CN109521001A - A kind of flying marking measuring method based on PSO and ε-SVR - Google Patents
A kind of flying marking measuring method based on PSO and ε-SVR Download PDFInfo
- Publication number
- CN109521001A CN109521001A CN201811376249.7A CN201811376249A CN109521001A CN 109521001 A CN109521001 A CN 109521001A CN 201811376249 A CN201811376249 A CN 201811376249A CN 109521001 A CN109521001 A CN 109521001A
- Authority
- CN
- China
- Prior art keywords
- sample
- flying
- svr
- flying dust
- pso
- 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
- 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
- G01N21/73—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light thermally excited using plasma burners or torches
Landscapes
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Plasma & Fusion (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses a kind of flying marking measuring methods based on PSO and ε-SVR, specific steps include: that (1) is chosen derived from the different flying dust sample of the phosphorus content of practical power plant as calibration sample, and the laser induced plasma emission spectrum of calibration sample is obtained using LIBS detection device；(2) to extracting relevant to flying marking measurement characteristic peaks spectroscopic data in acquired LIBS spectrum and be normalized；(3) PSO particle swarm optimization algorithm is combined to carry out the optimizing of the support vector machines relevant parameter based on flying marking measurement to the spectroscopic data after normalized；(4) ε-SVR calibration model is constructed according to identified optimal parameter and SVM；(5) flying dust sample to be measured is subjected to processing and obtains spectroscopic data, be input in the calibration model, obtain the phosphorus content of flying dust sample to be measured.The method of the present invention combination intelligent algorithm realizes that LIBS is applied to the measurement of unburned carbon in flue dust, and calculating is time-consuming short, and the accuracy and precision of measurement can be improved, applied widely.
Description
Technical field
The present invention relates to technical field of industrial measurement, in particular to a kind of flying marking measurement based on PSO and ε-SVR
Method.
Background technique
Unburned carbon in flue dust is that the important indicator of coal fired boiler of power plant efficiency of combustion illustrates coal when unburned carbon in flue dust is high
Consumption and cost of electricity-generating are excessively high.Unburned carbon in flue dust is detected in real time and accurately, is conducive to adjust coal-air ratio；Unburned carbon in flue dust is controlled
In optimum range, boiler combustion controlled level is improved, to guarantee economic, the safe and stable operation of unit, but since flying dust contains
Carbon amounts is by coal, boiler structure, the influence of many factors such as operation operation level, it is difficult to realize on-line measurement.
The generally existing analysis lag such as numerous off-line measurement methods such as calcination method, bounce technique and the representative difference of sample etc.
Problem is unable to satisfy field demand.And laser induced breakdown spectroscopy (Laser-Induced Breakdown
Spectroscopy, LIBS) technology is a kind of novel plasma emission spectroscopy technology, it can be achieved that the multicomponent of sample is synchronous
Quickly analysis, belongs to On-line Measuring Method, can be applied to the quality control and process monitoring of various industrial process, contain in flying dust
There is very big application potential on the on-line checking of carbon amounts.But the technology is a point-measurement technique, and obvious, cause is influenced by matrix
Keep precision not high, and then limits the development of the technology.How to realize that accurate measurement is that LIBS technology contains in flying dust
The premise and basis to play a role in carbon amounts detection.
Realize that the analysis method of flying dust unburned carbon detection mostly uses greatly the line of single argument or multivariable currently based on LIBS
Property regression analysis, although precision can be improved by means such as the recurrence of multivariable, effect is not able to satisfy industry spot still
Demand.In recent years, with the development of artificial intelligence technology, relevant application, such as people have also been obtained in the application field of LIBS
Artificial neural networks method (ANN), support vector machine method (SVM).Artificial intelligence technology is applied to unburned carbon in flue dust by the present invention
Measurement, improve unburned carbon in flue dust detection accuracy and precision.
Summary of the invention
It is an object of the invention to overcome shortcoming and deficiency in the prior art, provide a kind of based on PSO (particle group optimizing
Algorithm) and ε-SVR (support vector regression) flying marking measuring method, pass through LIBS laser induced breakdown spectroscopy
The spectroscopic data for obtaining the calibration flying dust sample of different carbon contents, then by PSO particle swarm optimization algorithm to pretreated
The training sample data of calibration flying dust sample are supported the optimizing of vector machine relevant parameter, and optimizing is selected optimal in different range
Parameter establishes corresponding ε-SVR Support vector regression calibration model, with the cross validation regressive mean error MSE of calibration sample
Minimum target determines optimal Quantitative Analysis Model, to improve the accuracy and standard that LIBS is applied to flying marking measurement
Exactness.
