclosed in the embodiment of the present disclosure, and an execution subject of the rolling die optimizing method provided in the embodiment of the present disclosure is generally a computer device having a certain computing capability, where the computer device includes, for example: the terminal device or server or other processing device may be a user device (UserEquipment, UE), a mobile device, a user terminal, a terminal, or the like. In some possible implementations, the roll die optimization method may be implemented by way of a processor invoking computer readable instructions stored in a memory.

Fig. 1 shows a schematic flow chart of a rolling die optimization method provided by an embodiment of the present disclosure, and as can be seen from fig. 1, the embodiment of the present disclosure at least includes steps S101 to S104:

step S101: and constructing an initial rolling die corresponding to the target blade by using finite element simulation software, and extracting initial die profile data corresponding to the initial rolling die, wherein the initial die profile data comprises an initial leaf basin profile matrix and an initial leaf back profile matrix.

The embodiments of the present disclosure are implemented based on finite element simulation software, which may be, for example, ANSYS, siemens NX Nastran, etc., and are not limited thereto.

Embodiments of the present disclosure contemplate optimizing a die for rolling a blade to obtain a high precision target blade. The initial rolling die can be built by setting parameters in a simulation software system according to the shape parameters, material properties and other information of the target blade. And then extracting an initial leaf basin profile matrix and an initial leaf back profile matrix corresponding to the initial rolling die. For example, the angular values corresponding to n sections may be extracted from the initial rolling die, and the contour lines corresponding to n sections of the leaf basin die may be extracted based on these angular valuesContour lines of n sections corresponding to blade back moldWherein n is a natural number,For the ith contour of the leaf basin mold,Is the ith contour line of the phyllotaxis mold. Then, for the leaf basin mold, the coordinates of a plurality of measuring points can be extracted on each contour line corresponding to the leaf basin mold, and the coordinates of the measuring points are extracted on the ith contour lineExtracting corresponding P measuring points, measuring the coordinates of each measuring point, and marking as Extracting y coordinates of P measuring points on the section to form an ith leaf basin contour lineY-coordinate matrix of (2)The y coordinate matrixes corresponding to the n leaf basin contour lines form a matrix for describing the profile of the leaf basin mould, namely an initial leaf basin profile matrix; for the back mold, the coordinates of a plurality of measuring points can be extracted on each contour line corresponding to the back mold, and the ith contour line is obtainedExtracting P measuring points, measuring the coordinates of each measuring point, and marking asExtracting y coordinates of P measuring points on the section to form an ith blade back contour lineY-coordinate matrix of (2)

And the y coordinate matrixes corresponding to the n blade back contour lines form a matrix for describing the profile of the blade back mould, namely an initial blade back profile matrix.

Exemplary, initial leaf basin profile matrixInitial leaf back profile matrixIt should be noted that, the method for extracting the initial mold profile data according to the embodiments of the present disclosure is not limited.

Step S102: and simulating the rolling process according to the initial mold profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade.

After the initial mold profile data is obtained, the rolling process can be simulated by using simulation software according to the initial mold profile data, and the rolled blade is obtained. In order to judge whether the precision of the rolled blade meets the standard or not, whether the initial rolling die is qualified or not is also judged, and in the step, the simulated blade profile matrix corresponding to the rolled blade can be extracted. Specifically, the profile line of the n sections corresponding to the rolled blade may be extracted and denoted as { L ₁,L₂,...,L_i,...,L_n }, where L _i represents the profile line of the i section of the rolled blade, P corresponding measurement points are extracted on the profile line L _i of the i section of the rolled blade, coordinates of each measurement point are measured and denoted as {(x_i,1,y_i,1′),(x_i,2,y_i,2'),...,(x_i,p,y_i,p'),...,(x_i,P,y_i,P')},p＝1,2,...,P,, y _i,p 'is the y coordinate value of the P corresponding measurement point on the profile line of the rolled blade, the y coordinates of the P corresponding measurement points on the P corresponding section are extracted to form the y coordinate matrix y _i′＝[y_i,1′ y_i,2′ ... y_i,p′ ... y_i,P' ] of the i rolled blade profile line L _i, and the y coordinate matrix corresponding to the n rolled blade profile lines forms a matrix describing the profile of the rolled blade, i.e., the simulated blade profile matrix.

