CN106756604B - Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof - Google Patents
Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof Download PDFInfo
- Publication number
- CN106756604B CN106756604B CN201611022159.9A CN201611022159A CN106756604B CN 106756604 B CN106756604 B CN 106756604B CN 201611022159 A CN201611022159 A CN 201611022159A CN 106756604 B CN106756604 B CN 106756604B
- Authority
- CN
- China
- Prior art keywords
- stainless steel
- svr
- improved corrosion
- nonstandard
- models
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- C—CHEMISTRY; METALLURGY
- C22—METALLURGY; FERROUS OR NON-FERROUS ALLOYS; TREATMENT OF ALLOYS OR NON-FERROUS METALS
- C22C—ALLOYS
- C22C38/00—Ferrous alloys, e.g. steel alloys
- C22C38/18—Ferrous alloys, e.g. steel alloys containing chromium
- C22C38/40—Ferrous alloys, e.g. steel alloys containing chromium with nickel
- C22C38/44—Ferrous alloys, e.g. steel alloys containing chromium with nickel with molybdenum or tungsten
-
- C—CHEMISTRY; METALLURGY
- C22—METALLURGY; FERROUS OR NON-FERROUS ALLOYS; TREATMENT OF ALLOYS OR NON-FERROUS METALS
- C22C—ALLOYS
- C22C38/00—Ferrous alloys, e.g. steel alloys
- C22C38/001—Ferrous alloys, e.g. steel alloys containing N
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
Abstract
The invention discloses nonstandard stainless composition of steel of a kind of improved corrosion based on SVR and preparation method thereof, it is supported vector regression (SVR) parameter optimization based on population (PSO), establish a kind of effective SVR models of the new stainless steel pitting potential based on stainless steel formula, and a kind of optimization formula of the nonstandard stainless steel of improved corrosion is predicted, component is by Cr 22~26%, Mo 2.9~3.3%, N 0.28~0.36%, Fe 60.31~64.8%, C in mass ratio<0.03% composition.The optimization formula that SVR models of the present invention provide, good reference can be provided for novel improved corrosion stainless steel research staff and industrial production, scientific guidance can be provided to improve steel corrosion resistance, reduction research and development test number and shortening R＆D cycle, to save a large amount of human and material resources, financial resources and time.
Description
Technical field
The invention belongs to alloying technology field more particularly to a kind of nonstandard stainless steel of improved corrosion based on PSO-SVR at
Point and preparation method thereof.
Background technology
With development of social progress, mineral resources are fewer and fewer.Stainless steel is as a kind of important basis in production and living
Material, application field are very extensive.Requirement of the people to stainless steel is higher and higher, under different environmental conditions using anti-
The different stainless steel material of corrosivity.Spot corrosion (pitting) is that the corrosion aperture developed in depth and breadth occur at metal surface position,
Leeway area does not corrode or corrodes slightly, and this etch state is called pitting or pitting corrosion spot corrosion.By taking steel as an example：It is stainless
The swift and violent increase of steel surface small " rust hole ", is to cause the main reason for stainless steel is by extensive corrosion.Currently, in the industry
Its resistance to corrosion usually can be indicated according to the spot corrosion equivalent value (PREN) of stainless steel material.For certain alloy,
PREN16Higher, resistance to spot corrosion is better.For example, the spot corrosion equivalent PREN of stainless steel16When value is more than 32, it is considered to be sea water resistance is rotten
Corrosion material；When two phase stainless steel PREN values are higher than 40, so that it may be suitable for H2Among the adverse circumstances such as S.But stainless steel exists
EN10088-1:It is always Standardization Order in 2005 and ASTM G48-11.People are in research and development new type stainless steel. corrosion resistance steel process
Middle most by empirical equation progress material chemical composition proportion design, this can undoubtedly cause the waste to resource and R＆D costs
Increase.Therefore how to search out that a kind of corrosion resistance is more preferable, stronger stainless steel material just seems particularly significant.Study people
Member payes attention to exploring the corrosion mechanism of stainless steel, and finds effective model to obtain the high material of corrosion resistance.All this
The not too many knowwhy of a little methods provides support, and result is also unsatisfactory, so having to look for a kind of new, more preferable
Method for stainless steel material chemical composition proportioning better theoretical direction is provided.
