Fast prediction method of variable parameter laser ablation effect based on finite element model and neural network model

Abstract

The application discloses a variable parameter laser ablation effect fast prediction method based on a finite element model and a neural network model, and solves the problems of limited application scenarios, limited application material ranges and insufficient real-time performance in the prior art. The application obtains simulation data (i.e. a second data set) through a laser ablation effect finite element model, trains and tests a neural network model through the simulation data, obtains an optimal neural network model, and then obtains corresponding data of a material to be measured according to actual working conditions, sets a calculation time step and a laser irradiation time, combines the optimal neural network model, and realizes real-time prediction of the laser ablation effect, so that the calculation precision is high, the time consumption is short, and the method is suitable for fast prediction of the laser ablation effect in a strong dynamic scene.

Description

Technical Field

[0001] This invention relates to a method for rapid prediction of laser ablation effects on materials, specifically a method for rapid prediction of variable parameter laser ablation effects based on finite element model and neural network model. Background Technology

[0002] Rapid prediction of laser ablation effects on materials is a crucial aspect of laser applications. Its core value lies in the rapid and accurate assessment of the ablation effects produced after laser irradiation, providing strong technical support and data references for numerous important engineering applications such as laser countermeasures and defense, and optimized design of laser manufacturing processes. This improves the implementation efficiency and reliability of related applications. In practical engineering scenarios, the core technical requirements for rapid prediction of laser ablation effects are reflected in two aspects: first, possessing efficient and rapid prediction capabilities to meet real-time application needs; and second, possessing good adaptability to adapt to complex scenarios with dynamically changing laser parameters.

[0003] Currently, the main methods for predicting the laser ablation effect of materials are traditional approaches such as finite element simulation and experimental pattern analysis. The core logic of these methods is as follows: by artificially changing influencing factors such as laser parameters, material parameters, and environmental parameters, a large number of repetitive experiments or simulation calculations are carried out to obtain massive amounts of laser ablation effect data. Based on this data, the patterns are summarized and inductively constructed to build engineering application models, thereby achieving the prediction of the laser ablation effect.

[0004] However, the aforementioned traditional methods have significant technical limitations. They are only applicable to relatively simple laser ablation prediction scenarios, such as static or quasi-static scenarios where the irradiated laser parameters remain stable and the environmental parameters change slowly. For dynamic application scenarios where laser parameters change dynamically and environmental conditions are complex and varied, these traditional methods not only have severely insufficient prediction capabilities and are difficult to adapt to the dynamic fluctuations of parameters, but their prediction accuracy cannot be effectively evaluated, and they cannot meet the requirements of actual engineering applications for prediction accuracy and dynamic adaptability.

[0005] Chinese patent CN113673127A discloses a method for rapidly obtaining the morphology of laser ablation of composite materials. This method can calculate the morphology of laser ablation of composite materials within minutes, which improves the prediction efficiency of laser ablation effects to a certain extent. However, the technical solution disclosed in this patent still has obvious limitations in its scope of application and performance shortcomings. It is only applicable to specific types of materials with low thermal conductivity, such as composite materials, and cannot be adapted to various materials with different thermal properties. At the same time, the method lacks real-time performance, making it difficult to meet the rapid prediction needs in scenarios with dynamic changes in laser parameters, and it cannot fully cover the diverse application needs in fields such as laser countermeasures and defense, and laser manufacturing and processing. Summary of the Invention

[0006] The purpose of this invention is to address the technical problems in existing technologies, such as limited applicable scenarios, limited applicable material range, and insufficient real-time performance, and to provide a rapid prediction method for variable parameter laser ablation effects based on finite element model and neural network model.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A rapid prediction method for variable-parameter laser ablation effects based on finite element model and neural network model is characterized by the following steps:

[0009] Step 1: Construct a finite element model of the laser ablation effect;

[0010] Step 2: Obtain the material parameters of the target material and the corresponding laser and environmental parameters; set up a parameter table for the laser and environmental parameters; input the parameter table into the finite element model to obtain the first dataset; organize the material parameters of the target material and the first dataset to form the second dataset; the target material includes the material to be tested;

[0011] Step 3: Build an initial neural network model, train and test the initial neural network model using the second dataset, until the optimal neural network model is obtained;

