CN112365045A - Main steam temperature intelligent prediction method based on big data - Google Patents
Main steam temperature intelligent prediction method based on big data Download PDFInfo
- Publication number
- CN112365045A CN112365045A CN202011240988.0A CN202011240988A CN112365045A CN 112365045 A CN112365045 A CN 112365045A CN 202011240988 A CN202011240988 A CN 202011240988A CN 112365045 A CN112365045 A CN 112365045A
- Authority
- CN
- China
- Prior art keywords
- main steam
- steam temperature
- data
- variables
- time
- 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
- 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
- G06N3/04—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
-
- 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
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
Abstract
The invention relates to a main steam temperature intelligent prediction method based on big data, which comprises the following steps: firstly, from the analysis of unit operation data and system characteristics, determining main factors of main steam temperature change by comprehensively applying principal component analysis and an improved grey correlation analysis technology; secondly, establishing a main steam temperature prediction model based on an LSTM long-short term memory neural network algorithm; and finally, predicting the main steam temperature by using historical and real-time segmented load data in the SIS. Compared with the prior art, the method has the advantages of reducing the fluctuation of the main steam temperature caused by the external influence, improving the control quality, improving the safety and the economy of the unit and the like.
Description
Technical Field
The invention relates to a main steam temperature prediction technology of a thermal power generating unit, in particular to a main steam temperature intelligent prediction method based on big data.
Background
The main steam temperature object of the motor set is a large-inertia, large-delay and nonlinear time-varying thermal object. The deformation of the steam heating surface and the metal material of the boiler pipeline can be accelerated by the overhigh temperature of the main steam, and the service life of a unit is further influenced. If the heated surface is in an overtemperature state for a long time, the strength of the metal material is rapidly reduced, and even a pipe explosion phenomenon can be generated, so that an accident is caused. And the main steam temperature is too low and can reduce the whole thermal efficiency of unit, influences the economic benefits of unit, can lead to the steam humidity of last several grades of blades of steam turbine too big simultaneously, and then influences the steam turbine blade life-span. Therefore, the temperature of the main steam of the thermal power generating unit is necessary to be controlled within a certain reasonable range, and the safe, stable and efficient operation of the thermal power generating unit is ensured.
The control method commonly used at present is mainly a cascade PID control strategy, but the conventional PID control has some problems for the main steam temperature and other objects. As the process characteristics determine that pipelines of all stages of superheaters are longer, the main steam temperature has slower reaction on the control input of the main steam temperature and the change of the desuperheating water amount of the water spray desuperheater, the combustion working condition of the boiler is often unstable, the disturbance on the flue gas side is frequent and has larger disturbance amount, and the factors cause the main steam temperature to change greatly.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide an intelligent main steam temperature prediction method based on big data.
The purpose of the invention can be realized by the following technical scheme:
a main steam temperature intelligent prediction method based on big data comprises the following steps:
firstly, from the analysis of unit operation data and system characteristics, determining main factors of main steam temperature change by comprehensively applying principal component analysis and an improved grey correlation analysis technology;
secondly, establishing a main steam temperature prediction model based on an LSTM long-short term memory neural network algorithm;
and finally, predicting the main steam temperature by using historical and real-time segmented load data in the SIS.
Preferably, the method specifically comprises the following steps:
step 1) selecting a plurality of related potential model input variables from SIS data of all process control parameters of a unit, calculating a Pearson correlation coefficient between the variables, finding out variables related to main steam temperature, marking the variables with high correlation, reducing input variable dimensions by using a principal component analysis method, and selecting final input variables of a main steam temperature prediction model by using an improved grey correlation analysis method;
step 2) acquiring original data of the SIS in the latest period of time t1 on line, preprocessing the data, segmenting the unit load according to different load segments, dividing the preprocessed data into a training set and a test set in each segment of segmented load according to the preprocessed data, establishing a segmented LSTM sub neural network model with the training set data as the main steam temperature, inputting the test set data into a main steam temperature prediction model, comparing the deviation with an actual value, judging whether the model can meet the prediction requirement, and if not, adjusting the parameters;
step 3) adjusting parameters of the LSTM sub-neural network of each load section, inputting training data again after adjustment to construct a new neural network model, enabling the model to meet the prediction requirement and storing the model;
and 4) inputting online SIS real-time input data into the LSTM sub-neural network model of each load section according to the current unit power, and predicting the main steam temperature after t2 time.
