CN109086547A - A kind of ageing of metal level measurement method - Google Patents
A kind of ageing of metal level measurement method Download PDFInfo
- Publication number
- CN109086547A CN109086547A CN201810963846.3A CN201810963846A CN109086547A CN 109086547 A CN109086547 A CN 109086547A CN 201810963846 A CN201810963846 A CN 201810963846A CN 109086547 A CN109086547 A CN 109086547A
- Authority
- CN
- China
- Prior art keywords
- ageing
- calibration
- signal
- sample
- metal
- 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.)
- Pending
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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/71—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light thermally excited
- G01N21/718—Laser microanalysis, i.e. with formation of sample plasma
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Optics & Photonics (AREA)
- Plasma & Fusion (AREA)
- Computer Hardware Design (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Geometry (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses a kind of ageing of metal level measurement methods, comprising steps of S1, the multiple calibration samples of selection, pass through LIBS system measurement and acquire plasma emission spectroscopy data；S2, calibration disaggregated model is constructed using Recursive Support Vector Machine algorithm and carries out feature selecting, determine that ageing of metal grade calibration disaggregated model is established in the merging of important feature group；S3, pass through LIBS system measurement and acquire the plasma emission spectroscopy data of metal sample to be measured；S4, the important feature combination determined according to S2 are extracted the characteristic signal of pre- sample, and are input in the disaggregated model of optimization, and the aging grade forecast value of metal sample is obtained.The present invention can be achieved the screening of characteristic variable and obtain corresponding calibration model, improve accuracy and accuracy that LIBS technology is applied to ageing of metal level measurement, and can promote the use of the real-time measurement assessment of the heat power equipment operating status of each industrial circle.
Description
Technical field
The present invention relates to ageing of metal level measurement fields, in particular to a kind of to be eliminated based on wavelet transformation and recursive feature
Achievable ageing of metal grade rapid survey method.
Background technique
Metal tube is as the large-scale warm of key that heat proof material is widely used in the industries such as electric power, metallurgy, chemical industry and machinery
In power equipment.Metal material performance in the environment of high temperature and pressure for a long time be on active service can change, it is microcosmic on be mainly reflected in
The distribution of microstructure changes, such as the precipitation of carbide and the variation of lattice size etc. in lattice；Macroscopically show material
Mechanical performance, i.e. the performances such as its hardness and tensile strength can be declined, and final material performance is unsatisfactory for requirement and leads
Cause cracks or even occurs pipe explosion accident, directly affects the operation and benefit of entire industrial production process.These metal materials
The phenomenon that performance is gradually degenerated is referred to as aging.Therefore, ageing of metal grade is measured and production equipment is safely operated
For be of great significance.
People's Republic of China's power industry standard " DL/T884-2004, thermal power plant's metal inspection and assessment technology instruct "
In point out, the degree of aging of qualitative analysis pipe fitting needs to carry out several experimental analyses, including structural aging, that is, observes
The roughening of Carbide Precipitation particle and distributional pattern variation；Embrittlement analysis, i.e. observation tissue signature composition and corresponding toughness with
Hardness performance variation；The situation of pipe fitting creep impairment is observed in the damage of creep hole under specified amplification factor.However these points
Analysis need to cut sample or it is in situ carry out surface replica inspection, it is complex for operation step and professional is needed to implement, lead to analysis week
Phase is longer, and work influence is examined normally to produce.It can be realized original position, micro-damage therefore, it is necessary to one kind and rapidly analyze pipe fitting
The method of aging grade.
Summary of the invention
The purpose of the present invention is to overcome the shortcomings of the existing technology and deficiency, provides a kind of ageing of metal level measurement side
Method, the method can realize the screening of characteristic variable, improve LIBS (laser induced breakdown spectroscopy) technology and are applied to ageing of metal etc.
The accuracy and accuracy of grade measurement.
