Surface integrity constrained aircraft engine blade grinding and polishing process optimization method and system

Abstract

The invention belongs to but not limited to the technical field of aviation, discloses a surface integrity constrained aircraft engine blade grinding and polishing process optimization method and system, and establishes a material removal model of a blade based on an orthogonal center combination experiment data set. And model input and output of machine learning are set, data are preprocessed, and the completeness of the grinding and polishing surface is predicted through a machine learning XGBoost algorithm. And obtaining multiple groups of solutions of the optimization variables according to an MSPSO particle swarm multi-objective optimization algorithm. And according to the multiple groups of solutions of the optimization variables, screening out a priority solution of each optimization target and a global optimal solution conforming to all constraints and optimization targets. According to the method, through analysis of a grinding and polishing experiment data set, a material removal model of the blade is established, and a material removal coefficient is calibrated. The model can quantitatively describe the influence of each process parameter on the material removal depth, and provides theoretical support for quantitative and accurate grinding and polishing processing of the aircraft engine blade.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of aviation technology, and in particular relates to a method and system for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints. Background Art

[0002] As the core component of advanced aero-engines, the machining accuracy and surface quality of blades directly determine the performance and service life of the whole machine. However, due to the characteristics of difficult material processing, complex curved surfaces, and high contour accuracy requirements of aero-engine blades, it is often difficult to ensure the surface quality and material removal accuracy after grinding and polishing. At present, the contour accuracy and surface quality of blades after grinding and polishing mainly depend on the technology of manual grinding and polishing, which makes it difficult to accurately control the grinding and polishing allowance and surface integrity of blades based on experience, resulting in large quality differences, poor surface consistency, and blade roughness and residual stress that cannot meet the requirements. The surface integrity and material removal of blade grinding and polishing are affected by many parameters such as contact force, feed speed, tool speed, abrasive grain size, tool curvature and workpiece curvature. The planning of machining parameters has a great influence on the surface integrity, material removal rate, quality consistency and contour removal accuracy of the workpiece.

[0003] In view of the above analysis, the technical problems that urgently need to be solved in the existing technology are: the process planning lacks consideration of the integrity of the grinding and polishing surface and the accuracy of material removal, making it difficult to ensure the quality and stability of the processed surface. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a method and system for optimizing the grinding and polishing process of aero-engine blades with surface integrity constraints. On the basis of ensuring the contour accuracy of aero-engine blades, the residual stress and surface roughness in the surface integrity of blade grinding and polishing are optimized to ensure the processing quality and stability.

[0005] The present invention is implemented as follows: a method for optimizing the grinding and polishing process of an aeroengine blade with surface integrity constraints, comprising:

[0006] S1: Based on the orthogonal central composite test CCD, the experimental data set between grinding and polishing process parameters and surface integrity and material removal depth was designed and obtained.

[0007] S2: Establish a material removal model for the blade based on the grinding and polishing experimental data set.

[0008] S3: Set the input and output of the machine learning model: Set the normal contact force F n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, and abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs.

[0009] S4: Preprocess the data: randomly divide the grinding and polishing experiment sample data into a training set of 90% samples and a test set of 10% samples. At the same time, use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure.

[0010] S5: Prediction of grinding and polishing surface integrity using machine learning XGBoost algorithm.

[0011] S6: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained. The positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched.

[0012] S7: Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing. is the surface roughness, is the surface residual stress.

[0013]

[0014] S8: Set the process parameters and material removal depth constraints, including the normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h.

[0015]

[0016] S9: According to the multiple groups of solutions for the optimization variables, the priority solutions of each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and a global optimal solution that meets all constraints and optimization objectives is selected.

[0017] Furthermore, the grinding and polishing process parameters in S1 include normal contact force, feed speed, tool rotation speed, abrasive grain size, tool curvature and workpiece curvature, and the experimental measurement results include surface roughness, residual stress and material removal depth.

[0018] Furthermore, the material removal model in S2 is:

[0019]

[0020] Where h is the removal depth, k h The material removal coefficient is calibrated by experiment, V s is the speed of the sanding machine, V f is the robot tangential feed speed, F n is the normal contact force, R1 is the radius of curvature of the contact wheel of the tool belt machine, and R2 is the radius of curvature of the workpiece contact point.

