CN112016250A - Flue gas SCR denitration system bad data identification method - Google Patents
Flue gas SCR denitration system bad data identification method Download PDFInfo
- Publication number
- CN112016250A CN112016250A CN202010900256.3A CN202010900256A CN112016250A CN 112016250 A CN112016250 A CN 112016250A CN 202010900256 A CN202010900256 A CN 202010900256A CN 112016250 A CN112016250 A CN 112016250A
- Authority
- CN
- China
- Prior art keywords
- samples
- bad
- neural network
- data
- sample
- 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
- 238000000034 method Methods 0.000 title claims abstract description 55
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 title claims description 25
- 239000003546 flue gas Substances 0.000 title claims description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000012216 screening Methods 0.000 claims abstract description 13
- 238000005070 sampling Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 21
- 230000005284 excitation Effects 0.000 claims description 9
- 238000013461 design Methods 0.000 claims description 8
- 239000003054 catalyst Substances 0.000 claims description 5
- 238000012886 linear function Methods 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 2
- 238000011478 gradient descent method Methods 0.000 claims 3
- 238000002939 conjugate gradient method Methods 0.000 claims 2
- 238000012546 transfer Methods 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 5
- 238000007689 inspection Methods 0.000 abstract 2
- MWUXSHHQAYIFBG-UHFFFAOYSA-N Nitric oxide Chemical compound O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 description 9
- 238000012795 verification Methods 0.000 description 5
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 229910021529 ammonia Inorganic materials 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000007789 gas Substances 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 238000003723 Smelting Methods 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 238000003915 air pollution Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010531 catalytic reduction reaction Methods 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Exhaust Gas Treatment By Means Of Catalyst (AREA)
Abstract
The invention provides a method for predicting and identifying bad data based on a BP neural network. The method comprises the steps of constructing a BP neural network and carrying out learning training on an original data sample obtained from an SCR denitration system, then randomly sampling the original sample for prediction and inspection, calculating the relative error between a predicted value and a true value of the BP neural network of the sample, screening the sample with a larger relative error, and adding the sample into a bad sample set. And repeatedly executing the process until certain requirements are met. And counting the occurrence frequency of the samples with the same value in the bad sample set, and marking the samples with the occurrence frequency exceeding a preset value as bad samples. Rejecting bad samples to reconstruct and train a BP neural network, inspecting the rejected bad samples, determining the bad samples as bad data if errors exceed a large value, and resolving the samples as normal data if the errors are smaller than a small value, and reconstructing the neural network for re-inspection under other conditions until all bad samples are verified, so that the accuracy of identifying the bad samples is ensured.
Description
Technical Field
The invention belongs to the field of nitrogen oxide treatment, and particularly relates to a method for identifying bad data of a flue gas SCR denitration system.
Background
With the development of economic society, the serious problem of air pollution caused by the emission of nitrogen oxides is gradually paid attention at home and abroad, the emission control of flue gas pollutants in coal-fired power plants, steel smelting, cement processing and other industries is increasingly strict, and the removal of the nitrogen oxides in the flue gas is increasingly popularized. At present, Selective Catalytic Reduction (SCR) is generally adopted at home and abroad to remove the flue gas so as to meet the requirement of environmental protection and emission. A large amount of data including system composition, catalyst design, flue gas operation conditions and the like can be formed in the design and operation processes of the SCR denitration system, the design of the SCR denitration system can be assisted through sampling, collecting and analyzing the data, and meanwhile, the method has important significance for stable operation and optimization of the SCR denitration system. However, due to uncontrollable factors existing in the processes of data detection, statistics, transmission and the like, some bad data often exist in a sample, and the accuracy of an analysis result is seriously influenced. Therefore, how to identify and reject the bad data is an inevitable problem in the optimization design of the SCR denitration system.
The traditional bad data identification method mainly comprises a zero residue method, an innovation graph method, an estimation identification method, a non-quadratic criterion method and the like, and the traditional bad data identification method has certain limitations and is mostly applied to the field of power systems. With the development of the fields of data mining technology, machine learning and artificial intelligence, bad data identification has made new progress. The chinese patent application CN201810059612.6 and the chinese patent application CN201811440318.6 respectively propose methods for identifying bad data of state estimation based on a BP neural network and an improved BP neural network, and improve the accuracy of identifying bad data in state estimation of a power system. The development of the SCR denitration technology is started late, and reports about an identification method of bad data of an SCR denitration system are not found at present.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a method for identifying bad data of a flue gas SCR denitration system.