Exemplary, simulated blade profile matrix

Step S103: obtaining an error function value according to the theoretical blade profile matrix corresponding to the target blade, the simulated blade profile matrix and a preset error function; the preset error function is generated according to a preset function and the theoretical blade profile matrix, and the preset function is used for representing a nonlinear mapping relation between the die profile data and the blade profile matrix.

Here, the theoretical blade profile matrix corresponding to the target blade may be extracted first. Specifically, n theoretical contour lines corresponding to n sections of the target blade may be extracted and denoted as { l ₁,l₂,...,l_i,...,l_n }, where l _i represents a theoretical contour line corresponding to the i-th section. P measuring points are extracted on a theoretical contour line l _i of the ith section of the target blade, and coordinates of the P measuring points are respectively recorded as {(x_i,1,y_i,1),(x_i,2,y_i,2),...,(x_i,p,y_i,p),...,(x_i,P,y_i,P)},p＝1,2,...,P,, wherein x _i,p and y _i,p are x coordinate values and y coordinate values of the P measuring points on the theoretical contour line. The y coordinates of P measuring points of the cross section are extracted to form a y coordinate matrix y _i＝[y_i,1y_i,2 ... y_i,p ... y_i,P of an ith theoretical contour line l _i, and a y coordinate matrix y _i of n theoretical contour lines to form a matrix describing the profile of the target blade, namely a theoretical blade profile matrix.

Exemplary, theoretical blade profile matrix

It should be noted that the number of the substrates,The section angle values corresponding to the four profile lines of L _i、l_i are consistent with the abscissa of each measuring point. The angle values and the abscissa of the measuring points can be set on the rolled blade according to actual measurement requirements, and the embodiment of the disclosure is not limited. Preferably, each angle value is an angle in consideration of forward slip compensation.

After the simulated blade profile matrix is obtained, the simulated blade profile matrix and the theoretical blade profile matrix can be used as independent variables of a preset error function to obtain an error function value through calculation in order to judge whether the simulated blade profile matrix is qualified or not. Here, the preset function is used to characterize the mapping between the mold profile data and the blade profile matrix. In the implementation, for example, the mapping relationship can be obtained by using the obtained multiple sets of mold profile data and the blade profile matrix of the simulation blade based on interpolation, fitting and other methods.

The preset error function is used for measuring the difference between the simulated blade profile matrix and the theoretical blade profile matrix and is generated according to the preset function and the theoretical blade profile matrix. In specific implementation, for example, an average absolute function, a root mean square error function, an absolute error function, and the like may be used as the preset error function, which is not limited in the embodiments of the present disclosure. Illustratively, the preset error function is a norm of a difference between the preset function and the theoretical blade profile matrix, i.e., the preset error function is the following formula (1):

w= |l' -l|| _F formula (1);

Step S104: and under the condition that the error function value is greater than or equal to a preset error value, updating the initial mold surface data by using a preset optimization algorithm, taking the obtained updated initial mold surface data as new initial mold surface data, and returning to the step: and simulating the rolling process according to the initial die profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade and the subsequent steps until the error function value is smaller than the preset error value.

The preset error value is a parameter for measuring whether the blade is qualified or not, and can be set according to actual needs when the blade is implemented.

After the error function value is obtained, the magnitude relation between the error function value and a preset error value can be compared, if the error function value is smaller than the preset error value, the difference between the simulated blade profile matrix and the theoretical blade profile matrix is smaller, the precision of the blade after rolling can meet the requirement through the initial die profile data, and the initial die is qualified; if the error function value is greater than or equal to the preset error value, the difference between the simulated blade profile matrix and the theoretical blade profile matrix is larger, the initial die profile data cannot enable the precision of the rolled blade to meet the requirement, and the initial die is unqualified, so that the initial die profile data can be updated by using a preset optimization algorithm, the obtained updated initial die profile data is used as new initial die profile data, and the new initial die profile data is returned to the step S102 and the subsequent steps until the error function value is smaller than the preset error value.

