CN112163335A - Training method, prediction method and device of NOx concentration prediction model - Google Patents
Training method, prediction method and device of NOx concentration prediction model Download PDFInfo
- Publication number
- CN112163335A CN112163335A CN202011025113.9A CN202011025113A CN112163335A CN 112163335 A CN112163335 A CN 112163335A CN 202011025113 A CN202011025113 A CN 202011025113A CN 112163335 A CN112163335 A CN 112163335A
- Authority
- CN
- China
- Prior art keywords
- training
- prediction model
- nox
- verification
- prediction
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Feedback Control In General (AREA)
Abstract
The application discloses a training method, a prediction method and a device of a NOx concentration prediction model, wherein the training method comprises the following steps: acquiring training data acquired from a decentralized control system of the SCR denitration system; based on an XGboost algorithm and correlation analysis, performing feature selection from training data to obtain a training feature set and a verification feature set; training a preset neural network by taking training characteristics in the training characteristic set as input parameters and training NOx prediction concentration corresponding to the training characteristics as output results to obtain an initial prediction model; inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model; calculating a prediction error between the actual concentration of the verified NOx and the predicted concentration of the verified NOx corresponding to the verification feature; and adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model. The technical problem that the prediction accuracy of the existing prediction model is not high when the concentration of the NOx is predicted is solved.
Description
Technical Field
The application relates to the technical field of SCR flue gas denitration, in particular to a training method, a prediction method and a device of a NOx concentration prediction model.
Background
With the increasingly prominent problem of atmospheric pollution, the emission of atmospheric pollutants of thermal power generating units is strictly regulated. Traditional low NOx (nitrogen Oxide) combustion control hardly satisfies emission standard, need cooperate the flue gas denitrification facility of afterbody, and wherein, SCR flue gas denitrification technique's use is comparatively extensive.
In the process of using the SCR flue gas denitration technology, the concentration of the emitted NOx needs to be predicted, and the adopted prediction models are usually an RBF prediction model, an LSSVM prediction model and an RNN prediction model, but the prediction accuracy of the prediction models is not high, and the predicted concentration of the NOx has large deviation from the actual concentration of the emitted NOx.
Disclosure of Invention
In view of this, the present application provides a training method, a prediction method, and an apparatus for a NOx concentration prediction model, which solve the technical problem that the prediction accuracy of the prediction model is not high when the NOx concentration is predicted in the prior art.
The first aspect of the present application provides a method for training a NOx concentration prediction model, including:
acquiring training data acquired from a decentralized control system of the SCR denitration system;
based on an XGboost algorithm and correlation analysis, performing feature selection from the training data to obtain a training feature set for training and a verification feature set for verification;
training a preset neural network by taking the training features in the training feature set as input parameters and the training NOx prediction concentration corresponding to the training features as an output result to obtain an initial prediction model;
inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model;
calculating a prediction error between a verification NOx actual concentration corresponding to the verification characteristic and the verification NOx predicted concentration;
and adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
Preferably, the preset neural network is an LSTM neural network.
Preferably, the LSTM neural network comprises: an input layer, a hidden layer and an output layer;
the input layer, the hidden layer, and the output layer each include 128 nodes.
Preferably, the adjusting the initial prediction model according to the prediction error to obtain the trained target prediction model specifically includes:
determining an adjustment parameter for adjusting the initial prediction model based on an Adam algorithm and the prediction error;
and adjusting the initial prediction model based on the adjustment parameters to obtain a trained target prediction model.
Preferably, the training features in the training feature set include: ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet NOx mass concentration and SCR denitration efficiency.
The second invention of the present application provides a method for predicting a NOx concentration prediction model, wherein the NOx concentration prediction model is obtained by training through the method for training the NOx concentration prediction model according to the first aspect;
acquiring a feature to be analyzed for prediction;
and inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
A third aspect of the present application provides a training apparatus for a NOx concentration prediction model, including:
the acquisition unit is used for acquiring training data acquired from a distributed control system of the SCR denitration system;
the selection unit is used for performing feature selection from the training data based on an XGboost algorithm and correlation analysis to obtain a training feature set for training and a verification feature set for verification;
the training unit is used for training a preset neural network by taking the training characteristics in the training characteristic set as input parameters and the training NOx prediction concentration corresponding to the training characteristics as an output result to obtain an initial prediction model;
the verification unit is used for inputting verification characteristics in the verification characteristic set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model;
a calculation unit for calculating a prediction error between a verification NOx actual concentration corresponding to the verification feature and the verification NOx predicted concentration;
and the adjusting unit is used for adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
Preferably, the preset neural network is an LSTM neural network.
