A multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm

The KDP crystal microdefect repair process parameters are optimized through the NSGA-II genetic algorithm, which solves the problem that the coupling influence of multi-process parameters is difficult to meet high-quality and high-efficiency repair, and optimizes for different needs, improving repair efficiency and quality.

CN115309108BActive Publication Date: 2024-11-12HARBIN INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210971683.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-08-12
Publication Date
2024-11-12
Estimated Expiration
2042-08-12

AI Technical Summary

Technical Problem

During the repair process of KDP crystal micro defects, the existing technology is difficult to effectively optimize the coupling effect of multiple process parameters, resulting in the difficulty of repairing surface quality and efficiency to meet the needs of high standards at the same time, and the existing methods fail to target different degrees of laser irradiation. Reasonable repair of micro defects.

Method used

Using a multi-objective optimization method based on NSGA-II genetic algorithm, the intrinsic relationship between process parameters and surface roughness and repair time is established through linear regression fitting, a multi-objective decision model is constructed, and the optimal process parameters are obtained by using NSGA-II algorithm. The solution is divided according to the optimal solution, and the quality-first solution and efficiency-first solution are generated for repair.

Benefits of technology

The process parameter optimization is achieved for different repair needs, the efficiency and quality of KDP crystal micro defect repair is improved, and the method of predicting surface quality and repair time and parameter selection is provided, which improves the efficiency and quality of the repair process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115309108B_ABST
    Figure CN115309108B_ABST
Patent Text Reader

Abstract

The present invention provides a multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm, which belongs to the technical field of optical element processing. In order to solve the problem that the existing micro-milling repair research lacks research on the coupling effect of multiple process parameters and simultaneously considers the needs of surface quality and repair efficiency. The method comprises the following steps: S1, taking layer milling allowance, feed speed, spindle speed and spiral pitch as decision variables, and taking surface roughness Sa and repair time T as optimization targets, constructing an objective function; S2, determining the constraints of the multi-objective decision model; S3, constructing a multi-objective decision model according to the constructed objective function and constraints; S4, solving the decision variables in the multi-objective decision model using the NSGA-II algorithm; S5, selecting the required priority solution according to the processing requirements for repair processing. The present invention provides an effective method for determining the actual repair processing parameters for different repair surface roughness and repair efficiency requirements.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of optical element processing, and in particular to a multi-objective optimization method for an optical crystal micro-defect repair process based on an NSGA-II genetic algorithm. Background Art

[0002] With the increasing depletion of natural resources such as coal and oil, countries around the world have built inertial confinement fusion devices (ICF) driven by strong lasers to obtain a safe, environmentally friendly and renewable energy. In ICF, KDP crystals have excellent nonlinear optical properties and can grow to a large diameter of more than 500mm under artificial conditions. Therefore, they are currently the only optical crystal material that can meet the requirements of ICF laser systems. At present, single-point diamond flycutting technology is the most ideal ultra-precision processing method for large-diameter KDP crystals, but it is easy to produce micro defects such as pits and cracks during the processing. At the same time, during the subsequent laser target service of the components, these micro defects will further expand due to the irradiation of high-energy lasers, thereby greatly reducing the laser damage threshold and service life of the crystal components.

[0003] In order to repair the micro defects on the surface of KDP crystal components, the common repair methods currently include CO2 laser repair, wet etching, femtosecond laser ablation and mechanical repair processing. Among them, the processing method of high-speed micro-milling repair using a ball-end milling cutter has been proven to be able to completely remove the damage points and obtain a stable repair surface quality, so that the KDP crystal can withstand multiple high-energy laser target shooting without brittleness and other phenomena. At this stage, the micro-machining method based on ball-end micro-milling is the preferred method for repairing large-diameter KDP crystal components. The current micro-milling repair of KDP crystal components adopts a combination of layer milling and spiral milling. Layer milling roughly mills the repair contour, and spiral milling optimizes the repair parameters to reduce the surface roughness of the repair contour. On the one hand, in the micro-milling repair process, the selection of each process parameter directly affects the surface roughness and repair efficiency of the milling profile, and the current research mainly focuses on the influence of a single process parameter on the surface quality or repair efficiency of micro-milling repair, while ignoring the joint influence of multiple process parameter coupling on the surface quality and repair efficiency of micro-milling repair; on the other hand, in engineering applications, after being irradiated with lasers of different energy levels, the micro-defects on the surface of KDP crystal components vary in degree, and the repair of KDP crystal components sometimes needs to meet high-quality repair and sometimes needs to meet high-efficiency repair, but the current work has not proposed a reasonable optimization method for the above repair requirements. Therefore, it is of great research significance and engineering value to determine the optimal combination of repair process parameters for different repair surface quality and repair efficiency requirements. Summary of the invention