As can be seen from the method shown in fig. 1, compared with the prior art, the embodiment of the disclosure does not treat the profile of the mold and the profile of the blade as a simple linear relationship, but uses a preset function for representing the nonlinear mapping relationship between the mold profile data and the blade profile matrix, obtains a preset error function based on the preset function, and searches for an optimal solution of the preset error function by using an optimization method, so that the simulated blade profile matrix converges along the trend of error reduction, thereby ensuring the effectiveness of optimization, reducing the iteration times, and avoiding negative optimization caused by excessive adjustment; meanwhile, the optimization method can be flexibly applied to different types of objective functions, the nonlinear optimization of the mold surface is well adapted, and the precision of the rolling mold can be greatly improved by the optimized mathematical model and the calculation method, so that the molding quality of the rolled blade is improved. In addition, the embodiment of the disclosure utilizes the computer simulation technology to model and optimize the blade rolling die, and the numerical calculation and analysis method avoids the defects of long experimental period, higher requirements on workers, high resource consumption and the like, reduces the production cost and improves the production efficiency.

Further, in order to better illustrate the process of the above-mentioned roll-die optimization method, as a refinement and extension to the above-mentioned embodiments, several alternative embodiments are provided in the present embodiment, but not limited thereto, and specifically shown as follows:

Considering the critical role of the preset function for achieving the optimization of the rolling die, in one possible embodiment of the present disclosure, the preset function is obtained according to the following method:

Step A1: acquiring a rolling dataset; the rolling data set comprises a plurality of samples, and the corresponding characteristics and labels of each sample are sample die profile data and sample blade profile matrix respectively.

Step A2: the rolling data set is normalized and divided into a training set, a validation set and a test set.

Step A3: and training the initial neural network model based on the training set and the verification set to obtain a trained neural network model.

Step A4: evaluating the model performance of the trained neural network model based on the test set, and determining the model as a target neural network model when the model performance reaches a preset model performance; and taking the target neural network model as the preset function.

In this embodiment, the rolling dataset may be acquired first. Specifically, different parameters of the leaf basin and the leaf back of the rolling die can be set in finite element simulation software to simulate the rolling process for a plurality of times, so as to obtain corresponding rolling blades; and extracting die profile data (a cone profile matrix and a back profile matrix) and a blade profile matrix of a corresponding rolled blade under any group of parameters of the rolling die cone and the back profile, and taking the die profile data and the blade profile matrix under the parameters as characteristics and labels of one sample respectively to finally obtain the rolling data set.

After the rolling dataset is obtained, the rolling dataset may be normalized and partitioned into a training set, a validation set, and a test set. The training set is used for training the model, the verification set is used for adjusting the super parameters and the structure of the model, and the test set is used for evaluating the performance of the model. For example, a division of 70% training set, 15% validation set and 15% test set may be used. It should be noted that the examples herein are illustrative only and are not limiting on the disclosed embodiments.

The initial neural network model in the embodiment of the disclosure mainly comprises an input layer, a hidden layer and an output layer. The input layer is responsible for receiving the mold surface data; the hidden layer carries out nonlinear transformation and feature extraction on the mold surface data of the mold, so as to learn the complex mode and structure of the data; the output layer then generates the final blade profile matrix prediction. As to the specific structure of the initial neural network model, the embodiments of the present disclosure are not limited as long as prediction of the blade profile matrix from the mold profile data can be achieved.

A loss function and optimizer may be defined to train the initial neural network model using a training set. And evaluating the performance of the model on the verification set, and adjusting the super parameters and the structure of the model according to the performance of the verification set so as to obtain the optimal performance of the model and obtain the trained neural network model.