Preferably, the LSTM neural network comprises: an input layer, a hidden layer and an output layer;
the input layer, the hidden layer, and the output layer each include 128 nodes.
A fourth aspect of the present application provides a prediction apparatus for a NOx concentration prediction model obtained by training with a training apparatus for a NOx concentration prediction model according to the third aspect;
an acquisition unit configured to acquire a feature to be analyzed for prediction;
and the prediction unit is used for inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
According to the technical scheme, the embodiment of the application has the following advantages:
the application provides a training method of a NOx concentration prediction model, which comprises the following steps: acquiring training data acquired from a decentralized control system of the SCR denitration system; based on an XGboost algorithm and correlation analysis, performing feature selection from training data to obtain a training feature set for training and a verification feature set for verification; training a preset neural network by taking training characteristics in the training characteristic set as input parameters and training NOx prediction concentration corresponding to the training characteristics as output results to obtain an initial prediction model; inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model; calculating a prediction error between the actual concentration of the verified NOx and the predicted concentration of the verified NOx corresponding to the verification feature; and adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
According to the method, after training data used for training are obtained, feature selection is conducted on the training data, a training feature set and a verification feature set are obtained respectively, then a preset neural network is trained through the training feature set to obtain an initial prediction model, then the initial prediction model is adjusted through the verification feature set, and a trained target prediction model can be obtained.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 is a schematic flow chart diagram illustrating a first embodiment of a method for training a NOx concentration prediction model according to an embodiment of the present disclosure;
FIG. 2 is a schematic flow chart diagram illustrating a second embodiment of a method for training a NOx concentration prediction model according to an embodiment of the present disclosure;
FIG. 3 is a schematic diagram of a training process of a method for training a NOx concentration prediction model according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of a training apparatus of a NOx concentration prediction model in an embodiment of the present application.
Detailed Description
The embodiment of the application provides a training method, a prediction method and a prediction device of a NOx concentration prediction model, and solves the technical problem that the prediction accuracy of the existing prediction model is not high when the NOx concentration is predicted.
In order to make the technical solutions of the present application better understood, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The application provides a training method of a NOx concentration prediction model in a first aspect.
Referring to fig. 1, a flowchart of a first embodiment of a method for training a NOx concentration prediction model according to an embodiment of the present application includes:
It can be understood that, in this embodiment, training data is first acquired from a distributed control system (DCS for short) of the SCR denitration system.
And 102, based on the XGboost algorithm and correlation analysis, performing feature selection from training data to obtain a training feature set for training and a verification feature set for verification.
The XGBoost algorithm may convert the data set into a subset having the selected characteristics. It can decide which features to select by a threshold that is used to consistently select the features to make up the training set and the test set.
In the embodiment, the selected features are divided according to 8:2 to form a training set and a test set.
And 103, training the preset neural network by taking the training characteristics in the training characteristic set as input parameters and the training NOx prediction concentration corresponding to the training characteristics as an output result to obtain an initial prediction model.
In the training stage, the training characteristics are used as input parameters, the corresponding training NOx prediction concentration of the training characteristics is used as an output result, and the preset neural network is trained. It can be understood that the neural network, when training, will adjust the model parameters according to the difference between the actual output and the target output. The training NOx actual concentration is also used in the training phase here.
The above training process for the preset neural network is similar to the existing training process for the neural network, and is not repeated herein.
And 104, inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model.
After the initial prediction model is obtained, the initial prediction model is verified through the verification feature set in this embodiment, so that after the output result of the initial prediction model to the verification feature set (i.e., the verification NOx predicted concentration) is obtained, the initial prediction model can be adjusted according to the output result and the actual result of the verification feature set (i.e., the verification NOx actual concentration). Therefore, in this embodiment, the verification features in the verification feature set are input to the initial prediction model as input parameters, so as to obtain the verification NOx predicted concentration output by the initial prediction model.
And 105, calculating a prediction error between the actual concentration of the verified NOx corresponding to the verification characteristic and the predicted concentration of the verified NOx.
And inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model, calculating the prediction error between the verification NOx actual concentration and the verification NOx prediction concentration corresponding to the verification features, and then adjusting the initial prediction model through the prediction error.
