CN108897925A - A kind of casting technological parameter optimization method based on casting defect prediction model - Google Patents
A kind of casting technological parameter optimization method based on casting defect prediction model Download PDFInfo
- Publication number
- CN108897925A CN108897925A CN201810592731.8A CN201810592731A CN108897925A CN 108897925 A CN108897925 A CN 108897925A CN 201810592731 A CN201810592731 A CN 201810592731A CN 108897925 A CN108897925 A CN 108897925A
- Authority
- CN
- China
- Prior art keywords
- casting
- data
- defect
- model
- prediction model
- 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.)
- Granted
Links
Classifications
-
- 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
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30116—Casting
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- General Factory Administration (AREA)
Abstract
The invention belongs to casting process optimization correlative technology fields, and disclose a kind of casting technological parameter optimization method based on casting defect prediction model, including:The defect information for reflecting its technological level is automatically extracted for the casting as test object, these defects are then distinguished into qualitative classification and quantitative management;It acquires and stores the data packet being made of creation data, process data and detection data three parts multi-source data;Based on above data packet, is constructed using deep neural network and optimize casting defect prediction model;Orthogonal test is designed by combination of process parameters input model, the insensitive mechanisms that research key parameter device to hole pine defect develops obtain key process parameter and optimize window.Through the invention, the intelligent predicting process of casting defect can be not only executed, but also can the multi-source data in entire casting technique be contacted and be used, optimizing result is accordingly obtained in a manner of higher precision, and then improve final casting quality.
Description
Technical field
The invention belongs to casting process optimization correlative technology fields, are predicted more particularly, to one kind based on casting defect
The casting technological parameter optimization method of model.
Background technique
Casting is used as a kind of widely used metal heat processing technique, and raw material sources are wide, adaptable, product scope
It is related to numerous areas, such as electronics, chemical industry, medical instrument and weapons, aviation etc., has important shadow to human production life
It rings.The level of casting quality and casting technique has close ties, and the quality of technological level directly determines the quality safety of casting.
Generally speaking, casting technique can be divided into three root phases, i.e., at casting metal preparation stage, casting mold preparation stage and casting
In the reason stage, since the process being related in whole process is numerous, supplemental characteristic is complicated, accordingly largely affects process water
Flat raising.
Traditional manufacture experience of the process improving method majority based on technologists, time-consuming and personnel for the process
Dependence is strong, and fitness is not high.With universal and science and technology the progress of the ideas such as " industry 4.0 " and " two change fusion ", casting
Enterprise is made also gradually to the development of information-based and intelligent direction, starts to carry out process optimization by non-experimental technology means.For example,
Casting defect prediction is carried out in advance using casting simulation software, carries out process optimization further according to prediction result adjusting parameter.But it should
There is also calculate the problems such as time is longer, technological parameter is simplified and is affected by mathematical model for method.In addition, based on casting
The support of resource management system, the data accumulated under mass production environment by arrangement, analysis and training, utilizes certain hand
Section can also realize that some processes improve function.Some relevant solutions have been proposed in the prior art.For example,
CN201710684770.6 discloses a kind of method using BP neural network prediction TC4 titanium alloy casting shrinkage cavity defect, wherein
By the way that BP neural network model to be applied to the prediction process of casting shrinkage cavity position, it is tired to solve titanium alloy casting shrinkage cavity detection
The problems such as difficult and casting pattern software prediction result accuracy rate is insufficient.However, further investigations have shown that, above-mentioned existing side
Case still has defect or deficiency below:Firstly, the program fails to have more influence factors in entire foundry technology process
It accounts for machine, can only often determine the defect of shrinkage cavity type, the synthesis of casting cannot be accurately reflected comprehensively in actual work
Quality problems;Secondly, the BP neural network algorithm selected by it can not handle nonlinear problem, classification, knowledge to multi-source data
Not and processing capacity is inadequate, this equally influences the efficiency and precision of prediction result;Finally, effective utilization to prediction result
Still insufficient.Correspondingly, this field needs to make further improvement, preferably to meet modernization casting technique to higher
Efficiency and the higher-quality demand of casting.