Researcher is added by element and is tested, and changes the pitting corrosion resistant performance of stainless steel, then by empirical equation/linear/
The methods of non-linear/ANN carries out the pitting potential of steel with addition each element component the modeling of indirectly/direct relation, to
Reach the pitting corrosion resistant performance for improving stainless steel, obtains the purpose of novel high stainless steel resisting spot corrosion.But often exist：
(1) accuracy is not high, stability is not strong；Training speed is slow；Learning effect is poor；Easily converge at local optimum
Deng；
(2) the problems such as small sample, dimension disaster, local minimum, not can solve；
(3) it is not easy to obtain optimal component and maximum spot corrosion potential value, each factor pair can not be understood and grasped well
The affecting laws of pitting potential；
(4) accurate theoretical reference cannot be provided for research staff, be easy to cause resource and waste of time, reduce and grind
Efficiency is sent out, R＆D costs are increased.
Invention content
The purpose of the present invention is to provide nonstandard stainless steels of a kind of improved corrosion and preparation method thereof, it is intended to solve background skill
The problem of art refers to.
The invention is realized in this way a kind of nonstandard stainless steel of improved corrosion, the nonstandard stainless steel of the improved corrosion presses quality
Than component by Cr 22~26%, Mo 2.9~3.3%, N 0.28~0.36%, Fe 60.31~64.8%, C<0.03% group
At.
Another object of the present invention is to provide a kind of preparation method of the nonstandard stainless steel of improved corrosion, the improved corrosion is non-
Mark stainless steel preparation method be to be combined support vector regression (SVR) with population optimizing (PSO), establish it is a kind of newly have
The SVR models of effect, and the thus optimal component of the nonstandard stainless steel of model prediction improved corrosion and corresponding maximum pitting potential, tool
Body includes the following steps：
Change element Cr, Mo, N, Fe, C content, being prepared using the methods of electric furnace, vaccum sensitive stove, there are different elements to contain
Several stainless steel samples of amount；
The pitting potential value of each sample is measured according to national standard electrochemical method, acquisition each sample Cr, Mo, N, Fe, C element contain
The pitting potential related data of amount and counter sample utilizes acquired training sample experimental data to build pitting potential and sample
SVR models between constituent content；
The accuracy or reliability of built SVR models are assessed using test samples data, analyzed, if built SVR
The average absolute percent error of model prediction reaches real requirement, then the model is reliable, otherwise changes training sample, re -training
To obtain new SVR models, the test samples data for having neither part nor lot in modeling training are recycled to test built SVR models, directly
Until the average absolute percent error of built SVR models reaches real requirement, SVR models at this time are optimal models；
Using the SVR models of above-mentioned optimization, changes independent variable (i.e. stainless steel each element percentage composition) value, swept by lattice point
It retouches, obtains each component content possessed by the nonstandard stainless steel of improved corrosion when there is highest pitting potential.
The appropriate model of establishing is：
(1) in formula, y is desired value (pitting potential), and l is supporting vector number, αi,For Lagrange multiplier, k (x,
xi) it is kernel function, b is deviation threshold, and x is sample independent variable (each composition quality percentage composition of stainless steel).
Each component content of the nonstandard stainless steel of improved corrosion when there is highest pitting potential is obtained using the SVR models of optimization
Afterwards, then carry out sample preparation, pitting potential measure etc. processes verified.
The SVR models of the optimization obtain each component content of the nonstandard stainless steel of improved corrosion using the SVR models of optimization,
The maximum spot corrosion potential value of the nonstandard stainless steel product of improved corrosion is obtained simultaneously.
