CN100383796C - Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method - Google Patents
Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method Download PDFInfo
- Publication number
- CN100383796C CN100383796C CNB2005100479036A CN200510047903A CN100383796C CN 100383796 C CN100383796 C CN 100383796C CN B2005100479036 A CNB2005100479036 A CN B2005100479036A CN 200510047903 A CN200510047903 A CN 200510047903A CN 100383796 C CN100383796 C CN 100383796C
- Authority
- CN
- China
- Prior art keywords
- parameter
- design
- neural network
- rolling
- finite element
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 91
- 229910000881 Cu alloy Inorganic materials 0.000 title claims abstract description 32
- 239000000463 material Substances 0.000 title claims description 46
- 238000005516 engineering process Methods 0.000 title claims description 39
- 238000003801 milling Methods 0.000 title claims description 14
- 238000013461 design Methods 0.000 claims abstract description 107
- 238000005096 rolling process Methods 0.000 claims abstract description 67
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 52
- 238000005457 optimization Methods 0.000 claims abstract description 46
- 230000002068 genetic effect Effects 0.000 claims abstract description 43
- 230000008569 process Effects 0.000 claims abstract description 39
- 238000004088 simulation Methods 0.000 claims abstract description 22
- 238000002474 experimental method Methods 0.000 claims abstract description 13
- 238000013528 artificial neural network Methods 0.000 claims description 88
- 238000005266 casting Methods 0.000 claims description 55
- 238000007667 floating Methods 0.000 claims description 51
- 238000012549 training Methods 0.000 claims description 30
- 239000011159 matrix material Substances 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 21
- 230000000977 initiatory effect Effects 0.000 claims description 19
- 238000001816 cooling Methods 0.000 claims description 17
- 230000007547 defect Effects 0.000 claims description 17
- 238000004519 manufacturing process Methods 0.000 claims description 15
- 238000009749 continuous casting Methods 0.000 claims description 14
- 238000004513 sizing Methods 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000010354 integration Effects 0.000 claims description 3
- 238000003754 machining Methods 0.000 abstract description 3
- 210000005036 nerve Anatomy 0.000 abstract 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 54
- 229910052802 copper Inorganic materials 0.000 description 54
- 239000010949 copper Substances 0.000 description 54
- 230000002950 deficient Effects 0.000 description 21
- 230000006870 function Effects 0.000 description 16
- 229910000570 Cupronickel Inorganic materials 0.000 description 10
- 208000037656 Respiratory Sounds Diseases 0.000 description 7
- 230000004075 alteration Effects 0.000 description 5
- 238000005304 joining Methods 0.000 description 5
- 239000000956 alloy Substances 0.000 description 4
- 210000004027 cell Anatomy 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 230000001537 neural effect Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000002203 pretreatment Methods 0.000 description 3
- 238000012827 research and development Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 241000196324 Embryophyta Species 0.000 description 2
- 238000012356 Product development Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 210000000349 chromosome Anatomy 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 238000002921 genetic algorithm search Methods 0.000 description 2
- 229910002804 graphite Inorganic materials 0.000 description 2
- 239000010439 graphite Substances 0.000 description 2
- 230000009931 harmful effect Effects 0.000 description 2
- 230000008676 import Effects 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 229910052751 metal Inorganic materials 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 206010006895 Cachexia Diseases 0.000 description 1
- 240000004859 Gamochaeta purpurea Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000000265 homogenisation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000001050 lubricating effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
- 208000016318 wasting Diseases 0.000 description 1
Images
Landscapes
- Continuous Casting (AREA)
Abstract
The invention discloses a method for designing and optimizing cast-rolling process parameters of copper alloy tubing, using database as design basic, using nerve network as design method of process parameters and indexes, and using genetic algorithm as process parameter optimizing means, integrating nerve network, genetic algorithm, finite element simulation, experiment design, CAD parameterized design and database technique into the process design and parameter optimization, designing and optimizing the cast-rolling process parameters of the copper alloy tubing. And the invention has high automation degree, and can be applied to machining deformation of copper alloy tubing, and make personnel short of special knowledge able to make accurate and standard machining process.
Description
Technical field
The present invention relates to the copper alloy tube material processing technology, specifically a kind of with neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization apply in process parameters design and the optimization with database technology, determine the method for copper-alloy pipe-material casting production optimum process parameter.
Technical background
Casting-rolling technology claims horizontal casting again---planetary rolling is for the base method, is the method that is used for producing accurate copper pipe that Finland OUTOKUMPU company the eighties in last century is developed mid-term.This technology has remarkable advantages such as flow process is short, yield rate is high, cost is low, equipment investment is few, is current advanced person's ARC copper pipe production technology.It has cancelled ingot casting heating, extruding etc., directly produce hollow bloom, adopt the rolling strand of three-roller planetary rolling mill, carry out tube drawing with floating plug after rolling by the horizontal casting unit, the online then dish that is curled into, the copper pipe that it makes the production substance surpass 500kg becomes possibility.This technology comprises three big master operations: horizontal casting, three-roller planetary is rolling and tube drawing with floating plug.
It is typical many kinds, many specifications that the copper-alloy pipe-material casting is produced, the processing technology of multiple operation, tubing changes through aspects such as the size of being everlasting, shape and materials, make the design of copper-alloy pipe-material processing technology that very big mobility be arranged, and workload is big, efficient is low, even there is the slip-stick artist of special knowledge and rich experiences also to be difficult to finish at short notice.Because the scale and the serialization of producing in case the technological design error occurs, cause a large amount of wastings of resources probably, have a strong impact on enterprise and produce.The design and the optimization method of research and development copper-alloy pipe-material casting-milling technology parameter can instruct the tubing processing, improve the quality of products, and speed production and R﹠D cycle, reduce cost, strengthen enterprise competitiveness.
The design of technological parameter in the past is to determine processing parameter according to slip-stick artist self experience and site test repeatedly, this method is intelligence and automaticity is low, method for designing is single, influence the ordinary production of enterprise not only, and be difficult to effectively carry out process parameters design, be difficult to particularly guarantee that obtaining technological parameter is optimized parameter.At present, the form of enterprise with inference machine arranged also, after processes such as experimental knowledge, formula are concluded and put in order, set up knowledge base.Expression model according to knowledge is mapped as discernible structure of computing machine or program with knowledge, enables the logically design and the optimization of reasoning technological parameter.But metal forming processing is a very complicated deformation process, and existing material nonlinearity has geometrical non-linearity again, adds the influence of complicated external constraint, causes forming process very complicated.The reciprocal effect of kinds of processes parameter makes and is difficult to seek optimum combination of process parameters with traditional inference machine.And because obtaining of experimental knowledge is indirect, the knowledge acquisition difficulty of inference machine like this, regular meeting causes the knowledge shot array because of the bad structure of knowledge base.