[0004] The present invention adopts the following technical solutions to solve the above technical problems:

[0005] The present invention provides a multi-objective optimization method for an optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm, comprising the following steps:

[0006] S1. According to the KDP crystal micro-defect repair processing, the layer milling allowance a p , feed speed f, spindle speed n and spiral pitch d are taken as decision variables, surface roughness Sa and repair time T are taken as optimization targets, and the objective function is constructed by linear regression fitting according to the intrinsic relationship between the optimization target and the decision variables;

[0007] S2. Determine the constraints of the multi-objective decision model based on the accuracy of the KDP crystal micro-defect repair motion system, the plasticity domain cutting parameter requirements, the surface quality requirements, and the current research on single process parameters;

[0008] S3, construct a multi-objective decision model based on the objective function constructed in S1 and the constraints determined in S2;

[0009] S4. Use the NSGA-II algorithm to solve the decision variables in the multi-objective decision model and obtain the optimal solution of the decision variables in the model;

[0010] S5. According to the distribution of the optimal solutions of the decision variables obtained, the quality priority solutions and the efficiency priority solutions are divided, the required priority solution is selected according to the processing requirements, and the NC machining NC code converted from the process parameters of the selected priority solution is used for repairing the micro defects of the KDP crystal;

[0011] The S4 comprises the following steps:

[0012] S41. Set the parameters of the NSGA-II algorithm, including the number of initialized populations N, the maximum number of iterations, the crossover probability, the mutation probability, the crossover distribution index, and the mutation distribution index, and obtain the parent population P. t , t = 1;

[0013] S42, sort the individuals non-dominatedly to obtain the non-dominated set, and calculate the crowding degree of each individual in each non-dominated set, and then obtain the offspring population Q through selection, crossover and mutation. t ;

[0014] S43, the offspring population Q t With the parent population P t Merge to form a population R of size 2N t , according to the elite strategy, the population is non-dominated and sorted to generate non-dominated sets and calculate the crowding degree to generate a new offspring population Q t+1 ;

[0015] S44, determine whether the number of iterations has reached the maximum number of iterations, if so, proceed to S25, if not, return to S22, set t=t+1, and perform iterative calculation;

[0016] S45. Output the optimal solution of the decision variables and the algorithm terminates.

[0017] Furthermore, in S1, the layer milling allowance a p , feed speed f and spiral pitch d are variables, and the function of the repair time T is constructed as follows:

[0018]

[0019] Layer milling allowance a p , feed speed f, spiral pitch d and spindle speed n are variables, and the function of surface roughness Sa is constructed as follows:

[0020] Sa=AK 5×5 A T

[0021] Where A=[1a p fnd];

[0022] Where, T 层 is the layer milling time, T 螺 is the spiral milling time, R is the radius of the repaired contour, k is the fitting coefficient, b is the feed and retract and acceleration and deceleration time coefficient; A is the process parameter matrix, K 5×5 is the coefficient matrix.

