CN109460403B - Real-time dynamic quantitative calibration method for ash blockage of air preheater - Google Patents
Real-time dynamic quantitative calibration method for ash blockage of air preheater Download PDFInfo
- Publication number
- CN109460403B CN109460403B CN201811569342.XA CN201811569342A CN109460403B CN 109460403 B CN109460403 B CN 109460403B CN 201811569342 A CN201811569342 A CN 201811569342A CN 109460403 B CN109460403 B CN 109460403B
- Authority
- CN
- China
- Prior art keywords
- air preheater
- sample
- model
- sample set
- delta
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E20/00—Combustion technologies with mitigation potential
- Y02E20/34—Indirect CO2mitigation, i.e. by acting on non CO2directly related matters of the process, e.g. pre-heating or heat recovery
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Air Supply (AREA)
- Gasification And Melting Of Waste (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
An air preheater ash blocking real-time dynamic quantitative calibration method relates to an air preheater operation condition monitoring method, in particular to data cleaning and identification of the ash blocking degree of an air preheater for modeling by a support vector machine, and belongs to the field of machine learning modeling. The invention obtains corresponding data under various operating conditions through the data input interface; a compressed neighbor method is adopted, samples with large information repetition degree are removed, and the samples are reduced; and obtaining a cleaned output sample D. Constructing a support vector machine model: obtaining a standard air preheater smoke pressure difference model; updating the ash blocking reference model; step five: calculating the ash blocking index: and setting a limit value of e according to actual conditions to correspond to different conditions of cleaning, early warning and alarming, and providing guidance suggestions for field operation and maintainers. The invention realizes the purposes of retaining the characteristic integrity of the sample, eliminating repeated information, realizing real-time quantitative monitoring of the ash blocking state, and effectively monitoring, diagnosing and quantitatively calibrating the ash blocking degree of the air preheater.
Description
Technical Field
The invention relates to a method for monitoring the running condition of an air preheater, in particular to the identification of the ash blocking degree of the air preheater for data cleaning and modeling of a support vector machine, and belongs to the field of machine learning modeling.
Background
With the development of science and technology, the researched data set is often characterized by high dimension and mass. This poses a challenge to data mining and data modeling of existing algorithms, and if high-dimensional massive data are processed directly by using them, it is often difficult to obtain a desired effect in practice due to limitations of objective conditions such as computation time and hardware devices. Aiming at the high-dimensional problem, feature extraction and selection are often adopted to realize dimensionality reduction; for the problem of large number of samples, a data cleaning method is often adopted. In order to eliminate redundant information in samples, many scholars have been continuously researching and discussing the problem of sample reduction and selection. Different selection strategies have been developed based on different considerations. Support Vector Machines (SVMs) were first proposed by cornna cortex and Vapnik, 1995, and show many unique advantages in solving small sample, nonlinear and high-dimensional pattern recognition, and can be generalized and applied to other Machine learning problems such as function fitting. In data modeling, the quality of a training sample is important, and the training sample is reduced, so that the modeling performance of a support vector machine can be effectively improved.
The operation condition of the air preheater of the power station affects the safe and economic operation of the unit, the ash blockage of the air preheater not only affects the safety of the operation of the boiler, so that the efficiency of the boiler is obviously reduced, the unit consumption of a fan is obviously increased, the temperature of exhaust smoke is increased, the temperature of the inlet smoke of a desulfurization system is overhigh when the temperature is serious, the MFT action of the desulfurization system is caused due to the high temperature of the outlet smoke, and even an anticorrosive coating in an absorption tower is damaged, so that the effective monitoring and the ash blockage prevention of the air preheater are very important.
In recent years, many scholars have studied the problems of prevention and inhibition of ash blockage of the air preheater from different angles, and have achieved certain results. However, the real-time monitoring, diagnosis and quantitative calibration of the ash blockage degree of the air preheater are less researched, and further research work needs to be carried out.