Nonstandard stainless steel of a kind of improved corrosion provided by the invention and preparation method thereof, applies it to non-standard two-phase not
It becomes rusty in the corrosion resistance research of steel, using built SVR models, the pitting potential maximum value of stainless steel is obtained by dot interlace scanning
And corresponding optimization formula (optimization formula (%) Cr:22~26, Mo:2.9~3.3, N:0.28~0.36, Fe:60.31~
64.8、C:<0.03)；Under optimum formula, stainless steel pitting potential Ep>1200mV；Its spot corrosion equivalent exponential PREN16>35,
PREN30>40, seawater corrosion resistance material is can be used as, and can be used for containing H2In the adverse circumstances of S gases；The SVR models provide
Optimization formula, good reference can be provided for novel improved corrosion stainless steel research staff and industrial production, can be improve
Steel corrosion resistance, reduce research and development test number and shorten the R＆D cycle provide scientific guidance, to save a large amount of manpower,
Material resources, financial resources and time.
Description of the drawings
Fig. 1 is the preparation method flow chart of the nonstandard stainless steel of improved corrosion provided in an embodiment of the present invention.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
The application principle of the present invention is explained in detail below in conjunction with the accompanying drawings.
The nonstandard stainless steel of improved corrosion provided in an embodiment of the present invention, the nonstandard stainless steel of improved corrosion component in mass ratio
By Cr 22~26%, Mo 2.9~3.3%, N 0.28~0.36%, Fe 60.31~64.8%, C<0.03% composition.
As shown in Figure 1：The preparation method of the nonstandard stainless steel of improved corrosion provided in an embodiment of the present invention, the improved corrosion are non-
The preparation method for marking stainless steel is to be combined SVR with PSO, establishes a kind of new effective SVR models, and thus model prediction
The optimal component of the nonstandard stainless steel of improved corrosion and corresponding maximum pitting potential, specifically include following steps：
S101：Change element Cr, Mo, N, Fe, C content, being prepared using the methods of electric furnace, vaccum sensitive stove has different members
Several stainless steel samples of cellulose content；
S102：The pitting potential value of each sample, acquisition each sample Cr, Mo, N, Fe, C are measured according to national standard electrochemical method
The pitting potential related data of constituent content and counter sample utilizes acquired training sample experimental data to build pitting potential
SVR models between sample constituent content；
S103：The accuracy or reliability of built SVR models are assessed using test samples data, analyzed, if institute
The average absolute percent error for building SVR model predictions reaches real requirement, then the model is reliable, otherwise changes training sample, weight
New training recycles the test samples data for having neither part nor lot in modeling training to examine built SVR models to obtain new SVR models
It tests, until the average absolute percent error of built SVR models reaches real requirement, SVR models at this time are optimal mould
Type；
S104：Using the SVR models of above-mentioned optimization, change independent variable (i.e. stainless steel each element percentage composition) value, passes through
Dot interlace scanning obtains each component content possessed by the nonstandard stainless steel of improved corrosion when having highest pitting potential；
S105：Optimizing application SVR models prediction with highest pitting potential when the nonstandard stainless steel of improved corrosion
Had optimal component prepares sample and measures its pitting potential, if reaching design standard or requirement, you can promotes and applies.
The appropriate model of establishing is：
(2) in formula, y is desired value (pitting potential), and l is supporting vector number, αi,For Lagrange multiplier, k (x,
xi) it is kernel function, b is deviation threshold, and x is sample independent variable (each composition quality percentage composition of stainless steel).
Each component content of the nonstandard stainless steel of improved corrosion when there is highest pitting potential is obtained using the SVR models of optimization
Afterwards, then carry out sample preparation, pitting potential measure etc. processes verified.
The SVR models of the optimization obtain each component content of the nonstandard stainless steel of improved corrosion using the SVR models of optimization,
The maximum spot corrosion potential value of the nonstandard stainless steel product of improved corrosion is obtained simultaneously.
Obtain after the nonstandard stainless steel product of improved corrosion concrete application in actual production.