Then, evaluating the model performance of the trained neural network model on the test set, and determining the model as a target neural network model when the model performance reaches the preset model performance; and when the model performance does not reach the preset model performance, training the model and adjusting the model parameters continuously until the model performance reaches the preset performance. Here, the preset model performance, for example, the prediction accuracy of the blade profile matrix, which is not limited to the embodiment of the disclosure, may be set according to actual needs. Finally, taking the target neural network model as a preset function, and taking the preset function L '=f (L ^e,L^q) as an example, where L' is a simulated blade profile matrix.

In the embodiment, the roll forming process is simulated for a plurality of times in finite element software to obtain a plurality of groups of roll die profile data and simulated blade profile data, based on the data, a nonlinear mapping relation between the die profile data and the blade profile matrix is fitted by using a neural network, and finally a target neural network model serving as a preset function is obtained.

In a possible implementation manner, the preset optimizing algorithm includes: newton's method and/or gradient descent method.

In this embodiment, the preset optimization algorithm includes, but is not limited to, newton's method, gradient descent method, for example, random gradient descent method, quasi-newton method, conjugate gradient method, and the like. For example, newton's method may be used, where an initial point is randomly selected as a starting point; calculating a gradient (first derivative) and a hessian matrix (second derivative) of the preset error function at the current point; calculating a moving direction by using information of the gradient and the hessian matrix; and updating the position of the current point by using the calculated moving direction and step length to obtain updated initial mold surface data. It should be noted that the examples herein are illustrative only and are not limiting on the disclosed embodiments.

In the actual optimization process of the rolling die, a large data volume is involved, and excessive storage resources are occupied. Based on this, in one possible embodiment of the present disclosure, the preset optimization algorithm is the gradient descent method; the updating the initial mold surface data by using a preset optimization algorithm, taking the obtained updated initial mold surface data as new initial mold surface data, comprising the following steps:

Step S1041: and respectively calculating a first partial derivative of the preset error function relative to the leaf basin profile matrix and a second partial derivative of the preset error function relative to the leaf back profile matrix.

Step S1042: and obtaining the updated leaf basin profile matrix according to the initial leaf basin profile matrix, the preset step length and the first partial derivative.

Step S1043: obtaining the updated phyllotame profile matrix according to the initial phyllotame profile matrix, the preset step length and the second partial derivative; and the updated leaf basin profile matrix and the updated leaf back profile matrix form the updated initial die profile data.

In the present embodiment, the predetermined function L '=f (L ^e,L^q), i.e., L' is a function related to L ^e and L ^q, so W is a function related to L ^e and L ^q, i.e., w=w (L ^e,L^q). In practice, a first partial derivative of the preset error function with respect to the basin profile matrix L ^e may be calculated, and a second partial derivative of the preset error function with respect to the back profile matrix L ^q may be calculated.

In particular, the method comprises the steps of,

And then, obtaining an updated leaf basin profile matrix according to the initial leaf basin profile matrix, the preset step length and the first partial derivative. The preset step length may be determined according to practical situations, which is not limited in the disclosure. For example, the updated leaf basin profile matrix may be determined according to the following equation (2):

Wherein L ^e is an initial leaf basin profile matrix, and eta is a preset step length.

And obtaining an updated phyllotame profile matrix according to the initial phyllotame profile matrix, the preset step length and the second partial derivative. In practice, the updated back profile matrix may be determined according to the following equation (3):

Wherein L ^q is an initial blade back profile matrix.

In particular, the method comprises the steps of,

After the updated leaf basin profile matrix and the updated leaf back profile matrix are obtained, the updated leaf basin profile matrix can be used as a new initial leaf basin profile matrix, and the updated leaf back profile matrix can be used as a new initial leaf back profile matrix.

In the embodiment, the gradient descent method in the optimization algorithm is selected to optimize the rolling die, and only the gradient is calculated by the gradient descent method without storing all data, so that the problems of large data volume and large occupied storage resources in the optimization process of the rolling die can be effectively solved. Meanwhile, the gradient descent method updates parameters along the negative gradient direction in the iteration process, so that the optimization result of the die can be converged rapidly, the die is optimized as soon as possible, the production efficiency is improved, and the material loss is reduced.