[0021] Furthermore, S5 specifically includes:

[0022] Objective function:

[0023]

[0024] in For training loss functions, common ones are square loss and logistic regression loss. is the sum of the complexity of all k regression trees, which is used as the regularization term in the objective function. The tree model trained in the tth iteration is f t ,but:

[0025]

[0026] In formula (13), only the t-th tree is an unknown variable. The structure of the first t-1 trees is determined, and the complexity and are known, which are replaced by constants. The loss function is expanded by the second order Taylor to obtain:

[0027]

[0028] Substituting equation (15) into equation (14) and removing the constant term, we get the simplified objective function:

[0029]

[0030] Furthermore, S6 specifically includes:

[0031] Usually, the velocity update is performed before the position update, and the update formulas for the two are as follows:

[0032]

[0033] Among them, N * is the population size; N D is the dimension of the search space, i.e. the number of variables; k is the current iteration number; I max is the maximum number of iterations; w is the inertia weight; is the velocity of particle p of dimension q in the kth iteration; c1 is the cognitive acceleration coefficient; c2 is the social acceleration coefficient; r1 and r2 are numbers generated randomly and uniformly in [0,1]; is the position value of particle p in dimension q in the kth iteration; is the value of the individual best position of particle p in dimension q until the kth iteration; is the value of the position of the global best particle in q dimension until the kth iteration.

[0034] Further, the MOPSO optimization algorithm process is as follows:

[0035] (a): Set the initial parameters of the MOPSO multi-objective particle swarm optimization algorithm, including the number of particles, particle search range, learning factor, inertia weight, and roulette selection operator;

[0036] (b): Evaluate particle fitness, obtain the initial non-dominated solution, update the individual optimum, and determine the global optimum;

[0037] (c): Update particle position and velocity;

[0038] (d): Determine whether to execute the mutation based on the demand;

[0039] (e): Calculate the function value of each optimization objective and update pbest;

[0040] (f): Update external files and select the global optimum;

[0041] (g): Determine the termination condition.

[0042] Another object of the present invention is to provide a surface integrity constrained aero-engine blade grinding and polishing process optimization system for realizing the surface integrity constrained aero-engine blade grinding and polishing process optimization method, comprising:

[0043] Dataset acquisition module: Based on the orthogonal central combination test CCD, the experimental data set between grinding and polishing process parameters and surface integrity and material removal depth is designed and acquired.

[0044] Material removal model building module: Establish the material removal model of the blade based on the grinding and polishing experiment data set.

[0045] Model input and output setting module: Set the model input and output of machine learning: set the normal contact force F n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, and abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs.

[0046] Data preprocessing module: Preprocess the data: randomly divide the grinding and polishing experiment sample data into a training set of 90% samples and a test set of 10% samples. At the same time, use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure.

[0047] Integrity prediction module: Predicts the integrity of the polished surface through the machine learning XGBoost algorithm.

[0048] Optimal solution search module: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained, the positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched.

[0049] Optimization target setting module: Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing.

[0050]

[0051] Constraint setting module: Set constraints for process parameters and material removal depth, including normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h.

[0052]

[0053] Global optimal solution selection module: based on the multiple sets of solutions of the optimization variables, the priority solutions of each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and the global optimal solution that meets all constraints and optimization objectives is selected.

[0054] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the surface integrity constrained aero-engine blade grinding and polishing process optimization method.

[0055] Another object of the present invention is to provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the surface integrity constrained aero-engine blade grinding and polishing process optimization method.

[0056] Another object of the present invention is to provide an information data processing terminal, which includes the surface integrity constrained aero-engine blade grinding and polishing process optimization system.

[0057] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0058] First, the advantages and positive effects of the technical solution to be protected by the present invention are:

[0059] (1) Based on the orthogonal central combination test CCD design, an experimental data set between grinding and polishing process parameters and surface integrity and material removal depth was obtained. A material removal model for the blade was established based on the grinding and polishing experimental data set, and the material removal coefficient was calibrated to determine the influence of each parameter on the material removal depth, thus ensuring the quantitative removal of material and the accuracy of the blade profile during the grinding and polishing process.