It is understood that the prediction error corresponds to a model evaluation index, and may be a root mean square error, an average absolute error, a coefficient of determination, or the like. The person skilled in the art can select the above-mentioned materials according to his own needs, and the above-mentioned materials are not limited or described herein.
And step 106, adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
In the embodiment, after training data used for training is obtained, feature selection is performed on the training data to respectively obtain a training feature set and a verification feature set, then a preset neural network is trained through the training feature set to obtain an initial prediction model, and then the initial prediction model is adjusted through the verification feature set, so that a trained target prediction model can be obtained.
The above is a first embodiment of a method for training a NOx concentration prediction model provided in the embodiment of the present application, and the following is a second embodiment of a method for training a NOx concentration prediction model provided in the embodiment of the present application.
Referring to fig. 2, a flowchart of a second embodiment of a method for training a NOx concentration prediction model according to an embodiment of the present application includes:
It should be noted that, the content of step 201 is the same as that of step 101 in the first embodiment, and reference may be specifically made to the content of step 101, which is not described herein again.
It should be noted that the training features in the training feature set include: ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet NOx mass concentration and SCR denitration efficiency.
Likewise, the training features in the training feature set also include: ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet NOx mass concentration and SCR denitration efficiency.
And step 203, training the preset neural network by taking the training characteristics in the training characteristic set as input parameters and the training NOx prediction concentration corresponding to the training characteristics as an output result to obtain an initial prediction model.
It is understood that the preset neural network in this embodiment is an LSTM neural network, which includes: an input layer, a hidden layer and an output layer; and the input layer, hidden layer, and output layer each contain 128 nodes.
An LSTM (long-short term memory) neural network, namely a long-short term memory neural network, is a time recursion neural network improved on the basis of RNN, is used for solving the problems of gradient disappearance, gradient explosion, lack of long-term memory capability and the like generated in the using process of RNN, and can be effectively applied to long-distance time sequence information.
Compared with the traditional model, the randomness, the hysteresis and the time series characteristics of the SCR system data cannot be deeply mined, the LSTM neural network is a time cycle neural network and is suitable for processing and predicting important events and regular characteristics with relatively long intervals and delays in time series, and meanwhile, the LSTM neural network is optimal in performance in predicting future data values according to historical data compared with the traditional model. Therefore, prediction of the denitration system based on the LSTM neural network has higher prediction accuracy.
As shown in FIG. 3, in training the LSTM neural network, the LSTM neural network is trained using a time-based back-propagation algorithm, which is similar to the classical back-propagation algorithm. Specifically, when training NOx predicted concentrations are obtained, an iterative point-by-point prediction method is employed.
And step 204, inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model.
It should be noted that the content of step 204 is the same as that of step 104 in the first embodiment, and reference may be specifically made to the content of step 104, which is not described herein again.
And step 205, calculating a prediction error between the actual concentration of the verified NOx corresponding to the verification characteristic and the predicted concentration of the verified NOx.
It should be noted that, the content of step 205 is the same as that of step 105 in the first embodiment, and reference may be specifically made to the content of step 105, which is not described herein again.
And step 206, determining an adjusting parameter for adjusting the initial prediction model based on the Adam algorithm and the prediction error.
The Adam algorithm is adopted in the parameter optimization algorithm, combines the advantages of a momentum gradient descent method and a root-mean-square back propagation algorithm, can calculate the adaptability of different parameters, occupies fewer resources of a processor, and shows greater advantages in practical application compared with other optimization algorithms.
And step 207, adjusting the initial prediction model based on the adjustment parameters to obtain a trained target prediction model.
In the embodiment, after training data used for training is obtained, feature selection is performed on the training data to respectively obtain a training feature set and a verification feature set, then a preset neural network is trained through the training feature set to obtain an initial prediction model, and then the initial prediction model is adjusted through the verification feature set, so that a trained target prediction model can be obtained.
In a second aspect of the present application, a method for predicting a NOx concentration prediction model is provided.
In an embodiment of the present application, in a prediction method of a NOx concentration prediction model, the NOx concentration prediction model is obtained by training through a training method of the NOx concentration prediction model according to the first aspect; the prediction method comprises the following steps:
acquiring a feature to be analyzed for prediction;
and inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
In this embodiment, since the prediction accuracy of the NOx concentration prediction model is high, the prediction accuracy is high because the predicted NOx concentration obtained by predicting the characteristic to be analyzed is matched with the actual NOx concentration.