Summary of the invention
In order to overcome existing method deficiency, the objective of the invention is to the method that proposes a kind of copper-alloy pipe-material casting-milling technology parameter design and optimize.Employing the present invention can be according to trimmed size and used alloy material, according to the copper pipe production procedure realize horizontal casting in the casting-rolling technology, three-roller planetary is rolling and the intelligent process parameters design and the optimization of three master operations of tube drawing with floating plug, its automaticity height, can be applicable to and make the machining deformation of various complexity to lack and enrich the processing technology that the professional knowledge personnel also can make accurate standard.
Technical solution of the present invention is to be design basis with the database, neural network is the method for designing of technological parameter and technic index, genetic algorithm is the process parameter optimizing means, comprehensive integration neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization and database technology, and apply in technological design and the parameter optimization, three processing steps of copper-alloy pipe-material casting are carried out process parameters design and optimization, to obtain optimum technological parameter.Specific as follows:
One, the design of database
Set up horizontal casting, three-roller planetary is rolling and the database of tube drawing with floating plug, with the basis of database as parameter designing and optimization, has stored the production data of these three kinds of operations of factory in the database, normal data and ephemeral data; The plant produced data include: specification storehouse, equipment library, operation storehouse, raw material store, mould storehouse, part warehouse, composition storehouse, technological design storehouse and design result storehouse; Normal data comprises the normative document of states such as the national standard of all size product and America and Europe, Japan; Ephemeral data comprises the initial designs data and calculates the intermediate data that produces.
Two, finite element analogy
Adopt finite element simulation can solve the border and the nonlinear problem of various complexity, be used for manufacture field and can improve the quality of products, reduce the research and development of products cycle, and can reduce cost, boost productivity.Along with the raising of computer level, become a kind of strong analysis means with commercial finite element software simulation forming processing technology, and precision can satisfy the application of engineering design, alternative on-the-spot engineer testing.The designer imports geometric parameter, technological parameter and material parameter in the finite element pre-treatment, after calculating, can obtain the result of calculation that needs in the aftertreatment environment.Adopt the finite element numerical simulation step, obtain horizontal casting temperature field, thermograde and cooling velocity value in the horizontal casting respectively, and derive the crack initiation propensity value; Roll-force evaluation and roll forming defects simulation result during three-roller planetary is rolling; And pulling capacity evaluation in the tube drawing with floating plug and drawing forming defects simulation result.The roll forming defective comprises fracture, tears and rolls card.The drawing forming defective is to break.
Because its modeling of finite element needs corresponding professional knowledge, and calculates length, for the layman can be designed fast, the present invention arranges the finite element numerical simulation scheme with the method for uniform experiment design.Uniform experiment design can economically, scientifically, reasonably be arranged the finite element numerical simulation number of times.Reasonably the testing program design can carried out under less numerical simulation time, the lower cost situation, is comprehensively reflected between input quantity and the output quantity the quantitatively information of rule.
Three, neural network
Multilayer neural network has height nonlinear fitting character and to the adaptability of multiple-input and multiple-output wide range of problem, it is longer than the reasoning of aspects such as handling associative memory, thinking in images, and has self-organization and self-learning capability.Can solve the bottleneck problem of obtaining knowledge.Can be regarded as a kind of implicit representation of rule based on neural network method.
For existing technological design scheme, the present invention learns by multi-layer artificial neural network, joins mould design technology parameter (total drawing passes and every time copper pipe wall thickness, external diameter value) with the drawing that the neural network after the training calculates in the pairing tube drawing with floating plug of corresponding product specification.
Neural network also can with the finite element numerical calculations incorporated, abundant finite element numerical calculation knowledge is excavated, therefrom find the Useful Information rule.Threshold values that obtains behind the neural metwork training and weight matrix can shine upon the relation of technological parameter and technic index as a kind of default rule.The technic index of copper-alloy pipe-material casting comprises the crack initiation propensity value of horizontal casting, the pulling capacity of roll-force that three-roller planetary is rolling and roll forming defective value, tube drawing with floating plug and drawing forming defective value.According to finite element numerical simulation result of calculation, the present invention learns by multi-layer artificial neural network, with the technic index value in the pairing horizontal continuous casting process of technological parameter, three-roller planetary rolling mill practice and the tube drawing with floating plug technology under different condition that calculates of neural network after the training.
Four, genetic algorithm
Genetic algorithm is different from traditional optimized Algorithm, it be a kind of use for reference highly-parallel that organic sphere natural selection and evolutionary mechanism grow up, at random, self-adaptive search algorithm, have big possibility to obtain globally optimal solution.It adopts artificial evolution's mode that object space is carried out the randomization search, body one by one or the chromosome of feasible solution in the Problem Areas being regarded as colony, and each individuality is encoded into the symbol string form, simulate the biological evolution process of Darwinian heredity selection and natural selection, colony is carried out repeatedly based on genetic operation (heredity, intersect and variation), according to predetermined target fitness function each individuality is estimated, according to the survival of the fittest, the evolutionary rule of the survival of the fittest, constantly obtain more excellent colony, search for the optimum individual of optimizing in the colony in overall parallel search mode simultaneously, try to achieve the optimum solution that meets the demands.The step of searching for optimized technological parameter with genetic algorithm is specially adapted to handle bad complexity and the non-linear problem of searching algorithm solution in the past, has obtained using widely in the engineering field.
To the technological parameter to be optimized in the technological parameter, comprise the throwing system parameter and the cooling system parameter of horizontal casting; Roll deflection angle, roller declination angle, opening degree and go-cart speed that three-roller planetary is rolling; Every time external mold cone angle, core print cone angle, drawing speed, external mold sizing section length and core print sizing section length of tube drawing with floating plug.The present invention searches for the optimum process parameter with genetic algorithm as optimization method.
At first, carry out parameter coding, constitute the initialization population according to corresponding optimization aim; Calculate each individual technic index value by neural network, just the fitness value in the genetic algorithm carries out the operation operator operation again; The generations of evolution of population up to searching optimum solution, is determined optimum process parameters, and the result is made process chart and design document.Wherein: operation operator is operated the three kinds of citation forms that comprise selection, intersect and make a variation.
Five, the graphic result of process parameters design and optimization is expressed
Adopt the CAD parameterization design method, with the optimal processing parameter that obtains in conjunction with CAD software carry out the three-roller planetary of mould rolling in the core print Die CAD parametrization design of rolling roll forming CAD, tube drawing with floating plug, designing and calculating, data processing and graphic plotting are carried out overall treatment.