[0060] (2) According to the machine learning XGBoost algorithm, the normal contact force F n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs. The data are preprocessed and divided, the influence of various grinding and polishing process parameters on surface integrity is explored and analyzed, and an accurate surface integrity prediction model is established.

[0061] (3) Taking into account the limited range of each processing parameter and the material removal constraint, the optimal surface integrity was taken as the optimization goal. The roughness and residual stress in the surface integrity were simultaneously optimized according to the MSPSO particle swarm multi-objective optimization algorithm, and the optimal grinding and polishing process parameters were obtained.

[0062] (4) The technical solution of the present invention is applicable to, but not limited to, the grinding and polishing of curved workpieces such as aircraft engine blades, and has strong adaptability and scalability. On the basis of ensuring the quantitative removal of workpiece material, the process parameters are optimized with the goal of optimal surface integrity, thus ensuring the contour accuracy, surface quality and consistency of aircraft engine blades.

[0063] Second, as auxiliary evidence of the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:

[0064] (1) The technical solution of the present invention fills the technical gap in the industry at home and abroad:

[0065] 1. Through the analysis of the grinding and polishing experimental data set, the blade material removal model was established and the material removal coefficient was calibrated. This model can quantitatively describe the influence of various process parameters on the material removal depth, providing theoretical support for the quantitative and precise grinding and polishing of aeroengine blades.

[0066] 2. The machine learning algorithm XGBoost is used to predict the surface integrity of the grinding and polishing, especially the surface roughness and residual stress. The accurate prediction of the surface integrity of the robot grinding and polishing is achieved.

[0067] 3. Through the MSPSO particle swarm multi-objective optimization algorithm, the optimization of grinding and polishing process parameters is achieved, especially in the simultaneous optimization of surface roughness and residual stress. The multi-objective optimization method of the present invention has important application value in the grinding and polishing of aeroengine blades, filling the gap in the existing technology in multi-objective optimization processing.

[0068] 4. Adaptability and scalability: The technical solution of the present invention is not only applicable to the grinding and polishing of aircraft engine blades, but can also be extended to the grinding and polishing needs of other curved surface workpieces, and has strong adaptability. This provides a new solution for the grinding and polishing of related industries and expands the application scope of the technology.

[0069] 5. Improvement of processing quality and stability: By optimizing the grinding and polishing process parameters, the contour accuracy, surface quality and consistency of the aero-engine blades are ensured, and the processing quality and stability are significantly improved. This achievement has important practical and economic value in the field of aero-engine manufacturing.

[0070] (2) The technical solution of the present invention solves a technical problem that people have been eager to solve but have never been able to solve successfully:

[0071] 1. Balance between surface integrity and material removal accuracy: In the grinding and polishing of aeroengine blades, how to optimize surface integrity (residual stress and surface roughness) while ensuring material removal accuracy has always been a difficult problem in the industry. Traditional grinding and polishing processes often find it difficult to meet these two requirements at the same time, resulting in unstable processing quality. The present invention successfully achieves a balance between the two by establishing a material removal model and optimizing grinding and polishing process parameters.

[0072] 2. The previous grinding and polishing process optimization mostly relied on experience and trial and error, which was inefficient and difficult to achieve ideal results. By introducing machine learning and multi-objective optimization algorithms, the present invention provides an efficient and scientific process optimization method, which enables various parameters in the grinding and polishing process to be accurately controlled, thereby significantly improving processing efficiency and quality.

[0073] 3. Difficulty in surface integrity control: Surface integrity has an important impact on the performance and life of aeroengine blades, but how to effectively control surface integrity during the grinding and polishing process has always been a technical challenge. The present invention successfully achieves effective control of surface integrity by optimizing the grinding and polishing process parameters, thus providing a guarantee for the long-term stability and reliability of aeroengine blades.

[0074] 4. Adaptability and scalability: The technical solution of the present invention not only solves the problem in the grinding and polishing of aircraft engine blades, but also has strong adaptability and scalability, and can be widely used in the grinding and polishing of other curved surface workpieces. This feature provides a new solution for the relevant grinding and polishing field and solves the limitations of the previous technology in terms of scope of application.