[0012] Step 4: Obtain the geometry of the material to be tested and mesh it, filter out the meshes of interest, and obtain the coordinates of each mesh of interest; initialize the temperature and survival status of each mesh of interest, and set the laser irradiation time;

[0013] Step 5: Set the calculation time step and assign the current time; calculate the distribution parameters of the laser beam used to ablate the material under test on the surface of the material under test;

[0014] Step 6: Combine the distribution parameters, the coordinates of the grid of interest, the time step, the material parameters of the material to be tested, and the current environmental parameters to obtain the third dataset;

[0015] Step 7: Input the third dataset into the optimal neural network model. The optimal neural network model will then output the predicted results of the material under test at the current time, which will be the temperature and survival status of the material under test.

[0016] Step 8: Repeat steps 5 to 7 to continuously predict the temperature and survival status of the material under test until the laser irradiation time is reached, and obtain all prediction results of the material under test throughout the entire laser irradiation time, thus completing the rapid prediction of the variable parameter laser ablation effect.

[0017] Further, in step 2, the acquisition of the material parameters of the target material and the corresponding laser parameters and environmental parameters, and the setting of a parameter table for the laser parameters and environmental parameters, specifically involves:

[0018] Obtain the material parameters of the target material, acquire the corresponding laser parameters and environmental parameters based on the actual working conditions, design the laser parameters and environmental parameters, and obtain the input parameter table P={I p , I a , d, x s , y s , t laser , T env The material parameters include heat capacity c (T), thermal conductivity k (T), density ρ (T), and absorptivity a (T); the laser parameters include peak power density I. p Average power density I a , Spot diameter d, Spot centroid coordinates (x s y s ) and laser irradiation time t laser The environmental parameters include ambient temperature T. env and the airflow velocity v on the material surface.

[0019] Further, in step 2, the step of inputting the parameter table into the finite element model to obtain the first dataset specifically involves:

[0020] Set the parameter table P={I p , I a , d, x s , y s , t laser , T env The finite element model is input, and it is calculated sequentially. The calculation results are then extracted to obtain the first dataset O corresponding to each mesh element in the finite element model at each time step. c ={T i,j , s i,j I p , I a , d, x s , y s , x j , y j , z j , T env , v,t i}; where i represents the i-th time step, j represents the j-th grid cell, and T i,j Let s be the temperature of the j-th grid cell at the i-th time step. i,j Let x be the survival state of the j-th grid cell at the i-th time step.j , y j , z j Let t be the coordinate of the j-th grid cell. i For the current time, t i ∈ (0, t) laser ].

[0021] Furthermore, in step 2, the material parameters of the target material and the first dataset are organized to form the second dataset, specifically as follows:

[0022] Organize the material parameters of the target material and the first dataset O c ={T i,j , s i,j I p , I a , d, x s , y s , x j ,y j , z j , T env , v, t i}, forming the second dataset D={I p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j ,k T,j , ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j ; T i,j , s i,j};

[0023] in: n is the total number of time steps, and N is the total number of grid cells.

[0024] Furthermore, step 3 specifically involves:

[0025] The second dataset D={I p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j , k T,j, ρ T,j , a T,j ,T env , v, T i-1,j , s i-1,j ; T i,j , s i,j The system is divided into training and testing sets to construct an initial neural network model. This initial model is then trained and tested using the training and testing sets until an optimal neural network model is obtained. The input parameters of the optimal neural network model are I. nn ={I p , I a , d, x s , y s , x j , y j , z j , Δt i ,c T,j , k T,j , ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j The output parameter is O. nn ={T i,j , s i,j}

[0026] Furthermore, step 4 specifically includes:

[0027] Step 4.1: Obtain the geometry of the material to be tested, mesh the geometry, select the meshes of interest from all the meshes, and obtain the coordinates (x, y) of each mesh of interest. j ', y j ', z j ');

[0028] Step 4.2: Initialize the temperature of each key grid to T. env The survival state of each key grid is initialized to 1, and the laser irradiation time is set to t. laser '.

[0029] Further, in step 5, setting the calculation time step and assigning the current time is specifically as follows:

[0030] Set the calculation time step Δt i ', and assign the current time t. i ', t i '∈(0,t) laser ').