The present invention has following advantage:
1. automaticity height.The present invention is with neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization and database are integrated, can avoid in the copper-alloy pipe-material casting technique forming process occurring harmful effects such as fracture, wrinkling, constriction, be a kind of copper-alloy pipe-material casting-milling technology parameter design and intelligent method of optimizing realized.
With traditional handicraft mainly be according to occur in the copper-alloy pipe-material casting-milling technology forming process in deviser's the experience fracture, plastic force excessive, join harmful effects such as the mould design is unreasonable, compare and revise some parameter of the processing that is shaped repeatedly or revise mold shape, the traditional handicraft process is expensive greatly, product development cycle is long, can not adapt to intense market competition and development of modern industry requirement in the world wide.The present invention can be applicable to complicated copper-alloy pipe-material casting forming technology because the automaticity height costs little, product development cycle weak point, makes to lack and enriches the processing technology that the professional knowledge personnel also can make accurate standard.
Description of drawings
Fig. 1-the 1st, the finite element model of horizontal continuous casting process.
Fig. 1-2 is the finite element model of three-roller planetary rolling mill practice.
Fig. 1-the 3rd, the finite element model of tube drawing with floating plug technology.
Fig. 2 is the neural network structure of horizontal casting crack initiation propensity value.
Fig. 3 is the neural network structure of rolling roll-force of three-roller planetary and roll forming failure prediction.
Fig. 4 is the neural network structure of tube drawing with floating plug power and drawing failure prediction.
Fig. 5 is the artificial neural network structure that tube drawing with floating plug is joined the mould design.
Fig. 6 is the cooling system and the continuous casting throwing system parameter of genetic algorithm optimization horizontal casting.
Fig. 7 is a genetic algorithm optimization three-roller planetary rolling parameter.
Fig. 8 is a genetic algorithm optimization floating core head parameter.
Fig. 9 is the operating process of copper-alloy pipe-material casting-milling technology parameter design and optimization method.
Embodiment
Further specify the present invention below in conjunction with accompanying drawing.
Embodiment
The present invention is with neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization, database technology apply in technological design and the parameter optimization, comprehensive integration is got up, and each processing step of copper-alloy pipe-material casting is optimized design, obtains optimum technological parameter.Process parameters design comprises: horizontal casting throwing system, cooling system optimal design; The rolling velocity field of three-roller planetary is calculated, the rolling parameter design; Mould design and the design of every time drawing parameter optimization are joined in the drawing of tube drawing with floating plug.
Specific as follows:
1) set up horizontal casting, three-roller planetary is rolling and the database of tube drawing with floating plug, with the basis of database as parameter designing and optimization.Production data, normal data and the ephemeral data of these three kinds of operations of factory have been stored in the database.The plant produced data include: specification storehouse, equipment library, operation storehouse, raw material store, mould storehouse, part warehouse, composition storehouse, technological design storehouse and design result storehouse; Normal data comprises the normative document of states such as the national standard of all size product and America and Europe, Japan; Ephemeral data comprises the initial designs data and calculates the intermediate data that produces.
2) set up horizontal casting, three-roller planetary is rolling and the finite element model of tube drawing with floating plug.
The finite element numerical calculation knowledge derives from commercial finite element pre-treatment input variable and corresponding numerical result.The designer imports geometric parameter, technological parameter and material parameter in the finite element pre-treatment, after calculating, can obtain the result of calculation that needs in the aftertreatment environment.Adopt the finite element numerical simulation step, obtain horizontal casting temperature field, thermograde and cooling velocity value in the horizontal casting respectively, and can derive the crack initiation propensity value; Roll-force evaluation and roll forming defects simulation result during three-roller planetary is rolling; And pulling capacity evaluation in the tube drawing with floating plug and drawing forming defects simulation result.Wherein roll forming defective comprises fracture, tears and rolls card, and the drawing forming defective is to break.
For under the situation that reduces finite element analogy number of times, time and reduction simulated cost, the information that is comprehensively reflected quantitative rule between finite element analogy input quantity and the finite element analogy output quantity, the present invention arranges the finite element numerical simulation scheme with the method for uniform experiment design.Uniform experiment design can economically, scientifically, reasonably be arranged the finite element numerical simulation number of times.
1. horizontal casting finite element analogy
Adopt the method for uniform experiment design to arrange the horizontal casting finite element analogy.By the horizontal casting finite element analogy, can obtain the pairing temperature field T of copper pipe strand, material parameter, throwing system and cooling system, thermograde G and cooling velocity value R at different specification size, (wherein 1 is cooling system to finite element model as Figure 1-1,2 is molten alloyed copper, 3 copper pipes that go out for continuous casting, 4 is graphite centre rod, the 5 graphite liners for the crystallizer employing).In horizontal casting, temperature, thermograde and cooling velocity are directly proportional with the probability that crackle produces, so, calculate the crack initiation propensity value of each finite element cell node by following formula according to temperature field, thermograde and the cooling velocity value that finite element analogy obtains.
In the formula, C
iBe the crack initiation propensity value, T is a temperature, and G is a thermograde, and R is a cooling velocity, and i represents cell node number.
In all cell nodes, find out maximum crackle rudiment propensity value C
Max, and the crackle rudiment that calculates tendency mean value, be expressed as follows:
In the formula,
Be crack initiation tendency mean value, i represents cell node number, and N is the node sum of finite element model.
Table 1 horizontal casting influences parameter:
So just having set up horizontal casting influences the quantitative relationship of parameter and crack initiation tendency.
2. the rolling finite element analogy of three-roller planetary
Adopt the method for uniform experiment design to arrange the rolling finite element analogy of three-roller planetary.By the rolling finite element analogy of three-roller planetary, can obtain copper pipe rolled blank, material parameter, rolling parameter and the roll-force value of different specification size and the quantitative relationship of rolling defect value, (wherein, 6 is roll to finite element model shown in Fig. 1-2,7 is copper pipe, and 8 is plug).Here the roll forming defective value is represented with 0 and 1, and 1 representative does not have defective to produce, and on behalf of defectiveness, 0 produce.
The rolling parameter that influences of table 2 three-roller planetary:
| Copper pipe rolled blank specification | Material parameter Y 7 | Rolling parameter |
| Copper pipe strand outside dimension X 1 | Red copper | Roll deflection angle Y 1 |
| Copper pipe strand wall thickness X 2 | Cupronickel B 10 | Roller declination angle Y 2 |
| The rolling back of copper pipe outside dimension D Roll | Cupronickel B 30 | Opening degree Y 3 |
| The rolling back of copper pipe wall thickness S Roll | ...... | Roll rotational speed Y 4 |
| ...... | The speed Y of go-cart 5 | |
| ...... | Friction factor Y 6 |
3. tube drawing with floating plug finite element analogy
Adopt the method for uniform experiment design to arrange the tube drawing with floating plug finite element analogy.By the tube drawing with floating plug simulation, can obtain copper pipe, material parameter, drawing parameter and the pulling capacity P of different specification size
DrawWith drawing forming defective value Q
DrawQuantitative relationship, finite element model is (wherein 9 is external mold, and 10 is copper pipe, and 11 is floating core head) as Figure 1-3.Here the drawing forming defective value is represented with 0 and 1, and 1 representative does not have defective to produce, and on behalf of defectiveness, 0 produce.