In order to achieve the above object, the present invention adopts the following technical scheme that:
A kind of flying marking measuring method based on PSO and ε-SVR, specific steps include:
(1) it calibration flying dust sample spectral data measurement: chooses and makees derived from the different flying dust sample of the phosphorus content of practical power plant
For calibration sample, the laser induced plasma emission spectrum of sample is obtained using laser induced breakdown spectroscopy detection device；
(2) characteristic peaks data are extracted: extracting m kind characteristic spectrum, including L kind from the spectroscopic data of each calibration sample
The characteristic spectrum of main matrix element, L >=2, n >=4 in characteristic spectrum relevant to carbon and n kind flying dust；
(3) place normalized and parameter optimization: is normalized to the peak value for the m kind characteristic spectrum that step (2) are extracted
Reason, carries out the support based on flying marking measurement to the spectroscopic data after normalized using PSO particle swarm optimization algorithm
The optimizing of vector machine relevant parameter；
(4) to the calibration sample after normalized, SVM support vector machines and step building calibration regression model: are utilized
(3) ε-SVR support vector machines for unburned carbon in flue dust of the optimal parameter building based on characteristic spectrum data that optimizing obtains is calibrated back
Return model；
(5) it predicts flying dust sample phosphorus content to be measured: obtaining flying dust sample to be measured using laser induced breakdown spectroscopy detection device
This spectroscopic data after extracting relevant characteristic peaks and normalized, is entered into the calibration that step (4) is established and returns
Return in model, and then obtains the phosphorus content of flying dust sample to be measured.
As a preferred technical solution, in the step (1), obtained using laser induced breakdown spectroscopy detection device carbon containing
The plasma emission spectroscopy of different flying dust samples is measured, method particularly includes:
Using Nd:YAG pulse laser, with the flying dust sample effect on sample translation stage after lens focus, lead to
Mobile example translation stage is crossed, changes the position of flying dust sample, and then change the active position of laser and flying dust sample, to obtain
The emission spectrum of same flying dust sample different location；Spectroscopic data obtained is averaging processing, and then obtains some
The laser induced breakdown spectroscopy data of flying dust sample；The above method is repeated to different phosphorus content flying dust samples, is obtained different carbon containing
Measure the spectroscopic data of flying dust sample.
As a preferred technical solution, in the step (2), L kind characteristic spectrum relevant to carbon, including C and CN are extracted
Characteristic spectrum；Extract the characteristic spectrum of the main matrix element of n kind flying dust sample, the feature including Si, Al, Mg, Fe element
Spectrum.
As a preferred technical solution, the step (3) specifically include the following steps:
The L kind characteristic peak spectroscopic data of different samples is pressed dimension normalized to [- 1,1], i.e., according to each column by (3-1)
In maximum original value be xmax, the smallest original value be xmin, it is x with original spectral data, then the numerical value after normalizing is
xnormalization, may be expressed as:
(3-2) is using the unburned carbon in flue dust of characteristic peaks spectroscopic data and calibration sample after normalized as calibration mould
Type data are allowed using PSO particle swarm optimization algorithm with the minimum target of cross validation regressive mean error MSE of calibration sample
Parameter-penalty factor c, kernel functional parameter g and loss function value p in ε-SVR support vector regression take in a certain range
Value, and then determine optimal c, g, p parameter；The calculation formula of regressive mean error MSE is as follows:
In formula, yreferenceFor carbon content reference value, ypredictedFor carbon content predicted value, in N cross validation regression process
Test set sample number.
As a preferred technical solution, in the step (3-2), the PSO particle swarm optimization algorithm principle of use are as follows: solving
Each particle is approached to 2 points simultaneously in space, and first point is that all particles were searched in each generation in entire population
Globally optimal solution q achieved in journeybest, another point is each particle itself individual achieved in each generation search process
Optimal solution pbest, pass through the position and speed of iteration more new particle.