Summary of the invention
For the above shortcoming and Improvement requirement of the prior art, the present invention provides one kind to be predicted based on casting defect
The casting technological parameter optimization method of model, wherein by Classification Management based on multi-source data and deep learning neural network come
The model for casting defect prediction is established, and subsequent optimization targetedly is carried out to model according to prediction result, accordingly not
The intelligent predicting process of casting defect can be only executed, and can be more scientificly by the multi-source data in entire casting technique
Contacted and used, and by higher precision and more targetedly in a manner of obtain optimizing result, and then significantly improve final
The casting quality of acquisition.
To achieve the above object, it is proposed, according to the invention, provide a kind of casting technique ginseng based on casting defect prediction model
Number optimization method, which is characterized in that this method includes the following steps:
(i) automatic defect classification and feature extraction
Casting is shot, its gray level image is obtained, and defect inspection is carried out to image, then will identify that and
Flaw discrimination be classified as being mingled with, shrinkage cavity and stomata three categories;Then, according to image internal reference object and pixel size, quantization is each
Position, shape and the size of a defect characteristic;
(ii) excavation and pretreatment of data packet
The data packet being made of following three parts multi-source data is acquired and stores, wherein first part's multi-source data belongs to life
Data are produced, and include penetrating wax temperature, penetrate wax pressure, the dwell time, formwork initial temperature, formwork thickness, do in casting process
The ginseng such as dry humidity, dewaxing temperature, maturing temperature, calcining time, smelting temperature, smelting time, pouring temperature, poring rate
Number;Second part multi-source data belongs to process data, and includes material composition, casting thickness, process parameter etc.;Part III is more
Source data belongs to detection data, and includes the corresponding data obtained after quantifying to the defect characteristic;Then, it will count above
According to extracting and form corresponding relationship with the number of casting;
(iii) foundation of casting defect prediction model
Based on above step (ii) data packet obtained, constructs the casting defect based on deep neural network and predict mould
Type, wherein using the creation data as the input of the model, the detection data as the output of the model, and in model structure
It is trained and optimizes after the completion of building;
(iv) acquisition of process parameter optimizing window
Orthogonal test is carried out to obtain multiple parameters combination to the technological parameter of proposed adoption, and is built using step (iii)
Vertical model obtains according to optimizing result to execute optimizing and reflects that the supplemental characteristic of current best production technology optimizes window
Mouthful.
As it is further preferred that in step (i), it is preferred to use convolutional neural networks come execute described image identification and
The process that defect characteristic extracts;Wherein, the input of model is set as the casting grayscale image, the output of model is set as defect
Type, middle layer include 5 layers of convolutional layer, 3 layers of pond layer and 3 layers of full articulamentum;It can be used to defects detection after model training is good,
The quantification treatment of defect characteristic is carried out further according to image information.
As it is further preferred that the training process model preferably includes following process in step (iii):It will be first
Beginning data, which are input in the prediction model, obtains corresponding prediction result, this result is compared with casting actual production result
Compared with, and difference between the two is fed back into network adjustment weight and threshold value, and until meeting error range, number needed for thus completing
The training of amount.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, mainly have below
Technological merit:
1, entered by the multi-source data formed from creation data, process data accumulated in casting process and detection data
Hand can more fully combine the actual production status of equipment and casting quality detection case of foundry enterprise, be effectively used
Corresponding data carries out the training of prediction model, and then obtains high-precision casting process prediction result based on this model;
2, the present invention has further selected deep neural network to construct the prediction model of casting defect, compared to pure
Other parameters fitting or autoregression model, the model have better accuracy and robustness, conform better to casting technique
With reality;
3, in addition, the present invention is on the basis of prediction model obtained, in a certain range by casting technique key parameter
It is fluctuated, and key parameter is obtained to the sensitivity effects of casting flaw by design orthogonal test, accordingly realized more
High-precision process parameter optimizing window finally significantly improves the directiveness and applicability of entire process optimization.