In optimizing the rolling die, an excessively large or small step size has a negative effect on the optimization speed and efficiency, and based on this, in one possible embodiment of the disclosure, the preset step size is obtained based on an adaptive learning rate algorithm.

In this embodiment, the adaptive learning rate algorithm includes, but is not limited to, the following: adaGrad Algorithm (ADAPTIVE GRADIENT Algorithm ), adaDelta Algorithm (ADAPTIVE DELTA, adaptive delta Algorithm), RMSProp Algorithm (Root Mean Square Propagation, root mean square propagation Algorithm), adam Algorithm (Adaptive Moment Estimation ), and the like.

In particular embodiments, for example, the adaptive learning rate algorithm may calculate and adjust the learning rate based on gradient information of a preset error function, historical update conditions, or other relevant indicators. For example, if the gradient of the preset error function is large, the algorithm increases its learning rate and increases the preset step length so as to update the argument of the preset error function more quickly; when the gradient is smaller, the algorithm can reduce the learning rate, reduce the preset step length and avoid excessive updating.

The implementation dynamically adjusts the preset step length by utilizing the self-adaptive learning rate algorithm, so that the optimization process can be greatly accelerated, and the optimization efficiency and the die precision can be improved while the convergence of the preset error function is ensured.

The embodiment of the disclosure also provides a rolling die optimization method, which comprises the following steps of S1-S10:

step S1: acquiring a rolling dataset; the rolling data set comprises a plurality of samples, and the corresponding characteristics and labels of each sample are sample die profile data and sample blade profile matrix respectively.

Step S2: the rolling data set is normalized and divided into a training set, a validation set and a test set.

Step S3: and training the initial neural network model based on the training set and the verification set to obtain a trained neural network model.

Step S4: evaluating the model performance of the trained neural network model based on the test set, and determining the model as a target neural network model when the model performance reaches a preset model performance; and taking the target neural network model as a preset function.

Step S5: and constructing an initial rolling die corresponding to the target blade by utilizing finite element simulation software, and extracting initial die profile data corresponding to the initial rolling die, wherein the initial die profile data comprises an initial leaf basin profile matrix and an initial leaf back profile matrix.

Step S6: and simulating the rolling process according to the initial mold profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade.

Step S7: obtaining an error function value according to a theoretical blade profile matrix, a simulated blade profile matrix and a preset error function corresponding to the target blade; the preset error function is generated according to the preset function and the theoretical blade profile matrix.

Step S8: and under the condition that the error function value is larger than or equal to the preset error value, respectively calculating a first partial derivative of the preset error function relative to the leaf basin profile matrix and a second partial derivative of the preset error function relative to the leaf back profile matrix.

Step S9: and obtaining an updated leaf basin profile matrix according to the initial leaf basin profile matrix, the preset step length and the first partial derivative.

Step S10: obtaining an updated phyllotame profile matrix according to the initial phyllotame profile matrix, the preset step length and the second partial derivative; the updated leaf basin profile matrix and the updated leaf back profile matrix form updated initial mold profile data, and the updated initial mold profile data is used as new initial mold profile data. The preset step length is obtained based on an adaptive learning rate algorithm.

Step S10: and returning to the step S6 until the error function value is smaller than the preset error value.

Referring to fig. 2, the embodiment of the present disclosure further provides a blade rolling method, which includes steps S201 to S202:

Step S201: obtaining a blade to be rolled and a blade rolling die, wherein the blade rolling die is obtained by the rolling die optimization method according to any embodiment;

step S202: and rolling the blade to be rolled by using the blade rolling die to obtain the rolled blade.

In this embodiment, the blade rolling die may be obtained according to the rolling die optimization method of any one of the above embodiments, for example, the blade rolling die may be obtained according to the above steps S101 to S104; receiving blade parameters input by a user, and constructing a blade to be rolled; and simulating the rolling process by using the profile matrix data of the back of the blade basin of the blade rolling die, and rolling the blade to be rolled to obtain the rolled blade. In this embodiment, the blade to be rolled is rolled by using the blade rolling die obtained by the rolling die optimizing method according to any one of the embodiments, and a rolled blade very close to the desired theoretical blade can be obtained.