Disclosure of Invention
The invention aims to provide a real-time dynamic quantitative calibration method for ash blockage of an air preheater, which not only retains the characteristic integrity of a sample, but also eliminates repeated information, realizes the real-time quantitative monitoring of the ash blockage state, effectively monitors, diagnoses and quantitatively calibrates the ash blockage degree of the air preheater, and provides a reference model for a high-level function module of a power plant monitoring information system.
A real-time dynamic quantitative calibration method for ash blockage of an air preheater comprises the following steps:
the method comprises the following steps: obtaining the corresponding load P under various boiler operation conditions through the data input interface e Oxygen amount O 2 The pressure value (P) of the two front flue gases from the outlet of the hearth to the air preheater 1 ,P 2 ) Inlet and outlet flue gas pressure (P) of air preheater 3 ,P 4 ) And calculates the two pressure difference values ap 12 =P 1 -P 2 ,ΔP 34 =P 3 -P 4 ;
Step two: eliminating samples with large information repetition degree by adopting a CNN method, and reducing the samples; the similarity of information is represented by Euclidean distance x between samples j -x i Taking | | I as the basis of evaluation; the sample set x after the air preheater is just cleaned is set as { x ═ x 1 ,x 2 ,…x n },x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 Using the obtained data as input quantity of a compressed neighbor method; obtaining a cleaned output sample D through a compressed neighbor method;
the specific process is as follows:
step 21, with time series sample set x ═ { x ═ x 1 ,x 2 ,…x n In which x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 The initial input quantity is used as the input quantity;
step 22, locate the first sample x 1 Denoted as sample set D, and the sample set x is updated to x ═ x-x 1 ;
Step 23, respectively calculating the ith sample x i Euclidean distances from each sample in the updated sample set x and the updated sample set D, i being 2, 3.
Step 24, x obtained in step 23 i Marking the sample with the minimum Euclidean distance as s, judging the relation between s and D, and if so, judging whether the sample has the minimum Euclidean distance with the minimum Euclidean distance as sThen step 24.1 is carried out, otherwise step 24.2 is carried out;
step 24.1, the sample set D is unchanged, and the original sample set is updated again with x ═ x-x i If x is phi, outputting a sample D; if not, i is i +1, go to step 23;
step 24.2, update sample set D, D ═ D + x i Updating the original sample set x ═ x-x i If x is phi, outputting a sample D; if not, i is i +1, go to step 23;
step three: output sample D with step two, { P } e ,O 2 ,ΔP 12 As input samples for the support vector machine model, { Δ P } 34 Get it as the output sample of the training modelA standard air preheater smoke pressure difference model;
sampling new samples P in time series e ,O 2 ,ΔP 12 ,ΔP 34 } new Model output air preheater inlet and outlet flue gas pressure difference { delta P 34 } model The difference value delta (delta P) between the sample smoke pressure difference and the model smoke pressure difference 34 )=(ΔP 34new -ΔP 34model )/ΔP 34new Quantifying the ash blockage degree, delta (delta P), of the air preheater corresponding to the current working condition of the new sample 34 ) Larger indicates more serious ash blockage;
step four: after the ash blockage degree is increased, the output of the model is unstable in large fluctuation, and then the difference value delta (delta P) 34 ) The reflected ash blocking degree has deviation; therefore, the standard air preheater smoke pressure difference model needs to be updated in real time when the value is delta (delta P) 34 ) When m times of continuous occurrence are more than 0.2, updating the ash blocking reference model, turning to the step two, recording the updating times count of the model, wherein the original value is 1, the updating time count is 2, and the like;
step five: defining a calculation formula of the gray blocking index e of the ith sample:
and setting a limit value of e according to actual conditions to correspond to different conditions of cleaning, early warning and alarming, and providing guidance suggestions for field operation and maintainers.
By adopting the technical scheme, compared with the prior art, the invention has the following advantages:
1. the method is used as a means for preprocessing data, not only retains the characteristic integrity of the sample, but also eliminates repeated information.
