CN112131679B - Reverse thrust adjustment method for initial residual stress of workpiece - Google Patents

Reverse thrust adjustment method for initial residual stress of workpiece Download PDF

Info

Publication number
CN112131679B
CN112131679B CN202010907299.4A CN202010907299A CN112131679B CN 112131679 B CN112131679 B CN 112131679B CN 202010907299 A CN202010907299 A CN 202010907299A CN 112131679 B CN112131679 B CN 112131679B
Authority
CN
China
Prior art keywords
stress
group
data
workpiece
residual stress
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010907299.4A
Other languages
Chinese (zh)
Other versions
CN112131679A (en
Inventor
郭江
贺增旭
景浩
白倩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian University of Technology
Original Assignee
Dalian University of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dalian University of Technology filed Critical Dalian University of Technology
Priority to CN202010907299.4A priority Critical patent/CN112131679B/en
Publication of CN112131679A publication Critical patent/CN112131679A/en
Application granted granted Critical
Publication of CN112131679B publication Critical patent/CN112131679B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/17—Mechanical parametric or variational design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P10/00—Technologies related to metal processing
    • Y02P10/20—Recycling

Abstract

A method for adjusting the reverse thrust of the initial residual stress of a workpiece belongs to the field of mechanical engineering. Firstly, dividing two identical workpieces into an experimental group A and a measurement group B, carrying out heat treatment, and removing materials of the workpiece A; and (4) measuring the stress of each layer of the group B workpiece by using a delamination method. Secondly, a nonlinear relation is constructed by using a support vector machine algorithm. And finally, adjusting the initial residual stress by using a genetic algorithm, and adopting the stress data of the individual with the maximum fitness as the finally adjusted data. The method adjusts the thought of the initial residual stress of the workpiece through the support vector machine algorithm and the genetic algorithm, and solves the problem of inaccurate simulation prediction result caused by limited measurement technology; constructing a nonlinear relation among residual stress, removal amount and deformation aiming at a simulation result of CAE software by using a support vector machine algorithm, and simplifying a calculation process; and the genetic algorithm is used for carrying out the iterative convergence of the final stress of each layer, and the convergence speed is high and the result is accurate.