The purpose of the present invention is realized by the following technical solution: a kind of ageing of metal level measurement method, comprising steps of
S1, multiple calibration samples are chosen, passes through LIBS system measurement and acquires plasma emission spectroscopy data；
S2, calibration disaggregated model is constructed using Recursive Support Vector Machine algorithm and carries out feature selecting, determine important feature
Group, which merges, establishes ageing of metal grade calibration disaggregated model；
S3, pass through LIBS system measurement and acquire the plasma emission spectroscopy data of metal sample to be measured；
S4, the important feature combination determined according to S2 extract the characteristic signal of metal sample to be measured, and are input to optimization
In disaggregated model, the aging grade forecast value of metal sample is obtained.
Preferably, the calibration sample in the step S1 is aging etc. after different ag(e)ing tests or different active times
The different metal sample of grade, the spectroscopic data of all calibration samples are that n × m ties up matrix, and wherein n is sample number, and m is Spectral Properties
Levy number, it may be assumed that
Preferably, de-noised signal processing is carried out to spectroscopic data obtained in step S1 using wavelet transformation, avoided
The influence of random noise signal in feature selecting.
Further, de-noised signal processing is carried out to spectroscopic data using wavelet transformation to specifically include:
S1.1, selection wavelet basis function carry out multilayer decomposition to spectroscopic data using Mallat algorithm, obtain corresponding low
Frequency subband signal coefficient and high frequency subband signals coefficient, the formula of wavelet transformation are as follows:
Wherein, WT (a, τ) is signal coefficient, and ψ () is wavelet function；A is scale factor, influences wavelet basis function
It is flexible；τ is shift factor, influences the translation scale of wavelet basis function；F (t) is variation of the spectral signal with wavelength；T is indicated
Wavelength；
S1.2, the white Gaussian noise feature according to zero-mean calculate the standard deviation of noise and further obtain small echo change
The threshold value of denoising is changed, calculation formula is as follows:
σ=MAD/0.6745
Wherein σ is the deviation of noise signal, and MAD is that spectroscopic data decomposes in medium-high frequency subband signal absolute coefficient
Value；
S1.3, the threshold value obtained using step S1.2 carry out soft-threshold denoising sonication to the high frequency subband signals of decomposition,
And signal is reconstructed, the spectral signal after obtaining denoising, the coefficient processing rule of soft-threshold denoising sound is as follows:
Wherein, WλFor the wavelet coefficient after soft-threshold denoising sound；W is coefficient of wavelet decomposition；λ is to determine in step S1.2
Threshold value；Sgn function is for returning to a parameter, it is indicated that the sign of input variable.
Preferably, the step S2 is specifically included:
S2.1, standard deviation standardization (abbreviation Z-score standardization) pretreatment is carried out to the spectroscopic data of calibration sample,
The middle standardized calculation formula of Z-score is as follows:
x*=(x- μ)/δ
Wherein, x* is the spectroscopic data carried out after Z-score standardization, and x is original spectroscopic data, and μ is this feature spectrum
The mean value of all samples data in line, δ are the standard deviation of all samples data in this feature spectral line；
S2.2, it is modeled using spectroscopic data of the supporting vector machine model to calibration sample, to current light spectrum feature group
Model consensus forecast ability under closing carries out weight evaluation；
S2.3, it is evaluated according to weight of the disaggregated model in S2.2 to each characteristic spectral line, by corresponding characteristic spectral line according to it
Weight order of magnitude is arranged, and the characteristic spectral line that weight coefficient ranks behind in proportion is rejected, remaining characteristic spectrum
Line is used for next layer of model construction as new feature combination；
S2.4, step S2.2-S2.3 is repeated, obtains the model prediction expressive ability under different characteristic combination, finally selects
The highest characteristic spectral line combination of prediction result accuracy rate is combined as important feature, and is established based on determining feature combination optimal
Calibration disaggregated model, guarantee ageing of metal grade separation model accurate and stablize.
Further, the weight evaluation method in the step S2.2 is cross validation method.
Preferably, it is carried out at de-noised signal using wavelet transformation metal sample spectroscopic data to be measured resulting to step S3
Reason.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1, the present invention combines Wavelet Transform Threshold denoising sound preconditioning and recursive feature removing method, avoids feature selecting
The influence of middle random noise signal, while the method for the elimination feature of recursion can be effective by way of screening and modeling layer by layer
Ground remains important feature, finally constructs an accurate and stable aging grade separation model.