Detailed description of the invention
Fig. 1 is the overall process schematic diagram according to casting technological parameter optimization method constructed by the present invention;
Fig. 2 be for more specifically illustrate casting quality testing result defect inspection according to the invention with
The logical schematic of characteristic quantification;
Fig. 3 is a specific embodiment according to the invention, for exemplary display casting defect predetermined depth nerve net
The schematic illustration of network model;
Fig. 4 is another specific embodiment according to the invention, establishes for exemplary display casting defect prediction model, is excellent
Change and parameter optimization flow chart.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
Not constituting a conflict with each other can be combined with each other.
Fig. 1 is the overall process schematic diagram according to casting technological parameter optimization method constructed by the present invention.Such as Fig. 1 institute
Show, which scientificlly and effectively can carry out technological parameter recommendation according to the mass historical data of actual production, form one
Process intelligent recommender system is covered, to improve production technology level, improves casting forming quality.In other words, key improvements
Place is, the data set of early period is interconnected and cast resource management system acquisition by equipment and image recognition is excavated,
Analysis and arrangement;By the training to mass data, intelligently casting defect can be predicted, it is true according to prediction result
The sensibility sequence of fixed each parameter, to control parameter value and obtain the recommendation of optimization, to actual production provide with reference to and
Guidance.Specific explanations explanation will be carried out one by one to these steps below.
Firstly, being automatic defect classification and characteristic extraction step.
In this step, for example casting is shot using non-destructive detecting device, obtains its gray level image, and to figure
As carrying out defect inspection, then will identify that the Flaw discrimination come be classified as being mingled with, shrinkage cavity and stomata three categories;Then,
According to image internal reference object and pixel size, quantify position, shape and the size of each defect.
More specifically, process flow is as shown in Fig. 2, existing casting quality testing result such as X-ray check can be based on
Instrument is taken a picture, and first to pretreatments such as the enhancing of gray level image degree of comparing, cuttings, is then sent into established deep learning figure
As identification CNN model is trained and optimizes.The defect identified is primarily based on its geometrical characteristic and carries out qualitative classification, specifically
Have be mingled with, shrinkage cavity (pine), stomata three categories, position, the shape of defect are calculated further according to image internal reference object and pixel size
And the information such as size, for example diameter and depth etc. are stored in database after drawbacks described above characteristic quantification.
It then, is the excavation and pre-treatment step of data packet.
The data packet being made of following three parts multi-source data is acquired and stores, wherein first part's multi-source data belongs to life
Data are produced, and include penetrating wax temperature, penetrate wax pressure, the dwell time, formwork initial temperature, formwork thickness, do in casting process
The ginseng such as dry humidity, dewaxing temperature, maturing temperature, calcining time, smelting temperature, smelting time, pouring temperature, poring rate
Number;Second part multi-source data belongs to process data, and includes material composition, casting thickness, process parameter etc.;Part III is more
Source data belongs to detection data, and includes the corresponding data obtained after quantifying to the defect characteristic;Then, it will count above
According to extracting and form corresponding relationship with the number of casting.
More specifically, data packet multi-source data consists of three parts in this patent:Creation data, process data and inspection
Measured data.Wherein creation data and detection data arrange input, output data set for model foundation and optimization respectively, and model is complete
Optimizing is carried out at later again process data is inputted;Creation data is acquired on the basis of equipment interconnects and is stored in database
It is interior, such as wax temperature, to penetrate wax pressure, dwell time, formwork initial temperature, formwork thickness, drying room dry for penetrating in hot investment casting
The parameters such as humidity, dewaxing temperature, maturing temperature, calcining time, smelting temperature, smelting time, pouring temperature, poring rate;Work
Skill data extract in casting resource management system, including material composition, casting thickness and specific process parameter value
Deng;Detection data is defect characteristic quantitative information, such as defects with diameters.By three parts data extract and with casting number pair
It answers, is managed using casting single-piece complete period administrative skill, form the Mechanical processing of casting flow data of complete set.
It again, is the establishment step of casting defect prediction model.