The computation model that the present invention is built is based on 36 with different Cr, Mo, N, Fe, C content and different pitting potentials
A stainless steel sample carries out SVR training modelings using 33 samples therein, and in addition 3 samples are as test sample.From training
From the point of view of test error, average absolute percent error (MAPE) all very littles of built SVR models, the MAPE of wherein training sample is
0.91%, most of result of calculation and experiment value are fairly close, or even have the error of 30 groups of training samples to tend to 0, are significantly better than
The MAPE (12.45%) of NRM models；In 3 test samples, the MAPE for the SVR models that the present invention is built is 16.39%, and NRM
The MAPE of model is 53.31%；The Statistical Comparison of the MAPE of SVR models and NRM model totality can be obtained, the meter that the present invention is built
Calculating model has quite high accuracy, this illustrates that built SVR models are reliable and effective；It can be with using built SVR models
It calculates and obtains Cr, Mo, N, Fe, C element content to the reciprocal effect rule of stainless steel pitting potential；Utilize built SVR models pair
Stainless steel pitting potential calculates the sensitivity analysis of each element, it is found that stainless steel pitting potential is most sensitive to Cr constituent contents,
And the content of N element is showed blunt at the latest.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
All any modification, equivalent and improvement etc., should all be included in the protection scope of the present invention made by within refreshing and principle.
Claims (3)
1. a kind of nonstandard stainless steel of improved corrosion, which is characterized in that the nonstandard stainless steel of improved corrosion component in mass ratio is by Cr
22~26%, Mo 2.9~3.3%, N 0.28~0.36%, Fe 60.31~64.8%, C<0.03% composition；
The preparation method of the nonstandard stainless steel of the improved corrosion is mutually to tie support vector regression (SVR) and population optimizing (PSO)
It closes, establishes a kind of effective SVR models of the new stainless steel pitting potential based on formula, and thus model prediction improved corrosion is non-
The optimal component of stainless steel and corresponding maximum pitting potential are marked, following steps are specifically included：
Change element Cr, Mo, N, Fe, C content, several with element different are prepared using electric furnace or vaccum sensitive stove
Stainless steel sample；
Measure the pitting potential value of each sample according to national standard electrochemical method, acquisition each sample Cr, Mo, N, Fe, C element content and
The pitting potential related data of counter sample utilizes acquired training sample experimental data to build pitting potential and sample element
SVR models between content；
The accuracy or reliability of built SVR models are assessed using test samples data, analyzed, if built SVR models
The average absolute percent error of prediction reaches real requirement, then the model is reliable, otherwise changes training sample, re -training is to obtain
SVR models newly are obtained, recycle the test samples data for having neither part nor lot in modeling training to test built SVR models, Zhi Daosuo
Build SVR models average absolute percent error reach real requirement until, SVR models at this time are optimal models；
Using above-mentioned optimal SVR models, change argument value, by dot interlace scanning, obtains when there is highest pitting potential
Each component content possessed by the nonstandard stainless steel of improved corrosion；
The SVR models are：
In formula, y is desired value, i.e. pitting potential, and l is supporting vector number, αi,For Lagrange multiplier, k (x, xi) it is core
Function, b are deviation threshold, and x is sample independent variable, i.e. each composition quality percentage composition of stainless steel；
It is modeled using machine Learning Theory and method, training sample.
2. the preparation method of the nonstandard stainless steel of improved corrosion as described in claim 1, which is characterized in that the optimal SVR moulds
Type, after each component content of the nonstandard stainless steel of improved corrosion when there is highest pitting potential is obtained using optimal SVR models, then
Carry out sample preparation, pitting potential measurement is verified.