Table 3 tube drawing with floating plug influences parameter:
| The drawing copper tubes specification | Material parameter Z 8 | The drawing parameter |
| Outside dimension D before the drawing copper tubes Draw0 | Red copper | Drawing speed Z 4 |
| Drawing copper tubes anterior wall thickness S Draw0 | Cupronickel B 10 | Core print cone angle Z 5 |
| Outside dimension D behind the drawing copper tubes Draw1 | Cupronickel B 30 | External mold cone angle Z 6 |
| Wall thickness S behind the drawing copper tubes Draw1 | ...... | Friction factor Z 7 |
| ...... | External mold sizing section length Z 9 | |
| ...... | Core print sizing section length Z 10 |
3) neural network
To arrange analog input amount that finite element analogy obtains and output quantity as a result with uniform experiment design, after arrangement and homogenization processing, as the training sample of neural network.At horizontal casting, three-roller planetary is rolling and these three kinds of operations of tube drawing with floating plug, uses multilayer (present embodiment adopt 3 layers) artificial neural network learning respective sample data respectively, trains and set up corresponding neural network.With the neural network after the training, after test, can calculate under different condition roll-force numerical value and rolling defect value during the crack initiation propensity value of pairing horizontal continuous casting process, three-roller planetary are rolling, pulling capacity numerical value in the tube drawing with floating plug and drawing defective value.
1. the neural network of horizontal casting crack initiation propensity value
The neural network structure of horizontal casting crack initiation propensity value is: 11 nodes of input layer, its parameter is the input quantity of finite element analogy, is respectively strand outside dimension X
1, strand wall thickness X
2, material parameter X
3, throwing time X
4, X between a stopping time
5, rise X
6, rise time X
7, X between two stopping times
8, casting temperature X
9, inlet water temperature X
10, hydraulic pressure X
11Wherein on behalf of red copper, 1, material parameter represent cupronickel B 10,2 to represent cupronickel B 30 with 0, arrives other Cu alloy material by that analogy.2 nodes of output layer, its parameter are the output quantity as a result of finite element analogy, are crackle rudiment tendency mean value
With crackle rudiment tendency maximal value C
MaxNeural network adopts the BP network based on Levenberg-Marquardt optimized Algorithm (LM algorithm), and hidden layer is a Sigmoid type activation function, and output layer is selected Purelin type activation function for use.The LM algorithm can shorten the training time greatly.
Finite element analogy input quantity and the output quantity that to arrange with uniform experiment design, do normalized after, generate the training sample file, preserve in the database.Behind neural metwork training, if error in tolerance band, deposits threshold values and the weight matrix that obtains neural network in the database in as neural network matrix file.Set up a mapping relations model like this, each influences relation between parameter and the crackle rudiment propensity value can to shine upon horizontal casting, and this relation is a kind of implication relation (referring to Fig. 2).
If the finite element analogy result who increases newly is arranged, but after the normalized, be integrated into original training sample file, after the training, neural network threshold values and the weight matrix that newly obtains deposited in the database as neural network matrix file again.This training sample file and neural network matrix file can upgrade along with increasing at any time of finite element analogy number.
2. the rolling roll-force of three-roller planetary and the neural network of roll forming failure prediction
The neural network structure of roll-force that three-roller planetary is rolling and roll forming failure prediction is: 10 nodes of input layer, its parameter is the input quantity of the rolling finite element analogy of three-roller planetary, is respectively roll deflection angle Y
1, roller declination angle Y
2, opening degree Y
3, roll rotational speed Y
4, go-cart speed Y
5, friction factor Y
6, material parameter Y
7, lengthening coefficient Y
8, initial diameter-wall-rate Y
9Subtract wall with tube reducing and compare Y
10Wherein on behalf of red copper, 1, material parameter represent cupronickel B 10,2 to represent cupronickel B 30 with 0, arrives other Cu alloy material by that analogy.According to similarity principle, lengthening coefficient, initial diameter-wall-rate and tube reducing subtract wall than the variation that can be used to represent rolling forward and backward copper pipe size, and three values are expressed as follows respectively:
Initial diameter-wall-rate
Subtract wall tube reducing ratio
In the formula, X
1, X
2Be respectively rolling preceding copper pipe external diameter and wall thickness, just copper pipe strand outside dimension and copper pipe strand wall thickness; D
Roll, S
RollBe respectively rolling back copper pipe external diameter and wall thickness.
2 nodes of output layer, its parameter are the output quantity as a result of finite element analogy, are pulling capacity P
RollWith drawing forming defective value Q
RollNeural network adopts the BP network based on Levenberg-Marquardt optimized Algorithm (LM algorithm), and hidden layer is a Sigmoid type activation function, and output layer is selected Purelin type activation function for use.
Finite element analogy input quantity and the output quantity that to arrange with uniform experiment design, do normalized after, generate the training sample file, preserve in the database.Behind neural metwork training, if error in tolerance band, deposits threshold values and the weight matrix that obtains neural network in the database in as neural network matrix file.Set up a mapping relations model like this, each influences relation between parameter and pulling capacity and the drawing forming defective value can to shine upon tube drawing with floating plug, and this relation is a kind of implication relation (referring to Fig. 3).
If the finite element analogy result who increases newly is arranged, but after the normalized, be integrated into original training sample file, after the training, neural network threshold values and the weight matrix that newly obtains deposited in the database as neural network matrix file again.This training sample file and neural network matrix file can upgrade along with increasing at any time of finite element analogy number.
3. the neural network of tube drawing with floating plug power and drawing failure prediction
The neural network structure of tube drawing with floating plug power and drawing failure prediction is: 10 nodes of input layer, its parameter is the input quantity of tube drawing with floating plug finite element analogy, is respectively initial diameter-wall-rate Z
1, tube reducing subtracts wall and compares Z
2, lengthening coefficient Z
3, drawing speed Z
4, core print cone angle Z
5, external mold cone angle Z
6,, friction factor Z
7, material parameter Z
8, external mold sizing section length Z
9, core print sizing section length Z
10Wherein on behalf of red copper, 1, material parameter represent cupronickel B 10,2 to represent cupronickel B 30 with 0, arrives other Cu alloy material by that analogy.According to similarity principle, lengthening coefficient, initial diameter-wall-rate and tube reducing subtract wall than the variation that can be used to represent the forward and backward copper pipe size of drawing, and three values are expressed as follows respectively:
Initial diameter-wall-rate
Subtract wall tube reducing ratio
In the formula, D
Draw0, S
Draw0Be respectively preceding copper pipe external diameter of drawing and wall thickness, D
Draw1, S
Draw1Be respectively copper pipe external diameter and wall thickness after the drawing.