Based on above step data packet obtained, the casting defect prediction model based on deep neural network is constructed,
It is middle using the creation data as the input of the model, the detection data as the output of the model.It should be noted that deep
The basic principle and treatment process for spending neural network are known in the art, for example, can also be closed by setting in this model
The hidden layer and each layer of neuron of suitable quantity, while improving weight, biasing etc. and network other hyper parameters, in addition appropriate
Activation primitive, input vector pass through certain calculating in each layer, successively by being reduced further according to prediction result until output
Loss carrys out feedback adjustment network parameter even structure, to reach effect of optimization.Its specific detailed implementation procedure is herein no longer
It repeats one by one.
It in a specific embodiment, can be according to depth nerve according to the inputoutput data packet put in order above
The principle of network algorithm establishes deep neural network model, such as specifically may include input layer parameter more than 20, hides the number of plies
The failure prediction result of layer and output layer.Data set is divided to training set, test set and verifying collection in proportion, using Python
Language realizes network model, by repeatedly learning and training, determines network parameter (weight corresponding with minimal error
With threshold value etc.), model, which meets optimization precision training, to be stopped.The concrete principle and implementation procedure of the deep neural network algorithm be
It is known in the art, therefore details are not described herein.
Finally, being the obtaining step of process parameter optimizing window.
Orthogonal test is carried out to obtain multiple parameters combination to the technological parameter of proposed adoption, and is established using above-mentioned steps
Model execute optimizing, window is optimized according to the supplemental characteristic that optimizing result obtains the current best production technology of reflection.
More specifically, data distribution modeling can be carried out key process parameter, by designing orthogonal test, by parameter combination
It is input in trained bug prediction model, prediction result is assessed, analyzing defect prediction result is preferentially chosen and corresponded to
Parameter, generate technique recommend window.
To sum up, technical solution proposed by the invention compared with prior art, can improve previous artificial data management
Mode improves efficiency, save the cost.In particular, it can obtain actual technical effect at aspect in detail below:1, in casting
Product defects automatic identification and quantization are realized on the basis of quality measurements, improve recognition efficiency, use manpower and material resources sparingly cost;
2, can the process data based on the magnanimity creation data acquired in system and storage data are carried out in conjunction with practical condition
The management and use of science, reduce management difficulty, improve data value;3, it by establishing bug prediction model, utilizes
The data of extraction determine key parameter to the sensibility of defective effect, thus obtain optimization technological parameter recommend window, can
Conscientiously valuable reference is provided for actual production.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should all include
Within protection scope of the present invention.
Claims (3)
1. a kind of casting technological parameter optimization method based on casting defect prediction model, which is characterized in that under this method includes
Column step:
(i) automatic defect classification and feature extraction
Casting is shot, obtains its gray level image, and defect inspection is carried out to image, then will identify that that comes lacks
Fall into qualitative classification be mingled with, shrinkage cavity and stomata three categories;Then, according to image internal reference object and pixel size, quantify each lack
Fall into position, shape and the size of feature;
(ii) excavation and pretreatment of data packet
The data packet being made of following three parts multi-source data is acquired and stores, wherein first part's multi-source data belongs to production number
According to, and include penetrating wax temperature, penetrating wax pressure, dwell time, formwork initial temperature, formwork thickness, drying room in casting process
The parameters such as humidity, dewaxing temperature, maturing temperature, calcining time, smelting temperature, smelting time, pouring temperature, poring rate;
Second part multi-source data belongs to process data, and includes material composition, casting thickness, process parameter etc.;Part III multi-source
Data belong to detection data, and include the corresponding data obtained after quantifying to the defect characteristic;Then, by above data
It extracts and forms corresponding relationship with the number of casting;
(iii) foundation of casting defect prediction model
Based on above step (ii) data packet obtained, the casting defect prediction model based on deep neural network is constructed,
It is middle using the creation data as the input of the model, the detection data as the output of the model, and it is complete in model construction
It is trained and optimizes after;
(iv) acquisition of process parameter optimizing window
Orthogonal test is carried out to obtain multiple parameters combination to the technological parameter of proposed adoption, and established using step (iii)
Model obtains according to optimizing result to execute optimizing and reflects that the supplemental characteristic of current best production technology optimizes window.