Preferably, the principal component analysis method is used for deleting repeated variables and simultaneously performing dimensionality reduction on the multiple variables, and a Pearson correlation coefficient r for measuring the linear correlation strength between continuous variables is adopted and has a value between-1 and 1; 0 ≦ r ≦ 0.3, low correlation; 0.3 ≦ r ≦ 0.8, moderately relevant; 0.8 ≦ r ≦ 1, highly relevant.
Preferably, the improved grey correlation analysis method specifically comprises:
the reference data column is often denoted x0The value at the 1 st time is recorded as x0(1) The value at the 2 nd time is x0(2) The value at the k-th time is x0(k) (ii) a Reference sequence x0Can be expressed as x0＝(x0(1)，x0(2)，…x0(n)), the compared series in the correlation analysis is often denoted as x1，x2，…，xkAnalogous reference sequence x0Is represented by the formula (I) having x1＝(x1(1)，x1(2)，…x1(n))…xk＝(xk(1)，xk(2)，…xk(n))；
For a reference data column x0There are several comparison series x1，x2，…，xnIn the case of (1), data is subjected to non-dimensionalization by using an averaging method, and the difference between each comparison curve and a reference curve at each point or moment is represented by the following relationship;
in the formula, xii(k) Is the comparison curve x at the kth momentiWith reference curve x0Is called xiFor x0A correlation coefficient at the time k, wherein 0.5 is a resolution coefficient, is marked as zeta, and is selected between 0 and 1;
in the improved grey correlation analysis method, the grey correlation coefficients of two data columns participating in comparison are weighted and calculated, compared with the traditional analysis method, the time and proximity principle is considered, and the weight proportion of the historical data closest to the predicted data is improved.
Preferably, the final input variables of the main steam temperature prediction model include: the system comprises overheating desuperheating water flow, steam temperature at the front opening of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal feeding quantity of each coal mill and unit power.
Preferably, the format of the original data in step 2) is in the form of csv or database.
Preferably, the raw data in step 2) includes superheated desuperheating water flow, steam temperature at the front inlet of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill and unit power.
Preferably, the preprocessing in step 2) includes normalization, deletion of outliers and supplementation of missing values by interpolation.
Preferably, the LSTM sub-neural network model in step 2) is specifically:
3 gates, namely an input gate, a forgetting gate and an output gate, and memory cells with the same shape as the hidden state are introduced;
assuming that the number of hidden units is h, inputting the small batch X with given time step tt∈Rn*dAnd previous time step hidden state Ht-1∈Rn*hN is the number of samples, d is the number of inputs; the formula for the three gates and the candidate memory cells is as follows:
an input gate: i ist＝σ(XtWxi+Ht-1Whi+bi)
Forget the door: ft＝σ(XtWxf+Ht-1Whf+bf)
An output gate: o ist＝σ(XtWxo+Ht-1Who+bo)
Alternative memory cells:
memory cell H at the current time stept∈Rn*hThe calculation combines the information of the memory cells of the last time step and the alternative memory cells of the current time step, and updates the memory cells of the current time step:
wherein XtIs a small batch input of time step t, Ht-1Is the hidden state of the last time step t-1; wxiAnd WhiIs inputting XtAnd hidden state Ht-1Corresponding input gate weight matrix, biIs the input gate offset; wxfAnd WhfIs inputting XtAnd hidden state Ht-1Corresponding forgetting gate weight matrix, bfIs a forgetting gate bias; wxoAnd WhoIs inputting XtAnd hidden state Ht-1Corresponding output gate weight matrix, boIs the output gate offset; wxcAnd WhcIs inputting XtAnd hidden state Ht-1Corresponding candidate memory cell weight matrix, bcIs an alternative memory cell bias; ct-1Is the memory cell of the previous time step t-1;
then, the output gate controls the memory cell to the hidden state HtFlow of information of (1): ht＝Ot⊙tanh(Ct)；
In the LSTM, selecting an activation function, and using a sigma function as the activation function by a forgetting gate, an input gate and an output gate; in generating the alternative memory, a hyperbolic tangent function tanh is used as the activation function.