A third aspect of the present application provides a training apparatus for a NOx concentration prediction model.
Referring to fig. 4, a training apparatus of a NOx concentration prediction model in the present embodiment includes:
an obtaining unit 401, configured to obtain training data acquired from a distributed control system of the SCR denitration system;
a selecting unit 402, configured to select features from training data based on an XGBoost algorithm and correlation analysis, to obtain a training feature set for training and a verification feature set for verification;
a training unit 403, configured to train a preset neural network by using the training features in the training feature set as input parameters and the training NOx predicted concentration corresponding to the training features as an output result, to obtain an initial prediction model;
a verification unit 404, configured to input verification features in the verification feature set as input parameters to the initial prediction model, so as to obtain a verification NOx prediction concentration output by the initial prediction model;
a calculation unit 405 for calculating a prediction error between the actual concentration of the verification NOx and the predicted concentration of the verification NOx corresponding to the verification feature;
and the adjusting unit 406 is configured to adjust the initial prediction model according to the prediction error to obtain a trained target prediction model.
Optionally, the pre-set neural network is an LSTM neural network.
Optionally, the LSTM neural network comprises: an input layer, a hidden layer and an output layer; the input layer, hidden layer, and output layer each contain 128 nodes.
In the embodiment, after training data used for training is obtained, feature selection is performed on the training data to respectively obtain a training feature set and a verification feature set, then a preset neural network is trained through the training feature set to obtain an initial prediction model, and then the initial prediction model is adjusted through the verification feature set, so that a trained target prediction model can be obtained.
A fourth aspect of the present application provides a prediction apparatus of a NOx concentration prediction model.
In the prediction device of the NOx concentration prediction model in the embodiment of the present application, the NOx concentration prediction model is obtained by training through a training device of a NOx concentration prediction model such as the third one; the prediction apparatus includes:
an acquisition unit configured to acquire a feature to be analyzed for prediction;
and the prediction unit is used for inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
In this embodiment, since the prediction accuracy of the NOx concentration prediction model is high, the prediction accuracy is high because the predicted NOx concentration obtained by predicting the characteristic to be analyzed is matched with the actual NOx concentration.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described device embodiments are merely illustrative, and for example, the division of the units is only one logical functional division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another grid network to be installed, or some features may be omitted or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed to by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.
Claims (10)
1. A method of training a NOx concentration prediction model, comprising:
acquiring training data acquired from a decentralized control system of the SCR denitration system;
based on an XGboost algorithm and correlation analysis, performing feature selection from the training data to obtain a training feature set for training and a verification feature set for verification;
training a preset neural network by taking the training features in the training feature set as input parameters and the training NOx prediction concentration corresponding to the training features as an output result to obtain an initial prediction model;
inputting the verification features in the verification feature set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model;
calculating a prediction error between a verification NOx actual concentration corresponding to the verification characteristic and the verification NOx predicted concentration;
and adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
2. The method of claim 1, wherein the pre-set neural network is an LSTM neural network.
3. The method of training a NOx concentration prediction model according to claim 2, wherein the LSTM neural network comprises: an input layer, a hidden layer and an output layer;
the input layer, the hidden layer, and the output layer each include 128 nodes.
4. The method for training the NOx concentration prediction model according to claim 1, wherein the adjusting the initial prediction model according to the prediction error to obtain the trained target prediction model specifically comprises:
determining an adjustment parameter for adjusting the initial prediction model based on an Adam algorithm and the prediction error;
and adjusting the initial prediction model based on the adjustment parameters to obtain a trained target prediction model.
5. The method of training a NOx concentration prediction model according to claim 1, wherein the training features in the training feature set include: ammonia injection mass flow, boiler load, SCR inlet flue gas temperature, SCR inlet flue gas oxygen content, SCR inlet NOx mass concentration and SCR denitration efficiency.