2. the method as described in claim 1, which is characterized in that in step (i), it is preferred to use convolutional neural networks execute
The process that described image identification and defect characteristic extract;Wherein, the input of model is set as the casting grayscale image, model
Output is set as defect type, and middle layer includes 5 layers of convolutional layer, 3 layers of pond layer and 3 layers of full articulamentum;After model training is good i.e.
It can be used for defects detection, the quantification treatment of defect characteristic carried out further according to image information.
3. the method as described in claim 1, which is characterized in that in step (iii), the training process model is preferably included
Following process:Primary data is input in the prediction model and obtains corresponding prediction result, this result is practical raw with casting
It produces result to be compared, and difference between the two is fed back into network adjustment weight and threshold value, until meeting error range, by
This completes required amount of training.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810592731.8A CN108897925B (en) | 2018-06-11 | 2018-06-11 | Casting process parameter optimization method based on casting defect prediction model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810592731.8A CN108897925B (en) | 2018-06-11 | 2018-06-11 | Casting process parameter optimization method based on casting defect prediction model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108897925A true CN108897925A (en) | 2018-11-27 |
CN108897925B CN108897925B (en) | 2020-09-08 |
Family
ID=64344705
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810592731.8A Active CN108897925B (en) | 2018-06-11 | 2018-06-11 | Casting process parameter optimization method based on casting defect prediction model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108897925B (en) |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492940A (en) * | 2018-12-06 | 2019-03-19 | 华中科技大学 | A kind of casting furnace lodge follow-up of quality method for moulding pouring product line |
CN109543361A (en) * | 2019-01-22 | 2019-03-29 | 东北大学 | A kind of continuous casting billet pressure process shrinkage cavity closure prediction calculation method |
CN109829907A (en) * | 2019-01-31 | 2019-05-31 | 浙江工业大学 | A kind of metal shaft surface defect recognition method based on deep learning |
CN109834050A (en) * | 2019-01-08 | 2019-06-04 | 西安科技大学 | A kind of varistor appearance delection device |
CN110222681A (en) * | 2019-05-31 | 2019-09-10 | 华中科技大学 | A kind of casting defect recognition methods based on convolutional neural networks |
CN110263640A (en) * | 2019-05-16 | 2019-09-20 | 东软集团股份有限公司 | Slab shrinkage cavity recognition methods, device, storage medium and electronic equipment |
CN112017172A (en) * | 2020-08-31 | 2020-12-01 | 佛山科学技术学院 | System and method for detecting defects of deep learning product based on raspberry group |
CN112417583A (en) * | 2020-08-28 | 2021-02-26 | 上海交通大学 | Quantitative prediction method for loose defect of large complex thin-wall high-temperature alloy casting |
CN112427624A (en) * | 2020-11-04 | 2021-03-02 | 中国科学院重庆绿色智能技术研究院 | Casting and forging double-control forming system based on digital twinning and parameter optimization method |
CN112485260A (en) * | 2020-11-26 | 2021-03-12 | 常州微亿智造科技有限公司 | Workpiece defect detection method and device |
CN112950071A (en) * | 2021-03-30 | 2021-06-11 | 北京瑞莱智慧科技有限公司 | Training method of process parameter adjustment model, and process parameter adjustment method and device |
CN112985318A (en) * | 2019-12-17 | 2021-06-18 | 财团法人金属工业研究发展中心 | Method and system for on-line prediction of fastener size |
CN113486457A (en) * | 2021-06-04 | 2021-10-08 | 宁波海天金属成型设备有限公司 | Die casting defect prediction and diagnosis system |