[0031] Further, in step 5, the calculation of the distribution parameters of the laser beam used for ablation of the test material on the surface of the test material specifically involves:

[0032] Based on the current time t i The spatial relationship between the laser beam used to ablate the test material and the surface of the test material is determined, and the distribution parameters of the laser beam on the surface of the test material are calculated. These distribution parameters include the peak power density I. p Average power density I a ', Spot diameter d' and Spot centroid coordinates (x, y) s ', y s ').

[0033] Furthermore, step 6 specifically includes:

[0034] For the aforementioned distribution parameters, the coordinates (x) of the grid of particular interest j ', y j ', z j '), Time step Δt i The material parameters of the material to be tested and the current environmental parameters are combined to obtain the third dataset I. nn '={I p ', I a ', d', x s ',y s ', x j ', y j ', z j �, Δt i ', c T,j ', k T,j ', ρ T,j ', a T,j ', T env ', v', T i-1,j ', s i-1,j '} .

[0035] Furthermore, step 7 specifically includes:

[0036] The third dataset I nn '={I p ', I a ', d', x s ', y s ', x j ', y j ', z j �, Δt i ', c T,j ',k T,j ', ρ T,j ', a T,j ', Tenv ', v', T i-1,j ', s i-1,j If the optimal neural network model is input, then the optimal neural network model will output the current time t. i The prediction result O under ' nn '={T i,j ', s i,j '}; where T i,j ' represents the temperature of the material under test in the j-th grid cell at the i-th time step, and s represents the temperature of the material under test in the j-th grid cell at the i-th time step. i,j ' represents the survival state of the material under test in the j-th grid cell at the i-th time step;

[0037] Step 8 specifically includes:

[0038] Repeat steps 5 through 7 to continuously predict the temperature and survival status of the material under test until the laser irradiation time t is reached. laser ', thus obtaining the material under test throughout the laser irradiation time t laser All prediction results within ' are used to complete the rapid prediction of the variable parameter laser ablation effect.

[0039] The beneficial effects of this invention are:

[0040] 1. This invention provides a rapid prediction method for variable parameter laser ablation effect based on finite element model and neural network model. Simulation data (i.e., the second dataset) is obtained through the finite element model of laser ablation effect. The neural network model is trained and tested using the simulation data to obtain the optimal neural network model. Then, the corresponding data of the material to be tested are obtained according to the actual working conditions. The calculation time step and laser irradiation time are set. Combined with the optimal neural network model, the laser ablation effect is predicted in real time. The calculation accuracy is high and the time consumption is short. It is suitable for rapid prediction of laser ablation effect in strong dynamic scenarios.

[0041] 2. The present invention provides a method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model. It can be encapsulated into a calculation module, and can be iteratively calculated on its own or embedded in other systems for use. It has a wide range of applications.

[0042] 3. The present invention provides a rapid prediction method for variable parameter laser ablation effect based on finite element model and neural network model. During the prediction process, the laser irradiation time and calculation time step can be modified in real time according to the actual working conditions, so as to make the prediction results more accurate. Attached Figure Description

[0043] Figure 1 This is a flowchart of an embodiment of the present invention, which is a method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model.

[0044] Figure 2This is a graph showing the calculation results of the finite element model in step 3 of this embodiment of the invention. The horizontal axis represents time, and the vertical axis represents the temperature at the center point of the target material surface.

[0045] Figure 3 The images show the temperature field and ablation morphology of the material under test in this embodiment of the invention. Detailed Implementation

[0046] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] This invention provides a method for rapid prediction of variable parameter laser ablation effects based on finite element model and neural network model, such as... Figure 1 As shown, the rapid prediction method for variable parameter laser ablation effect specifically includes the following steps:

[0048] Step 1: Construct a finite element model of the laser ablation effect; this finite element model has batch task processing capabilities, and can efficiently and automatically process a large number of computational tasks to generate corresponding computational data.

[0049] Step 2: Obtain the material parameters of the target material, and based on the actual working conditions, obtain the corresponding laser parameters and environmental parameters. Design the laser and environmental parameters to obtain the input parameter table P={I p , I a , d, x s , y s , t laser ,T env When designing parameters, various application scenarios can be covered.