[0075] Third, the present invention obtains experimental data sets of grinding and polishing process parameters (such as normal contact force, feed speed, tool speed, tool curvature, workpiece curvature and abrasive grain size) and surface integrity (such as surface roughness, residual stress) and material removal depth through orthogonal central composite experimental design (CCD). A material removal model is established based on the experimental data to describe the relationship between material removal depth and grinding and polishing parameters. The model comprehensively considers influencing factors such as tool curvature, workpiece curvature, normal contact force, etc., providing a data basis for optimizing the process.

[0076] The surface integrity prediction model was established through the machine learning XGBoost algorithm, with normal contact force, tool speed, feed speed, etc. as input, and surface roughness and residual stress as output. The data was normalized after preprocessing and randomly divided into training and test sets. The surface integrity was optimized by second-order Taylor expansion using the XGBoost algorithm to generate a high-precision prediction model. The model can quickly predict the surface quality under different grinding and polishing parameters, laying the foundation for process optimization.

[0077] The MSPSO (multi-objective particle swarm optimization) algorithm is used to optimize the process parameters, setting the surface roughness minimization and residual compressive stress maximization as the optimization goals, and setting the constraints of process parameters (such as normal contact force, tool speed, etc.) and material removal depth. The algorithm updates the speed and position of the particle swarm to find the Pareto frontier, and selects the global optimal solution through non-dominated sorting and external file updates to ensure that the optimization solution set meets the multi-objective constraints and achieves the optimal surface quality.

[0078] The optimized solutions selected through the Pareto frontier are concentrated, weights are set according to different processing objectives, and the priority solutions are flexibly selected. The optimization results show that the surface roughness is reduced and the absolute value of the residual compressive stress is increased, which significantly improves the surface quality of the blade. The technical solution of the present invention effectively solves the technical problem that it is difficult to balance surface integrity and processing efficiency in the prior art, and uses the particle swarm optimization algorithm to achieve precise optimization of complex process parameters, significantly improving processing efficiency and surface quality, and providing a new technical path for aero-engine blade processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is a schematic diagram of the overall process of the method for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints provided by an embodiment of the present invention;

[0080] Figure 2 It is a schematic diagram of an orthogonal experimental group based on an orthogonal central composite design provided by an embodiment of the present invention;

[0081] Figure 3 It is a flow chart of the MOPSO multi-objective optimization algorithm provided by an embodiment of the present invention;

[0082] Figure 4 It is a Pareto frontier graph obtained by the MOPSO multi-objective optimization algorithm provided by an embodiment of the present invention;

[0083] Figure 5 1. It is a diagram showing the effect of cascade blades before and after grinding and polishing provided by an embodiment of the present invention;

[0084] Figure 6 is a schematic diagram of finite element simulation of abrasive belt grinding and polishing provided by an embodiment of the present invention;

[0085] Figure 7 It is a structural diagram of an aero-engine blade grinding and polishing process optimization system with surface integrity constraints provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0086] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0087] Example 1: Optimization of grinding and polishing process for high temperature blades of aircraft engines

[0088] The high temperature blades of aircraft engines need to withstand extremely high temperatures and high-speed airflow during operation, and their surface integrity is crucial to the performance and life of the engine. The present invention is used to optimize the grinding and polishing process of blades to ensure that their surface quality meets high performance requirements.

[0089] 1) Data acquisition: The material removal depth, surface roughness and residual stress data under different normal contact forces, tool rotation speeds, abrasive grain sizes and workpiece curvatures are experimentally measured to construct a material removal model.

[0090] 2) Machine learning modeling: The XGBoost algorithm is used to establish a surface integrity prediction model to predict the surface roughness and residual stress under different grinding and polishing parameter combinations.

[0091] 3) Optimization calculation: The MSPSO algorithm is used to optimize the grinding and polishing process parameters, setting the goals of minimizing surface roughness and maximizing residual compressive stress.

[0092] 4) Application results: The optimal solution was screened out through the Pareto frontier surface. The optimized process parameters reduced the surface roughness of the blade by 20%, increased the residual compressive stress by 30%, and significantly improved the fatigue resistance and antioxidant capacity of the blade.

[0093] Example 2: Improving the surface processing quality of gas turbine blades

[0094] Gas turbine blades have high requirements for surface integrity, and their surface roughness and residual stress directly affect turbine efficiency and service life. The present invention is applied to the optimization of the grinding and polishing process of gas turbine blades to improve turbine performance.