6. A prediction method of a NOx concentration prediction model, characterized in that the NOx concentration prediction model is trained by a training method of a NOx concentration prediction model according to any one of claims 1 to 5;
acquiring a feature to be analyzed for prediction;
and inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
7. A training apparatus for a NOx concentration prediction model, comprising:
the acquisition unit is used for acquiring training data acquired from a distributed control system of the SCR denitration system;
the selection unit is used for performing feature selection from the training data based on an XGboost algorithm and correlation analysis to obtain a training feature set for training and a verification feature set for verification;
the training unit is used for training a preset neural network by taking the training characteristics in the training characteristic set as input parameters and the training NOx prediction concentration corresponding to the training characteristics as an output result to obtain an initial prediction model;
the verification unit is used for inputting verification characteristics in the verification characteristic set into the initial prediction model as input parameters to obtain the verification NOx prediction concentration output by the initial prediction model;
a calculation unit for calculating a prediction error between a verification NOx actual concentration corresponding to the verification feature and the verification NOx predicted concentration;
and the adjusting unit is used for adjusting the initial prediction model according to the prediction error to obtain a trained target prediction model.
8. Training apparatus of the NOx concentration prediction model according to claim 7, characterized in that the preset neural network is an LSTM neural network.
9. Training apparatus of the NOx concentration prediction model according to claim 8, characterized in that the LSTM neural network comprises: an input layer, a hidden layer and an output layer;
the input layer, the hidden layer, and the output layer each include 128 nodes.
10. A prediction apparatus of a NOx concentration prediction model, characterized in that the NOx concentration prediction model is trained by a training apparatus of the NOx concentration prediction model according to any one of claims 7 to 9;
an acquisition unit configured to acquire a feature to be analyzed for prediction;
and the prediction unit is used for inputting the characteristics to be analyzed into the NOx concentration prediction model to obtain the predicted NOx concentration output by the NOx concentration prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011025113.9A CN112163335A (en) | 2020-09-25 | 2020-09-25 | Training method, prediction method and device of NOx concentration prediction model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011025113.9A CN112163335A (en) | 2020-09-25 | 2020-09-25 | Training method, prediction method and device of NOx concentration prediction model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112163335A true CN112163335A (en) | 2021-01-01 |
Family
ID=73864030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011025113.9A Pending CN112163335A (en) | 2020-09-25 | 2020-09-25 | Training method, prediction method and device of NOx concentration prediction model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112163335A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113628694A (en) * | 2021-07-05 | 2021-11-09 | 国能国华(北京)电力研究院有限公司 | Method for predicting discharge amount of nitrogen oxides of boiler |
| CN113746899A (en) * | 2021-07-29 | 2021-12-03 | 济南浪潮数据技术有限公司 | Cloud platform access method and device |
| CN114159968A (en) * | 2021-12-10 | 2022-03-11 | 山西大学 | Prediction method for cooperative control of heavy metal pollutants in power plant flue gas |
| CN114191953A (en) * | 2021-12-07 | 2022-03-18 | 国网河北能源技术服务有限公司 | Flue gas desulfurization and denitrification control method based on convolutional neural network and XGboost |
| CN115732041A (en) * | 2022-12-07 | 2023-03-03 | 中国石油大学(北京) | Carbon dioxide capture amount prediction model construction method, intelligent prediction method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084424A (en) * | 2019-04-25 | 2019-08-02 | 国网浙江省电力有限公司 | A kind of Methods of electric load forecasting based on LSTM and LGBM |
| CN110245801A (en) * | 2019-06-19 | 2019-09-17 | 中国电力科学研究院有限公司 | A kind of Methods of electric load forecasting and system based on combination mining model |
| CN110263997A (en) * | 2019-06-19 | 2019-09-20 | 燕山大学 | A kind of cement producing line flue gas NO based on deep neural networkxConcentration prediction method |
| CN111369057A (en) * | 2020-03-05 | 2020-07-03 | 电子科技大学 | Air quality prediction optimization method and system based on deep learning |
-
2020
- 2020-09-25 CN CN202011025113.9A patent/CN112163335A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084424A (en) * | 2019-04-25 | 2019-08-02 | 国网浙江省电力有限公司 | A kind of Methods of electric load forecasting based on LSTM and LGBM |