[0023] Furthermore, the constraints of the multi-objective decision-making model in S2 are:

[0024] Layer milling allowance a p The value should be within the range allowed by the machine tool, that is,

[0025] a pmin ≤a p ≤a pmax

[0026] where a pmin 、a pmax They are the maximum and minimum values ​​of the layer milling allowance respectively;

[0027] The feed speed f should be within the range allowed by the machine tool, that is,

[0028] f min ≤f≤f max

[0029] where f min 、f max are the maximum and minimum values ​​of the feed speed respectively;

[0030] The value of the spiral pitch d should be within the range allowed by the machine tool, that is,

[0031] d min ≤d≤d max

[0032] where d min d max are the maximum and minimum values ​​of the spiral pitch respectively;

[0033] The spindle speed n should be within the range allowed by the machine tool, that is,

[0034] n min ≤n≤n max

[0035] Where n min 、n max are the maximum and minimum values ​​of the spindle speed respectively.

[0036] Furthermore, the multi-objective decision model is:

[0037]

[0038]

[0039] Furthermore, the surface roughness is measured by a white light interferometer, and the low-frequency information of the measured surface is filtered using a high-pass Gaussian filter to obtain the actual roughness of the repaired surface caused by the machining knife marks.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention discloses a multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm. First, the present invention fits a mathematical model through linear regression, establishes the intrinsic relationship between multiple process parameters and surface roughness and repair time, and provides a calculation model for the subsequent multi-objective optimization of KDP crystal micro-defect repair process parameters.

[0042] 2. The present invention uses the NSGA-Ⅱ genetic algorithm to calculate the optimal solution set of process parameters for surface roughness and repair time. Based on the optimal solution set, the actual KDP crystal micro-defect repair processing parameters can be determined according to different repair surface roughness and repair efficiency requirements, providing methods and means for predicting surface quality and repair time and selecting process parameters in the actual repair process.

[0043] 3. The present invention divides the optimal solution into quality priority solution and efficiency priority solution according to the distribution of the optimal solution, and analyzes the distribution law of different optimal solutions in the process range, and gives the recommended range of process parameters for the quality priority solution of KDP crystal repair under the premise of ensuring the repair efficiency, which is beneficial to improving the efficiency and quality of KDP crystal micro-defect repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a multi-objective optimization flow chart of the KDP crystal micro-milling repair process in an embodiment of the present invention;

[0045] Figure 2 Schematic diagram of layer milling allowance in the layer milling stage in an embodiment of the present invention;

[0046] Figure 3 Schematic diagram of the repair profile in the spiral milling stage in an embodiment of the present invention;

[0047] Figure 4 A distribution diagram of two optimization targets for KDP crystal repair in an embodiment of the present invention;

[0048] Figure 5 The repaired surface topography image and the filtered topography image are obtained by quality priority resolution through white light interferometry in an embodiment of the present invention;

[0049] Figure 6 The repaired surface topography image and the filtered topography image are obtained by taking pictures and processing with a white light interferometer in an embodiment of the present invention with a priority on efficiency;

[0050] Figure 7 This is a distribution diagram of repair time and surface roughness with layer milling allowance in an embodiment of the present invention;

[0051] Figure 8 The distribution diagram of the repair time and surface roughness with the feed speed in the embodiment of the present invention;

[0052] Fig. 9 The distribution diagram of repair time and surface roughness with spindle speed in the embodiment of the present invention;

[0053] Fig.10 This is a distribution diagram of repair time and surface roughness versus spiral pitch in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In the description of the present invention, it should be noted that the terms "first", "second", and "third" mentioned in the embodiments of the present invention are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second", and "third" may explicitly or implicitly include one or more of the features.

[0055] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0056] Combination Figure 1 As shown, the present invention provides a multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm, comprising the following steps:

[0057] S1. According to the KDP crystal micro-defect repair processing, the layer milling allowance a p , feed speed f, spindle speed n and spiral pitch d are taken as decision variables, surface roughness Sa and repair time T are taken as optimization targets, and the objective function is constructed by linear regression fitting according to the intrinsic relationship between the optimization target and the decision variables;

[0058] S2. Determine the constraints of the multi-objective decision model based on the accuracy of the KDP crystal micro-defect repair motion system, the plasticity domain cutting parameter requirements, the surface quality requirements, and the current research on single process parameters;

[0059] S3, construct a multi-objective decision model based on the objective function constructed in S1 and the constraints determined in S2;