2 nodes of output layer, its parameter are the output quantity as a result of finite element analogy, are pulling capacity P
DrawWith drawing forming defective value Q
DrawNeural network adopts the BP network based on Levenberg-Marquardt optimized Algorithm (LM algorithm), and hidden layer is a Sigmoid type activation function, and output layer is selected Purelin type activation function for use.
Finite element analogy input quantity and the output quantity that to arrange with uniform experiment design, do normalized after, generate the training sample file, preserve in the database.Behind neural metwork training, if error in tolerance band, deposits threshold values and the weight matrix that obtains neural network in the database in as neural network matrix file.Set up a mapping relations model like this, each influences relation between parameter and pulling capacity and the drawing forming defective value can to shine upon tube drawing with floating plug, and this relation is a kind of implication relation (referring to Fig. 4).
If the finite element analogy result who increases newly is arranged, but after the normalized, be integrated into original training sample file, after the training, neural network threshold values and the weight matrix that newly obtains deposited in the database as neural network matrix file again.This training sample file and neural network matrix file can upgrade along with increasing at any time of finite element analogy number.
4) tube drawing with floating plug to join the mould method for designing a lot, mainly contain the decline equation method, decline factor method, metal hardenability method, Geometric Sequence successive subtraction method and Kd-Ks method etc.These methods have certain practical value, but might not be fit to actual production.Join the mould design and need consider the various external factors such as material property of appointed condition, lubricating condition, pipe, so these theoretical methods do not possess versatility.Experienced slip-stick artist has the cover of oneself to be fit to self enterprises characteristics, the effective mould method for designing of joining usually.In order to make full use of this experience, adopt Artificial Neural Network model, train and learn join the mould design proposal from all size product of on-site collection.Neural network just can be joined the mould design automatically according to slip-stick artist's design philosophy like this.
The mould scheme is joined in drawing according to existing all size, therefrom selects and produces normal, the stable mould scheme of joining, as effectively joining the mould scheme, through the sample data of arrangement back as neural network.The mould scheme of joining with a kind of specification is an example, and method for sorting is as follows:
Total drawing passes of certain specification copper pipe is: n; Finished product copper pipe external diameter and wall thickness are: D
nAnd S
nCopper pipe external diameter and wall thickness after three-roller planetary rolls, the copper pipe external diameter and the wall thickness that just begin before the drawing are: D
0And S
0Copper pipe external diameter and wall thickness after the i passes of drawing are: D
iAnd S
iThe sample input quantity has 5 to be respectively: D
n, S
n, D
0, S
0And i/n, the sample output quantity has 2 to be respectively D
iAnd S
iThe drawing copper tubes of every kind of specification is joined the mould scheme and can be constituted n-1 sample like this.This kind specification is joined the sample of mould forecast scheme configuration and can be represented with following table:
The various mould schemes of effectively joining according to this method for sorting, are constituted a training sample file jointly, this document is preserved in the database, and come the learning training sample file with multi-layer artificial neural network.Carry out tube drawing with floating plug join the artificial neural network structure of mould design can be with as shown in Figure 5.
Artificial neural network after the training, deposits threshold values and the weight matrix that obtains neural network in the database in as neural network matrix file if satisfy test request through after testing.Set up a mapping relations model like this, can shine upon copper pipe dimensions and drawing and join relation between the mould.
Carry out at first will calculating total drawing passes n value before tube drawing with floating plug joins mould design with this neural network.Its computing method normally rule of thumb formula determine.
According to the copper pipe outer diameter D before the drawing
0With wall thickness S
0, finished product copper pipe outer diameter D
nWith wall thickness S
n, average passage lengthening coefficient
Determine total road number of times.Average passage lengthening coefficient is according to different E
nValue is determined, E
nValue is the material thickness index
In the drawbench better performances, under lubricated good, the normal situation of mould situation, E
nCan suitably get greatly, i.e. E
N+1Otherwise get littler, i.e. E
N-1
The total coefficient of elongation λ of drawing
∑Computing formula as follows:
The computing formula of the total road of drawing number of times n is:
The size of product that will join mould, just finished product copper pipe outer diameter D
nWith wall thickness S
n, and copper pipe external diameter and the wall thickness of three-roller planetary after rolling just begins the copper pipe outer diameter D before the drawing
0With wall thickness S
0, and i/n value, i is meant that (copper pipe size after the drawing of 1≤i＜n), n is total drawing passes value of calculating of formula rule of thumb to the current i passage that will predict.These numerical value as the neural network prediction input quantity, after calling in neural network matrix file, are joined mould design with neural network, can obtain copper pipe dimension D after the i passes of drawing
iAnd S
i
5) genetic algorithm
With genetic algorithm search for continuous casting cooling system and continuous casting throwing system parameter in the optimized horizontal casting, three-roller planetary in rolling rolling parameter and the drawing parameter of tube drawing with floating plug, carry out parameter coding, constitute the initialization population; Calculate each individual fitness value according to neural network, carry out the operation operator operation again; The generations of evolution of population up to searching optimum solution, is determined best process design parameter, and the result is made process chart and design document.Wherein: operation operator is operated the three kinds of citation forms that comprise selection, intersect and make a variation.
1. the cooling system of genetic algorithm optimization horizontal casting and continuous casting throwing system parameter
In the horizontal casting, between throwing time, a stopping time, between rise, rise time, two stopping times, casting temperature, inlet water temperature, hydraulic pressure be the key factor that influences horizontal casting crack initiation propensity value size.Being lower than safety value with horizontal casting crack initiation propensity value mean value minimum and maximal value is optimization aim, strand outside dimension, strand wall thickness, material parameter are as preset parameter, with genetic algorithm seek between with this understanding throwing time, a stopping time, between rise, rise time, two stopping times, the optimal value of casting temperature, inlet water temperature and hydraulic pressure, as shown in Figure 6.
The evaluation function of genetic algorithm claims fitness function again, converts optimization aim to be found the solution to fitness function, and the fitness value available horizontal continuous casting crackle germinating propensity value prediction neural network is here calculated each individual fitness value.Because optimization aim is the mean value minimum of crack initiation propensity value and the problem that maximal value is lower than safety value, constructive formula is:
Fitness value is:
Wherein, C
Max(X
i),
In the formula, X
iEach input quantity in the neural network of expression horizontal casting crack initiation propensity value prediction; C is the safety value of crack initiation propensity value, is constant coefficient, can select to determine according to actual conditions; C
Max(X
i),
The maximal value and the mean value of the crack initiation propensity value that expression obtains with neural network.