2, the present invention during ageing of metal level measurement, all data analysis processes in computer program from
Dynamic operation, extends to each industrial circle, carries out the real-time measurement assessment of heat power equipment operating status.
Detailed description of the invention
Fig. 1 is the flow chart of ageing of metal level measurement embodiment of the method 2 of the present invention.
Fig. 2 is the original band of light spectrogram in the embodiment of the present invention 2.
Fig. 3 is that the embodiment of the present invention 2 uses the spectrogram after Noise Elimination from Wavelet Transform sonication to band of light spectrogram.
Fig. 4 is that model described in the embodiment of the present invention 2 becomes to the predictive ability variation of forecast set under different characteristic combination
Gesture figure.
Specific embodiment
For a better understanding of the technical solution of the present invention, the implementation that the present invention is described in detail provides with reference to the accompanying drawing
Example, embodiments of the present invention are not limited thereto.
Embodiment 1
A kind of ageing of metal level measurement method eliminated based on wavelet transformation and recursive feature of the present embodiment, it is specific to walk
It is rapid as follows:
S1, the present embodiment choose multiple calibration samples using T91 heat resisting steel as measurement object first, pass through LIBS system
Plasma emission spectroscopy data are measured and acquire, LIBS technology passes through focusing pulse laser action to sample surfaces, formation etc.
Gas ions recycle spectrometer plasma emission spectrum to be acquired, and then carry out phase to spectroscopic data using computer
The analysis answered simultaneously identifies the element constituent in sample with this, and then may be implemented to carry out material qualitative and quantitative point
Analysis；
S2, calibration disaggregated model is constructed using Recursive Support Vector Machine algorithm and carries out feature selecting, determine important feature
Group, which merges, establishes optimal ageing of metal grade calibration disaggregated model；
S3, the heat-resisting steel sample of T91 to be measured is taken, utilizes the plasma emission spectroscopy of LIBS systematic survey metal sample to be measured
Data；
S4, according to step S3 determine important feature combination extract the heat-resisting steel sample of T91 characteristic signal, by feature believe
It number is input in ageing of metal grade calibration disaggregated model, obtains the aging grade forecast value of T91 heat resisting steel.
It specifically, is aging etc. after different ag(e)ing tests or different active times in the calibration sample of the step S1
The different metal sample of grade, the spectroscopic data of all metal samples are that n × m ties up matrix, and wherein n is sample number, and m is Spectral Properties
Levy number, it may be assumed that
The step S2 is specifically included:
S2.1, standard deviation standardization pretreatment is carried out to the spectroscopic data of calibration sample；The wherein standardized meter of Z-score
It is as follows to calculate formula:
x*=(x- μ)/δ
X* is the spectroscopic data carried out after Z-score standardization in formula, and x is original spectroscopic data, and μ is this feature spectral line
The mean value of middle all samples data, δ are the standard deviation of all samples data in this feature spectral line；
S2.2, the spectroscopic data of calibration sample is modeled using the algorithm of support vector machine model of linear kernel；It is fixed
The predictive ability that the spectroscopic data of standard specimen product folds cross-validation method verifying model using 5, i.e., every time by the spectrum of calibration sample
Data are divided into 5 parts by grade, and cyclically establish model with 4 parts of data sets therein and go to predict remaining number using the model
According to collection, the model consensus forecast ability under current signature spectral combination is evaluated；
S2.3, the weight of each characteristic spectral line is evaluated according in step S2.2, by corresponding characteristic spectral line according to its weight
Order of magnitude is arranged, and 40% feature that weight coefficient is ranked behind is rejected, and remaining characteristic spectral line is as newly
Feature combination is used for next layer of model construction；
S2.4, it is repeated in step S2.2 and S2.3, obtains the model prediction expressive ability under different characteristic combination, finally
The highest characteristic spectral line combination of cross validation prediction result accuracy rate is selected to combine as important feature, and based on determining feature
Optimal calibration disaggregated model is established in combination.