Preferably, the online SIS real-time input data of step 4) includes superheated desuperheating water flow, steam temperature at an advance port of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill and unit power.
Compared with the prior art, the invention provides the intelligent prediction analysis method of the main steam temperature based on the big data by combining the data analysis of the main steam temperature system and a plant monitoring information system (SIS) of a power plant and utilizing a deep learning related prediction technology aiming at the characteristics of large inertia, large delay, nonlinearity and strong coupling of the main steam temperature system of the supercritical (supercritical) unit. The LSTM neural network is used for predicting the main steam temperature and providing information for early intervention, so that fluctuation of the main steam temperature caused by external influence is reduced, the control quality is improved, and the safety and the economical efficiency of a unit are improved
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
The invention provides a main steam temperature intelligent prediction analysis method based on big data. The following description will be given by taking a 660MW thermal power plant as an example (see fig. 1).
The method mainly comprises the following steps:
1) firstly, all SIS point tables of the unit are obtained, corresponding data are searched from an SIS database, and relevant potential model input variables are selected according to different types of boiler of the unit. And then calculating a Pearson correlation coefficient between the variables, finding out the variable related to the temperature of the main steam, and marking the variable with high correlation. And reducing input variable dimensionality by applying a principal component analysis method, and finally selecting a final input variable of the main steam temperature prediction model by using an improved grey correlation analysis method. Default input variables are: the system comprises overheating desuperheating water flow, steam temperature at the front opening of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal feeding quantity of each coal mill and unit power. The input variables can be slightly adjusted according to actual results of different unit boiler types, principal component analysis methods and improved grey correlation analysis methods.
After the principal component analysis method, repeated variables (linearly related variables) are deleted, so that the key characteristics of the original data are ensured while input variables are reduced as much as possible. More than 200 potential input variables are subjected to dimensionality reduction, and a small number of variables are used for representing other variables, so that a large amount of computing resources can be saved. The pearson correlation coefficient r is used to measure the strength of linear correlation between successive variables, and has a value between-1 and 1. 0 ≦ r ≦ 0.3, low correlation; 0.3 ≦ r ≦ 0.8, moderately relevant; 0.8 ≦ r ≦ 1, highly relevant.
And then, judging the correlation degree among the variables according to the similarity degree among the curves by using an improved grey correlation analysis method. The larger the correlation coefficient of the two variables is, the larger the correlation degree is, and the smaller the correlation coefficient is otherwise. The reference data column is often denoted x0The value at the 1 st time is recorded as x0(1) The value at the 2 nd time is x0(2) The value at the k-th time is x0(k) In that respect Thus, reference sequence x0Can be expressed as x0＝(x0(1)，x0(2)，…x0(n)). The compared series in the correlation analysis is often denoted as x1，x2，…，xkAnalogous reference sequence x0Is represented by the formula (I) having x1＝(x1(1)，x1(2)，…x1(n))…xk＝(xk(1)，xk(2)，…xk(n)). For a reference data column x0There are several comparison series x1，x2，…，xnIn the case of (1), the data should be subjected to non-dimensionalization processing by using an averaging method. The difference between each comparison curve and the reference curve at each point (time) can be expressed by the following relationship.
In the formula, xii(k) Is the comparison curve x at the kth momentiWith reference curve x0Is called xiFor x0The correlation coefficient at time k. Wherein 0.5 is a resolution factor, denoted as ζ, generally selected between 0 and 1;
in the improved grey correlation analysis method, the grey correlation coefficients of two data columns participating in comparison are weighted, the time and proximity principle is considered, and the weight proportion of the historical data closest to the predicted data is improved.