CN113628211A (en) * | 2021-10-08 | 2021-11-09 | 深圳市信润富联数字科技有限公司 | Parameter prediction recommendation method, device and computer readable storage medium |
CN113642855A (en) * | 2021-07-26 | 2021-11-12 | 南京工业大学 | Knowledge model base-based method for optimizing differential pressure casting quality of automobile aluminum alloy steering knuckle |
CN114939643A (en) * | 2022-05-24 | 2022-08-26 | 昆山莱捷有色金属有限公司 | Die-casting control method and device |
CN115423218A (en) * | 2022-11-02 | 2022-12-02 | 江苏新恒基特种装备股份有限公司 | Prediction method and system for forge piece defect elimination process |
CN115494813A (en) * | 2022-11-16 | 2022-12-20 | 优铸科技(北京)有限公司 | Domain control method and domain controller for casting work part |
CN115615782A (en) * | 2022-11-10 | 2023-01-17 | 大连理工大学 | Evaluation and prediction method for fiber reinforced composite subsurface damage |
CN115625317A (en) * | 2022-12-21 | 2023-01-20 | 昆山市恒达精密机械工业有限公司 | Surface water wave optimization processing method and system for die-casting regulation |
CN115640933A (en) * | 2022-11-03 | 2023-01-24 | 昆山润石智能科技有限公司 | Method, device and equipment for automatically managing production line defects and storage medium |
CN116748352A (en) * | 2023-08-14 | 2023-09-15 | 江苏新恒基特种装备股份有限公司 | Metal pipe bending machine processing parameter monitoring control method, system and storage medium |
CN116822341A (en) * | 2023-06-12 | 2023-09-29 | 华中科技大学 | Defect prediction method and system based on three-dimensional casting model feature extraction |
CN117252486A (en) * | 2023-11-14 | 2023-12-19 | 长春师范大学 | Automobile part defect detection method and system based on Internet of things |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101853318A (en) * | 2010-05-11 | 2010-10-06 | 华中科技大学 | Method for quickly forecasting cooling time in injection molding of plastics |
JP2015030035A (en) * | 2013-08-07 | 2015-02-16 | 株式会社神戸製鋼所 | Determination method of plate crown prediction model |
CN104834794A (en) * | 2015-05-27 | 2015-08-12 | 中国人民解放军国防科学技术大学 | Improved method for prolonging product service life based on material ratio and key process parameter |
CN106055787A (en) * | 2016-05-30 | 2016-10-26 | 广西科技大学 | Automotive trim panel injection moulding technology based on BP (Back Propagation) neural network |
CN107577850A (en) * | 2017-08-11 | 2018-01-12 | 中国航发北京航空材料研究院 | Using the method for BP neural network prediction TC4 titanium alloy casting shrinkage cavity defects |
CN107609647A (en) * | 2017-10-16 | 2018-01-19 | 安徽工业大学 | One kind is based on BP neural network roll alloy mechanical property Forecasting Methodology |
-
2018
- 2018-06-11 CN CN201810592731.8A patent/CN108897925B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101853318A (en) * | 2010-05-11 | 2010-10-06 | 华中科技大学 | Method for quickly forecasting cooling time in injection molding of plastics |
JP2015030035A (en) * | 2013-08-07 | 2015-02-16 | 株式会社神戸製鋼所 | Determination method of plate crown prediction model |
CN104834794A (en) * | 2015-05-27 | 2015-08-12 | 中国人民解放军国防科学技术大学 | Improved method for prolonging product service life based on material ratio and key process parameter |
CN106055787A (en) * | 2016-05-30 | 2016-10-26 | 广西科技大学 | Automotive trim panel injection moulding technology based on BP (Back Propagation) neural network |
CN107577850A (en) * | 2017-08-11 | 2018-01-12 | 中国航发北京航空材料研究院 | Using the method for BP neural network prediction TC4 titanium alloy casting shrinkage cavity defects |
CN107609647A (en) * | 2017-10-16 | 2018-01-19 | 安徽工业大学 | One kind is based on BP neural network roll alloy mechanical property Forecasting Methodology |
Non-Patent Citations (2)
Title |
---|
刘剑龙: "《铝合金铸件检测中的X射线DR技术应用》", 《中国设备工程》 * |
张响: "《铝合金车轮数字化仿真及工艺优化》", 《中国博士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492940A (en) * | 2018-12-06 | 2019-03-19 | 华中科技大学 | A kind of casting furnace lodge follow-up of quality method for moulding pouring product line |