Embodiment 2
The present embodiment is equally using T91 heat resisting steel as measurement object, as shown in Figs 1-4.
A kind of ageing of metal level measurement method eliminated based on wavelet transformation and recursive feature, specific steps are as follows:
S1, multiple calibration samples are chosen, passes through LIBS system measurement and acquires plasma emission spectroscopy data, LIBS skill
Art forms plasma by focusing pulse laser action to sample surfaces, recycles spectrometer plasma emission spectrum
It is acquired, then spectroscopic data is analyzed accordingly using computer and is formed into this come the element identified in sample
Point, and then may be implemented to carry out qualitative and quantitative analysis to material；
S2, de-noised signal processing is carried out to spectroscopic data using discrete small wave converting method, avoided in feature selecting
The influence of random noise signal；
S3, calibration disaggregated model is constructed using Recursive Support Vector Machine algorithm and carries out feature selecting, determine important feature
Group, which merges, establishes optimal ageing of metal grade calibration disaggregated model；
S4, the heat-resisting steel sample of T91 to be measured is taken, utilizes the plasma emission spectroscopy number of LIBS systematic survey sample to be tested
According to, and noise signal is removed to spectroscopic data using discrete small wave converting method and is handled；
S5, further according to step S3 determine important feature combination extract the heat-resisting steel sample of T91 characteristic signal, by feature
Signal is input in ageing of metal grade calibration disaggregated model, obtains the aging grade forecast value of T91 heat resisting steel.
The specific steps of the step S1 the specific steps are the same as those in embodiment 1 step S1.
The step S2 is specifically included using the processing of discrete small wave converting method removal spectroscopic data noise:
S2.1, selection orthogonal wavelet basic function sym6 function carry out 6~8 layers points to spectroscopic data using Mallat algorithm
Solution, obtains corresponding low frequency sub-band signal coefficient and high frequency subband signals coefficient, wherein the formula of wavelet transformation is as follows:
Wherein, WT (a, τ) is signal coefficient；ψ () is wavelet function；A is scale factor, influences wavelet basis function
It is flexible；τ is shift factor, influences the translation scale of wavelet basis function；F (t) is variation of the spectral signal with wavelength；T is indicated
Wavelength；
F (t) signal has been changed into the form of WT () function signal coefficient by wavelet transformation.First according to scale factor and flat
The wavelet basis function of different frequency scale can be obtained by moving the factor, then using these wavelet basis functions fitting f (t), be obtained original
The wavelet coefficient signal of different frequency is corresponded at the different location of signal.Wherein according to the frequency of fitted signal just by signal system
Number is divided into high frequency subband signals and low frequency sub-band signal.It is important because low frequency signal is often the signal that represent gradual change
Signal, so decomposing to obtain important feature signal using multilayer.
S2.2, then according to the characteristic distributions of white Gaussian noise, calculate the standard deviation of noise signal and further obtain
The threshold value of Noise Elimination from Wavelet Transform sound, calculation formula are as follows；
σ=MAD/0.6745
Wherein σ is the standard deviation of noise signal, and MAD is that spectroscopic data first layer decomposition medium-high frequency subband signal coefficient is exhausted
To the intermediate value of value；
S2.3, finally the high frequency signal progress soft-threshold denoising sonication using the threshold value of S2.2 to decomposition, and it is right
Signal is reconstructed, the spectral signal after obtaining denoising；Wherein the coefficient processing rule of soft-threshold denoising sound is as follows:
Wherein, WλFor the wavelet coefficient after soft-threshold denoising sound, w is coefficient of wavelet decomposition, and λ is to determine in step S2.2
Threshold value；Sgn function is for returning to a parameter, it is indicated that the sign of input variable.
The step S3 specific steps are identical as the specific steps of step S2 in embodiment 1.