Genetic algorithm can adopt floating-point code and integer coding.The population scale of genetic algorithm generally gets 20～100, and in general, the initial population of selection greater number can be handled more simultaneously and separate, thereby finds globally optimal solution easily, and shortcoming is to have increased each iteration time.The hybrid rate of genetic algorithm generally gets 0.4～0.9, and the frequency of crossover operation is high more, can more quickly converge to most promising optimum solution zone, but too high frequency also may cause premature convergence.The general value 0.001～0.1 of the aberration rate of genetic algorithm, population size and chromosome length are big more, and aberration rate is chosen more little.The maximum evolutionary generation of genetic algorithm as a kind of simulation end condition, is decided according to repeatedly trying out on concrete condition, generally in 100～500 generations.
In the present embodiment, the genetic algorithms use floating-point code that horizontal casting optimization is used, the initial population value is 100, the hybrid rate value is 0.85, aberration rate value 0.08, maximum evolutionary generation value is 200.
2. genetic algorithm optimization three-roller planetary rolling parameter
During three-roller planetary is rolling, roll deflection angle, roller declination angle, opening degree, go-cart speed, these die parameters and technological parameter are the key factors of influence of rolled power, roll forming defective.With roll-force minimum and no roll forming defective is optimization aim, with material parameter, friction factor, roll rotational speed, lengthening coefficient, initial diameter-wall-rate with subtract the wall tube reducing than as preset parameter, with the optimal value of genetic algorithm searching roll deflection angle, roller declination angle, opening degree and go-cart speed with this understanding, as shown in Figure 7.
The evaluation function of genetic algorithm claims fitness function again, converts optimization aim to be found the solution to fitness function, and the fitness value here can calculate each individual fitness value with three-roller planetary rolling roll-force and forming defects prediction neural network.Because optimization aim is roll-force minimum value and no forming defects problem, constructive formula is:
Fitness value is:
Wherein, P
Roll(Y
i), S
Roll(Y
i)=ANN
Roll-force that three-roller planetary is rolling and forming defects prediction neural network(Y
i)
In the formula, Y
iEach input quantity in roll-force that the expression three-roller planetary is rolling and the forming defects prediction neural network; P
Roll(Y
i), S
Roll(Y
i) expression the roll-force value and the failure prediction value that obtain with neural network.
In the present embodiment, the genetic algorithm encoding of three-roller planetary rolling parameter optimization usefulness adopts floating-point code, and the initial population value is 80, and the hybrid rate value is 0.75, aberration rate value 0.08, and maximum evolutionary generation value is 100.
3. genetic algorithm optimization floating core head parameter
In the tube drawing with floating plug of aldary, external mold cone angle, core print cone angle, drawing speed, external mold sizing section length and core print sizing section length, these die parameters and technological parameter are the key factors that influences pulling capacity, drawing forming defective.With pulling capacity minimum and no forming defects is optimization aim, with material parameter, friction factor, drawing speed, lengthening coefficient, initial diameter-wall-rate with subtract the wall tube reducing than as preset parameter, with the optimal value of genetic algorithm searching external mold cone angle, core print cone angle, external mold sizing section length and core print sizing section length with this understanding, as shown in Figure 8.
The evaluation function of genetic algorithm claims fitness function again, converts optimization aim to be found the solution to fitness function, and the fitness value of present embodiment can calculate each individual fitness value with the pulling capacity and the forming defects prediction neural network of tube drawing with floating plug.Because optimization aim is pulling capacity minimum value and no forming defects problem, constructive formula is:
P
Draw(Z
i), S
Draw(Z
i)=ANN
The pulling capacity of tube drawing with floating plug and forming defects prediction neural network(Z
i)
In the formula, Z
iEach input quantity in the pulling capacity of expression tube drawing with floating plug and the forming defects prediction neural network; P
Draw(Z
i), S
Draw(Z
i) expression the pulling capacity value and the failure prediction value that obtain with neural network.
In the present embodiment, the genetic algorithm encoding that the floating core head parameter optimization is used adopts floating-point code, and the initial population value is 50, and the hybrid rate value is 0.8, aberration rate value 0.1, and maximum evolutionary generation value is 100.
6) adopt the CAD parameterization design method, with the optimal processing parameter that obtains in conjunction with CAD software carry out the three-roller planetary of mould rolling in the core print Die CAD parametrization design of rolling roll forming CAD, tube drawing with floating plug, designing and calculating, data processing and graphic plotting are carried out overall treatment.
The copper-alloy pipe-material casting-milling technology parameter design that the present invention set up and the operating process of optimization method are specially as shown in Figure 9:
(1) presses system prompt, product material type and light pipe dimensions that input will design;
(2) input level continuous casting copper pipes size;
(3) horizontal casting design: neural network calculated level casting temperature field crack initiation propensity value, and according to cooling system and continuous casting throwing system parameter with the genetic algorithm optimization horizontal casting, if what obtain is the final optimization pass parameter, execution in step (5) then, otherwise return step (3);
(4) input three-roller planetary rolling copper pipe size;
(5) the rolling design of three-roller planetary: calculate roll-force and roll forming defective value with neural network; According to using the genetic algorithm optimization rolling parameter; If what obtain is the final optimization pass parameter, execution in step (6) then, otherwise return step (5);
(6) the rolling roll forming of the roll CAD parametrization design of three-roller planetary;
(7) tube drawing with floating plug is joined the mould design: carry out drawing with neural network and join the mould design;
(8) calculate every time pulling capacity and drawing forming defective value of drawing with neural network; And with Genetic Algorithm optimized design tube drawing with floating plug parameter;
(9) the CAD parametrization of tube drawing with floating plug design; If what obtain is the final optimization pass parameter, termination routine then, otherwise return step (8).
In sum, copper-alloy pipe-material casting-milling technology parameter design that the present invention set up and the method for optimizing, with neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization apply in technological design and the parameter optimization with database technology, give full play on the one hand experience in the past, on the other hand with technology such as intellectual technology and finite element in conjunction with, realize process parameter optimizing design efficiently.Shortcomings such as method for designing is single have been overcome in the past like this.
Process the reinforcement that each operation is understood along with the improvement of the accumulation of experience, finite element simulation with to copper pipe, will further promote the validity of this system.