As a preferred technical solution, the step (4) specifically include the following steps:
(4-1) obtains training set sample (xi,yi), wherein i=1,2 ..., l, l are different calibration flying dust sample plasmas hair
Penetrate the sample size in spectrum data matrix for training；xiSpectroscopic data after being normalized for what is selected according to formula (1),
As input variable；yiFor carbon content, as corresponding output valve；
The estimation function of (4-2) ε-SVR support vector regression indicates are as follows:
WhereinIt is Nonlinear Mapping of the input space to higher dimensional space, ε-SVR support vector regression is exactly will be real
Border problem is transformed into high-dimensional feature space by Nonlinear Mapping, constructs linear regression function in high-dimensional feature space to realize
Nonlinear solshing in former space, coefficient ω and b can estimate that returning risk indicates by minimizing recurrence risk are as follows:
Wherein Γ () is loss function；C is penalty coefficient, is a constant；It can obtain being indicated with data point by formula (4)
Vector ω:
α in formulaiIndicate Lagrange multiplier α=(α1, α2... α1),For αiConjugation；
Wushu (5) substitutes into formula (3), and the estimation function of ε-SVR support vector regression can be expressed as follows:
In formula (6), dot product is by kernel function k (xi, x) and it replaces, kernel function can be without knowing mappingIn the case where benefit
It is inputted with lower dimensional space data and carries out dot product calculating in high-dimensional feature space.
As a preferred technical solution, in the step (4-2), the kernel function used is Radial basis kernel function, specific table
It is shown as:
K (x, xi)=exp (- γ ‖ x-xi‖2), γ > 0 (7)
Wherein, the σ of γ=1/22, σ is kernel function width parameter, controls the radial effect range of kernel function；
Used loss function is soft margin loss function, specific as follows:
Γ(xi, yi)=max { 0,1-yi((ω*xi)+b)} (8)。
The present invention has the following advantages compared with the existing technology and effect:
(1) the method for the present invention is on the basis of LIBS analytical technology, to the laser plasma emission spectrum of different flying dust samples
Data and phosphorus content are returned, and data analysis process is simple and fast, accuracy rate is high, can reach what unburned carbon in flue dust quickly detected
Effect.
(2) feature of ε-SVR Support vector regression calibration model maximum of the invention is using Statistical Learning Theory as base
Plinth realizes the learning method under condition of small sample and statistics rule in conjunction with the plasma emission spectroscopy characteristic of flying dust sample itself
Rule, be not only simple in structure, and performance it is more traditional BP neural network it is superior, therefore, this ε-SVR model is pre- in unburned carbon in flue dust
Surveying aspect has certain superiority.
(3) ε-SVR support vector machines under the method for the present invention is optimized based on PSO is during constructing prediction model, PSO
Optimization algorithm can seek optimal penalty factor c, kernel functional parameter g and penalty p, have global optimization ability, it is not only
ε-SVR model prediction accuracy can be improved, moreover it is possible to which the blindness tentative calculation for avoiding parameter reaches unburned carbon in flue dust prediction effect
It is best.
(4) data all in the method for the present invention analysis and statistic processes are automatically performed by computer program, time-consuming
It is short, it is applied widely, it is able to satisfy the real-time measurement of industry spot.
Detailed description of the invention
Fig. 1 is the flow chart of the unburned carbon in flue dust rapid survey in the present invention；
Fig. 2 is spectrogram of flying dust sample (#1) wavelength in the present embodiment in 230-400nm；
Fig. 3 is the comparison diagram of unburned carbon in flue dust predicted value and reference value under the optimum regression calibration model in the present embodiment.
Specific embodiment
In order to which the purpose of the present invention, technical solution and advantage is more clearly understood, with reference to the accompanying drawings and embodiments,
The present invention is further described in detail.It should be appreciated that described herein, the specific embodiments are only for explaining the present invention,
It is not limited to the present invention.
Embodiment
It is as shown in Figure 1 the flow chart of the unburned carbon in flue dust rapid survey in the present invention.
A kind of flying marking measuring method based on PSO and ε-SVR, includes the following steps:
(1) it chooses derived from the different flying dust sample of the phosphorus content of power plant as calibration sample, is obtained using LIBS detection device
The laser induced plasma emission spectrum for taking flying dust sample is illustrated in figure 2 single channel spectrometer flying dust sample (# obtained
1) in the spectrogram of 230-400nm；
In the present embodiment, 20 kinds of flying dust samples are chosen, the phosphorus content of flying dust sample is as shown in table 1, the data in table 1 by
Standard off-line analysis detection gained, as reference value.