Description

Reverse thrust adjustment method for initial residual stress of workpiece
Technical Field
The invention belongs to the field of mechanical engineering, relates to a method for adjusting measured initial residual stress, and particularly relates to a residual stress back-stepping adjusting method based on a support vector machine and a genetic algorithm.
Background
With the development of economy and science and technology in China, the demand of fields such as information electronics, aerospace, new energy industry and the like on high-end parts is higher and higher, so that the requirement on the manufacturing precision of core parts is higher and higher. Workpiece deformation is an important factor affecting the accuracy of workpiece manufacture. And the residual stress of the workpiece is generally considered to be the most important cause of deformation of the workpiece. The residual stress of the workpiece is generally classified into two types: the residual stress present in the workpiece itself, i.e., the initial residual stress and the residual stress introduced by the machining, i.e., the machining residual stress. The existence of residual stress brings great challenges to precision and ultra-precision machining.
The conventional residual stress measuring method mainly comprises a damaged test method and a nondestructive test method, and the conventional measuring method mainly comprises a strain gauge drilling method, a laser speckle pinhole method, a delamination method, a crack flexibility method, a ray diffraction method and the like. In the actual processing process, different stress measurement methods can be selected according to different material properties and processing requirements. However, the current stress measurement method has a minimum measurement error of only ± 20MPA due to the technical limitations, and for a soft metal material such as copper, the actual value of the residual stress in each direction inside the soft metal material after heat treatment is less than-20 MAP-20MPA, which causes a problem that the measured value of the residual stress of the soft metal is greatly deviated from the actual value.
The patent CN 103542961A discloses a milling residual stress prediction method based on deflection change and a reverse thrust method, wherein after a material is removed layer by layer, the stress corresponding to the layer is calculated by measuring the deflection of the last layer, and the relation among the stresses of the layers is established to reversely thrust the residual stress of the layers layer by layer. The method has higher measurement precision on the deformation of the workpiece, and sometimes the internal stress distribution state of the workpiece cannot be reflected more truly; a method for obtaining training data through Abaqus simulation, constructing a nonlinear relation by using a support vector machine and then optimizing machining parameters by using a genetic algorithm is provided in the simulation prediction and cutting parameter optimization of thin-wall parts. The method only optimizes the processing parameters and does not adjust the residual stress in the workpiece.
For the existing problems, the invention provides a residual stress back-stepping adjusting method based on a support vector machine and a genetic algorithm.
Disclosure of Invention
The invention provides an initial residual stress backstepping adjusting method based on a support vector machine and a genetic algorithm, aiming at the problem that the error of the initial residual stress is larger due to the limitation of the current measuring method, the adjusted initial residual stress can better reflect the initial stress distribution state in a workpiece compared with the unadjusted initial residual stress, and the stress redistribution state and the workpiece processing deformation after processing and removal have better prediction effects.
In order to achieve the purpose, the invention adopts the technical scheme that:
a residual stress back-stepping adjusting method based on a support vector machine and a genetic algorithm mainly comprises the following steps:
the method comprises the steps of firstly, dividing two workpieces with the same initial state into an experiment group A and a measurement group B, carrying out heat treatment, removing materials of the workpieces of the experiment group A, and recording the removal amount of the upper surface and the removal amount of the lower surface as r 0 And r 0 ', amount of deformation is denoted as Y 0 。
And step two, equally dividing the workpiece of the measurement group B into m parts in the thickness direction, measuring the stress of each layer by using a stripping method, and recording the stress of each layer as: sigma 1 ,σ 2 ,σ 3 ...σ m 。
Thirdly, using a support vector machine algorithm to construct a nonlinear relation:
randomly generating H groups of data within the range of +/-20 MPa based on the stress measurement result in the second step, wherein each group of data contains m parts; the H group data was randomly divided into L training sets and (H-L) test sets. The stress levels of the data in the ith group are respectively recorded as:
Figure BDA0002661877600000021
and (2) randomly generating H groups of data in the range of (0, 200), wherein each group of data comprises an upper surface removal amount and a lower surface removal amount. Wherein the data of the upper surface removal amount and the lower surface removal amount of the i-th group are recorded as (r) i ，r i ’)。
Step (3), H models are established in finite element simulation software, stress data generated in the step (1) are input into the models, and the models are simulated according to the removal quantity generated in the step (2)And really, the workpiece can deform after the material is removed, and the deformation of the ith model is recorded as Y i 。
And (4) aiming at the front L groups of training set data, establishing a relationship among initial stress, removal amount and deformation by using a support vector machine algorithm:
establishing a similar function f by using a Gaussian kernel function, and marking the similar function of the ith group as f (i) :
Figure BDA0002661877600000022
The input vector X of the ith group is noted as:
Figure BDA0002661877600000023
wherein the content of the first and second substances,
Figure BDA0002661877600000024
the stress data established in the step (1); r is a radical of hydrogen i ,r i ' removing amount data established for the step (2).
The parameter to be optimized theta is recorded as:
θ＝[θ 1 ,θ 2 ,θ 3 ...θ m ,θ m+1 ,θ m+2 ]
if the model material is removed to be a single surface, only the parameter theta is required to be removed m+2 And (5) deleting. The target parameters to be optimized are:
Figure BDA0002661877600000025
wherein, the function cost 0 And cost 1 The function is a cost function.
The optimized nonlinear relation is as follows:
Y(X)＝θ T f
wherein, theta T For the transposition of the parameters to be optimized, f is the kernel function established above.
And (5) detecting the optimized nonlinear relation by using the (H-L) group of test set data, wherein the nonlinear relation constructed by the evaluation algorithm is accurate.
And fourthly, adjusting the initial residual stress by using a genetic algorithm.
And (1) randomly generating a population of K individuals by using the method in the step (1) in the third step. Each individual also contains m parts of stress, and the method for generating the stress is the same as the step (1) in the third step, wherein the stress magnitude of the j group of data is respectively recorded as:
Figure BDA0002661877600000031