It will be appreciated by those skilled in the art that in the above-described method of the specific embodiment, the written order of steps does not imply a strict order of execution, but rather any limitations on the implementation, and that the specific order of execution of the steps should be determined by their functions and possibly inherent logic.

In practical application, all the possible embodiments may be combined in any combination manner to form possible embodiments of the disclosure, which are not described in detail herein.

Based on the same concept, the embodiments of the present disclosure also provide a rolling die optimizing apparatus, and fig. 3 illustrates a schematic structural diagram of the rolling die optimizing apparatus provided by the embodiments of the present disclosure, and referring to fig. 3, the rolling die optimizing apparatus 300 provided by the embodiments of the present disclosure includes:

the die construction module 301 is configured to construct an initial rolling die corresponding to a target blade by using finite element simulation software, and extract initial die profile data corresponding to the initial rolling die, where the initial die profile data includes an initial leaf basin profile matrix and an initial leaf back profile matrix;

the simulation module 302 is configured to simulate a rolling process according to the initial mold profile data to obtain a rolled blade, and extract a simulated blade profile matrix corresponding to the rolled blade;

The error generating module 303 is configured to obtain an error function value according to the theoretical blade profile matrix corresponding to the target blade, the simulated blade profile matrix, and a preset error function; the preset error function is generated according to a preset function and the theoretical blade profile matrix, and the preset function is used for representing a nonlinear mapping relation between the die profile data and the blade profile matrix;

The optimizing module 304 is configured to update the initial mold profile data with a preset optimizing algorithm when the error function value is greater than or equal to a preset error value, take the obtained updated initial mold profile data as new initial mold profile data, and return to the step: and simulating the rolling process according to the initial die profile data to obtain a rolled blade, and extracting a simulated blade profile matrix corresponding to the rolled blade and the subsequent steps until the error function value is smaller than the preset error value.

In a possible embodiment, the apparatus further comprises a training module for:

acquiring a rolling dataset; the rolling data set comprises a plurality of samples, and the corresponding characteristics and labels of each sample are sample die profile data and sample blade profile matrixes respectively;

Carrying out standardization processing on the rolling data set, and dividing the rolling data set into a training set, a verification set and a test set;

training the initial neural network model based on the training set and the verification set to obtain a trained neural network model;

Evaluating the model performance of the trained neural network model based on the test set, and determining the model as a target neural network model when the model performance reaches a preset model performance; and taking the target neural network model as the preset function.

In a possible implementation manner, in the device, the preset optimization algorithm includes newton method and/or gradient descent method.

In a possible implementation manner, in the device, the preset optimization algorithm is the gradient descent method; the optimizing module 304 is configured to:

respectively calculating a first partial derivative of the preset error function relative to the leaf basin profile matrix and a second partial derivative of the preset error function relative to the leaf back profile matrix;

Obtaining the updated leaf basin profile matrix according to the initial leaf basin profile matrix, a preset step length and the first partial derivative;

Obtaining the updated phyllotame profile matrix according to the initial phyllotame profile matrix, the preset step length and the second partial derivative; and the updated leaf basin profile matrix and the updated leaf back profile matrix form the updated initial die profile data.

In a possible implementation manner, in the above device, the preset step size is obtained based on an adaptive learning rate algorithm.

It should be noted that any of the above rolling die optimizing apparatuses may be in one-to-one correspondence with the foregoing rolling die optimizing method, and will not be described herein.

Based on the same concept, the embodiment of the present disclosure further provides a blade rolling apparatus, and fig. 4 illustrates a schematic structural diagram of the blade rolling apparatus provided by the embodiment of the present disclosure, referring to fig. 4, the blade rolling apparatus 400 provided by the embodiment of the present disclosure includes:

An obtaining module 401, configured to obtain a blade to be rolled and a blade rolling die, where the blade rolling die is obtained according to the rolling die optimization method of any one of the embodiments above;

And the rolling module 402 is used for rolling the blade to be rolled by using the blade rolling die to obtain a rolled blade.

