CN115330044A - Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network - Google Patents
Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network Download PDFInfo
- Publication number
- CN115330044A CN115330044A CN202210957885.9A CN202210957885A CN115330044A CN 115330044 A CN115330044 A CN 115330044A CN 202210957885 A CN202210957885 A CN 202210957885A CN 115330044 A CN115330044 A CN 115330044A
- Authority
- CN
- China
- Prior art keywords
- variable
- nitrogen oxide
- parameter
- data
- model
- 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
- MWUXSHHQAYIFBG-UHFFFAOYSA-N Nitric oxide Chemical compound O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 title claims abstract description 171
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000002068 genetic effect Effects 0.000 title claims abstract description 41
- 238000005457 optimization Methods 0.000 claims abstract description 40
- 238000005259 measurement Methods 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 36
- 238000012549 training Methods 0.000 claims description 21
- 230000002159 abnormal effect Effects 0.000 claims description 14
- 230000004913 activation Effects 0.000 claims description 13
- 239000003245 coal Substances 0.000 claims description 13
- 238000003860 storage Methods 0.000 claims description 13
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 claims description 12
- 239000003546 flue gas Substances 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 12
- 238000013507 mapping Methods 0.000 claims description 11
- 238000004519 manufacturing process Methods 0.000 claims description 8
- 238000011176 pooling Methods 0.000 claims description 8
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 7
- 239000001301 oxygen Substances 0.000 claims description 7
- 229910052760 oxygen Inorganic materials 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 5
- 238000013519 translation Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 239000007789 gas Substances 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000004806 packaging method and process Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 5
- 239000000779 smoke Substances 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000000738 capillary electrophoresis-mass spectrometry Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 229910021529 ammonia Inorganic materials 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010531 catalytic reduction reaction Methods 0.000 description 1
- 239000003638 chemical reducing agent Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000000415 inactivating effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005507 spraying Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000004804 winding 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
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Educational Administration (AREA)
- Computational Linguistics (AREA)
- Primary Health Care (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Operations Research (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Feedback Control In General (AREA)
Abstract
The invention provides a nitrogen oxide emission prediction method and equipment based on a genetic algorithm and a convolutional network, wherein the method is based on historical operating data, and is used for converting variable selection, delay lag characteristics of variable influence and a prediction model tuning process into a comprehensive function optimization problem, optimizing parameters of each link by using the genetic algorithm, determining a model structure according to a function solution space corresponding to an optimal solution, solving the problem of local optimal solution caused by variable selection and model tuning due to measurement time lag, so that a prediction model can more fully mine a NOx generation rule, and further providing more accurate support for denitration control.
Description
Technical Field
The invention relates to the technical field of coal-fired thermal power generation, in particular to a method, a device, equipment and a storage medium for predicting nitrogen oxide emission based on a genetic algorithm and a convolutional network.
Background
In the power production of China, the energy structure mainly comprising coal determines the dominant position of coal-fired thermal power generation, and with the development of the society, the environmental protection problem is more and more emphasized, the national NOx emission standard of a thermal power plant is also raised to a new level, so that the quality requirement of the thermal power plant on denitration control is higher. At present, a large thermal generator set generally adopts an SCR (selective catalytic reduction) mode, in order to reduce the emission of NOx and avoid the blockage of an air preheater caused by excessive ammonia spraying and the excessive emission of NOx exceeding caused by too few reducing agents, the emission concentration of nitrogen oxides must be measured and monitored in real time, and a denitration system is optimally controlled.
The automatic flue gas monitoring system (CEMS) which is widely used at present has a plurality of defects, the CEMS needs regular off-line maintenance, the workload is large, and the measured value is always accurate and effective in actual operation; the CEMS standard is measured by an extraction method, and longer measurement lag time is needed, so that the feedback of a control system has larger lag characteristic, and the control precision is influenced.
In order to overcome the defects, a prediction model based on data driving is widely applied to soft measurement and control optimization of the flue gas NOx, such as linear regression, a support vector machine, naive Bayes, a recurrent neural network and the like; most of the existing machine learning methods cannot learn the continuous influence on the time dimension, and although the recurrent neural network has strong nonlinear and time dimension learning capabilities, no good solution is provided for the delay of model data.
Disclosure of Invention
The invention provides a method, a device, equipment and a storage medium for predicting nitrogen oxide emission based on a genetic algorithm and a convolutional network, and aims to solve the problems of variable selection and model optimization caused by measurement data lag in the prior art.
To this end, a first object of the present invention is to provide a method for predicting nox emission based on genetic algorithm and convolutional network, comprising:
constructing a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimization module is used for carrying out parameter optimization on the variable selection hyperparameter set by the variable selection hyperparameter setting module, the variable lag hyperparameter set by the variable lag hyperparameter setting module and the convolution network regression model hyperparameter set by the convolution network regression model hyperparameter setting module based on a genetic algorithm, and determining a nitrogen oxide emission prediction model based on an optimization result;
acquiring historical boiler operation data, screening abnormal boiler operation data, and training a nitrogen oxide emission prediction model by using the abnormal boiler operation data as training data;
and acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model, and outputting a result as a nitrogen oxide emission prediction result.
The historical boiler operation data is data generated by boiler operation in a specified time interval and at least comprises coal feeding quantity, primary air quantity, secondary air quantity, smoke exhaust oxygen content, boiler load, hearth temperature, smoke quantity, smoke temperature and denitration reactor inlet NOx concentration.
Wherein, in the step of screening abnormal operation data of the boiler, the method comprises the following steps:
dividing the historical boiler operation data into a plurality of equal-length intervals along a time axis, and calculating the fluctuation amplitude of each type of data in the historical boiler operation data in any equal-length interval;
if the fluctuation amplitude of the corresponding data meets the preset condition, taking the corresponding type data in the corresponding equal-length interval as abnormal operation data of the boiler;
and repeatedly screening for multiple times to obtain first sample data.
The variable selection hyper-parameter setting module is used for constructing a variable combination mapping dictionary; the method specifically comprises the following steps:
defining model related variables of the nitrogen oxide emission prediction model as basic variables and optimizing variables; the basic variables comprise boiler load Steam, coal feeding amount Coal, total air volume Wind and exhaust oxygen content O2, and the optimization variables comprise primary air volume Wind1, secondary air volume Wind2, hearth temperature BoilerTem, flue Gas amount Gas and flue Gas temperature GasTem;
selecting a strategy for model related variables as a basic variable and an optimizing variable; determining the variable combination type of the model according to the combination recursion, and establishing a variable combination mapping dictionary;
and determining a variable selection hyper-parameter according to the variable combination mapping dictionary, and performing feature filtering on the first sample data to obtain second sample data.
The variable lag hyper-parameter setting module is used for eliminating the time delay characteristic generated in the running production process of the boiler;
if the current time is t moment, and the variable lag parameter in the second sample data is t1, it indicates that the variables at the [ t-t1, t ] moment all have influence on the generation of NOx at the current moment, and the variables at the [ t-t1, t ] moment all serve as input data;
if the current time is t, and the measurement delay t2 exists in the prediction variable NOx, the fact that the current time NOx measured value is actually the t-t2 time production data is indicated, and translation processing needs to be carried out on second sample data, namely the t-time NOx data corresponds to the t-t2 time production data;
and establishing a variable lag hyperparameter, representing a lag parameter set of each variable type in the second sample data, and preprocessing the second sample data according to the independent variable and the predictive variable delay parameter to obtain the sample data for training the model.
Wherein, the convolution network regression model comprises:
the packaging structure comprises a first coiling layer, a second coiling layer, a pooling layer, a Flatten layer, a Dropout layer, a full-connection layer and an output layer;
wherein the first convolution layer is a linear activation function, and the second convolution layer is a nonlinear activation function;
then setting the hyper-parameters of the regression model of the convolutional network comprises: the number of first convolution kernels, the size of the first convolution layer, the number of second convolution kernels, the size of the second convolution layer, the nonlinear activation function parameter and the pooling parameter.
The parameter optimizing module performs parameter optimizing based on a genetic algorithm, and the parameter optimizing module comprises the following steps:
taking an optimization function of the convolution network regression model as an optimization objective function of the genetic algorithm;
determining an optimization space of a variable selection hyperparameter, a variable lag hyperparameter and a convolution network regression model hyperparameter;
and randomly selecting and generating an initial population in the optimization space for each parameter, calculating a fitness value according to the optimization objective function, outputting an optimal solution to determine a final prediction model if a termination condition is met, otherwise, updating the parameter population, recalculating the fitness value, and performing iterative operation until the termination condition is met.
A second object of the present invention is to provide a device for predicting nox emission based on genetic algorithm and convolutional network, comprising:
the model construction module is used for constructing a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimizing module is used for optimizing the parameters of the variable selection superparameter set by the variable selection superparameter setting module, the variable lag superparameter set by the variable lag superparameter setting module and the convolutional network regression model superparameter set by the convolutional network regression model superparameter setting module based on a genetic algorithm, and determining a nitrogen oxide emission prediction model based on an optimizing result;
the model training module is used for acquiring historical boiler operation data, screening abnormal boiler operation data and training a nitrogen oxide emission prediction model as training data;
and the prediction module is used for acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model and outputting a result as a nitrogen oxide emission prediction result.
A third object of the present invention is to provide an electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method of the foregoing technical solution.
A fourth object of the present invention is to propose a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the steps of the method according to the aforementioned technical solution.
Different from the prior art, the nitrogen oxide emission prediction method based on the genetic algorithm and the convolutional network, provided by the invention, has the advantages that based on historical operating data, the variable selection, the delay lag characteristic of variable influence and the prediction model tuning process are converted into a comprehensive function optimization problem, the parameters of each link are optimized by using the genetic algorithm, the model structure is determined according to the function solution space corresponding to the optimal solution, the problem of local optimal solution caused by variable selection and model tuning due to measurement time lag is solved, the prediction model can more fully mine the NOx generation rule, and further more accurate support is provided for the denitration control.
Drawings
The invention and/or additional aspects and advantages will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic flow chart of a method for predicting nitrogen oxide emissions based on a genetic algorithm and a convolutional network according to the present invention.
FIG. 2 is a logic diagram of a method for predicting NOx emission based on a genetic algorithm and a convolutional network according to the present invention.
FIG. 3 is a logic diagram of a genetic algorithm in the method for predicting nitrogen oxide emissions based on the genetic algorithm and a convolutional network according to the present invention.
FIG. 4 is a schematic diagram of a model prediction effect in a nitrogen oxide emission prediction method based on a genetic algorithm and a convolutional network provided by the invention.
Fig. 5 is a schematic structural diagram of a nitrogen oxide emission prediction apparatus based on a genetic algorithm and a convolutional network according to the present invention.
Fig. 6 is a schematic structural diagram of a non-transitory computer-readable storage medium provided in the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
As shown in fig. 1, a method for predicting nox emission based on a genetic algorithm and a convolutional network according to an embodiment of the present invention includes:
s110: constructing a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimizing module is used for optimizing the parameters of the variable selection superparameter set by the variable selection superparameter setting module, the variable lag superparameter set by the variable lag superparameter setting module and the convolutional network regression model superparameter set by the convolutional network regression model superparameter setting module based on a genetic algorithm, and determining the nitrogen oxide emission prediction model based on an optimizing result.
S120: and obtaining historical boiler operation data, screening abnormal boiler operation data, and training the nitrogen oxide emission prediction model as training data.
S130: and acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model, and outputting a result as a nitrogen oxide emission prediction result.
The historical boiler operation data in the invention is data generated by boiler operation in a specified time interval, and at least comprises coal feeding quantity, primary air quantity, secondary air quantity, smoke exhaust oxygen content, boiler load, hearth temperature, smoke quantity, smoke temperature and denitration reactor inlet NOx concentration.
The logic of the method of the present invention is shown in FIG. 2.
Specifically, the step of screening abnormal operation data of the boiler comprises the following steps:
dividing historical boiler operation data into a plurality of equal-length intervals along a time axis, and calculating the fluctuation amplitude of each type of data in the historical boiler operation data in any equal-length interval;
a fluctuation width v of
If the fluctuation range of the corresponding data meets the preset condition, taking the corresponding type data in the corresponding equal-length interval as abnormal operation data of the boiler;
illustratively, the fluctuation amplitude v _ steam of the boiler load is more than or equal to 0.03; the fluctuation amplitude v _ coal of the coal feeding amount is more than or equal to 0.03; the fluctuation range v _ o2 of the oxygen content of the discharged smoke is more than or equal to 0.05; and the abnormal operation data of other types of boilers are set according to actual conditions.
The screening was repeated a plurality of times to obtain first sample data D1 (L1, L2, L3 \8230; ln).
The variable selection hyper-parameter setting module is used for constructing a variable combination mapping dictionary; the method specifically comprises the following steps:
defining model related variables of the nitrogen oxide emission prediction model as basic variables and optimizing variables; the basic variables comprise boiler load Steam, coal feeding amount Coal, total air volume Wind and exhaust oxygen content O2, and the optimization variables comprise primary air volume Wind1, secondary air volume Wind2, hearth temperature BoilerTem, flue Gas amount Gas and flue Gas temperature GasTem;
selecting a strategy for model related variables as a basic variable and an optimizing variable; determining the variable combination type of the model according to the combination recursion, and establishing a variable combination mapping dictionary;
the combined recursion formula is expressed as
Model number 2 n And (4) various variable combinations.
The variable combination mapping dictionary is represented as:
dict = {1: basic variable
2: base variable + optimization variable 1
3: base variable + optimization variable 2
……
2 n : basic variable + optimizing variable 1 \8230 \ 8230; optimizing variable n }.
Determining a variable selection hyper-parameter according to the variable combination mapping dictionary, wherein the variable selection hyper-parameter is expressed as: params1= [1,2,3 ] \8230; \82302 n ]；
And when the variable parameter is i, performing feature filtering on the sample data D1 according to a variable corresponding to the mapping dictionary Dict to obtain second sample data D2 (D21, D22, \8230;, D2 n).
Due to the delay characteristic existing in the operation and production process of the boiler, media such as coal, air and the like are sent into a hearth to be combusted, inertia and pure lag characteristics exist between the generation of flue gas, and the variable lag super-parameter setting module is used for eliminating the delay characteristic generated in the operation and production process of the boiler.
If the current time is t moment, and the variable lag parameter in the second sample data is t1, it indicates that the variables at the [ t-t1, t ] moment all have influence on the generation of NOx at the current moment, and the variables at the [ t-t1, t ] moment all serve as input data;
if the current time is t, the measurement delay t2 of the prediction variable NOx exists, the fact that the measured value of the current time NOx is actually the production data at the time t-t2 is indicated, and translation processing needs to be carried out on second sample data, namely the NOx data at the time t corresponds to the data produced at the time t-t 2;
and establishing a variable lag hyperparameter, representing a lag parameter set of each variable type in the second sample data, and preprocessing the second sample data according to the independent variable and the predictive variable delay parameter to obtain the sample data for training the model.
Establishing variable lag super parameter params2= [ T1, T2, T3 \8230; tn, tnox ].
The convolutional network is a feedforward neural network which comprises convolutional calculation and has a deep structure, is one of representative algorithms for deep learning, and has the characteristic learning capability of performing translation invariant classification on input information according to a hierarchical mechanism, so that the sequential data has better learning capability.
The convolution network regression model comprises:
a first winding layer: linear activation function (number of convolution kernels k11, convolution kernel size k 12);
a second convolution layer: the nonlinear activation function (number of convolution kernels k21, convolution kernel size k22, nonlinear activation function parameter fun);
a pooling layer: (pooling parameter pool, including maximum pooling and average pooling);
flatten layer: processing the multidimensional data into one-dimensional data;
dropout layer: randomly inactivating neurons in proportion;
full connection layer: (linear activation function);
and (6) an output layer.
The loss function of the convolution network regression model is used for averaging absolute errors, and the formula is expressed as follows:
wherein y' is a predicted value, y is an actual value, and n is the data volume of the test set;
the activation function comprises a linear activation function, a nonlinear activation function tanh, relu and sigmoid; wherein the content of the first and second substances,
linear(x)＝x
parameters set in the network structure are combined into a model hyper-parameter params3= [ k11, k12, k21, k22, fun, pool ], wherein k11, k21 optimize space [1 32], k12, k22 optimize space [ 1.
Dividing sample data serving as a model training set into K parts, training the model by using (K-1) parts, and using the remaining 1 part of data to evaluate the quality of the model; the invention utilizes k-fold cross validation, performs model training through a neural network mechanism, and sets an optimization function as an average value of average absolute errors on different test sets:
wherein, mae k The mean absolute error of the model on the kth test set is used.
The Genetic Algorithm (GA) is a self-adaptive random search heuristic algorithm designed and proposed according to the evolution rule of organisms in the nature, and the basic framework of the algorithm is a self-organizing and self-adaptive artificial intelligence technology which is based on the natural selection rule and the genetic theory and is used for solving problems by simulating the evolution mode and the inheritance of the organisms in the nature, and the algorithm is widely applied to the optimization of complex function systems, machine learning, system identification, fault diagnosis, classification systems, controller design, neural network design, self-adaptive filter design and the like.
The parameter optimizing module carries out parameter optimizing based on a genetic algorithm, and the parameter optimizing module comprises the following steps:
taking an optimization function of the convolution network regression model as an optimization objective function of the genetic algorithm; the optimization function fitness is an optimization objective function in the genetic algorithm;
determining an optimization space of a variable selection hyperparameter, a variable lag hyperparameter and a convolution network regression model hyperparameter;
namely to
Variable selection of superparameter params1= [1,2,3 ] \8230; \82302 n ]、
Variable lag super parameter params2= [ T1, T2, T3 \8230; \8230Tn, tnox ], (in the alternative, the values of Tn, tnox;),
Model hyper-parameters params3= [ k11, k12, k21, k22, fun, pool ] were optimized.
And randomly selecting and generating an initial population in the optimization space for each parameter, calculating a fitness value according to the optimization objective function, outputting an optimal solution to determine a final prediction model if a termination condition is met, otherwise, updating the parameter population, recalculating the fitness value, and performing iterative operation until the termination condition is met. The genetic algorithm flow is shown in figure 3.
In summary, the NOx prediction method combining the genetic algorithm and the convolutional neural network provided in this embodiment converts the variable selection, the delay lag characteristic of the variable influence, and the prediction model tuning process into a comprehensive function optimization problem, optimizes the parameters of each link by using the genetic algorithm, and determines the model structure according to the function solution space corresponding to the optimal solution, thereby solving the problem of local optimal solution caused by the variable selection and model tuning due to measurement time lag, so that the prediction model can more fully mine the NOx generation rule, and further provide more accurate support for the denitration control. The predicted effect is shown in fig. 4.
As shown in fig. 5, the present invention also provides a nox emission prediction apparatus based on a genetic algorithm and a convolutional network, comprising:
a model construction module 310, configured to construct a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimizing module is used for optimizing the parameters of the variable selection superparameter set by the variable selection superparameter setting module, the variable lag superparameter set by the variable lag superparameter setting module and the convolutional network regression model superparameter set by the convolutional network regression model superparameter setting module based on a genetic algorithm, and determining a nitrogen oxide emission prediction model based on an optimizing result;
the model training module 320 is used for acquiring historical boiler operation data, screening abnormal boiler operation data and training a nitrogen oxide emission prediction model as training data;
and the prediction module 330 is used for acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model, and outputting a result as a nitrogen oxide emission prediction result.
In order to implement the embodiment, the present invention further provides an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the steps of the method for predicting nitrogen oxide emissions of the present disclosure.
As shown in fig. 6, the non-transitory computer readable storage medium includes a memory 810 of instructions executable by the nox emission prediction processor 820 to perform a method, and an interface 830. Alternatively, the storage medium may be a non-transitory computer readable storage medium, for example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
To achieve the embodiments, the present invention also proposes a non-transitory computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements a nitrogen oxide emission prediction as an embodiment of the invention.
In the description of the specification, reference to the description of "one embodiment," "some embodiments," "an example," "a specific example," or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, a schematic representation of the terms does not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
The logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Further, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the described embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
One of ordinary skill in the art will appreciate that all or part of the steps carried out in the method of implementing the embodiments described herein may be implemented by hardware associated with instructions of a program, which may be stored in a computer-readable storage medium, and which, when executed, includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The mentioned storage medium may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present invention have been shown and described above, it is understood that the embodiments are illustrative and not restrictive, and that those skilled in the art may make changes, modifications, substitutions and alterations to the embodiments described herein without departing from the scope of the invention.
Claims (10)
1. A method for predicting nitrogen oxide emission based on a genetic algorithm and a convolutional network is characterized by comprising the following steps:
constructing a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimizing module is used for optimizing the parameters of the variable selection superparameter set by the variable selection superparameter setting module, the variable lag superparameter set by the variable lag superparameter setting module and the convolutional network regression model superparameter set by the convolutional network regression model superparameter setting module based on a genetic algorithm, and determining a nitrogen oxide emission prediction model based on an optimizing result;
acquiring historical boiler operation data, screening abnormal boiler operation data, and training the nitrogen oxide emission prediction model as training data;
and acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model, and outputting a result as a nitrogen oxide emission prediction result.
2. The method for predicting nitrogen oxide emission based on genetic algorithm and convolutional network as claimed in claim 1, wherein the historical boiler operation data is data generated by boiler operation in a specified time interval, and at least comprises coal feeding amount, primary air volume, secondary air volume, flue gas oxygen content, boiler load, furnace temperature, flue gas amount, flue gas temperature, and outlet NOx concentration of a denitration reactor.
3. The method of predicting nitrogen oxide emissions based on genetic algorithm and convolutional network as claimed in claim 2, wherein in the step of screening abnormal operation data of the boiler, comprising the steps of:
dividing the historical boiler operation data into a plurality of equal-length intervals along a time axis, and calculating the fluctuation amplitude of each type of data in the historical boiler operation data in any equal-length interval;
if the fluctuation amplitude of the corresponding data meets the preset condition, taking the corresponding type data in the corresponding equal-length interval as abnormal operation data of the boiler;
and repeatedly screening for multiple times to obtain first sample data.
4. The method of predicting nitrogen oxide emissions based on genetic algorithms and convolutional networks of claim 3, wherein said variable selection hyper-parameter setting module is used to construct a variable combination mapping dictionary; the method specifically comprises the following steps:
defining model-related variables of the nitrogen oxide emission prediction model as basic variables and optimizing variables; the basic variables comprise boiler load Steam, coal feeding amount Coal, total air volume Wind and exhaust oxygen content O2, and the optimization variables comprise primary air volume Wind1, secondary air volume Wind2, hearth temperature BoilerTem, flue Gas amount Gas and flue Gas temperature GasTem;
selecting a strategy for the relevant variables of the model as a basic variable and an optimizing variable; determining the variable combination type of the model according to the combination recursion, and establishing a variable combination mapping dictionary;
and determining a variable selection hyper-parameter according to the variable combination mapping dictionary, and performing feature filtering on the first sample data to obtain second sample data.
5. The method of claim 4, wherein the variable lag superparameter setting module is configured to eliminate a delay characteristic generated during the operation of the boiler;
if the current time is t moment and the variable lag parameter in the second sample data is t1, indicating that the variables at the [ t-t1, t ] moment all influence the generation of NOx at the current moment, and taking the variables at the [ t-t1, t ] moment as input data;
if the current time is t, the measurement delay t2 of the prediction variable NOx exists, the fact that the measured value of the current time NOx is actually the production data at the time t-t2 is indicated, and translation processing needs to be carried out on second sample data, namely the NOx data at the time t corresponds to the data produced at the time t-t 2;
and establishing a variable lag hyperparameter, representing a lag parameter set of each variable type in the second sample data, and preprocessing the second sample data according to the independent variable and the predictive variable delay parameter to obtain the sample data for training the model.
6. The method of claim 5, wherein the convolutional network regression model comprises:
the packaging structure comprises a first coiling layer, a second coiling layer, a pooling layer, a Flatten layer, a Drapout layer, a full-connection layer and an output layer;
wherein the first convolution layer is a linear activation function, and the second convolution layer is a nonlinear activation function;
then setting the hyper-parameters of the regression model of the convolutional network comprises: the number of first convolution kernels, the size of the first convolution layer, the number of second convolution kernels, the size of the second convolution layer, the nonlinear activation function parameter and the pooling parameter.
7. The method of claim 6, wherein the step of performing parameter optimization by the parameter optimization module based on genetic algorithm comprises:
taking an optimization function of the convolution network regression model as an optimization objective function of a genetic algorithm;
determining an optimization space of a variable selection hyperparameter, a variable lag hyperparameter and a convolution network regression model hyperparameter;
and randomly selecting and generating an initial population in the optimization space for each parameter, calculating a fitness value according to the optimization objective function, outputting an optimal solution to determine a final prediction model if a termination condition is met, otherwise, updating the parameter population, recalculating the fitness value, and performing iterative operation until the termination condition is met.
8. A nitrogen oxide emission prediction device based on a genetic algorithm and a convolutional network, comprising:
the model construction module is used for constructing a nitrogen oxide emission prediction model;
the nitrogen oxide emission prediction model comprises a variable selection hyper-parameter setting module, a variable lag hyper-parameter setting module, a convolution network regression model hyper-parameter setting module and a parameter optimization module; the parameter optimizing module is used for optimizing the parameters of the variable selection superparameter set by the variable selection superparameter setting module, the variable lag superparameter set by the variable lag superparameter setting module and the convolutional network regression model superparameter set by the convolutional network regression model superparameter setting module based on a genetic algorithm, and determining a nitrogen oxide emission prediction model based on an optimizing result;
the model training module is used for acquiring historical boiler operation data, screening abnormal boiler operation data and training the nitrogen oxide emission prediction model as training data;
and the prediction module is used for acquiring boiler operation data in real time, inputting the trained nitrogen oxide emission prediction model and outputting a result as a nitrogen oxide emission prediction result.
9. An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the method of any one of claims 1-7.
10. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the steps of the method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210957885.9A CN115330044A (en) | 2022-08-10 | 2022-08-10 | Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210957885.9A CN115330044A (en) | 2022-08-10 | 2022-08-10 | Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115330044A true CN115330044A (en) | 2022-11-11 |
Family
ID=83922601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210957885.9A Pending CN115330044A (en) | 2022-08-10 | 2022-08-10 | Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115330044A (en) |
-
2022
- 2022-08-10 CN CN202210957885.9A patent/CN115330044A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111260107B (en) | Boiler combustion optimization system and method | |
| CN107694337A (en) | Coal unit SCR denitrating flue gas control methods based on network response surface | |
| CN110263395A (en) | The power plant's denitration running optimizatin method and system analyzed based on numerical simulation and data | |
| CN113433911B (en) | Accurate control system and method for ammonia spraying of denitration device based on accurate concentration prediction | |
| CN111860701B (en) | Denitration system working condition discrimination preprocessing method based on clustering method | |
| CN111522290A (en) | Denitration control method and system based on deep learning method | |
| US20200173649A1 (en) | System and method for optimizing combustion of boiler | |
| CN105697166A (en) | Systems and Methods for Controlling Air-to-Fuel Ratio Based on Catalytic Converter Performance | |
| CN112364562B (en) | Flue gas environment-friendly island cooperative control method and system | |
| CN113469449B (en) | Optimization control method and system for desulfurization system | |
| CN111589302A (en) | Method, device, equipment and storage medium for predicting SCR denitration performance of coal-fired power plant | |
| CN116688754A (en) | Ship flue gas desulfurization automatic control system and method thereof | |
| CN110059821A (en) | Neural network optimization, device, computer equipment and storage medium | |
| CN116665808A (en) | NOx emission prediction method and device based on neighborhood rough set and fuzzy neural network | |
| CN113628694A (en) | Method for predicting discharge amount of nitrogen oxides of boiler | |
| CN115330044A (en) | Nitrogen oxide emission prediction method and equipment based on genetic algorithm and convolution network | |
| CN109933884B (en) | Neural network inverse control method for SCR denitration system of coal-fired unit | |
| CN112183872A (en) | Blast furnace gas generation amount prediction method combining generation of countermeasure network and neural network | |
| CN116050643A (en) | Method for predicting emission concentration of process industrial pollutants based on integrated model | |
| CN115591378A (en) | Feedforward compensation and disturbance suppression control system and method for SCR denitration of thermal power generating unit | |
| CN115113519A (en) | Coal-gas co-combustion boiler denitration system outlet NO x Concentration early warning method | |
| CN114089636A (en) | SCR denitration external hanging type intelligent ammonia spraying closed-loop control method and equipment | |
| CN116312869A (en) | Method, device and system for predicting nitrogen oxides in catalytic cracking regenerated flue gas | |
| CN113935230A (en) | Implementation of NO based on attention mechanism LSTM modelxEmission amount prediction method | |
| Tingting et al. | Modeling on SCR process of a coal-fired boiler using LSSVM method |
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 |