step (2) of calculating removal amounts r of the respective populations j on the upper and lower surfaces based on the above-mentioned relationship by using the nonlinear relationship Y (X) established in the third step 0 、r 0 The amount of deformation in the case of is noted as Y (j) ’。
And (3) taking the stress value and the removal quantity of each layer as chromosomes of different individuals, and carrying out floating point number coding.
And (4) designing a fitness function S (j), and taking the fitness function as the distribution selection probability of different individuals in the population. The individual probabilities are ranked from large to small. Wherein S (j) is:
Figure BDA0002661877600000032
step (5), adopting roulette method to make selection operation, and using support vector machine to calculate obtained deformation value Y (j) ' with the deformation value Y measured in the experimental group in the first step 0 The approaching individuals are passed on to the next generation.
And (6) setting the crossing probability as a and the mutation probability as b, and crossing and mutating the population individuals to form a new population.
And (7) repeating the steps (4) to (6) until the iteration times are finished, and recording the individual with the maximum fitness.
And fifthly, using the stress data of the individual with the maximum fitness as the finally adjusted data.
The invention has the following beneficial effects: (1) The idea of adjusting the initial residual stress of the workpiece through the support vector machine algorithm and the genetic algorithm is innovatively provided, and the problem of inaccurate simulation prediction results caused by limited measurement technologies is solved.
(2) And a support vector machine algorithm is used for constructing a nonlinear relation among residual stress, removal amount and deformation aiming at a simulation result of finite element simulation software, so that the calculation process is simplified.
(3) And the genetic algorithm is used for carrying out the iterative convergence of the final stress of each layer, and the convergence speed is high and the result is accurate.
(4) The method is not only suitable for the material double-sided removal model, but also suitable for the single-sided removal model, and has a wide application range.
Drawings
FIG. 1 is a flow chart of stress adjustment.
Fig. 2 is a machine learning flow diagram.
FIG. 3 is a flow chart of genetic algorithm adjustment.
Fig. 4 model removal schematic.
In the figure: 1, an original model; 2, removing the upper surface of the model; 3, removing the lower surface of the model; 4 removing the back model.
Detailed description of the invention
In order to further understand the method for adjusting the residual stress back-stepping based on the support vector machine and the genetic algorithm, the following embodiment is used to describe the invention in detail, and the method comprises the following steps:
the workpiece is a copper sheet with the diameter of 100mm and the thickness of 2 mm.
The method comprises the steps of dividing two workpieces with the same initial state into an experiment group A and a measurement group B, carrying out heat treatment, removing materials from the workpieces of the experiment group A, and measuring the removal amount of the upper and lower surfaces to be r0=10 μm and r 0 ' =15 μm, and the amount of deformation is represented as Y 0 ＝26.12μm。
And step two, equally dividing the workpieces of the measurement group B into 20 parts in the thickness direction, measuring the stress of each layer by using a stripping method, and recording the stress of each layer as: sigma 1 ,σ 2 ,σ 3 ...σ 20 。
Thirdly, using a support vector machine algorithm to construct a nonlinear relation:
and (1) randomly generating 400 groups of data within the range of +/-20 MPA based on the stress measurement result in the second step, wherein each group of data comprises 20 parts, and randomly dividing the 400 groups of data into 300 training sets and 100 testing sets. The stress levels of the data in the ith group are respectively recorded as:
Figure BDA0002661877600000041
and (2) randomly generating 400 groups of data in the range of (0, 200), wherein each group of data comprises an upper surface removal amount and a lower surface removal amount. Wherein the data of the upper surface removal amount and the lower surface removal amount of the i-th group are recorded as (r) i ，r i ’)。
Step (3), 400 models are built in CAE software, stress data generated in the step (1) are input into the models, the models are simulated according to the removal amount generated in the step (2), the workpiece can deform after the material is removed, and the deformation generated by the ith model is recorded as Y i 。
And (4) aiming at the first 300 groups of training set data, establishing the relationship between initial stress, removal amount and deformation by using a support vector machine algorithm:
establishing a similar function f by using a Gaussian kernel function, and marking the similar function of the ith group as f (i) :
Figure BDA0002661877600000042
The input vector X of the ith group is noted as:
Figure BDA0002661877600000043
wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0002661877600000044
the stress data established in the step (1); r is i ,r i ' removing amount data established for the step (2).
The parameter to be optimized theta is recorded as:
θ＝[θ 1 ,θ 2 ,θ 3 ...θ m ,θ m+1 ,θ m+2 ]
the target parameters to be optimized are as follows:
Figure BDA0002661877600000051
wherein, the function cost 0 And cost 1 The function is a cost function.
The optimized nonlinear relation is as follows:
Y(X)＝θ T f
wherein theta is T For the transposition of the parameters to be optimized, f is the kernel function established above.
And (5) detecting the optimized nonlinear relation by using 100 groups of test set data, wherein the nonlinear relation constructed by the evaluation algorithm is accurate.
And fourthly, adjusting the initial residual stress by using a genetic algorithm.
And (1) randomly generating 100 groups of data as 100 individuals by using the method in the step (1) in the third step.
Step (2), calculating the removal amounts of the upper and lower surfaces of each population j to r0=10 μm and r, respectively, by using the nonlinear relation Y (X) established in the third step 0 ' =15 μm deformation, and is denoted as Y (j) ’。
And (3) taking the stress value and the removal quantity of each layer as chromosomes of different individuals, and carrying out floating point number coding.
And (4) designing a fitness function S (j), and taking the fitness function as the distribution selection probability of different individuals in the population. The individual probabilities are ranked from large to small. Wherein S (j) is:
Figure BDA0002661877600000052
and (5) reselecting individuals according to the individual probability by using a roulette method, and ensuring that the population total number is 100.
And (6) setting the crossing probability as a =0.6 and the mutation probability as b =0.1, and crossing and mutating the population individuals to form a new population.
And (7) repeating the steps (4) to (6) until the iteration times are finished, and recording the individual with the maximum fitness.
And fifthly, using the stress data of the individual with the maximum fitness as the finally adjusted data.
The above-mentioned embodiments only express the embodiments of the present invention, but not should be understood as the limitation of the scope of the invention patent, it should be noted that, for those skilled in the art, many variations and modifications can be made without departing from the concept of the present invention, and these all fall into the protection scope of the present invention.