CN109834050A (en) * | 2019-01-08 | 2019-06-04 | 西安科技大学 | A kind of varistor appearance delection device |
CN109543361A (en) * | 2019-01-22 | 2019-03-29 | 东北大学 | A kind of continuous casting billet pressure process shrinkage cavity closure prediction calculation method |
CN109543361B (en) * | 2019-01-22 | 2023-05-16 | 东北大学 | Shrinkage cavity closure degree prediction calculation method in continuous casting billet pressing process |
CN109829907A (en) * | 2019-01-31 | 2019-05-31 | 浙江工业大学 | A kind of metal shaft surface defect recognition method based on deep learning |
CN110263640A (en) * | 2019-05-16 | 2019-09-20 | 东软集团股份有限公司 | Slab shrinkage cavity recognition methods, device, storage medium and electronic equipment |
CN110222681A (en) * | 2019-05-31 | 2019-09-10 | 华中科技大学 | A kind of casting defect recognition methods based on convolutional neural networks |
CN112985318A (en) * | 2019-12-17 | 2021-06-18 | 财团法人金属工业研究发展中心 | Method and system for on-line prediction of fastener size |
CN112985318B (en) * | 2019-12-17 | 2022-11-22 | 财团法人金属工业研究发展中心 | Method and system for on-line prediction of fastener size |
CN112417583A (en) * | 2020-08-28 | 2021-02-26 | 上海交通大学 | Quantitative prediction method for loose defect of large complex thin-wall high-temperature alloy casting |
CN112017172A (en) * | 2020-08-31 | 2020-12-01 | 佛山科学技术学院 | System and method for detecting defects of deep learning product based on raspberry group |
CN112427624A (en) * | 2020-11-04 | 2021-03-02 | 中国科学院重庆绿色智能技术研究院 | Casting and forging double-control forming system based on digital twinning and parameter optimization method |
CN112427624B (en) * | 2020-11-04 | 2021-11-09 | 中国科学院重庆绿色智能技术研究院 | Casting and forging double-control forming system based on digital twinning and parameter optimization method |
CN112485260A (en) * | 2020-11-26 | 2021-03-12 | 常州微亿智造科技有限公司 | Workpiece defect detection method and device |
CN112485260B (en) * | 2020-11-26 | 2023-01-03 | 常州微亿智造科技有限公司 | Workpiece defect detection method and device |
CN112950071A (en) * | 2021-03-30 | 2021-06-11 | 北京瑞莱智慧科技有限公司 | Training method of process parameter adjustment model, and process parameter adjustment method and device |
CN113486457A (en) * | 2021-06-04 | 2021-10-08 | 宁波海天金属成型设备有限公司 | Die casting defect prediction and diagnosis system |
CN113642855B (en) * | 2021-07-26 | 2023-10-10 | 南京工业大学 | Method for optimizing differential pressure casting quality of automobile aluminum alloy steering knuckle based on knowledge model base |
CN113642855A (en) * | 2021-07-26 | 2021-11-12 | 南京工业大学 | Knowledge model base-based method for optimizing differential pressure casting quality of automobile aluminum alloy steering knuckle |
CN113628211B (en) * | 2021-10-08 | 2022-02-15 | 深圳市信润富联数字科技有限公司 | Parameter prediction recommendation method, device and computer readable storage medium |
CN113628211A (en) * | 2021-10-08 | 2021-11-09 | 深圳市信润富联数字科技有限公司 | Parameter prediction recommendation method, device and computer readable storage medium |
CN114939643A (en) * | 2022-05-24 | 2022-08-26 | 昆山莱捷有色金属有限公司 | Die-casting control method and device |
CN115423218B (en) * | 2022-11-02 | 2023-03-24 | 江苏新恒基特种装备股份有限公司 | Prediction method and system for forge piece defect elimination process |
CN115423218A (en) * | 2022-11-02 | 2022-12-02 | 江苏新恒基特种装备股份有限公司 | Prediction method and system for forge piece defect elimination process |
CN115640933A (en) * | 2022-11-03 | 2023-01-24 | 昆山润石智能科技有限公司 | Method, device and equipment for automatically managing production line defects and storage medium |
CN115640933B (en) * | 2022-11-03 | 2023-10-13 | 昆山润石智能科技有限公司 | Automatic defect management method, device and equipment for production line and storage medium |
CN115615782A (en) * | 2022-11-10 | 2023-01-17 | 大连理工大学 | Evaluation and prediction method for fiber reinforced composite subsurface damage |