The phosphorus content of 1 flying dust sample of table
(2) characteristic peaks data are extracted: extracting 12 kinds of different wave lengths of calibration flying dust sample original spectral data matrix
Characteristic spectrum extracts spectral wavelength as shown in table 2 and figure 2.
2 characteristic spectrum wavelength (nm) of table
(3) characteristic spectrum data matrix is normalized, spectroscopic data is carried out based on phosphorus content using PSO
The optimizing of support vector machines relevant parameter；
In the present embodiment, collected 20 flying dusts sample plasma emission spectra data matrix is divided into two groups: one
Group 17, to calibrate sample；One group 3, be sample to be tested；The phosphorus content of sample to be tested is evenly distributed in calibration sample, point
Not Yong Yu spectroscopic data and unburned carbon in flue dust reference value correlativity foundation and inspection.
Calibration sample, sample to be tested are normalized into [- 1,1] by dimension, to the training set sample after normalized
Product carry out the optimizing of the Support vector regression relevant parameter based on flying marking measurement using PSO particle swarm optimization algorithm,
The relevant parameter is penalty factor c, RBF kernel functional parameter g and loss function value p.
(4) to the calibration sample after normalized, the ε-SVR using SVM and optimal parameter building unburned carbon in flue dust is fixed
Mark regression model；
Specifically, it by the characteristic spectrum data matrix and phosphorus content after normalized, is converted by RBF kernel function
Project high-dimensional feature space.
(5) sample to be tested spectrum data matrix is obtained using LIBS detection device, by extracted characteristic peaks and normalizing
Change that treated that spectrum data matrix to be measured is input in the calibration regression model, obtains the carbon containing magnitude of sample to be tested.
In the present embodiment, it is when penalty factor c, kernel functional parameter g and loss function value p take optimized parameter c respectively
When 77.4279, g be 0.1 and p is 0.053997, best ε-SVR calibration model is obtained.At this point, the model training collection cross validation
Mean square error is 0.0209, and the mean square error of sample to be tested is 0.0146, squared correlation coefficient 0.9997.Such as table 3 and Fig. 3 institute
Show, the mean square error of gained model is small, related coefficient close to 1, the predicted value of sample to be tested phosphorus content and reference value relatively partially
Difference is smaller, and model accuracy rate is high.
Phosphorus content quantitative analysis results under 3 optimal models of table
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 flying marking measuring method based on PSO and ε-SVR, which is characterized in that specific steps include:
(1) it calibration flying dust sample spectral data measurement: chooses derived from the different flying dust sample of the phosphorus content of practical power plant as fixed
Standard specimen product obtain the laser induced plasma emission spectrum of sample using laser induced breakdown spectroscopy detection device；
(2) characteristic peaks data are extracted: extracting m kind characteristic spectrum, including L kind and carbon from the spectroscopic data of each calibration sample
The characteristic spectrum of main matrix element, L >=2, n >=4 in relevant characteristic spectrum and n kind flying dust；
(3) normalized and parameter optimization: being normalized the peak value for the m kind characteristic spectrum that step (2) are extracted, benefit
The support vector machines based on flying marking measurement is carried out to the spectroscopic data after normalized with PSO particle swarm optimization algorithm
The optimizing of relevant parameter；
(4) it building calibration regression model: to the calibration sample after normalized, is sought using SVM support vector machines and step (3)
ε-SVR support vector machines the calibration of excellent obtained unburned carbon in flue dust of the optimal parameter building based on characteristic spectrum data returns mould
Type；
(5) it predicts flying dust sample phosphorus content to be measured: obtaining flying dust sample to be measured using laser induced breakdown spectroscopy detection device
Spectroscopic data after extracting relevant characteristic peaks and normalized, is entered into the calibration that step (4) is established and returns mould
In type, and then obtain the phosphorus content of flying dust sample to be measured.