[0050] The aforementioned material parameters include heat capacity c(T), thermal conductivity k(T), density ρ(T), and absorptivity a(T); the aforementioned laser parameters include peak power density I. p Average power density I a , Spot diameter d, Spot centroid coordinates (x s y s ) and laser irradiation time t laser The above environmental parameters include ambient temperature T. env and the airflow velocity v on the material surface;

[0051] In this embodiment, the target material is selected as iron-based material, including Q460 high-strength steel, 304 stainless steel, 316 stainless steel, No. 42 steel, etc.; the airflow velocity v is subsonic, ranging from (0~200) m / s.

[0052] Step 3: Obtain the material parameters of the target material and the parameter table P={I p , I a , d, x s , y s , t laser , T env Input the finite element model described above, and the finite element model will perform calculations on it sequentially. The calculation results (such as...) Figure 2 The curve shown is a typical calculation result curve. By traversing and extracting data, the first dataset O corresponding to each mesh element in the finite element model at each time step is obtained. c ={T i,j , s i,j I p , I a , d, x s , y s , x j , y j , z j , T env , v, t i}; where i represents the i-th time step, j represents the j-th grid cell, and T i,j Let s be the temperature of the j-th grid cell at the i-th time step. i,j Let x be the survival state of the j-th grid cell at the i-th time step. j , y j , z j Let t be the coordinate of the j-th grid cell. i For the current time, t i ∈ (0, t) laser ];

[0053] Step 4: Organize the material parameters of the target material and the first dataset O c ={T i,j , s i,j I p , I a , d, x s ,y s , x j , y j , z j , T env , v, t i}, forming the second dataset D={I p , I a , d, x s , y s , x j , y j , zj , Δt i ,c T,j , k T,j , ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j ; T i,j , s i,j};

[0054] in: n is the total number of time steps, and N is the total number of grid cells;

[0055] In this embodiment, the total number of time steps n is 100, and the total number of grid cells N is 130,000;

[0056] Step 5: Convert the second dataset D={I} p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j , k T,j , ρ T,j ,a T,j , T env , v, T i-1,j , s i-1,j ; T i,j , s i,j The dataset is divided into training and testing sets to construct an initial neural network model. This initial model is then trained and tested using the training and testing sets until an optimal neural network model is obtained. The input parameters of this optimal neural network model are I. nn ={I p , I a , d, x s , y s , x j , y j , z j ,Δt i , c T,j , k T,j , ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j The output parameter is O. nn ={T i,j , si,j The initial and optimal neural network models mentioned above include, but are not limited to, backpropagation neural network models, recurrent neural network models, and graph neural network models. In this embodiment, a backpropagation neural network model is used.

[0057] Step 6: Obtain the geometry of the material to be tested, mesh the geometry, select the meshes of interest from all the meshes, and obtain the coordinates (x, y) of each mesh of interest. j ', y j ', z j In this embodiment, the material to be tested is Q460 stainless steel;

[0058] Step 7: Initialize the temperature of each key grid to T. env The survival state of each key grid is initialized to 1, and the laser irradiation time is set to t. laser In this embodiment, T env Given the current room temperature, the laser irradiation time t laser ' is 7 seconds;

[0059] Step 8: Set the calculation time step Δt i ', and assign the current time t. i ', t i '∈(0,t) laser [ ], so that subsequent iterative calculations can verify whether the set laser irradiation time t has been reached. laser In this embodiment, the time step Δt is calculated. i ' is 0.1 seconds;

[0060] Step 9: Based on the current time t i The spatial relationship between the laser beam used to ablate the test material and the surface of the test material is determined, and the distribution parameters of the laser beam on the test material surface are calculated. These distribution parameters include the peak power density I0. p Average power density I a ', Spot diameter d' and Spot centroid coordinates (x, y) s ', y s ');

[0061] Step 10: Obtain the material parameters of the material under test based on the temperature of each grid cell. This includes the distribution parameters, the coordinates of the grid cells of particular interest, and the time step Δt. i The material parameters of the material to be tested and the current environmental parameters are combined to obtain the third dataset I. nn '={I p ', I a ', d', x s ', ys ', x j ', y j ', z j �, Δt i ',c T,j ', k T,j ', ρ T,j ', a T,j ', T env ', v', T i-1,j ', s i-1,j �} ;