2. By utilizing a combination method of compression neighbor and a support vector machine and updating the model for quantifying the ash blockage degree in time, the calibration value of the ash blockage degree can be dynamically obtained in real time, so that the real-time quantitative monitoring of the ash blockage state is realized.
3. The running performance of the air preheater is reflected more timely and accurately, the reduction of the running performance of the boiler caused by serious ash blockage of the air preheater can be prevented, and a new thought and a new method are provided for the optimization of the performance of the boiler and the monitoring of the state of the boiler.
4. And a referential model is provided for advanced function modules (such as operation optimization, state monitoring, fault diagnosis and the like) of the monitoring information system of the power plant.
Description of the drawings:
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a flow chart of the CNN algorithm.
The specific implementation mode is as follows:
the technical scheme of the invention is explained in detail by combining the drawings as follows:
as shown in fig. 1, the method for dynamically and quantitatively calibrating the ash blockage of the air preheater in real time comprises the following steps:
the method comprises the following steps: obtaining corresponding load P under various operation conditions via data input interface e Oxygen amount O 2 The pressure value (P) of the two front flue gases from the outlet of the hearth to the air preheater 1 ,P 2 ) Inlet and outlet flue gas pressure (P) of air preheater 3 ,P 4 ) And calculates the two pressure difference values ap 12 =P 1 -P 2 ,ΔP 34 =P 3 -P 4 。
Step two: and (3) eliminating samples with large information repetition degree by adopting a compressed neighbor method, and reducing the samples:
the similarity of information is represented by Euclidean distance x between samples j -x i And | l is taken as the basis of evaluation. The sample set x after the air preheater is just cleaned is set as { x ═ x 1 ,x 2 ,...x n },x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 And the input quantity of the compression neighbor method is used. The compressed neighbor method utilizes the original sample set T to gradually generate a new sample subset D, so that D can still correctly classify each sample in T by the nearest neighbor method under the condition of reducing the number of samples. Compressed neighbor algorithm flow chart as shown in FIG. 2, cleaningThe latter output sample is D.
Step three: constructing a support vector machine model:
output sample D with step two, { P } e ,O 2 ,ΔP 12 As input samples for the support vector machine model, { Δ P } 34 Taking the smoke pressure difference as an output sample of the training model to obtain a standard air preheater smoke pressure difference model;
sampling new samples P in sequence according to time sequence e ,O 2 ,ΔP 12 ,ΔP 34 } new Model output smoke pressure difference [ delta P ] 34 } model The difference value (Δ (Δ P)) between the sample smoke pressure difference and the model smoke pressure difference 34 )=(ΔP 34new -ΔP 34model )/ΔP 34new ) Indicating the ash blockage degree, delta (delta P), of the air preheater corresponding to the current working condition of the new sample 34 ) Larger indicates more serious ash blockage.
Step four: updating a dust blocking reference model:
after the ash blockage degree is increased, the output of the model is unstable in large fluctuation, and then the difference value delta (delta P) 34 ) The reflected degree of ash blockage may deviate. Therefore, the standard air preheater smoke pressure difference model needs to be updated in real time when the value is delta (delta P) 34 ) And (3) when the number of the continuous m times is more than 0.2, updating the ash blocking reference model, turning to the step (2), and recording the number of times of model updating count (the number is 1 originally, the number is 2 when the model is updated once, and the like).
Step five: calculating the ash blocking index:
defining a calculation formula of the gray blocking index e of the ith sample:
and setting a limit value of e according to actual conditions to correspond to different conditions of cleaning, early warning and alarming, and providing guidance suggestions for field operation and maintainers.
The compressed neighbor method of the invention comprises the following steps:
step 21, with time series sample set x ═ { x ═ x 1 ,x 2 ,...x n In which x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 The initial input quantity is used as the input quantity;
step 22, locate the first sample x 1 Denoted as sample set D, and the sample set x is updated to x ═ x-x 1 ;
Step 23, respectively calculating the ith sample x i Euclidean distances from each sample in the updated sample set x and the updated sample set D, i being 2, 3.