2. the flying marking measuring method according to claim 1 based on PSO and ε-SVR, it is characterised in that: the step
Suddenly in (1), the plasma emissioning light of the different flying dust sample of phosphorus content is obtained using laser induced breakdown spectroscopy detection device
Spectrum, method particularly includes:
Using Nd:YAG pulse laser, with the flying dust sample effect on sample translation stage after lens focus, pass through shifting
Dynamic sample translation stage changes the position of flying dust sample, and then changes the active position of laser and flying dust sample, to obtain same
The emission spectrum of flying dust sample different location；Spectroscopic data obtained is averaging processing, and then obtains some flying dust
The laser induced breakdown spectroscopy data of sample；The above method is repeated to different phosphorus content flying dust samples, different phosphorus content is obtained and flies
The spectroscopic data of grey sample.
3. a kind of flying marking measuring method based on PSO and ε-SVR according to claim 1, it is characterised in that: institute
It states in step (2), extracts L kind characteristic spectrum relevant to carbon, the characteristic spectrum including C and CN；It is main to extract n kind flying dust sample
Matrix element characteristic spectrum, the characteristic spectrum including Si, Al, Mg, Fe element.
4. a kind of flying marking measuring method based on PSO and ε-SVR according to claim 1, which is characterized in that institute
The step of stating (3) specifically include the following steps:
The L kind characteristic peak spectroscopic data of different samples is pressed dimension normalized to [- 1,1], i.e., according in each column by (3-1)
Maximum original value is xmax, the smallest original value be xmin, it is x with original spectral data, then the numerical value after normalizing is
xnormalization, may be expressed as:
(3-2) is using the unburned carbon in flue dust of characteristic peaks spectroscopic data and calibration sample after normalized as calibration model number
According to allowing ε-SVR using PSO particle swarm optimization algorithm with the minimum target of cross validation regressive mean error MSE of calibration sample
Parameter-penalty factor c, kernel functional parameter g and loss function value p in support vector regression value in a certain range, into
And determine optimal c, g, p parameter；The calculation formula of regressive mean error MSE is as follows:
In formula, yreferenceFor carbon content reference value, ypredictedSurvey for carbon content predicted value, in N cross validation regression process
Examination collection sample number.
5. a kind of flying marking measuring method based on PSO and ε-SVR according to claim 4, which is characterized in that institute
It states in step (3-2), the PSO particle swarm optimization algorithm principle of use are as follows: each particle is carried out to 2 points simultaneously in solution space
It approaches, first point is all particles globally optimal solution q achieved in each generation search process in entire populationbest, separately
One point is each particle itself individual optimal solution p achieved in each generation search processbest, pass through iteration more new particle
Position and speed.
6. a kind of flying marking measuring method based on PSO and ε-SVR according to claim 4, which is characterized in that institute
The step of stating (4) specifically include the following steps:
(4-1) obtains training set sample (xi,yi), wherein i=1,2 ..., l, l are different calibration flying dust sample plasma optical emissions
Sample size in modal data matrix for training；xiSpectroscopic data after being normalized for what is selected according to formula (1), as
Input variable；yiFor carbon content, as corresponding output valve；
The estimation function of (4-2) ε-SVR support vector regression indicates are as follows:
WhereinIt is Nonlinear Mapping of the input space to higher dimensional space, ε-SVR support vector regression exactly will be asked actually
Topic is transformed into high-dimensional feature space by Nonlinear Mapping, constructs linear regression function in high-dimensional feature space to realize former sky
Between in nonlinear solshing, coefficient ω and b can return risk and be estimated by minimizing, and returning risk indicates are as follows:
Wherein Γ () is loss function；C is penalty coefficient, is a constant；The vector indicated with data point can be obtained by formula (4)
ω:
α in formulaiIndicate Lagrange multiplier α=(α1, α2... αl),For αiConjugation；
Wushu (5) substitutes into formula (3), and the estimation function of ε-SVR support vector regression can be expressed as follows:
In formula (6), dot product is by kernel function k (xi, x) and it replaces, kernel function can be without knowing mappingIn the case where using low
The input of dimension space data carries out dot product calculating in high-dimensional feature space.