Fig. 2 and Fig. 3 is the wave band 555nm-585nm spectrogram carried out before and after Denoising disposal using wavelet transformation, from figure
In it can be seen that the spectrogram after denoising eliminates the noise signal that high frequency changes at random, while characteristic spectral line signal is still intact
Ground is kept down.Fig. 4 is showed in conjunction with the model prediction ability of Noise Elimination from Wavelet Transform sonication and recursive feature removing method
Trend chart, wherein solid line is the cross-validation result curve of different characteristic combined down molds type, and dotted line is corresponding special
Sign combines lower calibration model to the prediction result curve of external test set.As can be seen from the figure the model table after Denoising disposal
Existing ability is obviously got well than the model without denoising, i.e. Denoising disposal can effectively improve model accuracy, and recurrence
Supporting vector machine model can improve simultaneously the accuracy rate of its forecast set in gradually rejecting extraneous features.Finally, according to cross validation
As a result it selects optimal feature to combine, while establishing model using this feature combination and external test set is predicted, by
Figure, can be according to cross validation curve it is found that cross validation results curve and external testing collection prediction result curve are with uniformity
Determine that optimal feature number of combinations is 207, corresponding model measurement collection predictablity rate reaches 0.93.
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, other any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present invention,
It should be equivalent substitute mode, be included within the scope of the present invention.
Claims (7)
1. a kind of ageing of metal level measurement method, which is characterized in that comprising steps of
S1, multiple calibration samples are chosen, passes through LIBS system measurement and acquires plasma emission spectroscopy data；
S2, calibration disaggregated model is constructed using Recursive Support Vector Machine algorithm and carries out feature selecting, determine that important feature combines
And establish ageing of metal grade calibration disaggregated model；
S3, pass through LIBS system measurement and acquire the plasma emission spectroscopy data of metal sample to be measured；
S4, the important feature combination determined according to step S2 extract the characteristic signal of metal sample to be measured, and are input to optimization
In disaggregated model, the aging grade forecast value of metal sample is obtained.
2. ageing of metal level measurement method according to claim 1, it is characterised in that: the calibration sample in the step S1
Product are the metal sample that aging grade is different after different ag(e)ing tests or different active times, the spectrum of all calibration samples
Data are that n × m ties up matrix, and wherein n is sample number, and m is spectral signature number, it may be assumed that
3. ageing of metal level measurement method according to claim 1, it is characterised in that: using wavelet transformation to step S1
Resulting spectroscopic data carries out de-noised signal processing.
4. ageing of metal level measurement method according to claim 3, it is characterised in that: using wavelet transformation to spectrum number
Specific steps are handled according to de-noised signal is carried out are as follows:
S1.1, selection wavelet basis function carry out multilayer decomposition to spectroscopic data using Mallat algorithm, obtain corresponding low frequency
Band signal coefficient and high frequency subband signals coefficient, the formula of wavelet transformation are as follows:
Wherein, WT (a, τ) is the small echo signal coefficient after wavelet transformation, and ψ () is wavelet function；A be scale because
Son influences the flexible of wavelet basis function；τ is shift factor, influences the translation scale of wavelet basis function；F (t) be spectral signal with
The variation of wavelength；T indicates wavelength；
S1.2, the white Gaussian noise feature according to zero-mean calculate the standard deviation of noise and further obtain wavelet transformation and go
The threshold value of noise, calculation formula are as follows:
σ=MAD/0.6745
Wherein σ is the standard deviation of noise signal, and MAD is that spectroscopic data decomposes in medium-high frequency subband signal absolute coefficient
Value；
S1.3, the threshold value obtained using step S1.2 carry out soft-threshold denoising sonication to the high frequency subband signals of decomposition, and right
Signal is reconstructed, the spectral signal after obtaining denoising, and the coefficient processing rule of soft-threshold denoising sound is as follows:
Wherein, WλFor the wavelet coefficient after soft-threshold denoising sound；W is coefficient of wavelet decomposition；λ is the threshold value determined in step S1.2；
Sgn function is for returning to a parameter, it is indicated that the sign of input variable.