[0060] S4. Use the NSGA-II algorithm to solve the decision variables in the multi-objective decision model and obtain the optimal solution of the decision variables in the model;

[0061] S5. According to the distribution of the optimal solutions of the decision variables obtained, the quality priority solutions and the efficiency priority solutions are divided, the required priority solution is selected according to the processing requirements, and the NC machining NC code converted from the process parameters of the selected priority solution is used for repairing the micro defects of the KDP crystal;

[0062] The S4 comprises the following steps:

[0063] S41. Set the parameters of the NSGA-II algorithm, including the number of initialized populations N, the maximum number of iterations, the crossover probability, the mutation probability, the crossover distribution index, and the mutation distribution index, and obtain the parent population P. t , t = 1;

[0064] S42, sort the individuals non-dominatedly to obtain the non-dominated set, and calculate the crowding degree of each individual in each non-dominated set, and then obtain the offspring population Q through selection, crossover and mutation. t ;

[0065] S43, the offspring population Q t With the parent population P t Merge to form a population R of size 2N t, according to the elite strategy, the population is non-dominated and sorted to generate non-dominated sets and calculate the crowding degree to generate a new offspring population Q t+1 ;

[0066] S44, determine whether the number of iterations has reached the maximum number of iterations, if so, proceed to S25, if not, return to S22, set t=t+1, and perform iterative calculation;

[0067] S45. Output the optimal solution of the decision variables and the algorithm terminates.

[0068] Since the length of the repair time is directly related to the actual cutting path length and feed speed of layer milling and spiral milling; Figure 2 As shown in the figure, in the layer milling stage, since the ball end milling cutter processes layer by layer from top to bottom, the layer milling time and the layer milling allowance a p Into a linear relationship;

[0069] like Figure 3 As shown, in the spiral milling stage, based on the concept of microelement, the repair contour area can be approximated into multiple rectangles, and the area of ​​each rectangle is equal to the spiral step multiplied by the feed speed;

[0070] The tool feed and retract as well as acceleration and deceleration time should also be considered during the actual machining process.

[0071] Therefore, S1 is based on the layer milling allowance a p , feed speed f and spiral pitch d are variables, and the function of the repair time T is constructed as follows:

[0072]

[0073] Due to the layer milling allowance a p , feed speed f, spindle speed n and spiral pitch d all have an impact on surface roughness. Here, the function of surface roughness Sa is constructed by using the quaternary quadratic linear regression fitting method;

[0074] Therefore, the layer milling allowance a p , feed speed f, spiral pitch d and spindle speed n are variables, and the function of surface roughness Sa is constructed as follows:

[0075] Sa=AK 5×5 A T

[0076] Among them, A=[1 a p fnd];

[0077] Where, T 层 is the layer milling time, T 螺 is the spiral milling time, R is the radius of the repaired contour, k is the fitting coefficient, b is the feed and retract and acceleration and deceleration time coefficient; A is the process parameter matrix, K 5×5is the coefficient matrix.

[0078] The constraints of the multi-objective decision-making model described in S2 are:

[0079] Layer milling allowance a p The value should be within the range allowed by the machine tool, that is,

[0080] a pmin ≤a p ≤a pmax

[0081] where a pmin 、a pmax They are the maximum and minimum values ​​of the layer milling allowance respectively;

[0082] The feed speed f should be within the range allowed by the machine tool, that is,

[0083] f min ≤f≤f max

[0084] where f min 、f max are the maximum and minimum values ​​of the feed speed respectively;

[0085] The value of the spiral pitch d should be within the range allowed by the machine tool, that is,

[0086] d min ≤d≤d max

[0087] where d min d max are the maximum and minimum values ​​of the spiral pitch respectively;

[0088] The spindle speed n should be within the range allowed by the machine tool, that is,

[0089] n min ≤n≤n max

[0090] Where n min 、n max are the maximum and minimum values ​​of the spindle speed respectively.

[0091] The multi-objective decision model is:

[0092]

[0093]

[0094] Since low-frequency surface errors will be introduced when measuring the roughness of the repaired surface using white light interferometer, it is necessary to use a high-pass Gaussian filter function to filter the low-frequency information of the measured surface in order to obtain the actual roughness of the repaired surface caused by the machining knife marks.

[0095] Specific implementation plan:

[0096] Firstly, according to the accuracy of micro-defect repair, plasticity domain cutting parameter requirements, surface quality requirements and existing research on single process parameters, the process parameter constraints of KDP crystal micro-milling repair are determined as shown in Table 1;

[0097] Table 1

[0098]

[0099] 100 sets of machining parameter combinations were randomly generated within the process range, and then the UG software established a cone-shaped repair contour model with a width of 800μm and a depth of 40μm. The CNC machining codes under different process parameters were generated, and the machining experiments were carried out on the repair machine tool. The machining time during the repair process was recorded. The surface roughness of the repair contour was obtained using a white light interferometer, and 100 sets of data on four process parameters, surface roughness and repair time were established.

[0100] According to the 100 sets of repair time and corresponding process parameter data, the function of the repair time T is obtained by data fitting:

[0101]

[0102] 100 data sets of surface roughness, repair time and corresponding process parameters were established. The four-variable quadratic linear regression equation of surface roughness was obtained by fitting the established data sets through the regress function of MATLAB. The code is as follows:

[0103] A = xlsread('data set.xlsx');

[0104] a p =A(1:100,1);% layer milling allowance

[0105] f=A(1:100,2); % feed rate

[0106] n=A(1:100,3);%spindle speed

[0107] d=A(1:100,4);% spiral milling step

[0108] T=A(1:100,5);%Repair time

[0109] Sa=A(1:100,6);% roughness

[0110] X=[ones(size(T))a p.^2a.*f a.*n a.*d f.^2f.*n f.*d n.^2n.*d d.^2a f nd];

[0111] [b,bint,r,rint,stats]=regress(Sa,X);

[0112] The coefficients of the quaternary quadratic linear regression equation calculated by the software are shown in the following table:

[0113] Table 2

[0114] A B C D E F G H 549.46634 -4.219344 -7.951739 -0.000436 -0.083095 163.49265 0.0003584 9.6079628 I J K L M N O 6.50E-09 -1.79E-05 0.5966724 36.147519 -510.7599 -7.14E-05 -28.70442

[0115] The function of the surface roughness Sa obtained by fitting is:

[0116] Sa=549.46634-4.219344a p 2 -7.951739apf -0.000436a p n-0.083095a p d+163.49265f 2 +0.0003584fn+9.6079628fd+6.50×10 -9 n 2 -1.79×10 -5 nd+0.5966724d 2 +36.147519a p -510.7599f-7.14×10 -5 n-28.70442d

[0117] The goodness determination coefficient R of the roughness equation obtained by fitting 2 =0.9451>0.85, P in the significance test of the equation = 2.47×10 -47 <0.001, indicating that the fitting effect of this equation is good.

[0118] According to the obtained functions of repair time and surface roughness and the constraints of various process parameters, a multi-objective decision model is constructed; the constraint range of surface roughness is set to 25-85nm, and the constraint range of repair time is set to 60-300s. The NSGA-II algorithm is used to solve the decision variables in the multi-objective decision model. The specific steps are as follows:

[0119] Set the parameters of the NSGA-II algorithm, including the initial population number N as 100, the maximum number of iterations as 500, the crossover probability as 0.9, the mutation probability as 0.1, the crossover distribution index as 20, and the mutation distribution index as 20, and obtain the parent population P t , t = 1;

[0120] Perform non-dominated sorting on the individuals to obtain a non-dominated set. Calculate the crowding degree of each individual in each non-dominated set, and then obtain the offspring population Q through selection, crossover, and mutation. t ;

[0121] The offspring population Q t With the parent population P t Merge to form a population R of size 2N t , according to the elite strategy, the population is non-dominated and sorted to generate non-dominated sets and calculate the crowding degree to generate a new offspring population Q t+1 ;

[0122] Perform iterative calculation until the number of iterations reaches 500;

[0123] The optimal solution of the decision variables is output and the algorithm terminates.