The method comprises the steps of firstly obtaining an original data sample, randomly sampling a trained BP neural network from the original sample, then pre-screening a bad sample by utilizing a neural network prediction result, establishing a bad sample set by repeatedly executing the operations, then carrying out statistical analysis on the bad sample set to determine the bad sample, then removing the bad sample from the original sample, reconstructing the trained BP neural network, and utilizing the reconstructed BP neural network to test the bad sample to prevent error identification, thereby improving the accuracy of the neural network in identifying the bad data.
In order to achieve the above objects and achieve the above technical effects, the present invention is implemented by the following technical solutions: a flue gas SCR denitration system bad data identification method comprises the following steps:
the method comprises the following steps of firstly, obtaining an original data sample of an SCR denitration system;
establishing a BP neural network model, wherein the main contents comprise a BP neural network topological structure, an excitation function and a learning training algorithm;
randomly selecting a certain number of samples from original data samples as training samples to perform BP neural network training, using the rest samples as prediction samples to perform BP neural network prediction, and calculating the relative error between a predicted value and a true value; the certain number of samples refers to 70% -95% of the original sample capacity;
screening samples with relative errors larger than a preset value, adding the samples into a bad sample set, and repeatedly executing the step three and the step four to a certain number of times;
counting the occurrence frequency of the same samples in the bad sample set, and screening the samples with the occurrence frequency exceeding a preset value to be marked as bad samples;
removing marked bad samples, reconstructing and training a BP neural network aiming at the optimized samples, predicting the marked bad samples by using the trained neural network, and calculating relative errors;
and step seven, screening samples with prediction errors exceeding a limited range to determine bad data, correcting the samples with the prediction errors smaller than a given value to non-bad data, repeating the step six and the step seven for at least 3 times aiming at the residual samples, and determining all the residual samples to be the bad data.
Preferably, the identification method for the bad data of the flue gas SCR denitration system comprises a second step of constructing a three-layer BP neural network structure, determining the number of nodes of an input layer and an output layer according to an original data sample, and determining the number of nodes of a hidden layer according to the number of nodes of the input layer and the output layer.
Preferably, the identification method for the bad data of the flue gas SCR denitration system comprises a second step of adopting a hyperbolic tangent (Tanh) function as a hidden layer excitation function
Sigmoid function
Or ReLU function
One of them, the output layer excitation function adopts a linear function
。
Preferably, the identification method for the bad data of the flue gas SCR denitration system comprises a step II of learning and training an algorithm by adopting a Levenberg-Marquardt method.
Preferably, in the method for identifying the bad data of the flue gas SCR denitration system, the preset value and the certain times can be fixed values or determined according to actual prediction errors and training sample capacity. For example: the preset value comprises 20-50% or more than 50%, and the certain times comprise 50-100 times or more than 100 times.
Preferably, in the fifth step of the method for identifying the bad data of the flue gas SCR denitration system, the preset value may be a fixed value or determined according to the frequency distribution of the bad samples. For example: the preset value is 3.
Preferably, in the seventh step of the method for identifying the bad data of the flue gas SCR denitration system, the limited range and the given value may be fixed values or determined according to an error range of an actual prediction result. For example: the limit range is 50%, and the given value is 10%.
The invention has the beneficial effects that:
at present, denitration data can only be analyzed in a manual verification mode, and the method has the defects of low efficiency and certain experience required for workers.
The invention provides a method for identifying bad data of an SCR denitration system for the first time. The concrete advantages are as follows: firstly, under the condition that bad data is not known, training neural networks are constructed by randomly selecting training samples, then, the rest samples are predicted, and a bad sample set is constructed according to errors, so that the subjectivity of manual selection is avoided; secondly, the omission of bad samples is prevented through repeated construction, training, prediction and identification of a neural network; thirdly, screening samples with high frequency in the bad sample set as bad samples, thereby overcoming the defect that the bad samples are identified by mistake due to possible local optimization of the BP neural network; fourthly, bad data correction is carried out by optimizing sample reconstruction and training the BP neural network, and therefore the neural network in the identification process is prevented from utilizing a training set containing bad data, and therefore error identification of the bad data is prevented.
Drawings
FIG. 1 is a diagram of a BP neural network topology with a single hidden layer;
FIG. 2 is a schematic flow chart of the method of the present invention.
Detailed Description
The invention is further described with reference to the following figures and specific embodiments. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
In this embodiment, the volume design data of 200 catalysts in the SCR denitration process of the coal-fired power plant is taken as an example, and the method of the present invention is used to detect the bad data in these samples.
Example 1
A method for identifying bad data of a flue gas SCR denitration system, as shown in fig. 1 and 2, comprises the following steps:
step one, obtaining original data samples, wherein each sample comprises unit capacity, smoke gas amount and inlet NOxConcentration, outlet NOxConcentration, flue gas temperature, ash concentration, design denitration efficiency, SO2Content, SO2The conversion, ammonia slip concentration and total volume were 11 parameters total, with a total sample volume of 200.
Step two, constructing a three-layer neural network structure shown in fig. 1, taking the first 10 parameters in an original data sample as input parameters, taking the total volume of a catalyst as output parameters, and calculating and selecting a proper number of nodes of a hidden layer by using a formula 1; a hyperbolic tangent function is adopted as an excitation function of the hidden layer, and a linear function is adopted as an excitation function of the output layer; BP neural network learning training is carried out by adopting a Levenberg-Marquardt method.
And step three, randomly selecting 90% of samples from original samples by adopting a random number method as training samples to perform learning training on the constructed BP neural network, predicting the rest samples as prediction samples, and calculating the relative error of the prediction samples.
And step four, screening samples with the relative error larger than 50% and adding the samples into the bad sample set, and repeatedly executing the step three and the step four for 100 times.
And step five, counting the occurrence frequency of the same bad samples in the bad sample set, and screening the samples with the occurrence frequency exceeding 3 to mark as the bad samples.
And step six, removing the marked bad samples from the original data samples, performing learning training on the neural network again, predicting the marked bad samples by using the trained neural network, and calculating relative errors.
Selecting samples with relative errors larger than 50% to determine as bad data, and selecting samples with relative errors smaller than 10% to correct as non-bad data; and (4) repeating the sixth step and the seventh step for 4 times aiming at the bad samples with the relative error of 10-50%, and determining all the samples with the error still in the range as the bad data.
Through the operation of the steps, 3 groups of bad samples are identified, theoretical analysis and verification are carried out on the original samples in a manual verification mode, the original samples are found to contain 4 groups of bad data, all the bad samples identified by the method are the bad data, and therefore the identification rate of the bad data identified by the method is 75%, and the accuracy rate is 100%.
Example 2
A method for identifying bad data of a flue gas SCR denitration system, as shown in fig. 1 and 2, comprises the following steps:
step one, obtaining original data samples, wherein each sample comprises unit capacity, smoke gas amount and inlet NOxConcentration, outlet NOxConcentration, flue gas temperature, ash concentration, design denitration efficiency, SO2Content, SO2The conversion, ammonia slip concentration and total volume were 11 parameters total, with a total sample volume of 200.
Step two, constructing a three-layer neural network structure shown in fig. 1, taking the first 10 parameters in an original data sample as input parameters, taking the total volume of a catalyst as output parameters, and calculating and selecting a proper number of nodes of a hidden layer by using a formula 1; a hyperbolic tangent function is adopted as an excitation function of the hidden layer, and a linear function is adopted as an excitation function of the output layer; BP neural network learning training is carried out by adopting a Levenberg-Marquardt method.
And step three, selecting 80% of samples from original samples randomly by a random number method as training samples to perform learning training on the constructed BP neural network, predicting the rest samples as prediction samples, and calculating the relative error of the prediction samples.
And step four, screening samples with the relative error larger than 40% and adding the samples into the bad sample set, and repeatedly executing the step three and the step four for 100 times.
And step five, counting the occurrence frequency of the same bad samples in the bad sample set, and screening the samples with the occurrence frequency exceeding 3 to mark as the bad samples.
And step six, removing the marked bad samples from the original data samples, performing learning training on the neural network again, predicting the marked bad samples by using the trained neural network, and calculating relative errors.
Selecting samples with relative errors larger than 50% to determine as bad data, and selecting samples with relative errors smaller than 10% to correct as non-bad data; and (4) repeating the sixth step and the seventh step for 4 times aiming at the bad samples with the relative error of 10-50%, and determining all the samples with the error still in the range as the bad data.
Through the operation of the steps, 4 groups of bad samples are identified, theoretical analysis and verification are carried out on the original samples in a manual verification mode, the original samples are found to contain 4 groups of bad data, all the bad samples identified by the method are the bad data, and therefore the identification rate of the bad data identified by the method is 100%, and the accuracy rate is 100%.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention are intended to be included in the scope of the present invention.