[0062] Step 11: Transfer the third dataset I nn '={I p ', I a ', d', x s ', y s ', x j ', y j ', z j �, Δt i ',c T,j ', k T,j ', ρ T,j ', a T,j ', T env ', v', T i-1,j ', s i-1,j If the optimal neural network model is input, then the optimal neural network model will output the current time t. i The prediction result O under ' nn '={T i,j ', s i,j '}; where T i,j ' represents the temperature of the material under test in the j-th grid cell at the i-th time step, and s represents the temperature of the material under test in the j-th grid cell at the i-th time step. i,j ' represents the survival state of the material under test in the j-th mesh element at the i-th time step; s i,j A value of '1' indicates that the mesh cell is alive, s i,j A value of 0 indicates that the grid cell is dead;

[0063] Step 12: Repeat steps 8 to 11 to continuously predict the temperature and survival status of the material under test until the laser irradiation time t is reached. laser ', thus obtaining the material under test throughout the laser irradiation time t laser All prediction results within the specified timeframe enable rapid prediction of variable-parameter laser ablation effects. This iterative calculation module features a user interface that can be embedded into other software systems that require prediction of laser ablation effects on materials.

[0064] When repeating steps 7 and 9, the laser irradiation time and calculation time step can be modified according to the actual working conditions.

[0065] like Figure 3 As shown, further predictions can be made based on the above prediction results O. nn '={T i,j ', s i,j Output the temperature field distribution and ablation morphology of the material under test.

[0066] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present invention should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A rapid prediction method for variable parameter laser ablation effects based on finite element model and neural network model, characterized in that, Includes the following steps: Step 1: Construct a finite element model of the laser ablation effect; Step 2: Obtain the material parameters of the target material and the corresponding laser and environmental parameters, and set up a parameter table for the laser and environmental parameters; Input the parameter table into the finite element model to obtain the first dataset; organize the material parameters of the target material and the first dataset to form the second dataset; the target material includes the material to be tested. Step 3: Build an initial neural network model, train and test the initial neural network model using the second dataset, until the optimal neural network model is obtained; Step 4: Obtain the geometry of the material to be tested and mesh it, filter out the meshes of interest, and obtain the coordinates of each mesh of interest; initialize the temperature and survival status of each mesh of interest, and set the laser irradiation time; Step 5: Set the calculation time step and assign the current time as the value; Calculate the distribution parameters of the laser beam used to ablate the material under test on the surface of the material under test; Step 6: Combine the distribution parameters, the coordinates of the grid of interest, the time step, the material parameters of the material to be tested, and the current environmental parameters to obtain the third dataset; Step 7: Input the third dataset into the optimal neural network model. The optimal neural network model will then output the predicted results of the material under test at the current time, which will be the temperature and survival status of the material under test. Step 8: Repeat steps 5 to 7 to continuously predict the temperature and survival status of the material under test until the laser irradiation time is reached, and obtain all prediction results of the material under test throughout the entire laser irradiation time, thus completing the rapid prediction of the variable parameter laser ablation effect.

2. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 1, characterized in that, In step 2, the acquisition of the material parameters of the target material and the corresponding laser and environmental parameters, and the setting of a parameter table for the laser and environmental parameters, specifically involves: Obtain the material parameters of the target material, acquire the corresponding laser parameters and environmental parameters based on the actual working conditions, design the laser parameters and environmental parameters, and obtain the input parameter table P={I p , I a , d, x s , y s , t laser , T env The material parameters include heat capacity c(T), thermal conductivity k(T), density ρ(T), and absorptivity a(T); the laser parameters include peak power density I. p Average power density I a , Spot diameter d, Spot centroid coordinates (x s y s ) and laser irradiation time t laser The environmental parameters include ambient temperature T. env and the airflow velocity v on the material surface.

3. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 2, characterized in that, In step 2, the step of inputting the parameter table into the finite element model to obtain the first dataset is specifically as follows: Set the parameter table P={I p , I a , d, x s , y s , t laser , T env The finite element model is input, and it is calculated sequentially. The calculation results are then extracted to obtain the first dataset O corresponding to each mesh element in the finite element model at each time step. c ={T i,j , s i,j I p , I a , d, x s , y s , x j , y j , z j , T env , v, t i }; where i represents the i-th time step, j represents the j-th grid cell, and T i,j Let s be the temperature of the j-th grid cell at the i-th time step. i,j Let x be the survival state of the j-th grid cell at the i-th time step. j , y j , z j Let t be the coordinate of the j-th grid cell. i For the current time, t i ∈ (0, t) laser ].

4. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 3, characterized in that, In step 2, the material parameters of the target material and the first dataset are organized to form the second dataset, specifically: Organize the material parameters of the target material and the first dataset O c ={T i,j , s i,j I p , I a , d, x s , y s , x j , y j ,z j , T env , v, t i }, forming the second dataset D={I p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j , k T,j ,ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j ; T i,j , s i,j }; in: n is the total number of time steps, and N is the total number of grid cells.

5. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 4, characterized in that, Step 3 specifically involves: The second dataset D={I p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j , k T,j , ρ T,j , a T,j , T env ,v, T i-1,j , s i-1,j ; T i,j , s i,j The system is divided into training and testing sets to construct an initial neural network model. This initial model is then trained and tested using the training and testing sets until an optimal neural network model is obtained. The input parameters of the optimal neural network model are I. nn ={I p , I a , d, x s , y s , x j , y j , z j , Δt i , c T,j ,k T,j , ρ T,j , a T,j , T env , v, T i-1,j , s i-1,j The output parameter is O. nn ={T i,j , s i,j } 6. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 5, characterized in that, Step 4 specifically includes: Step 4.1: Obtain the geometry of the material to be tested, mesh the geometry, select the meshes of interest from all the meshes, and obtain the coordinates (x, y) of each mesh of interest. j ', y j ', z j '); Step 4.2: Initialize the temperature of each key grid to T. env The survival state of each key grid is initialized to 1, and the laser irradiation time is set to t. laser '.

7. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 6, characterized in that, In step 5, setting the calculation time step and assigning the current time is specifically as follows: Set the calculation time step Δt i ', and assign the current time t. i ', t i '∈(0,t) laser ').

8. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 7, characterized in that, In step 5, the calculation of the distribution parameters of the laser beam used for ablation on the surface of the test material is specifically as follows: Based on the current time t i The spatial relationship between the laser beam used to ablate the test material and the surface of the test material is determined, and the distribution parameters of the laser beam on the surface of the test material are calculated. These distribution parameters include the peak power density I. p Average power density I a ', Spot diameter d' and Spot centroid coordinates (x, y) s ', y s ').

9. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 8, characterized in that, Step 6 specifically involves: For the aforementioned distribution parameters, the coordinates (x) of the grid of particular interest j ', y j ', z j '), Time step Δt i The material parameters of the material to be tested and the current environmental parameters are combined to obtain the third dataset I. nn '={I p ', I a ', d', x s ', y s ',x j ', y j ', z j �, Δt i ', c T,j ', k T,j ', ρ T,j ', a T,j ', T env ', v', T i-1,j ', s i-1,j '} .

10. The method for rapid prediction of variable parameter laser ablation effect based on finite element model and neural network model according to claim 9, characterized in that, Step 7 specifically includes: The third dataset I nn '={I p ', I a ', d', x s ', y s ', x j ', y j ', z j �, Δt i ', c T,j ', k T,j ',ρ T,j ', a T,j ', T env ', v', T i-1,j ', s i-1,j If the optimal neural network model is input, then the optimal neural network model will output the current time t. i The prediction result O under ' nn '={T i,j ', s i,j '}; where T i,j ' represents the temperature of the material under test in the j-th grid cell at the i-th time step, and s represents the temperature of the material under test in the j-th grid cell at the i-th time step. i,j ' represents the survival state of the material under test in the j-th grid cell at the i-th time step; Step 8 specifically includes: Repeat steps 5 through 7 to continuously predict the temperature and survival status of the material under test until the laser irradiation time t is reached. laser ', thus obtaining the material under test throughout the laser irradiation time t laser All prediction results within ' are used to complete the rapid prediction of the variable parameter laser ablation effect.

Images