5. ageing of metal level measurement method according to claim 1, it is characterised in that: the step S2 is specifically included:
S2.1, Z-score standardization pretreatment, the wherein standardized calculating of Z-score are carried out to the spectroscopic data of calibration sample
Formula is as follows:
x*=(x- μ)/δ
Wherein, x* is the spectroscopic data carried out after Z-score standardization, and x is original spectroscopic data, and μ is in this feature spectral line
The mean value of all samples data, δ are the standard deviation of all samples data in this feature spectral line；
S2.2, it is modeled using spectroscopic data of the supporting vector machine model to calibration sample, under the combination of current light spectrum feature
Model consensus forecast ability carry out weight evaluation；
S2.3, the weight of each characteristic spectral line is evaluated according in step S2.2, corresponding characteristic spectral line is absolute according to its weight
Value size is arranged, and the characteristic spectral line that weight coefficient ranks behind in proportion is rejected, and remaining characteristic spectral line is as new
Feature combination be used for next layer of model construction；
S2.4, step S2.2-S2.3 is repeated, obtains the model prediction ability under different characteristic combination, finally selects prediction result
The highest characteristic spectral line combination of accuracy rate is combined as important feature, and establishes optimal calibration point based on determining feature combination
Class model.
6. ageing of metal level measurement method according to claim 5, it is characterised in that: the weight in the step 2.2
Evaluation method is the model evaluation method based on cross validation.
7. ageing of metal level measurement method according to claim 1, it is characterised in that: using wavelet transformation to step S3
Resulting metal sample spectroscopic data to be measured carries out de-noised signal processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810963846.3A CN109086547A (en) | 2018-08-23 | 2018-08-23 | A kind of ageing of metal level measurement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810963846.3A CN109086547A (en) | 2018-08-23 | 2018-08-23 | A kind of ageing of metal level measurement method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109086547A true CN109086547A (en) | 2018-12-25 |
Family
ID=64794477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810963846.3A Pending CN109086547A (en) | 2018-08-23 | 2018-08-23 | A kind of ageing of metal level measurement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109086547A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111198165A (en) * | 2020-01-14 | 2020-05-26 | 重庆理工大学 | Method for measuring water quality parameters based on spectral data standardization |
| CN112001446A (en) * | 2020-08-25 | 2020-11-27 | 中国特种设备检测研究院 | Method and device for determining aging grade of high-chromium martensite heat-resistant steel structure |
| CN112816553A (en) * | 2021-01-22 | 2021-05-18 | 国电锅炉压力容器检验有限公司 | Heat-resistant steel aging grade evaluation method based on support vector machine |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6937659B1 (en) * | 1997-11-14 | 2005-08-30 | Ac Capital Management, Inc. | Apparatus and method for compressing video information |
| CN102818629A (en) * | 2012-05-04 | 2012-12-12 | 浙江大学 | Micro-spectrometer signal denoising method based on stable wavelet transform |
| CN103543132A (en) * | 2013-10-17 | 2014-01-29 | 清华大学 | Coal property measuring method based on wavelet transform |
| CN105701336A (en) * | 2015-12-31 | 2016-06-22 | 深圳先进技术研究院 | Traditional Chinese medicine syndrome differentiation typing system based on electroencephalogram data and building method and system of traditional Chinese medicine syndrome differentiation typing mode based on electroencephalogram data |
| CN107976432A (en) * | 2017-10-16 | 2018-05-01 | 华南理工大学 | A kind of heat resisting steel aging level measurement method based on support vector machines |
-
2018
- 2018-08-23 CN CN201810963846.3A patent/CN109086547A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6937659B1 (en) * | 1997-11-14 | 2005-08-30 | Ac Capital Management, Inc. | Apparatus and method for compressing video information |
| CN102818629A (en) * | 2012-05-04 | 2012-12-12 | 浙江大学 | Micro-spectrometer signal denoising method based on stable wavelet transform |