[0095] 1) Experimental design: Through orthogonal central composite design (CCD), the surface integrity experimental data of gas turbine blades under different grinding and polishing parameters were collected.

[0096] 2) Material removal model construction: Combined with experimental data, a material removal model is established to analyze the influence of grinding and polishing parameters on material removal depth and surface integrity.

[0097] 3) Optimization and implementation: Use the MSPSO algorithm to optimize process parameters, set constraints on material removal depth and surface roughness, and screen out the global optimal solution.

[0098] 4) Optimization effect: After optimization, the surface roughness of the blade is reduced by 15%, and the absolute value of the residual compressive stress is increased by 25%. The optimized process significantly improves the surface quality of the blade, making its performance more stable and reliable under high temperature and high pressure conditions.

[0099] The two embodiments demonstrate the application of the present invention in the grinding and polishing process of aircraft engine and gas turbine blades, which achieves dual improvements in surface integrity and material removal by optimizing process parameters, providing an efficient and precise technical solution for the manufacture of high-performance blades.

[0100] The present invention is implemented as follows: a method for optimizing the grinding and polishing process of an aeroengine blade with surface integrity and surface integrity constraints is as follows: Figure 1 As shown, it is characterized in that a material removal model for blades is established based on an orthogonal central combination experimental data set. The model input and output of machine learning are set and the data is preprocessed, and the surface integrity of the polishing surface is predicted by the machine learning XGBoost algorithm. According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained. According to the multiple sets of solutions for the optimization variables, the priority solutions for each optimization objective and the global optimal solution that meets all constraints and optimization objectives are screened out.

[0101] S1.1: Based on the orthogonal central composite test CCD design and obtain the experimental data set between grinding and polishing process parameters and surface integrity and material removal depth. Figure 2 As shown in the figure, the grinding and polishing process parameters include normal contact force, feed speed, tool rotation speed, abrasive grain size, tool curvature and workpiece curvature, and the experimental measurement results include surface roughness, residual stress and material removal depth.

[0102] S1.2: Establish a material removal model for the blade based on the grinding and polishing experimental data set.

[0103]

[0104] Where h is the removal depth, k h The material removal coefficient is calibrated by experiment, V s is the belt sander speed, V f is the robot tangential feed speed, F n is the normal contact force, R1 is the radius of curvature of the contact wheel of the tool belt machine, and R2 is the radius of curvature of the workpiece contact point.

[0105] S2.1: Set the model input and output of machine learning. n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, and abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs.

[0106] S2.2: Preprocess the data. Randomly divide the grinding and polishing test sample data into a training set of 90% samples and a test set of 10% samples. At the same time, use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure.

[0107] S2.3: Predict grinding and polishing surface integrity using machine learning XGBoost algorithm.

[0108] Objective function:

[0109]

[0110] in For training loss functions, common ones are square loss and logistic regression loss. is the sum of the complexity of all k regression trees, which is used as the regularization term in the objective function. The tree model trained in the tth iteration is f t ,but:

[0111]

[0112]

[0113] In formula (13), only the t-th tree is an unknown variable. The structure of the first t-1 trees is determined, and the complexity and are known, which are replaced by constants. The loss function is expanded by the second order Taylor to obtain:

[0114]

[0115] Substituting equation (15) into equation (14) and removing the constant term, we get the simplified objective function:

[0116]

[0117] S3.1: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained. The positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched. Usually, the speed update is performed before the position update, and the update formulas of the two are as follows:

[0118]

[0119] Among them, N * is the population size; N D is the dimension of the search space, i.e. the number of variables; k is the current iteration number; I max is the maximum number of iterations; w is the inertia weight; is the velocity of particle p of dimension q in the kth iteration; c1 is the cognitive acceleration coefficient; c2 is the social acceleration coefficient; r1 and r2 are numbers generated randomly and uniformly in [0,1]; is the position value of particle p in dimension q in the kth iteration; is the value of the individual best position of particle p in dimension q until the kth iteration; is the value of the position of the global best particle in q dimension until the kth iteration.