7. a kind of flying marking measuring method based on PSO and ε-SVR according to claim 1, which is characterized in that institute
It states in step (4-2), the kernel function used is embodied as Radial basis kernel function:
K (x, xi)=exp (- γ ‖ x-xi‖2), γ > 0 (7)
Wherein, the σ of γ=1/22, σ is kernel function width parameter, controls the radial effect range of kernel function；
Used loss function is soft margin loss function, specific as follows:
Γ(xi, yi)=max { 0,1-yi((ω*xi)+b)} (8)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811376249.7A CN109521001A (en) | 2018-11-19 | 2018-11-19 | A kind of flying marking measuring method based on PSO and ε-SVR |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811376249.7A CN109521001A (en) | 2018-11-19 | 2018-11-19 | A kind of flying marking measuring method based on PSO and ε-SVR |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109521001A true CN109521001A (en) | 2019-03-26 |
Family
ID=65776299
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811376249.7A Pending CN109521001A (en) | 2018-11-19 | 2018-11-19 | A kind of flying marking measuring method based on PSO and ε-SVR |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109521001A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111060497A (en) * | 2019-12-31 | 2020-04-24 | 西安交通大学 | LIBS (laser induced breakdown spectroscopy) measuring method for unburned carbon content of mixed-type fly ash based on SVM (support vector machine) cascade model |
| CN113916860A (en) * | 2021-11-02 | 2022-01-11 | 淮阴工学院 | Pesticide residue type identification method based on fluorescence spectrum |
| CN116777537A (en) * | 2023-05-26 | 2023-09-19 | 华能国际电力股份有限公司上海石洞口第二电厂 | Method and system for calculating and optimizing electricity-measuring coal cost and blending coal based on coal quality characteristics |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101915753A (en) * | 2010-07-30 | 2010-12-15 | 浙江师范大学 | Genetic Neural NetworkQuantitative analysis method for laser induced breakdown spectroscopy based on gGenetic Neural Networkgenetic neural network |
| CN102778538A (en) * | 2012-07-06 | 2012-11-14 | 广东电网公司电力科学研究院 | Soft measuring method based on improved SVM (Support Vector Machine) for measuring boiler unburned carbon content in fly ash |
| CN104198396A (en) * | 2014-07-30 | 2014-12-10 | 江苏大学 | Method for diagnosing nitrogen, phosphorus and potassium deficiency of crops by using polarization-hyperspectral technique |
| CN106442470A (en) * | 2016-08-31 | 2017-02-22 | 广州博谱能源科技有限公司 | Coal quality characteristic quantitative analysis method based on LIBS (laser induced breakdown spectrum) and genetic neural network |
| CN108627500A (en) * | 2018-04-16 | 2018-10-09 | 华南理工大学 | A kind of solid fuel measurement of caloric value method |
| CN108960492A (en) * | 2018-06-20 | 2018-12-07 | 上海电力学院 | A kind of exhaust enthalpy of turbine prediction technique based on PSO-SVR soft-sensing model |
-
2018
- 2018-11-19 CN CN201811376249.7A patent/CN109521001A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101915753A (en) * | 2010-07-30 | 2010-12-15 | 浙江师范大学 | Genetic Neural NetworkQuantitative analysis method for laser induced breakdown spectroscopy based on gGenetic Neural Networkgenetic neural network |
| CN102778538A (en) * | 2012-07-06 | 2012-11-14 | 广东电网公司电力科学研究院 | Soft measuring method based on improved SVM (Support Vector Machine) for measuring boiler unburned carbon content in fly ash |
| CN104198396A (en) * | 2014-07-30 | 2014-12-10 | 江苏大学 | Method for diagnosing nitrogen, phosphorus and potassium deficiency of crops by using polarization-hyperspectral technique |
| CN106442470A (en) * | 2016-08-31 | 2017-02-22 | 广州博谱能源科技有限公司 | Coal quality characteristic quantitative analysis method based on LIBS (laser induced breakdown spectrum) and genetic neural network |
| CN108627500A (en) * | 2018-04-16 | 2018-10-09 | 华南理工大学 | A kind of solid fuel measurement of caloric value method |
| CN108960492A (en) * | 2018-06-20 | 2018-12-07 | 上海电力学院 | A kind of exhaust enthalpy of turbine prediction technique based on PSO-SVR soft-sensing model |
Non-Patent Citations (7)
| Title |
|---|
| SHUNCHUN YAO等: "Rapidly Measuring Unburned Carbon in Fly Ash Using Molecular CN by Laser-Induced Breakdown Spectroscopy", 《ENERGY FUELS》 * |
| 张天龙等: "化学计量学在激光诱导击穿光谱分析中的研究进展", 《分析化学》 * |
| 张雷等: "激光诱导击穿光谱精确测定燃煤工业分析指标的研究", 《光谱学与光谱分析》 * |