The present invention has remedied the deficiency of traditional design method in the past with neural network, finite element analogy, genetic algorithm, cad technique and database techniques.Adopt finite element software that forming process is simulated, can accurately reflect tubing actual production process, Products Development, development and processing are instructed and predicted.Utilize height nonlinear fitting character that neural network has that finite element analogy input parameter and corresponding simulating result and empirical data are learnt, train, threshold values that obtains after the training and weight matrix can shine upon the relation of technological parameter and technic index as default rule.Overcome in the past that inference machine obtains the bottleneck problem and the reasoning shot array problem of knowledge, and be easy to upgrade.To reach the technic index of expectation, solved the problem that optimization method in the past easily is absorbed in local extremum with the optimum technological parameter of the global optimizing characteristic of genetic algorithm search.The copper pipe production data storehouse of setting up, the technical experience of production technology data and long-term accumulation is integrated, be used to instruct technological design.According to design result, carry out the parametrization design in conjunction with CAD software, generate die drawing automatically.This system is used for the copper alloy tube horizontal casting, three-roller planetary is rolling and the technological design of tube drawing with floating plug production run, solves the various practical problemss in the copper pipe process, makes the processing technology of accurate standard.
Claims (5)
1. copper-alloy pipe-material casting-milling technology parameter design and the method optimized, it is characterized in that: be design basis with the database, neural network is the method for designing of technological parameter and technic index, genetic algorithm is the process parameter optimizing means, comprehensive integration neural network, genetic algorithm, finite element analogy, test design, the design of CAD parametrization and database technology design and optimize the technological parameter of copper-alloy pipe-material casting in technological design and parameter optimization; Specifically comprise:
1) database
Set up horizontal casting, three-roller planetary is rolling and the database of tube drawing with floating plug, as the basis of parameter designing and optimization, stores production data in the database, normal data and ephemeral data;
2) finite element analogy
Adopt the finite element numerical simulation step, obtain horizontal casting temperature field, thermograde and cooling velocity value in the horizontal casting respectively, and derive the crack initiation propensity value; Roll-force evaluation and roll forming defects simulation result during three-roller planetary is rolling; And pulling capacity evaluation in the tube drawing with floating plug and drawing forming defects simulation result;
3) neural network
Learn finite element numerical simulation result of calculation by multi-layer artificial neural network, calculate technic index value in the pairing horizontal continuous casting process of technological parameter, three-roller planetary rolling mill practice and the tube drawing with floating plug technology of copper-alloy pipe-material casting under different condition with the neural network after the training;
By training and learn, join the mould design proposal with what the neural network after the training obtained required specification to join the mould scheme from all size drawing of on-site collection;
4) genetic algorithm
The rolling parameter according to continuous casting cooling system in the horizontal casting and continuous casting throwing system parameter, three-roller planetary in rolling and the drawing Parameter Optimization target of tube drawing with floating plug are carried out parameter coding, constitute the initialization population; Calculate each individual technic index value by neural network, just the fitness value in the genetic algorithm carries out the operation operator operation again; The generations of evolution of population up to searching optimum solution, is determined optimum process parameters, and the result is made process chart and design document;
5) avatars
Adopt the CAD parameterization design method, with the optimal processing parameter that obtains in conjunction with CAD software carry out the three-roller planetary of mould rolling in the core print Die CAD parametrization design of rolling roll forming CAD, tube drawing with floating plug, with designing and calculating, data processing and graphic plotting carry out overall treatment.
2. according to the method for described copper-alloy pipe-material casting-milling technology parameter design of claim 1 and optimization, it is characterized in that: described finite element numerical simulation adopts the method for uniform experiment design.
3. according to design of the described copper-alloy pipe-material casting-milling technology parameter of claim 1 and the method optimized, it is characterized in that: if the finite element analogy result who increases newly is arranged, normalized, be integrated into original training sample file, again after the training, neural network threshold values and the weight matrix that newly obtains deposited in the database as neural network matrix file.
4. according to the method for described copper-alloy pipe-material casting-milling technology parameter design of claim 1 and optimization, it is characterized in that: described technological parameter comprises: the throwing system parameter of horizontal casting and cooling system parameter; Roll deflection angle, roller declination angle, opening degree and go-cart speed that three-roller planetary is rolling; Every time external mold cone angle, core print cone angle, drawing speed, external mold sizing section length and core print sizing section length of tube drawing with floating plug.
5. according to design of the described copper-alloy pipe-material casting-milling technology parameter of claim 1 and the method optimized, it is characterized in that: described neural network is the BP network based on the Levenberg-Marquardt optimized Algorithm, hidden layer is a Sigmoid type activation function, and output layer is selected Purelin type activation function for use.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100479036A CN100383796C (en) | 2005-12-02 | 2005-12-02 | Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100479036A CN100383796C (en) | 2005-12-02 | 2005-12-02 | Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1979496A CN1979496A (en) | 2007-06-13 |
| CN100383796C true CN100383796C (en) | 2008-04-23 |
Family
ID=38130659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100479036A Expired - Fee Related CN100383796C (en) | 2005-12-02 | 2005-12-02 | Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100383796C (en) |
Families Citing this family (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101201871B (en) * | 2007-12-18 | 2010-06-09 | 东北大学 | Rigid plasticity finite element method for estimation of roll-force in hot rolling process |
| CN101221416B (en) * | 2007-12-28 | 2010-12-15 | 东北大学 | Finite element method for on-line board temperature calculation in course of hot rolling |
| CN101414321B (en) * | 2008-11-14 | 2010-11-10 | 西安交通大学 | Design method for evaporation type cooler/condenser for chemical industry |