[0120] like Figure 3 As shown, the MOPSO optimization algorithm process is as follows:

[0121] (a): Set the initial parameters of the MOPSO multi-objective particle swarm optimization algorithm, including the number of particles, particle search range, learning factor, inertia weight, and roulette selection operator;

[0122] (b): Evaluate particle fitness, obtain the initial non-dominated solution, update the individual optimum, and determine the global optimum;

[0123] (c): Update particle position and velocity;

[0124] (d): Determine whether to execute the mutation based on the demand;

[0125] (e): Calculate the function value of each optimization objective and update pbest;

[0126] (f): Update external files and select the global optimum;

[0127] (g): Determine the termination condition.

[0128] S3.2: Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing.

[0129]

[0130] S3.3: Set the process parameters and material removal depth constraints, including the normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h.

[0131]

[0132] S3.4: Based on the multiple sets of solutions for the optimization variables, the priority solutions for each optimization objective are screened out to obtain the following: Figure 4 The Pareto frontier surface diagram shown in the figure selects the global optimal solution that meets all constraints and optimization objectives. In this embodiment, the lower the roughness in the surface integrity and the larger the absolute value of the residual compressive stress, the better the surface quality of the blade after grinding and polishing. The embodiment of the present invention uses a multi-objective particle swarm optimization algorithm to calculate the optimal solution set for surface integrity. Based on the optimal solution set, different weights can be set for different optimization objectives to solve the different requirements for each optimization objective in the actual processing process. The optimization results obtained using the particle swarm optimization algorithm are as follows:

[0133]

[0134]

[0135] like Figure 5 The figure shows the effect of blade grinding and polishing before and after. From the perspective of the integrity of the grinding and polishing surface, the process parameters obtained by the multi-objective optimization method of the embodiment of the present invention are compared with the original empirical process parameters. On the basis of ensuring the same material removal depth, the optimized grinding and polishing process parameters reduce the roughness of the workpiece surface and increase the absolute value of the residual compressive stress, thereby improving the surface integrity of the blade after grinding and polishing, and is more conducive to improving the service performance and fatigue life of the blade.

[0136] like Figure 7 As shown in the figure, the surface integrity constrained aero-engine blade grinding and polishing process optimization system includes:

[0137] Dataset acquisition module: Based on the orthogonal central combination test CCD, the experimental data set between grinding and polishing process parameters and surface integrity and material removal depth is designed and acquired;

[0138] Material removal model building module: build the material removal model of the blade according to the grinding and polishing experiment data set;

[0139] Model input and output setting module: Set the model input and output of machine learning: set the normal contact force F n , Tool speed V s , robot tangential feed speed V f, tool curvature radius R1, workpiece curvature radius R2, abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs;

[0140] Data preprocessing module: preprocess the data: randomly divide the grinding and polishing experiment sample data into a training set of 90% samples and a test set of 10% samples, and use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure;

[0141] Integrity prediction module: predicts the integrity of the grinding and polishing surface through the machine learning XGBoost algorithm;

[0142] Optimal solution search module: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained, the positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched;

[0143] Optimization target setting module: Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing;

[0144]

[0145] Constraint setting module: Set constraints for process parameters and material removal depth, including normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h;

[0146]

[0147] Global optimal solution selection module: based on the multiple sets of solutions of the optimization variables, the priority solutions of each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and the global optimal solution that meets all constraints and optimization objectives is selected.

[0148] An application embodiment of the present invention provides a computer device, which includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the processor executes the steps of a method for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints.

[0149] An application embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of a method for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints.

[0150] An application embodiment of the present invention provides an information data processing terminal, which includes an aero-engine blade grinding and polishing process optimization system with surface integrity constraints.

[0151] Aircraft engine blades work under high temperature and high pressure and are required to have excellent fatigue resistance and surface roughness. In the manufacturing process of aircraft engines, the grinding and polishing process is an important link that affects the surface integrity of the blades. Reasonable process parameters can effectively improve the surface quality of the blades and extend their service life. To this end, this embodiment will optimize the parameters of the grinding and polishing process of aircraft engine blades.

[0152] Application Example 1: Optimization of Grinding and Polishing Process Parameters for Cascade Blades of Aircraft Engines

[0153] The cascade blades of an aero-engine bear important functions of airflow guidance and compression in the engine, and their surface quality directly affects the performance and efficiency of the engine. The surface integrity constraint aero-engine blade grinding and polishing process optimization method of the present invention can improve the surface quality and performance of the blade after grinding and polishing.