| 梁循著: "《支持向量机算法及其金融应用》", 31 January 2012, 知识产权出版社 * |
| 王涛等: "基于改进粒子群算法优化的飞灰含碳量建模研究", 《黑龙江电力》 * |
| 王红军著: "《基于知识的机电***故障诊断与预测技术》", 31 January 2014, 中国财富出版社 * |
| 王鹏: "基于多元回归的LIBS钢液成分定量分析", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111060497A (en) * | 2019-12-31 | 2020-04-24 | 西安交通大学 | LIBS (laser induced breakdown spectroscopy) measuring method for unburned carbon content of mixed-type fly ash based on SVM (support vector machine) cascade model |
| CN111060497B (en) * | 2019-12-31 | 2020-11-17 | 西安交通大学 | LIBS (laser induced breakdown spectroscopy) measuring method for unburned carbon content of mixed-type fly ash based on SVM (support vector machine) cascade model |
| CN113916860A (en) * | 2021-11-02 | 2022-01-11 | 淮阴工学院 | Pesticide residue type identification method based on fluorescence spectrum |
| CN116777537A (en) * | 2023-05-26 | 2023-09-19 | 华能国际电力股份有限公司上海石洞口第二电厂 | Method and system for calculating and optimizing electricity-measuring coal cost and blending coal based on coal quality characteristics |
| CN116777537B (en) * | 2023-05-26 | 2024-04-12 | 华能国际电力股份有限公司上海石洞口第二电厂 | Method and system for calculating and optimizing electricity-measuring coal cost and blending coal based on coal quality characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111369070B (en) | Multimode fusion photovoltaic power prediction method based on envelope clustering | |
| CN105894130B (en) | A kind of optimization placement method for public supply mains monitoring point | |
| CN109521001A (en) | A kind of flying marking measuring method based on PSO and ε-SVR | |
| CN105486655B (en) | The soil organism rapid detection method of model is intelligently identified based on infrared spectroscopy | |
| CN101726451A (en) | Method for measuring viscosity index of internal combustion engine oil | |
| CN103487411A (en) | Method for recognizing steel grade by combining random forest algorithm with laser-induced breakdown spectroscopy | |
| CN104483292B (en) | A kind of method that use multiline ratio method improves laser microprobe analysis accuracy | |
| KR101914770B1 (en) | Predicting System Of Energy Efficiency For Ships And Predicting Method In Using Same | |
| CN102830096A (en) | Method for measuring element concentration and correcting error based on artificial neural network | |
| CN109324013A (en) | A method of it is quickly analyzed using Gaussian process regression model building oil property near-infrared | |
| CN104730042A (en) | Method for improving free calibration analysis precision by combining genetic algorithm with laser induced breakdown spectroscopy | |
| CN107132266A (en) | A kind of Classification of water Qualities method and system based on random forest | |
| CN110208252A (en) | A kind of coal ash fusion temperature prediction technique based on laser induced breakdown spectroscopy analysis | |
| CN112884012A (en) | Building energy consumption prediction method based on support vector machine principle | |
| CN116632823A (en) | Short-term photovoltaic power prediction method based on power conversion model and multi-layer perceptron | |
| CN114240003A (en) | New energy output prediction method, system, storage medium and equipment | |
| CN110264006B (en) | Wind power probabilistic prediction method based on chaotic firefly algorithm and Bayesian network | |
| CN105277531B (en) | A kind of coal characteristic measuring method based on stepping | |
| CN109886314B (en) | Kitchen waste oil detection method and device based on PNN neural network | |
| CN109142251B (en) | LIBS quantitative analysis method of random forest auxiliary artificial neural network | |
| CN108627500A (en) | A kind of solid fuel measurement of caloric value method | |
| CN105913161A (en) | Method of acquiring maximum power point of photovoltaic system based on multi-objective optimization | |
| CN116029160B (en) | Method and system for constructing mapping model of defects and power generation efficiency loss of photovoltaic module | |
| CN110321528A (en) | A kind of Hyperspectral imaging heavy metal-polluted soil concentration evaluation method based on semi-supervised geographical space regression analysis | |
| CN115965177A (en) | Improved autoregressive error compensation wind power prediction method based on attention mechanism |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190326 |
|
| WD01 | Invention patent application deemed withdrawn after publication |