Claims (1)

1. A method for adjusting the initial residual stress of a workpiece in a reverse thrust manner is characterized by comprising the following steps of:
the first step, dividing two workpieces with the same initial state into an experiment group A and a measurement group B, carrying out heat treatment, removing materials of the workpieces of the experiment group A, and recording the removal amount of the upper and lower surfaces as r 0 And r 0 ', amount of deformation is denoted as Y 0 ；
And step two, equally dividing the workpiece of the measurement group B into m parts in the thickness direction, measuring the stress of each layer by using a stripping method, and recording the stress of each layer as: sigma 1 ,σ 2 ,σ 3 ...σ m ；
Thirdly, using a support vector machine algorithm to construct a nonlinear relation:
randomly generating H groups of data within the range of +/-20 MPa based on the stress measurement result in the second step, wherein each group of data contains m parts; randomly dividing the H group data into an L group training set and an (H-L) group testing set; the stress levels of the data in the ith group are respectively recorded as:
Figure FDA0002661877590000011
randomly generating H groups of data within the range of (0, 200), wherein each group of data comprises an upper surface removal amount and a lower surface removal amount; wherein the data of the upper surface removal amount and the lower surface removal amount of the i-th group are recorded as (r) i ，r i ’)；
Step (3), H models are established in CAE simulation software, stress data generated in the step (1) are input into the models, the models are simulated according to the removal amount generated in the step (2), the workpiece can deform after the material is removed, and the deformation generated by the ith model is recorded as Y i ；
And (4) aiming at the front L groups of training set data, establishing a relationship among initial stress, removal amount and deformation by using a support vector machine algorithm:
establishing a similar function f by using a Gaussian kernel function, and marking the similar function of the ith group as f (i) :
Figure FDA0002661877590000012
The input vector X of the ith group is noted as:
Figure FDA0002661877590000013
wherein the content of the first and second substances,
Figure FDA0002661877590000014
the stress data established in the step (1); r is i ,r i ' removing amount data established for the step (2);
the parameter to be optimized theta is recorded as:
θ＝[θ 1 ,θ 2 ,θ 3 ...θ m ,θ m+1 ,θ m+2 ]
if the model material is removed as a single surface, only the parameter theta needs to be removed m+2 Deleting;
the target parameters to be optimized are as follows:
Figure FDA0002661877590000015
wherein the function cost 0 And cost 1 The function is a cost function;
the optimized nonlinear relation is as follows:
Y(X)＝θ T f
wherein, theta T F is the kernel function established for the transposition of the parameters to be optimized;
step 5, detecting the optimized nonlinear relation by using the (H-L) group of test set data, wherein the nonlinear relation constructed by the evaluation algorithm is accurate;
fourthly, adjusting the initial residual stress by using a genetic algorithm;
randomly generating a population of K individuals by using the method in the third step (1); each individual also contains m parts of stress, and the method for generating the stress is the same as the step (1) in the third step, wherein the stress magnitude of the j group of data is respectively recorded as:
Figure FDA0002661877590000021
step (2) of calculating removal amounts r of the respective populations j on the upper and lower surfaces based on the above-mentioned relationship by using the nonlinear relationship Y (X) established in the third step 0 、r 0 The amount of deformation in the case of is noted as Y (j) ’；
Step (3), taking stress values and removal quantities of all layers as chromosomes of different individuals, and carrying out floating point number coding;
designing a fitness function S (j), and taking the fitness function as the distribution selection probability of different individuals of the population; arranging the individual probabilities from large to small; wherein S (j) is:
Figure FDA0002661877590000022
step (5), adopting roulette method to make selection operation, and using support vector machine to calculate the obtained deformation value Y (j) ' with the deformation value Y measured in the experimental group in the first step 0 Close individuals are inherited by the next generation;
step (6), setting the cross probability as a and the mutation probability as b, and forming a new population after crossing and mutating the population individuals;
step (7), repeating the steps (4) - (6) until the iteration times are finished, and recording the individual with the maximum fitness;
and fifthly, using the stress data of the individual with the maximum fitness as the final adjustment data.
CN202010907299.4A 2020-09-02 2020-09-02 Reverse thrust adjustment method for initial residual stress of workpiece Active CN112131679B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010907299.4A CN112131679B (en) 2020-09-02 2020-09-02 Reverse thrust adjustment method for initial residual stress of workpiece