Step 24, x obtained in step 23 i Marking the sample with the minimum Euclidean distance as s, judging the relation between s and D, if so, judging the relation between s and DThen step 24.1 is carried out, otherwise step 24.2 is carried out;
step 24.1, the sample set D is unchanged, and the original sample set is updated again with x ═ x-x i If x is phi, outputting a sample D; if not, i is i +1, go to step 23;
step 24.2, update the sample set D, D ═ D + x i Updating the original sample set x ═ x-x i If x is phi, outputting a sample D; if not, i is i +1, go to step 23;
the output sample set D is a sample after only CNN distance compression.
The following data come from the performance parameters of a certain power station from 1 month and 1 day 0 point to 5 months and 1 day 0 point in 2014, the sampling interval is 3 minutes, and the load, the oxygen content, the furnace outlet pressure, the SCR inlet pressure, the air preheater inlet flue gas pressure and the air preheater outlet flue gas pressure are 6. The whole process mainly comprises core modules of input data preprocessing, original data sample compression, support vector machine modeling, quantification of ash blocking degree, monitoring management and the like. The detailed flow is shown in figure 1:
1. the performance parameters are obtained through a DCS database, a data serial interface and a data acquisition system, and simple difference processing is carried out to obtain x ═ P e ,O 2 ,ΔP 12 ,ΔP 34 -samples;
2. and after the data enters an input data preprocessing link, removing unstable working condition data according to the correlation of each parameter and load or the characteristics of the data to obtain an original stable working condition sample.
3. And the data from 1 month 1 to 1 month 5 days in the original steady-state working condition sample represents the sample with the lightest ash plugging degree of the air preheater, and the boiler operation working condition span is large, so that the data is used as the input of a compression neighbor model to obtain a reduced ash plugging data sample of the air preheater.
4. Using the samples as training samples for modeling of the support vector machine, { P } e ,O 2 ,ΔP 12 Is input, { Δ P 34 And (6) outputting to obtain a standard air preheater smoke pressure difference model.
5. Sampling new samples P in time series e ,O 2 ,ΔP 12 ,ΔP 34 } new Model output smoke pressure difference [ delta P ] 34 } model The difference value (Δ (Δ P)) between the sample smoke pressure difference and the model smoke pressure difference 34 )=(ΔP 34new -ΔP 34model )/ΔP 34new )。
6. When Δ (Δ P) 34 ) And (3) when the number of times of continuous occurrence is more than 0.2, updating the ash blocking reference model, using the data sample of the current time period, switching to the step (3), and recording the number of times of model updating count (the number is 1 originally, the number is 2 when the number is updated once, and the like).
7. And calculating the ash blocking index e in real time, and setting a cleaning range of 0-0.2 and an early warning range of 0.2-0.5. Emergency alert range > 0.5.
And integrating the ash plugging index quantitative calibration library for the ash plugging index to be used as a basis for monitoring and diagnosing the ash plugging state of the air preheater.
Claims (1)
1. The method for dynamically and quantitatively calibrating the ash blockage of the air preheater in real time is characterized by comprising the following steps of:
the method comprises the following steps: obtaining the corresponding load P under various boiler operation conditions through the data input interface e Oxygen amount O 2 The pressure value (P) of the two front flue gases from the outlet of the hearth to the air preheater 1 ,P 2 ) Inlet and outlet flue gas pressure (P) of air preheater 3 ,P 4 ) And calculates the two pressure difference values ap 12 =P 1 -P 2 ,ΔP 34 =P 3 -P 4 ;
Step two: eliminating samples with large information repetition degree by adopting a CNN method, and reducing the samples; the similarity of information is represented by Euclidean distance x between samples j -x i Taking | | I as the basis of evaluation; the sample set x after the air preheater is just cleaned is set as { x ═ x 1 ,x 2 ,...x n },x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 Using the obtained data as input quantity of a compressed neighbor method; obtaining a cleaned output sample set D through a compressed neighbor method;
the compression neighbor method comprises the following specific processes:
step 21, with time series sample set x ═ { x ═ x 1 ,x 2 ,...x n In which x i ={P e ,O 2 ,ΔP 12 ,ΔP 34 The initial input quantity is used as the input quantity;
step 22, locate the first sample x 1 Denoted as sample set D, and the sample set x is updated to x ═ x-x 1 ;
Step 23, respectively calculating the ith sample x i Euclidean distances from each sample in the updated sample set x and the updated sample set D, i being 2, 3.