[0154] 1) Based on the orthogonal central combination test CCD, an experimental data set between grinding and polishing process parameters and surface integrity and material removal depth was obtained.

[0155] 2) Establish a material removal model for the blade based on the grinding and polishing experimental data set.

[0156] 3) Set the model input and output for machine learning.

[0157] 4) Predict the surface integrity of blade grinding and polishing through machine learning XGBoost algorithm.

[0158] 5) According to the MSPSO particle swarm multi-objective optimization algorithm, multiple groups of solutions for the optimization variables are obtained.

[0159] 6) Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing.

[0160] 7) Set constraints on process parameters and material removal depth.

[0161] 8) According to the multiple sets of solutions for the optimization variables, the priority solutions for each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and a global optimal solution that meets all constraints and optimization objectives is selected.

[0162] Application Example 2: Optimization of grinding and polishing process parameters for aircraft engine blisk blades

[0163] After the integral blade disk of an aero-engine is milled and formed, the surface quality and roughness of the blade still cannot meet the manufacturing precision requirements. In order to improve the surface integrity of the blade, the integral blade disk blade needs to be ground and polished. The surface integrity constraint aero-engine blade grinding and polishing process optimization method of the present invention can improve the surface quality and performance of the blade after grinding and polishing.

[0164] 1) Based on the orthogonal central combination test CCD, an experimental data set between grinding and polishing process parameters and surface integrity and material removal depth was obtained.

[0165] 2) Establish a material removal model for the blade based on the grinding and polishing experimental data set.

[0166] 3) Set the model input and output for machine learning.

[0167] 4) Predict the surface integrity of blade grinding and polishing through machine learning XGBoost algorithm.

[0168] 5) According to the MSPSO particle swarm multi-objective optimization algorithm, multiple groups of solutions for the optimization variables are obtained.

[0169] 6) Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing.

[0170] 7) Set constraints on process parameters and material removal depth.

[0171] 8) According to the multiple sets of solutions for the optimization variables, the priority solutions for each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and a global optimal solution that meets all constraints and optimization objectives is selected.

[0172] like Figure 5 As shown in the figure, by comparing the surface of the blade before and after grinding and polishing, the grinding and polishing optimization parameter group adopted in this example can significantly improve the surface finish of the blade. Figure 6 The figure shows the finite element simulation results of the surface integrity of belt grinding and polishing. By comparing the empirical parameter group with the optimized parameter group under constraints, it can be seen that, while ensuring consistent material removal, the roughness of the grinding and polishing surface integrity of the optimized parameter group selected in this example is lower and the absolute value of the residual compressive stress is larger, indicating that the surface quality of the blade after grinding and polishing is better, and it has better fatigue resistance and performance.

[0173] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0174] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A surface integrity constrained aeroengine blade grinding and polishing process optimization method, characterized in that: include: S1: Design and obtain experimental data sets based on orthogonal central composite test CCD between grinding and polishing process parameters and surface integrity and material removal depth; S2: Establishing the material removal model of the blade based on the grinding and polishing experimental data set; S3: Set the input and output of the machine learning model: Set the normal contact force F n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs; S4: Preprocess the data: randomly divide the grinding and polishing test sample data into a training set of 90% samples and a test set of 10% samples, and use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure; S5: Prediction of grinding and polishing surface integrity using machine learning XGBoost algorithm; S6: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained; the positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched; S7: Setting the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing; S8: Set the process parameters and material removal depth constraints, including the normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h; S9: According to the multiple groups of solutions for the optimization variables, the priority solutions of each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and a global optimal solution that meets all constraints and optimization objectives is selected.

2. The surface integrity constrained aeroengine blade grinding and polishing process optimization method according to claim 1, characterized in that: The grinding and polishing process parameters in S1 include normal contact force, feed speed, tool rotation speed, abrasive grain size, tool curvature and workpiece curvature, and the experimental measurement results include surface roughness, residual stress and material removal depth.

3. The surface integrity constrained aeroengine blade grinding and polishing process optimization method according to claim 1, characterized in that: The material removal model in S2 is: Where h is the removal depth, k h The material removal coefficient is calibrated by experiment, V s is the belt sander speed, V f is the robot tangential feed speed, F n is the normal contact force, R1 is the radius of curvature of the contact wheel of the tool belt machine, and R2 is the radius of curvature of the workpiece contact point.