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010907299.4A CN112131679B (en) 2020-09-02 2020-09-02 Reverse thrust adjustment method for initial residual stress of workpiece

Publications (2)

Publication Number Publication Date
CN112131679A CN112131679A (en) 2020-12-25
CN112131679B true CN112131679B (en) 2022-10-14

Family

ID=73847222

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010907299.4A Active CN112131679B (en) 2020-09-02 2020-09-02 Reverse thrust adjustment method for initial residual stress of workpiece

Country Status (1)

Country Link
CN (1) CN112131679B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113191050B (en) * 2021-04-28 2024-02-02 大连理工大学 Workpiece deformation simulation prediction method based on MSVR-GA (minimum shift register-GA) optimized initial residual stress

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102456109A (en) * 2011-08-01 2012-05-16 中国人民解放军国防科学技术大学 Training method for Trojan horse incident prediction least squares support vector machine, and prediction method
CN104951834A (en) * 2015-06-24 2015-09-30 上海大学 LSSVM (least squares support vector machine) wind speed forecasting method based on integration of GA (genetic algorithm) and PSO (particle swarm optimization)
CN108509335B (en) * 2018-01-31 2021-03-19 浙江理工大学 Software test data generation method based on genetic algorithm optimization
CN110837929A (en) * 2019-11-06 2020-02-25 苏文电能科技股份有限公司 Least square support vector machine electricity utilization prediction method based on adaptive genetic algorithm

Also Published As

Publication number Publication date
CN112131679A (en) 2020-12-25

Similar Documents

Publication Publication Date Title
CN107357992B (en) Composite structure correction method for finite element model based on cluster analysis
Cho et al. Investigation on influence of infill pattern and layer thickness on mechanical strength of PLA material in 3D printing technology
CN112668227A (en) Thin-wall part cutter relieving deformation error prediction model establishing method and application thereof
Ragavendran et al. Sensitivity analysis and optimization of EDM process parameters
CN113191050B (en) Workpiece deformation simulation prediction method based on MSVR-GA (minimum shift register-GA) optimized initial residual stress
CN112131679B (en) Reverse thrust adjustment method for initial residual stress of workpiece
Luo et al. Optimization of process parameters for the minimization of surface residual stress in turning pure iron material using central composite design
CN101738949A (en) System and method for controlling membrane-base reflecting mirror surface shape
Wang et al. Development of a prediction model based on RBF neural network for sheet metal fixture locating layout design and optimization
Liu et al. Small data-driven modeling of forming force in single point incremental forming using neural networks
CN107292029A (en) A kind of determination method that sheet forming technological parameter is predicted based on forming defects
CN114861519B (en) Initial ground stress field acceleration optimization inversion method under complex geological conditions
CN109739178B (en) Workpiece machining deformation online prediction method driven by monitoring data
Guo et al. A novel method for workpiece deformation prediction by amending initial residual stress based on SVR-GA
Shibahara et al. Approach to automation of line heating by combination of reinforcement learning and finite element method simulation
CN111444619B (en) Online analysis method and equipment for injection mold cooling system
Bai et al. A hybrid physics-data-driven surface roughness prediction model for ultra-precision machining
Gao et al. A 6-sigma robust optimization method for stamping forming of automobile covering parts based on residual error and radial basis interpolation
Rahmani Dehaghani et al. Iterative uncertainty calibration for modeling metal additive manufacturing processes using statistical moment-based metric
CN112179541B (en) Initial residual stress adjusting method based on deformation reverse thrust
CN114813347A (en) Method for predicting mechanical property of ultrathin niobium strip
CN114386188A (en) Kergin agent model-based optimization design method for valve clack of nuclear power steam safety valve
Mantari et al. New methodology for the construction of best theory diagrams using neural networks and multi-objective genetic algorithm
Glänzel et al. Parameterization of environmental influences by automated characteristic diagrams for the decoupled fluid and structural-mechanical simulations
Deng et al. On-Line surface roughness classification for multiple CNC milling conditions based on transfer learning and neural network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant