CN114186708A - Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method - Google Patents
Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method Download PDFInfo
- Publication number
- CN114186708A CN114186708A CN202111190859.XA CN202111190859A CN114186708A CN 114186708 A CN114186708 A CN 114186708A CN 202111190859 A CN202111190859 A CN 202111190859A CN 114186708 A CN114186708 A CN 114186708A
- Authority
- CN
- China
- Prior art keywords
- elm
- pso
- fluidized bed
- circulating fluidized
- bed unit
- 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 39
- 239000002245 particle Substances 0.000 claims abstract description 49
- 239000003245 coal Substances 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 19
- 239000003344 environmental pollutant Substances 0.000 claims abstract description 15
- 231100000719 pollutant Toxicity 0.000 claims abstract description 15
- 235000019738 Limestone Nutrition 0.000 claims abstract description 14
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims abstract description 14
- 239000006028 limestone Substances 0.000 claims abstract description 14
- 229910052760 oxygen Inorganic materials 0.000 claims abstract description 14
- 239000001301 oxygen Substances 0.000 claims abstract description 14
- 230000006870 function Effects 0.000 claims abstract description 8
- 210000002569 neuron Anatomy 0.000 claims description 20
- 230000003044 adaptive effect Effects 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 11
- 238000005457 optimization Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 7
- 238000002485 combustion reaction Methods 0.000 claims description 6
- 230000002068 genetic effect Effects 0.000 claims description 5
- 238000012795 verification Methods 0.000 claims description 4
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 claims description 3
- NINIDFKCEFEMDL-UHFFFAOYSA-N Sulfur Chemical compound [S] NINIDFKCEFEMDL-UHFFFAOYSA-N 0.000 claims description 3
- 239000011575 calcium Substances 0.000 claims description 3
- 229910052791 calcium Inorganic materials 0.000 claims description 3
- 229910052717 sulfur Inorganic materials 0.000 claims description 3
- 239000011593 sulfur Substances 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 claims 1
- RAHZWNYVWXNFOC-UHFFFAOYSA-N sulfur dioxide Inorganic materials O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 abstract description 31
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 abstract description 2
- 239000003546 flue gas Substances 0.000 abstract description 2
- 230000002349 favourable effect Effects 0.000 abstract 1
- 238000006477 desulfuration reaction Methods 0.000 description 5
- 230000023556 desulfurization Effects 0.000 description 5
- OSGAYBCDTDRGGQ-UHFFFAOYSA-L calcium sulfate Chemical group [Ca+2].[O-]S([O-])(=O)=O OSGAYBCDTDRGGQ-UHFFFAOYSA-L 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000033116 oxidation-reduction process Effects 0.000 description 2
- 230000001590 oxidative effect Effects 0.000 description 2
- 238000009628 steelmaking Methods 0.000 description 2
- 238000010977 unit operation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000356 contaminant Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Biology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a circulating fluidized bed unit SO based on PSO-ELM, belonging to the technical field of pollutant control of thermal power generating units2And (3) a concentration prediction method. This SO2The concentration prediction method is based on SO2In the generation and removal processes, primary air quantity, secondary air quantity, oxygen quantity, coal quantity, load, bed temperature and limestone feeding quantity are selected as the SO based on the PSO-ELM circulating fluidized bed unit2Inputting a concentration prediction model; on the basis of the extreme learning machine model, a particle swarm algorithm is adopted, the minimum square difference is used as a target function, and the connection weight and the threshold parameter between the input layer and the hidden layer of the extreme learning machine are optimized. The method is favorable for solving the problems of over-standard sulfur dioxide concentration, difficult control and the like caused by difficult modeling of the sulfur dioxide in the original flue gas, and the accuracy of the model is improved.
Description
Technical Field
The invention belongs to the technical field of pollutant control of thermal power generating units, and particularly relates to a circulating fluidized bed unit SO based on PSO-ELM2And (3) a concentration prediction method.
Background
The circulating fluidized bed unit has the advantages of simple desulfurization equipment and low cost, and can realize in-furnace desulfurization, so the circulating fluidized bed unit is considered to be one of the most promising combustion technologies. However, with the stricter pollutant discharge index of the coal-fired power plant in China and the gradual improvement of environmental protection, certain difficulty is brought to the pollutant control of the dynamic process of the circulating fluidized bed unit, and simultaneously, the load control difficulty of the CFB unit is increased because the furnace state of the dynamic combustion process of the circulating fluidized bed is more complex and has the characteristics of large delay, strong coupling, nonlinearity, time variation and the like. Currently for SO in CFB unit dynamic process2The modeling research of concentration is less, and the pollutant generation condition is comparatively complicated in the CFB unit combustion process, and the emission control of pollutant is unstable, causes the instantaneous excessive easily, influences the economic nature of unit operation, can make the variable load capacity of unit receive the restriction of pollutant emission simultaneously.
The patent CN106096788 discloses a converter steelmaking process cost control method and system based on a PSO _ ELM neural network, wherein the method comprises the steps ofSelecting control parameters affecting cost; constructing a modeling sample set; obtaining a normalized sample set; constructing a feedforward neural network; training the ELM neural network parameters by adopting a PSO algorithm to obtain the neural network parameters; optimizing the model constructed by the PSO _ ELM neural network by using a genetic algorithm, acquiring the most value of the constructed model, and determining the optimal control parameter according to the most value of the constructed model; and determining the minimum cost value of the converter steelmaking process according to the comparison result of the optimal control parameter cost value and the minimum cost value in the modeling sample set. Solving the problem of SO2The pollutant emission concentration exceeds the standard and the control precision is not good,
the invention provides a raw flue gas SO based on PSO-ELM2The dynamic concentration modeling method is used for optimizing the hyper-parameters of the extreme learning machine by applying a particle swarm algorithm, and the model precision is improved.
Disclosure of Invention
The invention aims to provide a circulating fluidized bed unit SO based on PSO-ELM2The concentration prediction method is characterized by comprising the following steps:
s1, according to SO2The primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are selected as the SO based on the PSO-ELM circulating fluidized bed unit2Inputting a concentration prediction model;
s2, in a circulating fluidized bed unit SO based on PSO-ELM2In the process of establishing the concentration prediction model, considering that the CFB unit has large inertia and delay, and the coal quantity and the air quantity of the current input parameters may influence the long-time pollutant change in the future, therefore, the current time and the previous two sampling points of the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are introduced to be used as the SO based on the PSO-ELM circulating fluidized bed unit2Input of concentration prediction model, and SO of first two sampling points2The data of (a);
s3, after the input variable is determined, further determining the network model structure parameter of the ELM according to the input variable; the number of neurons in the hidden layer of the ELM is very important; a small number of neurons may cause large errors in the testing process, and a large number of neurons may cause overfitting and increase training time;
and S4, optimizing the connection weight and the threshold parameter between the input layer of the limit learning machine and the ELM hidden layer by utilizing a particle swarm optimization.
In the step S3, a small number of neurons may cause a large error in the test process, a large number of neurons may cause overfitting and increase the training time, and the number of iterations of the genetic algorithm also affects the SO based PSO-ELM circulating fluidized bed unit2Accuracy of the concentration prediction model. Too few iteration times can reduce the SO of the circulating fluidized bed unit based on the PSO-ELM2Accuracy of concentration prediction model, and excessive iteration times of the concentration prediction model2The accuracy of the concentration prediction model has little influence, but the SO of the circulating fluidized bed unit based on the PSO-ELM is increased2The complexity of the concentration prediction model needs to be selected according to the quantity of input data.
In the step S4, the model used by the particle swarm algorithm is an extreme learning machine model after particle swarm optimization.
Determination of input variables in said step S2, according to SO2The circulating fluidized bed unit has the capability of in-furnace desulfurization, so the limestone flow is considered as one of the input variables; SO generated during operation of CFB device2Mainly from the combustion of coal; furthermore, according to the results of the prior studies, the molar ratio of calcium to sulfur to SO2The amount of coal to be fed is important and therefore needs to be taken into account during the modeling process.
The step S5 calculating particle swarm optimization
(1) Initializing a particle swarm and endowing each particle with a random initial position and speed;
(2) calculating the adaptive value of each particle according to the fitness function;
(3) comparing the adaptive value of the current position of each particle with the adaptive value corresponding to the historical optimal position of each particle, and if the adaptive value of the current position is more in line with the requirement, selecting the current position for updating;
(4) comparing the fitness value of the current position of each particle with the global optimal fitness value, and if the fitness value of the current position is more in line with the requirement, updating the global optimal position by using the current position;
(5) updating the particle position and velocity;
(6) and if the end condition is met, ending the algorithm to obtain a global optimal position, namely a global optimal solution.
The PSO-ELM algorithm steps are described in detail as follows:
(1) determining a network structure of the ELM;
(2) initializing connection weight and threshold of ELM to form initial particle group Q (t)0) Randomly generating n particles;
(3) with Q (t)0) As a connection weight between input and output layers, and a threshold value of the ELM to train the neural network;
(4) calculating errors and fitness, reserving the optimal individual, and taking the optimal individual as the evolution target of the next generation; the fitness function is chosen as shown below,
wherein, is Δ VtTo verify the absolute error of the set of verification results, VtAn output that is a validation set;
(5) the particle group Q (t) obtains a better fitness value after updating the position and the speed of the particles;
(6) and (4) increasing the iteration number by 1 and returning to the step (3).
The circulating fluidized bed unit SO based on the PSO-ELM2Concentration prediction model parameter selection
In a circulating fluidized bed unit SO based on PSO-ELM2In the process of establishing the concentration prediction model, the CFB unit is considered to have large inertia and large delay, and the current input parameters of the coal quantity and the air quantity can possibly influence the long-time pollutant change in the future; therefore, the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are introduced at the current moment and two sampling points before, and two sampling points beforeSampling point SO2The data of (A) are used as models based on a PSO-ELM circulating fluidized bed unit SO2And inputting a concentration prediction model.
The beneficial effects of the invention include:
(1) according to the mechanism analysis, a variable related to the emission concentration of sulfur dioxide of the circulating fluidized bed unit is selected.
(2) On the basis, a sulfur dioxide emission concentration model of the circulating fluidized bed unit based on the extreme learning machine is established.
(3) And optimizing the connection weight and the threshold parameter between the input layer and the hidden layer of the limit learning machine by adopting a particle swarm algorithm and taking the minimum square difference as a target function.
Drawings
FIG. 1 is a flow chart of the prediction of sulfur dioxide concentration in a circulating fluidized bed unit.
Detailed Description
The invention provides a circulating fluidized bed unit SO based on PSO-ELM2The concentration prediction method, as shown in fig. 1, includes the following steps:
s1, according to SO2The primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are selected as the SO based on the PSO-ELM circulating fluidized bed unit2Inputting a concentration prediction model;
s2, in a circulating fluidized bed unit SO based on PSO-ELM2In the process of establishing the concentration prediction model, considering that the CFB unit has large inertia and delay, and the coal quantity and the air quantity of the current input parameters may influence the long-time pollutant change in the future, the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are introduced, two sampling points at the current moment and before, and two sampling points SO2As a PSO-ELM based circulating fluidized bed unit SO2Inputting a concentration prediction model;
s3, after the input variable is determined, further determining the network model structure parameter of the ELM according to the input variable; the number of neurons in the hidden layer of the ELM is very important; a small number of neurons may cause large errors in the testing process, and a large number of neurons may cause overfitting and increase training time;
and S4, optimizing the connection weight and the threshold parameter between the input layer of the limit learning machine and the ELM hidden layer by utilizing a particle swarm optimization.
In the step S3, a small number of neurons may cause a large error in the test process, a large number of neurons may cause overfitting and increase the training time, and the number of iterations of the genetic algorithm also affects the SO based PSO-ELM circulating fluidized bed unit2Accuracy of the concentration prediction model. Too few iteration times can reduce the SO of the circulating fluidized bed unit based on the PSO-ELM2Accuracy of concentration prediction model, and excessive iteration times of the concentration prediction model2The accuracy of the concentration prediction model has little influence, but the SO of the circulating fluidized bed unit based on the PSO-ELM is increased2The complexity of the concentration prediction model needs to be selected according to the quantity of input data.
In the step S3, a small number of neurons may cause a large error in the test process, a large number of neurons may cause overfitting and increase the training time, and the number of iterations of the genetic algorithm also affects the SO based PSO-ELM circulating fluidized bed unit2Accuracy of the concentration prediction model. Too few iteration times can reduce the SO of the circulating fluidized bed unit based on the PSO-ELM2Accuracy of concentration prediction model, and excessive iteration times of the concentration prediction model2The accuracy of the concentration prediction model has little influence, but the SO of the circulating fluidized bed unit based on the PSO-ELM is increased2The complexity of the concentration prediction model needs to be selected according to the quantity of input data.
In the step S4, the model used by the particle swarm algorithm is an extreme learning machine model after particle swarm optimization.
Determination of the input variables
According to SO2The circulating fluidized bed unit has the capability of in-furnace desulfurization, thus taking limestone flow as one of the input variables. SO generated during operation of CFB device2Mainly from combustion of coal. Furthermore, according to the results of the prior studies, the molar ratio of calcium to sulfur to SO2The amount of coal to be fed is important and therefore needs to be taken into account during the modeling process. The increase of the excess air coefficient can increase the oxidizing atmosphere in the furnace, reduce the reducing atmosphere in a dense phase zone and inhibit the decomposition of calcium sulfate. Meanwhile, the proportion of primary air and secondary air can change the oxidizing atmosphere in the furnace and influence SO2And (4) concentration. To study SO in dynamic processes2Variation, the load was used as a circulating fluidized bed unit SO based on PSO-ELM2One of the inputs to the concentration prediction model. The coal quality remains substantially unchanged during the unit operation. In summary, the primary air flow, secondary air flow, oxygen flow, coal flow and load were assigned to the PSO-ELM-based circulating fluidized bed unit SO2Input parameters of the concentration prediction model. The oxygen amount can reflect the oxidation-reduction atmosphere in the furnace and is the key to the generation of pollutants. By installing two measuring points, more accurate oxygen amount can be obtained, and the oxidation-reduction atmosphere in the furnace can be known. Thus, two measurement points are used as the PSO-ELM-based circulating fluidized bed unit SO2And inputting a concentration prediction model. Bed temperature is also one of the key factors affecting the production of contaminants. The desulfurization reaction speed is accelerated along with the increase of the bed temperature, thereby affecting the SO2The amount of production of (c).
In conclusion, the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are selected as the PSO-ELM-based circulating fluidized bed unit SO2And inputting a concentration prediction model.
The particle swarm algorithm calculating step
(1) Initializing a particle swarm and endowing each particle with a random initial position and speed;
(2) calculating the adaptive value of each particle according to the fitness function;
(3) comparing the adaptive value of the current position of each particle with the adaptive value corresponding to the historical optimal position of each particle, and if the adaptive value of the current position is more in line with the requirement, selecting the current position for updating;
(4) comparing the fitness value of the current position of each particle with the global optimal fitness value, and if the fitness value of the current position is more in line with the requirement, updating the global optimal position by using the current position;
(5) updating the particle position and velocity;
(6) and if the end condition is met, ending the algorithm to obtain a global optimal position, namely a global optimal solution.