4. The surface integrity constrained aeroengine blade grinding and polishing process optimization method according to claim 1, characterized in that: S5 specifically includes: Objective function: in For training loss functions, common ones are square loss and logistic regression loss. is the sum of the complexity of all k regression trees, which serves as the regularization term in the objective function; the tree model trained in the tth iteration is f t ,but: In formula (13), only the t-th tree is an unknown variable. The structures of the first t-1 trees are determined, and the complexity and are known, which are replaced by const constants. The loss function is expanded by the second-order Taylor to obtain: Substituting equation (15) into equation (14) and removing the constant term, we get the simplified objective function:

5. The surface integrity constrained aeroengine blade grinding and polishing process optimization method according to claim 1, characterized in that: S6 specifically includes: Usually, the velocity update is performed before the position update, and the update formulas for the two are as follows: Among them, N * is the population size; N D is the dimension of the search space, i.e. the number of variables; k is the current iteration number; I max is the maximum number of iterations; w is the inertia weight; is the velocity of particle p of dimension q in the kth iteration; c1 is the cognitive acceleration coefficient; c2 is the social acceleration coefficient; r1 and r2 are numbers generated randomly and uniformly in [0,1]; is the position value of particle p in dimension q in the kth iteration; is the value of the individual best position of particle p in dimension q until the kth iteration; is the value of the position of the global best particle in q dimension until the kth iteration.

6. The surface integrity constrained aeroengine blade grinding and polishing process optimization method according to claim 1, characterized in that: The MOPSO optimization algorithm process is as follows: (a): Set the initial parameters of the MOPSO multi-objective particle swarm optimization algorithm, including the number of particles, particle search range, learning factor, inertia weight, and roulette selection operator; (b): Evaluate particle fitness, obtain the initial non-dominated solution, update the individual optimum, and determine the global optimum; (c): Update particle position and velocity; (d): Determine whether to execute the mutation based on the demand; (e): Calculate the function value of each optimization objective and update pbest; (f): Update external files and select the global optimum; (g): Determine the termination condition.

7. A surface integrity constrained aero-engine blade grinding and polishing process optimization system that implements the surface integrity constrained aero-engine blade grinding and polishing process optimization method as claimed in any one of claims 1 to 6, characterized in that: include: Dataset acquisition module: Based on the orthogonal central combination test CCD, the experimental data set between grinding and polishing process parameters and surface integrity and material removal depth is designed and acquired; Material removal model building module: build the material removal model of the blade according to the grinding and polishing experiment data set; Model input and output setting module: Set the model input and output of machine learning: set the normal contact force F n , Tool speed V s , robot tangential feed speed V f , tool curvature radius R1, workpiece curvature radius R2, abrasive grain size P are used as model feature inputs, and surface roughness and residual stress are used as model outputs; Data preprocessing module: preprocess the data: randomly divide the grinding and polishing experiment sample data into a training set of 90% samples and a test set of 10% samples, and use Max-Min standardization to perform linear transformation on the original data, normalize it to [0,1], and ensure the original data structure; Integrity prediction module: predicts the integrity of the grinding and polishing surface through the machine learning XGBoost algorithm; Optimal solution search module: According to the MSPSO particle swarm multi-objective optimization algorithm, multiple sets of solutions for the optimization variables are obtained, the positions of the swarm particles are continuously updated, and then the optimal solution is continuously searched; Optimization target setting module: Set the optimization target as the surface roughness and residual stress of the workpiece after grinding and polishing; Constraint setting module: Set constraints for process parameters and material removal depth, including normal contact force F n , Tool speed V s , robot tangential feed speed V f The limited range of each process parameter and the material removal depth constraint h; Global optimal solution selection module: based on the multiple sets of solutions of the optimization variables, the priority solutions of each optimization objective are screened out, a Pareto frontier surface diagram is obtained, and the global optimal solution that meets all constraints and optimization objectives is selected.

8. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to execute the steps of the method for optimizing the grinding and polishing process of an aero-engine blade with surface integrity constraints as described in any one of claims 1 to 6.

10. An information data processing terminal, comprising the surface integrity constrained aero-engine blade grinding and polishing process optimization system according to claim 7.

Images