Claims (6)
1. A flue gas SCR denitration system bad data identification method is characterized by comprising the following steps:
the method comprises the following steps of firstly, obtaining an original data sample of an SCR denitration system;
step two, establishing a BP neural network model, including a BP neural network topological structure, an excitation function and a learning training algorithm;
randomly selecting samples with the quantity of 70% -95% of the original sample capacity from the original data samples in the step one as training samples, carrying out BP neural network training, taking the rest samples as prediction samples to carry out BP neural network prediction, and calculating the relative error between a predicted value and a true value;
screening samples with relative errors larger than 20%, adding the samples into a bad sample set, and repeatedly executing the step three and the step four to more than or equal to 50 times;
counting the occurrence frequency of the same samples in the bad sample set, and screening samples with the occurrence frequency exceeding 2 to mark as bad samples;
removing the marked bad samples, reconstructing and training a BP neural network aiming at the optimized samples, predicting the marked bad samples by using the trained neural network, and calculating relative errors;
and step seven, screening the samples with the relative error of more than 10 percent to determine the samples as bad data, correcting the samples with the relative error of less than 5 percent to be non-bad data, and repeating the step six and the step seven for at least 3 times aiming at the rest samples to determine all the rest samples as the bad data.
2. The method for identifying the bad data of the flue gas SCR denitration system according to claim 1, wherein the BP neural network topology structure of the second step comprises a hidden layer, an input layer, an output layer and the number of nodes of each layer; the number of nodes of the input layer and the output layer is determined by the original data sample, and the number of nodes of the hidden layer is calculated according to the following formula:
in the formula:qin order to imply the number of layer nodes,mas the number of nodes of the output layer,nin order to input the number of nodes of the layer,vis constant and satisfies 1<v<10。
3. The method for identifying the bad data of the flue gas SCR denitration system according to claim 1, wherein the excitation function of the second step comprises a transfer function of a hidden layer and an output layer, the output function of the hidden layer adopts a Sigmoid function, a hyperbolic tangent function or a ReLU function, the output layer function adopts a linear function, and the function expression is as follows:
sigmoid function:
；
hyperbolic tangent function:
；
ReLU function:
；
linear function:
。
4. the method for identifying the bad data of the flue gas SCR denitration system as claimed in claim 1, wherein the learning training algorithm of the second step comprises at least one of a gradient descent method, a momentum gradient descent method, a self-adaptive learning gradient descent method, a Fletcher-Reeves conjugate gradient method, a Powell-Belle conjugate gradient method, a quasi-Newton algorithm, a one-step secant algorithm and a Levenberg-Marquardt method.
5. The method for identifying the bad data of the flue gas SCR denitration system of claim 1, wherein the raw data of the SCR denitration system comprises design parameters of the denitration system, catalyst parameters or historical operation data of the system.
6. The method for identifying the bad data of the flue gas SCR denitration system according to claim 1, wherein in the third step, the random sampling method comprises at least one of a draw method and a random number method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010900256.3A CN112016250A (en) | 2020-08-31 | 2020-08-31 | Flue gas SCR denitration system bad data identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010900256.3A CN112016250A (en) | 2020-08-31 | 2020-08-31 | Flue gas SCR denitration system bad data identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112016250A true CN112016250A (en) | 2020-12-01 |
Family
ID=73515254
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010900256.3A Pending CN112016250A (en) | 2020-08-31 | 2020-08-31 | Flue gas SCR denitration system bad data identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112016250A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016248A (en) * | 2020-08-31 | 2020-12-01 | 华北电力大学 | Elman neural network based SCR denitration system bad data identification method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133721A (en) * | 2017-04-17 | 2017-09-05 | 国网江苏省电力公司 | Power distribution network based on gray theory measures bad data relation factor analysis method |
| US20180046924A1 (en) * | 2015-08-31 | 2018-02-15 | Guangzhou Institute Of Energy Conversion, Chinese Academy Of Sciences | Whole-life-cycle power output classification prediction system for photovoltaic systems |
| CN108491404A (en) * | 2018-01-22 | 2018-09-04 | 国电南瑞科技股份有限公司 | A kind of state estimation bad data recognition method based on BP neural network |