[0124] Result analysis:

[0125] like Figure 4 As shown in the figure, the repair time of the quality priority solution is longer but the roughness is smaller, while the processing time of the efficiency priority solution is shorter but the roughness is significantly increased. The congestion of the two solutions is similar and there is no complete dominance relationship. Figure 5 As shown in the figure, the surface roughness of the machined surface is 26 nm under the quality priority solution. Figure 6 It is shown that the roughness of the machined surface under the efficiency priority solution is 39 nm, which can verify the accuracy of the quality priority solution and efficiency priority solution obtained by the present invention.

[0126] like Figure 7 As shown in the figure, in the range of 0.5 to 3 μm of layer milling allowance, the efficiency priority solutions are distributed, and most of them are concentrated at 3 μm; the quality priority solutions are all distributed at 0.5 μm and 3 μm, which shows that in the process of population variable variation, a p <0.5μm and a p >3μm can calculate a large number of better solutions, the best quality solution a p The values ​​are 0.5μm and 3μm, and note that a p =3μm roughness is basically greater than a p = Roughness of 0.5μm.

[0127] like Figure 8As shown in the figure, in the range of feed speed 0.2-1.5 mm / s, the optimal solution is distributed in 0.4-1.5 mm / s, indicating that 0.2-0.4 mm / s is an overflow interval. The decision variables in this interval have low efficiency due to the slow feed speed, but the roughness does not increase. They belong to the dominated population and have been eliminated in the evolution process. The processing time decreases with the increase of feed speed, and the roughness increases with the increase of feed speed. It can be seen that the feed speed has a greater impact on the processing time and surface quality.

[0128] like Fig. 9 As shown in the figure, in the range of spindle speed of 35000 ~ 60000rpm, the quality priority solution is mainly distributed at 36000 ~ 46000rpm and 60000rpm, but when the spindle speed is close to 60000rpm, the speed fluctuation frequency of the machine tool movement will be greater than 150Hz, and the vibration amplitude will reach 30μm, which will have an adverse effect on the stability of the machine tool. Therefore, the quality priority solution is selected at 36000 ~ 46000rpm.

[0129] like Fig.10 As shown in the figure, in the range of spiral pitch of 5μm to 20μm, the efficiency priority solution is mainly concentrated at 20μm, while the quality priority solution is mainly concentrated at 5μm; the optimal solution is mainly concentrated at the upper and lower limits of the interval, indicating that a better solution can be calculated outside the interval based on the established mathematical model, but in the actual KDP repair process, when the spiral pitch is greater than 20μm, the roughness value of the crystal repair surface no longer meets the engineering requirements; when too small a tool spacing is used, it will not only extend the processing time and reduce the processing efficiency, but also cause plastic plowing effect due to the soft crystal.

[0130] In order to verify the effectiveness of the prediction model and algorithm results, 10 solutions were randomly selected from the 100 optimal solutions evolved by genetic evolution for experimental verification. The specific values ​​of the decision variables and optimization targets of the experiment are shown in Table 3. Among them, the calculated value is the result calculated by the fitting function in the NSGA-Ⅱ algorithm operation, and the actual value is the measured result. From the data comparison, the measured results are consistent with the algorithm prediction results. The error between the measured results and the predicted results of the machining time is small, only within the 1s error allowed by manual timing. The measured roughness is slightly larger than the predicted roughness as a whole, and the error does not exceed 4nm when the layer milling allowance is 0.5μm; the error does not exceed 6nm when the allowance is 3μm. In summary, the optimal solution calculated by the NSGA-Ⅱ genetic algorithm meets better quality and efficiency requirements and has a good predictive effect on the actual optimization target value.

[0131] Table 3

[0132]

[0133] The repair efficiency of this implementation can be controlled within a few minutes, meeting the repair efficiency requirement, and further improving the quality of the repair according to the distribution of quality-priority solutions.

[0134] In summary, based on the distribution of quality-first solutions and combined with the physical properties of KDP crystals, machine tool performance, etc., a set of high-quality repair parameters can be summarized: a p =0.5μm, 0.4mm / s≤f≤0.9mm / s, 43000r / min≤n≤46000r / min, d=5μm.