| CN103543132A (en) * | 2013-10-17 | 2014-01-29 | 清华大学 | Coal property measuring method based on wavelet transform |
| CN105701336A (en) * | 2015-12-31 | 2016-06-22 | 深圳先进技术研究院 | Traditional Chinese medicine syndrome differentiation typing system based on electroencephalogram data and building method and system of traditional Chinese medicine syndrome differentiation typing mode based on electroencephalogram data |
| CN107976432A (en) * | 2017-10-16 | 2018-05-01 | 华南理工大学 | A kind of heat resisting steel aging level measurement method based on support vector machines |
Non-Patent Citations (1)
| Title |
|---|
| 谭跃刚: "《机械装置的光纤光栅动态检测技术及应用》", 31 January 2018, 武汉理工大学出版社 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111198165A (en) * | 2020-01-14 | 2020-05-26 | 重庆理工大学 | Method for measuring water quality parameters based on spectral data standardization |
| CN112001446A (en) * | 2020-08-25 | 2020-11-27 | 中国特种设备检测研究院 | Method and device for determining aging grade of high-chromium martensite heat-resistant steel structure |
| CN112816553A (en) * | 2021-01-22 | 2021-05-18 | 国电锅炉压力容器检验有限公司 | Heat-resistant steel aging grade evaluation method based on support vector machine |
| CN112816553B (en) * | 2021-01-22 | 2023-04-07 | 国能锅炉压力容器检验有限公司 | Heat-resistant steel aging grade evaluation method based on support vector machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086547A (en) | A kind of ageing of metal level measurement method | |
| CN110865092A (en) | In-situ analysis method for representing component distribution of high-temperature alloy by utilizing X-ray fluorescence spectrum | |
| CN103487411B (en) | A kind of random forests algorithm is in conjunction with the method for Laser-induced Breakdown Spectroscopy identification type of steels | |
| CN113155809B (en) | Novel spectral detection method for ore classification and real-time quantitative analysis | |
| CN104697966A (en) | Method for quantitatively analyzing chromium and manganese in steel based on least square support vector machine algorithm combined with laser-induced breakdown spectroscopy | |
| Li et al. | Fatigue crack sizing in rail steel using crack closure-induced acoustic emission waves | |
| CN110569566B (en) | Method for predicting mechanical property of plate strip | |
| CN107976432B (en) | Heat-resistant steel aging grade measuring method based on support vector machine | |
| Huang et al. | A hybrid model combining wavelet transform and recursive feature elimination for running state evaluation of heat-resistant steel using laser-induced breakdown spectroscopy | |
| CN112834485A (en) | Non-calibration method for quantitative analysis of laser-induced breakdown spectroscopy elements | |
| Ruan et al. | A novel hybrid filter/wrapper method for feature selection in archaeological ceramics classification by laser-induced breakdown spectroscopy | |
| Cai et al. | Estimating the aging grade of heat-resistant steel by using portable laser-induced breakdown spectroscopy | |
| CN105181676A (en) | Method for measuring depth and quality of plating of galvanized plate through glow discharge spectrometer | |
| CN106679911B (en) | Beam type structure damage identification method based on multi-scale data fusion theory | |
| CN117233347B (en) | Carbon steel spheroidization grade measuring method, system and equipment | |
| CN102103079B (en) | Spectrum analysis method | |
| CN110672586B (en) | Concrete corrosion state detection method based on LIBS | |
| CN105717094B (en) | A kind of metal element content analysis method based on large database concept identification | |
| CN107941758B (en) | Coal quality heat value measuring method based on spectrum pretreatment | |
| Bashkov et al. | The use of acoustic emission for the construction of a generalized fatigue diagram of metals and alloys | |
| CN105510305A (en) | Method for detecting content of aluminum in potato starch | |
| CN111624194B (en) | Blade element nondestructive measurement method based on laser-induced breakdown spectroscopy | |
| CN101666746B (en) | Laser induced spectrum data processing method based on wavelet analysis | |
| Yu et al. | Convolutional neural network-based quantitative evaluation for corrosion cracks in oil/gas pipeline by millimeter-wave imaging | |
| CN109668862B (en) | Aluminum electrolyte molecular ratio detection method based on laser-induced breakdown spectroscopy |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181225 |
|
| RJ01 | Rejection of invention patent application after publication |