| CN109376850A (en) * | 2018-11-29 | 2019-02-22 | 国网辽宁省电力有限公司抚顺供电公司 | A kind of detection method based on bad data in improved BP state estimation |
| CN110435412A (en) * | 2019-08-26 | 2019-11-12 | 吉林大学 | One kind being based on the more aperture control methods of BP neural network commercial vehicle intelligence grid |
| CN110544047A (en) * | 2019-09-10 | 2019-12-06 | 东北电力大学 | Bad data identification method |
-
2020
- 2020-08-31 CN CN202010900256.3A patent/CN112016250A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180046924A1 (en) * | 2015-08-31 | 2018-02-15 | Guangzhou Institute Of Energy Conversion, Chinese Academy Of Sciences | Whole-life-cycle power output classification prediction system for photovoltaic systems |
| CN107133721A (en) * | 2017-04-17 | 2017-09-05 | 国网江苏省电力公司 | Power distribution network based on gray theory measures bad data relation factor analysis method |
| CN108491404A (en) * | 2018-01-22 | 2018-09-04 | 国电南瑞科技股份有限公司 | A kind of state estimation bad data recognition method based on BP neural network |
| CN109376850A (en) * | 2018-11-29 | 2019-02-22 | 国网辽宁省电力有限公司抚顺供电公司 | A kind of detection method based on bad data in improved BP state estimation |
| CN110435412A (en) * | 2019-08-26 | 2019-11-12 | 吉林大学 | One kind being based on the more aperture control methods of BP neural network commercial vehicle intelligence grid |
| CN110544047A (en) * | 2019-09-10 | 2019-12-06 | 东北电力大学 | Bad data identification method |
Non-Patent Citations (3)
| Title |
|---|
| H. SALEHFAR, R. ZHAO: "A neural network preestimation filter for bad-data detection and identification in power system state estimation", ELECTRIC POWER SYSTEMS RESEARCH, vol. 978, no. 2, pages 127 - 134 * |
| 唐诗洁，陆强，曲艳超，任翠涛，杨勇平: "基于遗传算法优化BP 神经网络的 SCR 脱硝系统催化剂体积设计", 发电技术, vol. 978, no. 3, pages 247 - 13 * |
| 梁海文，周 栋: "基于流程模拟与Ｅｌｍａｎ神经网络的燃煤电厂 ＣＯ2 吸收系统动态数据检验方法", 华东电力, vol. 38, no. 7, pages 1087 - 1090 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016248A (en) * | 2020-08-31 | 2020-12-01 | 华北电力大学 | Elman neural network based SCR denitration system bad data identification method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109978048B (en) | Fault analysis and diagnosis method for slurry circulating pump of desulfurizing tower | |
| CN110716512A (en) | Environmental protection equipment performance prediction method based on coal-fired power plant operation data | |
| CN110580936B (en) | Method and system for predicting service life of medium-low temperature SCR denitration catalyst | |
| CN113792988B (en) | Enterprise online monitoring data anomaly identification method | |
| CN113223628B (en) | Method for assisting in rapid screening of metal oxide SCR (selective catalytic reduction) catalyst based on machine learning technology | |
| CN107798416B (en) | Method for predicting water pollution treatment effect of newly-built industrial park based on Monte Carlo method | |
| CN112465239B (en) | Desulfurization system operation optimization method based on improved PSO-FCM algorithm | |
| CN112613237B (en) | CFB unit NOx emission concentration prediction method based on LSTM | |
| CN112016248A (en) | Elman neural network based SCR denitration system bad data identification method | |
| CN109740301A (en) | A kind of accounting method of the Gas Generator Set carbon emission amount based on BP neural network | |
| CN112016249A (en) | SCR denitration system bad data identification method based on optimized BP neural network | |
| CN112034801A (en) | Method, system, medium and terminal for calculating pollution discharge coefficient of total amount of volatile organic pollutants | |
| CN110135466A (en) | A kind of exceeded vehicle judgment method of pollutant emission and system | |
| CN112016250A (en) | Flue gas SCR denitration system bad data identification method | |
| CN116228017A (en) | Intelligent overall process control method based on enterprise environment-friendly big data aggregation | |
| CN115171798A (en) | Method for predicting low-temperature desulfurization and/or denitration performance of activated carbon based on Na content in coal | |
| CN118280472A (en) | Thermal power unit SCR reactor nitrogen oxide concentration prediction method based on snow melt optimizer | |
| CN112037106B (en) | Data anomaly analysis method based on feature cross-correlation and probability density | |
| CN115113519A (en) | Coal-gas co-combustion boiler denitration system outlet NO x Concentration early warning method | |
| CN112766790A (en) | Steel enterprise green degree evaluation method based on AHP-entropy | |
| CN110703701A (en) | Efficient data preprocessing method suitable for operating data of coal-fired power plant environment-friendly equipment | |
| CN112379035B (en) | System and method for evaluating performance of SCR (selective catalytic reduction) denitration device catalyst on line | |
| CN116312869A (en) | Method, device and system for predicting nitrogen oxides in catalytic cracking regenerated flue gas | |
| CN115310813A (en) | Environment-friendly treatment method and system for thermoelectric industry | |
| CN113947142A (en) | Method and system for predicting emission concentration of acid gas and storage medium |
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 |