Step 24, mixing the obtained product in step 23 with x i Marking the sample with the minimum Euclidean distance as s, judging the relation between s and D, and if so, judging the relation between s and DThen step 24.1 is carried out, otherwise step 24.2 is carried out;
step 24.1, the sample set D is unchanged, and the original sample set is updated again with x ═ x-x i If x is phi, outputting a sample set D; if not, i is i +1, go to step 23;
step 24.2, update the sample set D, D ═ D + x i Updating the original sample set x ═ x-x i If x is phi, outputting a sample set D; if not, i is i +1, go to step 23;
step three: in the output sample set D in step two, { P } e ,O 2 ,ΔP 12 As input samples for the support vector machine model, { Δ P } 34 Taking the smoke pressure difference as an output sample of the training model to obtain a standard air preheater smoke pressure difference model;
sampling new samples P in time series e ,O 2 ,ΔP 12 ,ΔP 34 } new Model output air preheater inlet and outlet flue gas pressure difference { delta P 34 } model The difference value delta (delta P) between the sample smoke pressure difference and the model smoke pressure difference 34 )=(ΔP 34new -ΔP 34model )/ΔP 34new Quantifying the ash blockage degree, delta (delta P), of the air preheater of the new sample corresponding to the current working condition 34 ) Larger indicates more serious ash blockage;
step four: after the ash blockage degree is increased, the output of the model is unstable in large fluctuation, and then the difference value delta (delta P) 34 ) The reflected ash blocking degree has deviation; therefore, the standard air preheater smoke pressure difference model needs to be updated in real time when the value is delta (delta P) 34 ) When m times of continuous occurrence are more than 0.2, updating the ash blocking reference model, turning to the step two, recording the updating times count of the model, wherein the original value is 1, the updating time count is 2, and the like;
step five: defining a calculation formula of the gray blocking index e of the ith sample:
according to the actual situation, the limit value of e is set to correspond to different situations of cleaning, early warning and alarming, and guidance suggestions are provided for field operation and maintenance personnel.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811569342.XA CN109460403B (en) | 2018-12-21 | 2018-12-21 | Real-time dynamic quantitative calibration method for ash blockage of air preheater |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811569342.XA CN109460403B (en) | 2018-12-21 | 2018-12-21 | Real-time dynamic quantitative calibration method for ash blockage of air preheater |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109460403A CN109460403A (en) | 2019-03-12 |
CN109460403B true CN109460403B (en) | 2022-08-16 |
Family
ID=65614126
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811569342.XA Active CN109460403B (en) | 2018-12-21 | 2018-12-21 | Real-time dynamic quantitative calibration method for ash blockage of air preheater |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109460403B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109460403B (en) * | 2018-12-21 | 2022-08-16 | 南京工程学院 | Real-time dynamic quantitative calibration method for ash blockage of air preheater |
CN110132342A (en) * | 2019-05-22 | 2019-08-16 | 江苏方天电力技术有限公司 | Power station air preheater dust stratification monitoring method based on Clustering |
CN111554077B (en) * | 2020-05-14 | 2023-04-14 | 润电能源科学技术有限公司 | Blockage early warning method, system, equipment and storage medium of air pre-heater |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104215546A (en) * | 2014-08-15 | 2014-12-17 | 国家电网公司 | Monitoring system for ash fouling in power station boiler air preheater and working method of system |