CN115494813A (en) * | 2022-11-16 | 2022-12-20 | 优铸科技(北京)有限公司 | Domain control method and domain controller for casting work part |
CN115625317A (en) * | 2022-12-21 | 2023-01-20 | 昆山市恒达精密机械工业有限公司 | Surface water wave optimization processing method and system for die-casting regulation |
CN116822341A (en) * | 2023-06-12 | 2023-09-29 | 华中科技大学 | Defect prediction method and system based on three-dimensional casting model feature extraction |
CN116748352A (en) * | 2023-08-14 | 2023-09-15 | 江苏新恒基特种装备股份有限公司 | Metal pipe bending machine processing parameter monitoring control method, system and storage medium |
CN116748352B (en) * | 2023-08-14 | 2023-11-07 | 江苏新恒基特种装备股份有限公司 | Metal pipe bending machine processing parameter monitoring control method, system and storage medium |
CN117252486A (en) * | 2023-11-14 | 2023-12-19 | 长春师范大学 | Automobile part defect detection method and system based on Internet of things |
CN117252486B (en) * | 2023-11-14 | 2024-02-02 | 长春师范大学 | Automobile part defect detection method and system based on Internet of things |
Also Published As
Publication number | Publication date |
---|---|
CN108897925B (en) | 2020-09-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108897925A (en) | A kind of casting technological parameter optimization method based on casting defect prediction model | |
CN110378252A (en) | A kind of distress in concrete recognition methods based on depth migration study | |
CN104134101B (en) | Low permeability reservoir Gas Productivity Forecasting Methodology | |
Manouchehrian et al. | Application of artificial neural networks and multivariate statistics to estimate UCS using textural characteristics | |
CN110222681A (en) | A kind of casting defect recognition methods based on convolutional neural networks | |
CN109214437A (en) | A kind of IVF-ET early pregnancy embryonic development forecasting system based on machine learning | |
CN111881495A (en) | Prestress steel structure safety assessment method based on digital twinning | |
CN107169205B (en) | Iron ore classification modeling method | |
CN107133400A (en) | A kind of aircraft structure fatigue reliability Bayes's combination forecasting method | |
CN104809311A (en) | Structural part remaining life predicting method based on multi-factor fusion correction | |
CN112948932A (en) | Surrounding rock grade prediction method based on TSP forecast data and XGboost algorithm | |
CN115879654B (en) | Underground roadway stability prediction and reliability evaluation method | |
CN112614021A (en) | Tunnel surrounding rock geological information prediction method based on built tunnel information intelligent identification | |
CN113918538A (en) | Newly-built road maintenance data migration system based on artificial neural network | |
CN110489844B (en) | Prediction method suitable for uneven large deformation grade of soft rock tunnel | |
CN106568647A (en) | Nerve network-based concrete strength predication method | |
CN109920489A (en) | It is a kind of that model and method for building up are hydrocracked based on Lasso-CCF-CNN | |
CN112947080B (en) | Scene parameter transformation-based intelligent decision model performance evaluation system | |
CN116128165A (en) | MIV-BP-based building element quality prediction method and system | |
CN115565669B (en) | Cancer survival analysis method based on GAN and multitask learning | |
CN115526114B (en) | Hydraulic fracture morphology inversion method and system based on ensemble Kalman filtering method | |
Choudhury et al. | Prediction of parallel clay cracks using neural networks–a feasibility study | |
CN116340770A (en) | Intelligent identification method for tunnel lining geological radar data | |
CN116721720A (en) | Composite structure reliability digital twin parameter updating process and method | |
CN106096274A (en) | A kind of sectional linear fitting health evaluating method based on normal distribution |
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 |