| CN101850410B (en) * | 2010-06-22 | 2012-06-20 | 攀钢集团钢铁钒钛股份有限公司 | Continuous casting breakout prediction method based on neural network |
| CN102172812B (en) * | 2010-12-31 | 2012-11-07 | 马志强 | Hollow aluminium division bar continuous moulding method |
| CN102393884B (en) * | 2011-10-12 | 2015-04-15 | 杭州电子科技大学 | Hot continuous rolling electromagnetic induction heating temperature prediction method based on BP (back-propagation) neural network |
| US9317626B2 (en) | 2011-11-16 | 2016-04-19 | Wai Man Chan | Method and system for combinatorial layout design |
| EP2633929A1 (en) | 2012-03-01 | 2013-09-04 | Siemens Aktiengesellschaft | Modelling of a casting-rolling assembly |
| CN102819651A (en) * | 2012-08-20 | 2012-12-12 | 西北工业大学 | Simulation-based parameter optimizing method for precise casting process of single crystal turbine blade |
| CN103042366B (en) * | 2013-01-14 | 2015-04-08 | 湖南金龙国际铜业有限公司 | Manufacturing technique for copper pipes for magnetrons |
| CN104408259B (en) * | 2014-12-03 | 2017-11-17 | 南通友星线束有限公司 | A kind of automatic drafting method of crimp blade figure |
| CN106295795A (en) * | 2016-08-09 | 2017-01-04 | 衢州学院 | A kind of copper pipe interior whorl forming qualitative forecasting method based on BP neural network algorithm |
| CN107784142B (en) * | 2016-08-31 | 2021-10-19 | 株洲中车时代半导体有限公司 | Thermal simulation method of semiconductor power component |
| CN106446413A (en) * | 2016-09-26 | 2017-02-22 | 北京无线电测量研究所 | Parameterization technological designing method for waveguide |
| CN108345749B (en) * | 2018-02-11 | 2021-06-15 | 中国电子科技集团公司第二十九研究所 | Modeling and packaging method for radio frequency integrated process tolerance and electrical performance coupling characteristics |
| CN108334712A (en) * | 2018-02-14 | 2018-07-27 | 武汉科技大学 | A kind of process parameter optimizing method of three-roller planetary rolling mill rolled bar material |
| CN109101712B (en) * | 2018-07-27 | 2023-06-20 | 石家庄创天电子科技有限公司 | Product model design system and method based on graph network |
| CN109241686A (en) * | 2018-11-02 | 2019-01-18 | 中国航空工业集团公司西安飞机设计研究所 | A kind of pull rod parameter optimization method for installing aircraft airborne equipment |
| CN110008636A (en) * | 2019-04-20 | 2019-07-12 | 福州大学 | A kind of computer-implemented method of thin-walled copper alloy water meter case casting technique |
| CN110457824B (en) * | 2019-08-13 | 2023-04-18 | 沈阳理工大学 | Method for calculating size of multi-pass drawn copper pipe blank based on tensile strength |
| CN111027240B (en) * | 2019-11-14 | 2023-05-09 | 哈尔滨工业大学（深圳） | Buried pipeline safety assessment method and related equipment |
| CN111061219B (en) * | 2019-12-16 | 2021-03-23 | 南京航空航天大学 | Method for rapidly determining forming process parameters |
| CN111681607B (en) * | 2020-08-17 | 2020-11-13 | 武汉精立电子技术有限公司 | Gamma adjusting method and system based on genetic algorithm |
| CN114798779B (en) * | 2022-03-23 | 2023-06-02 | 中南大学 | Intelligent metal drawing processing device and method based on neural network |
| CN115330091A (en) * | 2022-10-12 | 2022-11-11 | 邢台纳科诺尔精轧科技股份有限公司 | Precision rolling process optimization method based on artificial neural network and genetic algorithm |
| CN117350598B (en) * | 2023-11-16 | 2024-04-02 | 张家港广大特材股份有限公司 | Gear steel process control method and system |
| CN117399435B (en) * | 2023-12-11 | 2024-03-15 | 常州润来科技有限公司 | Self-adaptive compensation adjustment method and system for copper pipe rolling process |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1190518A (en) * | 1997-09-18 | 1999-04-06 | Nkk Corp | Method for controlling continuous hot rolling mill |
| JP2003266157A (en) * | 2002-03-14 | 2003-09-24 | Furukawa Electric Co Ltd:The | Method for manufacturing low oxygen copper wire rod with belt and wheel type continuous casting and rolling method |
| CN1589986A (en) * | 2003-08-29 | 2005-03-09 | 东北大学 | Automatic controlling technical parameter optimization method of metal plate rolling |
-
2005
- 2005-12-02 CN CNB2005100479036A patent/CN100383796C/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH1190518A (en) * | 1997-09-18 | 1999-04-06 | Nkk Corp | Method for controlling continuous hot rolling mill |
| JP2003266157A (en) * | 2002-03-14 | 2003-09-24 | Furukawa Electric Co Ltd:The | Method for manufacturing low oxygen copper wire rod with belt and wheel type continuous casting and rolling method |
| CN1589986A (en) * | 2003-08-29 | 2005-03-09 | 东北大学 | Automatic controlling technical parameter optimization method of metal plate rolling |
Non-Patent Citations (1)
| Title |
|---|
| 铜合金管坯旋轧成形的三维热力耦合有限元模拟. 李冰，李章刚，张士宏，刘化民，张光亮，张海渠，张金利.科学技术与工程，第5卷第17期. 2005 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1979496A (en) | 2007-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100383796C (en) | Copper-alloy pipe-material casting-milling technology parameter designing and optimizing method | |
| US11755976B2 (en) | Method for predicting oxygen load in iron and steel enterprises based on production plan | |
| CN103226741B (en) | Public supply mains tube explosion prediction method | |
| CN109902861A (en) | A kind of order manufacturing schedule real-time predicting method based on the double-deck transfer learning | |
| CN109647899B (en) | Method for forecasting power consumption of multi-specification rolled pieces in hot rolling and finish rolling process of strip steel | |
| CN103942422B (en) | Granular-computation-based long-term prediction method for converter gas holder positions in metallurgy industry | |
| CN106295795A (en) | A kind of copper pipe interior whorl forming qualitative forecasting method based on BP neural network algorithm | |
| CN102880908B (en) | Method for calculating remanufactured part environmental loss based on back propagation (BP) neural network | |
| CN110287583A (en) | Industrial equipment method for predicting residual useful life based on Recognition with Recurrent Neural Network | |
| CN111651929A (en) | Multi-objective optimization method based on fusion of Dynaform and intelligent algorithm | |
| CN105975701A (en) | Parallel scheduling disassembly path forming method based on mixing fuzzy model | |
| Liu et al. | [Retracted] Evaluation of Mechanical Properties of Materials Based on Genetic Algorithm Optimizing BP Neural Network | |
| Ciancio et al. | Design of a high performance predictive tool for forging operation | |
| CN114418177A (en) | New product material distribution prediction method for digital twin workshop based on generation of countermeasure network | |
| CN113705897A (en) | Product quality prediction method and system for industrial copper foil production | |
| CN106326677A (en) | Soft measurement method of acetic acid consumption in PTA device | |
| CN116757059A (en) | Carbon accounting method, device and related medium based on product type and process | |
| CN110738307A (en) | Application of BP neural network model in prediction of water quality of peach forest mouth reservoir | |
| Butdee | Adaptive aluminum extrusion die design using case-based reasoning and artificial neural networks | |
| Wu et al. | Optimal shape design of an extrusion–forging die using a polynomial network and a genetic algorithm | |
| CN102479259A (en) | Optimization of laser strengthening process for mold surface | |
| CN112100909A (en) | Parallel configurable intelligent optimization method based on collaborative optimization strategy | |
| Wu et al. | Prediction of mechanical properties of hot rolled strips by BP artificial neural network | |
| Liu et al. | Process optimal control of sheet metal forming springback based on evolutionary strategy | |
| Zhang et al. | Optimizing casting parameters of ingot based on neural network and genetic algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20080423 Termination date: 20111202 |