CN105930629A (en) * | 2016-07-13 | 2016-09-07 | 大唐南京环保科技有限责任公司 | On-line fault diagnosis method based on massive amounts of operating data |
CN109460403A (en) * | 2018-12-21 | 2019-03-12 | 南京工程学院 | The real-time dynamic quantitative scaling method of the stifled ash of air preheater |
-
2018
- 2018-12-21 CN CN201811569342.XA patent/CN109460403B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104215546A (en) * | 2014-08-15 | 2014-12-17 | 国家电网公司 | Monitoring system for ash fouling in power station boiler air preheater and working method of system |
CN105930629A (en) * | 2016-07-13 | 2016-09-07 | 大唐南京环保科技有限责任公司 | On-line fault diagnosis method based on massive amounts of operating data |
CN109460403A (en) * | 2018-12-21 | 2019-03-12 | 南京工程学院 | The real-time dynamic quantitative scaling method of the stifled ash of air preheater |
Non-Patent Citations (2)
Title |
---|
基于烟气压差的空气预热器灰污监测模型在计算机中的应用;谢婷等;《安徽冶金科技职业学院学报》;20121031;第22卷(第4期);第13-16页 * |
基于烟气差压的回转式空预器积灰实时监测研究;蓝晓村等;《机电信息》;20041231(第19期);第4-5页 * |
Also Published As
Publication number | Publication date |
---|---|
CN109460403A (en) | 2019-03-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109460403B (en) | Real-time dynamic quantitative calibration method for ash blockage of air preheater | |
US20190015806A1 (en) | Catalyst Transfer Pipe Plug Detection | |
CN111752147B (en) | Multi-working-condition process monitoring method with continuous learning capability and improved PCA (principal component analysis) | |
Perry et al. | Estimating the change point of a Poisson rate parameter with a linear trend disturbance | |
US8768664B2 (en) | Predictive emissions monitoring using a statistical hybrid model | |
US20120278026A1 (en) | Predictive emissions monitoring system and method | |
CN108446529B (en) | Organic Rankine cycle system fault detection method based on generalized mutual entropy-DPCA algorithm | |
CN111159844B (en) | Abnormity detection method for exhaust temperature of gas turbine of power station | |
CN110413601B (en) | Generator data screening method based on Gauss naive Bayes and prediction error method | |
CN113011010A (en) | Boiler fault diagnosis method and diagnosis system based on structural mechanism and operation data | |
CN113433911B (en) | Accurate control system and method for ammonia spraying of denitration device based on accurate concentration prediction | |
CN111522230B (en) | MIMO different factor tight format model-free control method | |
Ruiz et al. | Fault diagnosis support system for complex chemical plants | |
US20180307997A1 (en) | Systems and methods for improved quantification of uncertainty in turbomachinery | |
CN113792988A (en) | Online monitoring data anomaly identification method for enterprise | |
CN111259523B (en) | Process monitoring method based on KPCA-CVA model and random algorithm | |
Kini et al. | Kantorovich distance based fault detection scheme for non-linear processes | |
CN112418306B (en) | Gas turbine compressor fault early warning method based on LSTM-SVM | |
CN111648992B (en) | Gas turbine compressor fault identification early warning method | |
CN111582639B (en) | Denitration system running state evaluation system | |
Chen et al. | Online support vector based gas emission prediction system for generation power plant | |
CN110762081A (en) | Large hydraulic machine fault diagnosis method and device based on HSMM-SVM model | |
CN115496188A (en) | Coal mill fault early warning method based on deep learning convolutional neural network | |
CN114742300B (en) | Denitration device boiler flue resistance early warning method and system | |
CN115146703A (en) | Multi-model-based on-line monitoring method for wall temperature of furnace chamber under wide load of coal-fired unit |
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 |