3. the preparation method of the nonstandard stainless steel of improved corrosion as described in claim 1, which is characterized in that the optimal SVR moulds
Type obtains each component content of the nonstandard stainless steel of improved corrosion using optimal SVR models, while it is nonstandard stainless to obtain improved corrosion
The maximum spot corrosion potential value of steel product.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611022159.9A CN106756604B (en) | 2016-11-16 | 2016-11-16 | Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611022159.9A CN106756604B (en) | 2016-11-16 | 2016-11-16 | Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106756604A CN106756604A (en) | 2017-05-31 |
| CN106756604B true CN106756604B (en) | 2018-07-31 |
Family
ID=58969671
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611022159.9A Active CN106756604B (en) | 2016-11-16 | 2016-11-16 | Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106756604B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902872B (en) * | 2019-02-28 | 2021-10-01 | 北京首钢股份有限公司 | Method and device for combining performance prediction model with product research and development |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH02153044A (en) * | 1988-12-02 | 1990-06-12 | Kubota Ltd | Duplex stainless steel having high cracking resistance |
| US7807028B2 (en) * | 2005-03-09 | 2010-10-05 | Xstrata Queensland Limited | Stainless steel electrolytic plates |
| CN102383058A (en) * | 2011-10-31 | 2012-03-21 | 江苏亚盛金属制品有限公司 | High-nitrogen low-nickel super two-phase stainless steel |
| CN103382540A (en) * | 2012-05-02 | 2013-11-06 | 由国峰 | Antifatigue stainless steel wire preparation method |
-
2016
- 2016-11-16 CN CN201611022159.9A patent/CN106756604B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106756604A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sun et al. | Water resource utilization efficiency and spatial spillover effects in China | |
| Blanco et al. | EIS study of passivation of austenitic and duplex stainless steels reinforcements in simulated pore solutions | |
| Pargar et al. | Determination of chloride content in cementitious materials: from fundamental aspects to application of Ag/AgCl chloride sensors | |
| Du et al. | Does international trade promote CO2 emission performance? An empirical analysis based on a partially linear functional-coefficient panel data model | |
| Karthick et al. | Long-term relative performance of embedded sensor and surface mounted electrode for corrosion monitoring of steel in concrete structures | |
| Angst et al. | Beyond the chloride threshold concept for predicting corrosion of steel in concrete | |
| Chandwani et al. | Applications of artificial neural networks in modeling compressive strength of concrete: a state of the art review | |
| Xu et al. | Comparing monofractal and multifractal analysis of corrosion damage evolution in reinforcing bars | |
| Vélez et al. | Electrochemical characterization of early corrosion in prestressed concrete exposed to salt water | |
| CN106756604B (en) | Nonstandard stainless steel of a kind of improved corrosion based on PSO-SVR and preparation method thereof | |
| Song et al. | Identifying characteristic changes in club convergence of China's urban pollution emission: A spatial-temporal feature analysis | |
| Chen et al. | Port cargo throughput forecasting based on combination model | |
| CN114861496B (en) | Concrete mesoscopic structure chloride ion erosion model based on cold region natural environment effect | |
| Khan et al. | Towards sustainable construction: machine learning based predictive models for strength and durability characteristics of blended cement concrete | |
| Méndez-Figueroa et al. | Neural networks and correlation analysis to improve the corrosion prediction of SiO2-nanostructured patinated bronze in marine atmospheres | |
| CN103353296A (en) | Cold pressing head deformation quantity predicating method based on deformation-induced martensite phase transformation | |
| Yin et al. | Prediction of the vertical vibration of ship hull based on grey relational analysis and SVM method | |
| Ogunsanya et al. | The critical chloride concentration of austenitic and duplex stainless steel reinforcing bars | |
| Zhang et al. | Electrochemical noise analysis on the crevice corrosion behavior of Ni–Cr–Mo–V high strength steel using recurrence plots | |
| Al-Martini et al. | Genetic algorithm rheological equations for cement paste | |
| Dauji | Prediction of concrete spall damage under blast: Neural approach with synthetic data | |
| Qiuhua et al. | Use of principal component scores in multiple linear regression models for simulation of chlorophyll-a and phytoplankton abundance at a karst deep reservoir, southwest of China | |
| CN109344558A (en) | A kind of matter-element model evaluation method of the reinforced concrete corrosion based on ideal point method | |
| Liao et al. | Investigation of chloride diffusion in cement mortar via statistical learning theory | |
| Dhanpat et al. | Estimation of the effect of bio-admixtures on concrete workability using linear regression and support vector machines |
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 |