[0135] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the protection scope of the present invention.

Claims

1. A multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm, characterized in that The steps include: S1. According to the KDP crystal micro-defect repair processing, the layer milling allowance a p , feed speed f, spindle speed n and spiral pitch d are taken as decision variables, surface roughness Sa and repair time T are taken as optimization targets, and the objective function is constructed by linear regression fitting according to the intrinsic relationship between the optimization target and the decision variables; S2. Determine the constraints of the multi-objective decision model based on the accuracy of the KDP crystal micro-defect repair motion system, the plasticity domain cutting parameter requirements, the surface quality requirements, and the current research on single process parameters; S3, construct a multi-objective decision model based on the objective function constructed in S1 and the constraints determined in S2; S4. Use the NSGA-II algorithm to solve the decision variables in the multi-objective decision model and obtain the optimal solution of the decision variables in the model; S5. According to the distribution of the optimal solutions of the decision variables obtained, the quality priority solutions and the efficiency priority solutions are divided, the required priority solution is selected according to the processing requirements, and the NC machining NC code converted from the process parameters of the selected priority solution is used for repairing the micro defects of the KDP crystal; The S4 comprises the following steps: S41. Set the parameters of the NSGA-II algorithm, including the number of initialized populations N, the maximum number of iterations, the crossover probability, the mutation probability, the crossover distribution index, and the mutation distribution index, and obtain the parent population P. t , t = 1; S42, sort the individuals non-dominatedly to obtain the non-dominated set, and calculate the crowding degree of each individual in each non-dominated set, and then obtain the offspring population Q through selection, crossover and mutation. t ; S43, the offspring population Q t With the parent population P t Merge to form a population R of size 2N t , according to the elite strategy, the population is non-dominated and sorted to generate non-dominated sets and calculate the crowding degree to generate a new offspring population Q t+1 ; S44, determine whether the number of iterations has reached the maximum number of iterations, if so, proceed to S25, if not, return to S22, set t=t+1, and perform iterative calculation; S45. Output the optimal solution of the decision variables and the algorithm terminates.

2. The multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm according to claim 1, characterized in that In S1, the layer milling allowance a p , feed speed f and spiral pitch d are variables, and the function of the repair time T is constructed as follows: Layer milling allowance a p , feed speed f, spiral pitch d and spindle speed n are variables, and the function of surface roughness Sa is constructed as follows: That=AK 5×5 A T Among them, A=[1 a p fnd]; Where, T 层 is the layer milling time, T 螺 is the spiral milling time, R is the radius of the repaired contour, k is the fitting coefficient, b is the feed and retract and acceleration and deceleration time coefficient; A is the process parameter matrix, K5 ×5 is the coefficient matrix.

3. According to claim 2, a multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm, It is characterized in that the constraints of the multi-objective decision-making model in S2 are: Layer milling allowance a p The value should be within the range allowed by the machine tool, that is, a pmin ≤a p ≤a pmax where a pmin 、a pmax They are the maximum and minimum values ​​of the layer milling allowance respectively; The feed speed f should be within the range allowed by the machine tool, that is, f min ≤f≤f max where f min 、f max are the maximum and minimum values ​​of the feed speed respectively; The value of the spiral pitch d should be within the range allowed by the machine tool, that is, d min ≤d≤d max where d min ,d max are the maximum and minimum values ​​of the spiral pitch respectively; The spindle speed n should be within the range allowed by the machine tool, that is, n min ≤n≤n max where n min 、n max are the maximum and minimum values ​​of the spindle speed respectively.

4. The multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm according to claim 3 is characterized in that The multi-objective decision model is:

5. The multi-objective optimization method for optical crystal micro-defect repair process based on NSGA-Ⅱ genetic algorithm according to claim 1, characterized in that The surface roughness is measured by white light interferometer, and the low-frequency information of the measured surface is filtered using a high-pass Gaussian filter to obtain the actual roughness of the repaired surface caused by the machining knife marks.