The PSO-ELM algorithm step and the PSO-ELM step are described in detail as follows:
(1) the network structure of the ELM is determined.
(2) Initializing connection weight and threshold of ELM to form initial particle group Q (t)0) N particles are randomly generated.
(3) With Q (t)0) As a connection weight between input and output layers, and as a threshold value for the ELM.
(4) Calculating errors and fitness, reserving the optimal individual, and taking the optimal individual as the evolution target of the next generation; the fitness function is chosen as shown below,
wherein, is Δ VtTo verify the absolute error of the set of verification results, VtIs the output of the verification set.
(5) The particle population q (t) gets better fitness values after updating the particle position and velocity.
(6) The iteration number is increased by 1, and the step (3) is returned.
PSO-ELM model parameter selection
In a circulating fluidized bed unit SO based on PSO-ELM2In the process of establishing the concentration prediction model, the condition that the CFB unit has large inertia and delay and the current input parameters such as coal quantity and air quantity possibly influence the pollutant change for a long time in the future is considered; therefore, the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are respectively sampled at the current moment and two previous sampling points, and the two previous sampling points are respectively provided with SO2As a PSO-ELM based circulating fluidized bed unit SO2And inputting a concentration prediction model.
After the input variable is determined, further determining the circulating fluidized bed unit SO based on the PSO-ELM according to the input variable2The concentration prediction model structure parameters, wherein the number of neurons of an ELM hidden layer is very important, a small number of neurons may cause large errors in the test process, and a large number of neurons may cause overfitting and increase the training time. Meanwhile, the iteration times of the particle swarm optimization also influence the SO of the circulating fluidized bed unit based on the PSO-ELM2The accuracy of the concentration prediction model is reduced due to the fact that the number of iterations is too small, and the SO of the circulating fluidized bed unit based on the PSO-ELM is reduced2Accuracy of concentration prediction model, and excessive iteration times of the concentration prediction model2The accuracy of the concentration prediction model has little influence, but the SO of the circulating fluidized bed unit based on the PSO-ELM is increased2Complexity of the concentration prediction model; circulating fluidized bed unit SO based on PSO-ELM2The structure parameters of the concentration prediction model are selected as shown in table 1.
TABLE 1 PSO-ELM circulating fluidized bed unit SO2Structural parameters of concentration prediction model
Claims (7)
1. Circulating fluidized bed unit SO based on PSO-ELM2The concentration prediction method is characterized by comprising the following steps:
s1, according to SO2The primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are selected as the SO based on the PSO-ELM circulating fluidized bed unit2Inputting a concentration prediction model;
s2, circulating fluidized bed unit SO in PSO-ELM2In the process of establishing the concentration prediction model, considering that the CFB unit has large inertia and delay, and the coal quantity and the air quantity of the current input parameters can influence the long-time pollutant change in the future, therefore, the primary air quantity, the secondary air quantity, the oxygen quantity, the coal quantity, the load, the bed temperature and the limestone feeding quantity are introduced at the current moment and two previous sampling pointsAnd the first two sampling points SO2As a PSO-ELM based circulating fluidized bed unit SO2Inputting a concentration prediction model;
s3, after the input variable is determined, further determining the network model structure parameter of the ELM according to the input variable; the number of neurons in the hidden layer of the ELM is very important; a small number of neurons can cause large errors in the testing process, and a large number of neurons can cause overfitting and increase training time;
and S4, optimizing the connection weight and the threshold parameter between the input layer of the limit learning machine and the ELM hidden layer by utilizing a particle swarm optimization.
2. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The concentration prediction method is characterized in that in the step S3, a small number of neurons can cause larger errors in the test process, a large number of neurons can cause overfitting and increase the training time, and meanwhile, the iteration number of the genetic algorithm also influences the PSO-ELM-based circulating fluidized bed unit SO2The accuracy of the concentration prediction model; too few iteration times can reduce the SO of the circulating fluidized bed unit based on the PSO-ELM2Accuracy of concentration prediction model, and excessive iteration times of the concentration prediction model2The accuracy of the concentration prediction model has little influence, but the SO of the circulating fluidized bed unit based on the PSO-ELM is increased2The complexity of the concentration prediction model needs to be selected according to the quantity of input data.
3. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The concentration prediction method is characterized in that in step S4, the model used in the particle swarm optimization is an extreme learning machine model after particle swarm optimization.
4. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The method for predicting concentration is characterized in that the determination of the input variable in the step S2 is based on SO2With in-furnace desulphurisationCapacity, thus considering limestone flow as one of the input variables; SO generated during operation of CFB device2Mainly from the combustion of coal; furthermore, according to the results of the prior studies, the molar ratio of calcium to sulfur to SO2The amount of coal to be fed is important and therefore needs to be taken into account during the modeling process.
5. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The concentration prediction method, wherein the step S5 is a particle swarm algorithm calculation step
(1) Initializing a particle swarm and endowing each particle with a random initial position and speed;
(2) calculating the adaptive value of each particle according to the fitness function;
(3) comparing the adaptive value of the current position of each particle with the adaptive value corresponding to the historical optimal position of each particle, and if the adaptive value of the current position is more in line with the requirement, selecting the current position for updating;
(4) comparing the fitness value of the current position of each particle with the global optimal fitness value, and if the fitness value of the current position is more in line with the requirement, updating the global optimal position by using the current position;
(5) updating the particle position and velocity;
(6) and if the end condition is met, ending the algorithm to obtain a global optimal position, namely a global optimal solution.
6. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The concentration prediction method is characterized in that the PSO-ELM algorithm steps are described in detail as follows:
(1) determining a network structure of the ELM;
(2) initializing connection weight and threshold of ELM to form initial particle group Q (t)0) Randomly generating n particles;
(3) with Q (t)0) As a connection weight between input and output layers, and a threshold value of the ELM to train the neural network;
(4) calculating errors and fitness, reserving the optimal individual, and taking the optimal individual as the evolution target of the next generation; the fitness function is chosen as shown below,
wherein, is Δ VtTo verify the absolute error of the set of verification results, VtAn output that is a validation set;
(5) the particle group Q (t) obtains a better fitness value after updating the position and the speed of the particles;
(6) and (4) increasing the iteration number by 1 and returning to the step (3).
7. The PSO-ELM-based circulating fluidized bed unit SO according to claim 12The concentration prediction method is characterized in that the PSO-ELM model parameter selection
In a circulating fluidized bed unit SO based on PSO-ELM2In the process of establishing the concentration prediction model, the CFB unit is considered to have large inertia and large delay, and the current input parameters of the coal quantity and the air quantity can possibly influence the long-time pollutant change in the future; thus, the primary air flow, the secondary air flow, the oxygen amount, the coal amount, the load, the bed temperature, the limestone feed amount and the SO are introduced2Taking two sampling points before and at the current time of concentration as a circulating fluidized bed unit SO based on PSO-ELM2And inputting a concentration prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111190859.XA CN114186708A (en) | 2021-10-13 | 2021-10-13 | Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111190859.XA CN114186708A (en) | 2021-10-13 | 2021-10-13 | Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114186708A true CN114186708A (en) | 2022-03-15 |