It should be noted that any of the blade rolling apparatuses may be used to implement the blade rolling method according to the foregoing one-to-one correspondence, and will not be described herein.

Fig. 5 shows a schematic structural diagram of an electronic device according to an embodiment of the disclosure. As shown in fig. 5, at the hardware level, the electronic device comprises a processor, optionally together with an internal bus, a network interface, a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.

The processor, network interface, and memory may be interconnected by an internal bus, which may be an ISA (Industry Standard Architecture ) bus, a PCI (PERIPHERAL COMPONENT INTERCONNECT, peripheral component interconnect standard) bus, or EISA (Extended Industry Standard Architecture ) bus, among others. The buses may be divided into address buses, data buses, control buses, etc. For ease of illustration, only one bi-directional arrow is shown in FIG. 5, but not only one bus or type of bus.

And the memory is used for storing programs. In particular, the program may include program code including computer-operating instructions. The memory may include memory and non-volatile storage and provide instructions and data to the processor.

The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to form the roll-die optimizing apparatus or blade rolling apparatus on a logic level. And the processor is used for executing the program stored in the memory and particularly used for executing the method.

The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but may also be a digital signal Processor (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components. The various methods, steps and logic blocks of the disclosure in the embodiments of the disclosure may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present disclosure may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software modules in a decoded processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method.

The electronic device may perform the rolling die optimizing method/blade rolling method provided in the embodiments of the present disclosure, and implement the functions of the rolling die optimizing device in the embodiment shown in fig. 3, or the functions of the blade rolling device in the embodiment shown in fig. 4, which are not described herein.

The disclosed embodiments also provide a computer-readable storage medium storing one or more programs, the one or more programs including instructions, which when executed by an electronic device comprising a plurality of application programs, enable the electronic device to perform the rolling die optimization method/blade rolling method provided by the various embodiments of the disclosure.

It will be apparent to those skilled in the art that embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

It will be appreciated by those skilled in the art that embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The foregoing is merely exemplary of the present disclosure and is not intended to limit the present disclosure. Various modifications and variations of this disclosure will be apparent to those skilled in the art. Any modifications, equivalent substitutions, improvements, or the like, which are within the spirit and principles of the present disclosure, are intended to be included within the scope of the claims of the present disclosure.

Claims

1. A method of optimizing a roll die, the method comprising:

2. The method according to claim 1, wherein the predetermined function is obtained according to the following method:

3. The method according to claim 1, wherein the preset optimization algorithm comprises newton's method and/or gradient descent method.

4. A method according to claim 3, wherein the preset optimization algorithm is the gradient descent method; the updating the initial mold surface data by using a preset optimization algorithm, taking the obtained updated initial mold surface data as new initial mold surface data, comprising the following steps:

5. The method of claim 4, wherein the preset step size is based on an adaptive learning rate algorithm.

6. A method of rolling a blade, the method comprising:

obtaining a blade to be rolled and a blade rolling die, wherein the blade rolling die is obtained according to the method of any one of claims 1-5;

7. A roll die optimizing apparatus, the apparatus comprising:

The error calculation module is used for obtaining an error function value according to the theoretical blade profile matrix corresponding to the target blade, the simulated blade profile matrix and a preset error function; the preset error function is generated according to a preset function and the theoretical blade profile matrix, and the preset function is used for representing a nonlinear mapping relation between the die profile data and the blade profile matrix;

8. A blade rolling apparatus, the apparatus comprising:

An acquisition module for acquiring a blade to be rolled and a blade rolling die, wherein the blade rolling die is obtained according to the method of any one of claims 1-5;

9. An electronic device, comprising:

A processor; and

A memory arranged to store computer executable instructions which, when executed, cause the processor to perform the method of any one of claims 1-5/the steps of the method of claim 6.

10. A computer readable storage medium storing one or more programs, which when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the method of any of claims 1-5/the steps of the method of claim 6.