| CN110245801A (en) * | 2019-06-19 | 2019-09-17 | 中国电力科学研究院有限公司 | A kind of Methods of electric load forecasting and system based on combination mining model |
| CN110263997A (en) * | 2019-06-19 | 2019-09-20 | 燕山大学 | A kind of cement producing line flue gas NO based on deep neural networkxConcentration prediction method |
| CN111369057A (en) * | 2020-03-05 | 2020-07-03 | 电子科技大学 | Air quality prediction optimization method and system based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 李元浩: "基于神经网络的电站锅炉NOx排放预测方法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113628694A (en) * | 2021-07-05 | 2021-11-09 | 国能国华(北京)电力研究院有限公司 | Method for predicting discharge amount of nitrogen oxides of boiler |
| CN113746899A (en) * | 2021-07-29 | 2021-12-03 | 济南浪潮数据技术有限公司 | Cloud platform access method and device |
| CN114191953A (en) * | 2021-12-07 | 2022-03-18 | 国网河北能源技术服务有限公司 | Flue gas desulfurization and denitrification control method based on convolutional neural network and XGboost |
| CN114191953B (en) * | 2021-12-07 | 2024-02-20 | 国网河北能源技术服务有限公司 | Flue gas desulfurization and denitrification control method based on convolutional neural network and XGBoost |
| CN114159968A (en) * | 2021-12-10 | 2022-03-11 | 山西大学 | Prediction method for cooperative control of heavy metal pollutants in power plant flue gas |
| CN114159968B (en) * | 2021-12-10 | 2023-09-26 | 山西大学 | Prediction method for cooperative control of heavy metal multi-pollutants in power plant flue gas |
| CN115732041A (en) * | 2022-12-07 | 2023-03-03 | 中国石油大学(北京) | Carbon dioxide capture amount prediction model construction method, intelligent prediction method and device |
| CN115732041B (en) * | 2022-12-07 | 2023-10-13 | 中国石油大学(北京) | Carbon dioxide capture quantity prediction model construction method, intelligent prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112163335A (en) | Training method, prediction method and device of NOx concentration prediction model | |
| Luo et al. | AccuAir: Winning solution to air quality prediction for KDD Cup 2018 | |
| Scholz et al. | A cyclic time-dependent Markov process to model daily patterns in wind turbine power production | |
| CN112667613B (en) | Flue gas NOx prediction method and system based on multi-delay characteristic multivariable correction | |
| CN111589301A (en) | Method, device, equipment and storage medium for predicting SCR denitration performance of coal-fired power plant | |
| CN108288111B (en) | Thermal power plant exhaust smoke temperature reference value determining method and device based on association rules | |
| US6910192B2 (en) | Method of robust technology design using rational robust optimization | |
| CN116760006A (en) | Wind power prediction method, device, equipment and storage medium | |
| Weber et al. | Multicomponent reaction-diffusion processes on complex networks | |
| CN115630316A (en) | Ultrashort-term wind speed prediction method based on improved long-term and short-term memory network | |
| Wang et al. | Investigation of RBFNN Based on Improved PSO Optimization Algorithm for Performance and Emissions Prediction of a High‐Pressure Common‐Rail Diesel Engine | |
| CN114038513A (en) | Method, device and terminal for predicting mass concentration of hydrogen sulfide in coal-fired boiler | |
| Mariani et al. | Spoilers optimization to reduce the induced stresses on a racing helmet | |
| CN112712202A (en) | Short-term wind power prediction method and device, electronic equipment and storage medium | |
| Perešíni et al. | DAG-Sword: A Simulator of Large-Scale Network Topologies for DAG-Oriented Proof-of-Work Blockchains | |
| CN115527078A (en) | Data prediction method, device and equipment and model training method | |
| Reifman et al. | Identification and control of NOx emissions using neural networks | |
| Alur et al. | Relating average and discounted costs for quantitative analysis of timed systems | |
| CN111985681A (en) | Data prediction method, model training method, device and equipment | |
| Su | World CO2 Emissions: Simple Analysis and its Relationship with Global Temperature Change | |
| Shawky et al. | Modeling clones evolution in open source systems through chaos theory | |
| CN112361376B (en) | Control method and device for boiler combustion process and storage medium | |
| CN116362415B (en) | Airport ground staff oriented shift scheme generation method and device | |
| Mahmoodi et al. | A Multi-Objective Water Cycle Algorithm for the BI-Objective Multi-Mode Project Resource Renting Problem. | |
| KR102148777B1 (en) | Statistical overpower penalty calculation system for generic thermal margin analysis model |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room 501-503, annex building, Huaye building, No.1-3 Chuimao new street, Xihua Road, Yuexiu District, Guangzhou City, Guangdong Province 510000 Applicant after: China Southern Power Grid Power Technology Co.,Ltd. Address before: Room 501-503, annex building, Huaye building, No.1-3 Chuimao new street, Xihua Road, Yuexiu District, Guangzhou City, Guangdong Province 510000 Applicant before: GUANGDONG DIANKEYUAN ENERGY TECHNOLOGY Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210101 |