Family
ID=80539482
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111190859.XA Pending CN114186708A (en) | 2021-10-13 | 2021-10-13 | Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114186708A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113908673A (en) * | 2021-09-30 | 2022-01-11 | 湖北华电襄阳发电有限公司 | Wet desulphurization efficiency prediction system and method based on extreme learning machine |
| CN114880919A (en) * | 2022-03-24 | 2022-08-09 | 华北电力大学 | Method for calculating optimal furnace internal and external desulfurization proportion of circulating fluidized bed unit |
-
2021
- 2021-10-13 CN CN202111190859.XA patent/CN114186708A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113908673A (en) * | 2021-09-30 | 2022-01-11 | 湖北华电襄阳发电有限公司 | Wet desulphurization efficiency prediction system and method based on extreme learning machine |
| CN114880919A (en) * | 2022-03-24 | 2022-08-09 | 华北电力大学 | Method for calculating optimal furnace internal and external desulfurization proportion of circulating fluidized bed unit |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110368808B (en) | Ammonia spraying amount control method and system for SCR flue gas denitration system | |
| CN114186708A (en) | Circulating fluidized bed unit SO based on PSO-ELM2Concentration prediction method | |
| CN107016455B (en) | Prediction system and method for oxygen content of flue gas at hearth outlet of circulating fluidized bed household garbage incineration boiler | |
| CN106202946A (en) | Clinker free calcium levels Forecasting Methodology based on degree of depth belief network model | |
| CN108549792B (en) | Soft measurement method for dioxin emission concentration in solid waste incineration process based on latent structure mapping algorithm | |
| CN111921377B (en) | SCR denitration ammonia injection control method and system based on mechanism and data driving | |
| CN113433911B (en) | Accurate control system and method for ammonia spraying of denitration device based on accurate concentration prediction | |
| CN104715142B (en) | A kind of station boiler NOxDischarge dynamic soft-measuring method | |
| CN110675920A (en) | MI-LSTM-based boiler NOxPrediction method | |
| CN112742187A (en) | Method and device for controlling pH value in desulfurization system | |
| CN114225662A (en) | Flue gas desulfurization and denitrification optimization control method based on hysteresis model | |
| CN104696952A (en) | Fluidized bed boiler air input capacity control method based on model online identification | |
| CN109992844A (en) | A kind of boiler flyash carbon content prediction technique based on ADQPSO-SVR model | |
| CN115034129A (en) | NOx emission concentration soft measurement method for thermal power plant denitration device | |
| CN115392629A (en) | Coal gangue heat productivity soft measurement method based on improved SVR | |
| CN112613237B (en) | CFB unit NOx emission concentration prediction method based on LSTM | |
| Chen et al. | Prediction method of sulfur dioxide emission | |
| CN116050643A (en) | Method for predicting emission concentration of process industrial pollutants based on integrated model | |
| CN113488111B (en) | Ammonia injection amount optimization modeling method for SCR denitration system | |
| CN113780639B (en) | Urban solid waste incineration nitrogen oxide NOx emission prediction method based on multitask learning framework | |
| CN116312869A (en) | Method, device and system for predicting nitrogen oxides in catalytic cracking regenerated flue gas | |
| CN117190173B (en) | Optimal control method and control system for flue gas recirculation and boiler coupling system | |
| CN116092594A (en) | Elman neural network inlet NOx concentration prediction method based on polymerization characteristics | |
| CN113869359B (en) | Method for predicting nitrogen oxides in urban solid waste incineration process based on modularized neural network | |
| CN112488145B (en) | NO based on intelligent methodxOnline prediction method and system |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20231212 Address after: 100096 No. 2, Beinong Road, Huilongguan, Changping District, Beijing Applicant after: NORTH CHINA ELECTRIC POWER University Applicant after: HUANENG TAICANG POWER GENERATION Co.,Ltd. Applicant after: Xi'an Thermal Power Research Institute Co.,Ltd. Applicant after: HUANENG GROUP TECHNOLOGY INNOVATION CENTER Co.,Ltd. Address before: 102206 No. 2 Nong Road, Zhu Xin Zhuang, Beijing, Changping District Applicant before: NORTH CHINA ELECTRIC POWER University Applicant before: Xi'an Thermal Power Research Institute Co.,Ltd. Applicant before: HUANENG GROUP TECHNOLOGY INNOVATION CENTER Co.,Ltd. |