2) And (3) acquiring original data (csv or in a database form) of the SIS at the latest time t1 on line, wherein the original data comprises overheating temperature-reducing water flow, steam temperature of an advancing opening of each stage of superheater temperature reducer, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill and unit power. These data are preprocessed, including normalization, outlier deletion, missing value interpolation, etc. And (3) segmenting the load of the unit according to every 100MW, dividing the load of each segment into a training set and a test set according to the preprocessed data, and establishing a segmented LSTM sub-neural network model with the output as the main steam temperature by using the training set data. And inputting the test set data into a main steam temperature prediction model, comparing the deviation with an actual value, and judging whether the model can meet the prediction requirement. And if the requirements are not met, performing parameter adjustment.
LSTM: a Long Short-Term Memory network (LSTM) is a special RNN (recurrent neural network), mainly aims to solve the problems of gradient elimination and gradient explosion in the Long sequence training process, and compared with the common RNN, the LSTM can have better performance in a longer sequence. The LSTM incorporates 3 gates, an input gate, a forgetting gate and an output gate, and memory cells of the same shape as the hidden state.
Setting the number of hidden units as h, and setting the small batch input X of time step tt∈Rn*d(number of samples n, number of inputs d) and hidden state H of last time stept-1∈Rn*h. The formula for the three gates and the candidate memory cells is as follows:
an input gate: i ist＝σ(XtWxi+Ht-1Whi+bi)
Forget the door: ft＝σ(XtWxf+Ht-1Whf+bf)
An output gate: o ist＝σ(XtWxo+Ht-1Who+bo)
Alternative memory cells:
memory cell H at the current time stept∈Rn*hThe calculation combines the information of the memory cells of the last time step and the alternative memory cells of the current time step, and updates the memory cells of the current time step:
then, the output gate controls the memory cell to the hidden state HtFlow of information of (1): ht＝Ot⊙tanh(Ct)。
In the LSTM, selecting an activation function, and using a sigma (Sigmoid) function as the activation function for a forgetting gate, an input gate and an output gate; in generating the alternative memory, a hyperbolic tangent function tanh is used as the activation function.
3) And adjusting the super parameters such as LSTM sub neural network learning rate of each load section, and inputting training data again to construct a new neural network model after adjustment, so that the model meets the prediction requirement and is stored.
4) And data are input in real time through an online SIS, and the data comprise overheating desuperheating water flow, steam temperature of an advancing opening of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill and unit power. And inputting the data into an LSTM sub neural network model of each load section according to the current unit power, and predicting the main steam temperature after t2 time.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A main steam temperature intelligent prediction method based on big data is characterized by comprising the following steps:
firstly, from the analysis of unit operation data and system characteristics, determining main factors of main steam temperature change by comprehensively applying principal component analysis and an improved grey correlation analysis technology;
secondly, establishing a main steam temperature prediction model based on an LSTM long-short term memory neural network algorithm;
and finally, predicting the main steam temperature by using historical and real-time segmented load data in the SIS.
2. The intelligent main steam temperature prediction method based on big data as claimed in claim 1, wherein the method specifically comprises the following steps:
step 1) selecting a plurality of related potential model input variables from SIS data of all process control parameters of a unit, calculating a Pearson correlation coefficient between the variables, finding out variables related to main steam temperature, marking the variables with high correlation, reducing input variable dimensions by using a principal component analysis method, and selecting final input variables of a main steam temperature prediction model by using an improved grey correlation analysis method;
step 2) acquiring original data of the SIS in the latest period of time t1 on line, preprocessing the data, segmenting the unit load according to different load segments, dividing the preprocessed data into a training set and a test set in each segment of segmented load according to the preprocessed data, establishing a segmented LSTM sub neural network model with the training set data as the main steam temperature, inputting the test set data into a main steam temperature prediction model, comparing the deviation with an actual value, judging whether the model can meet the prediction requirement, and if not, adjusting the parameters;
step 3) adjusting parameters of the LSTM sub-neural network of each load section, inputting training data again after adjustment to construct a new neural network model, enabling the model to meet the prediction requirement and storing the model;
and 4) inputting online SIS real-time input data into the LSTM sub-neural network model of each load section according to the current unit power, and predicting the main steam temperature after t2 time.
3. The intelligent big-data-based main steam temperature prediction method according to claim 2, wherein the principal component analysis method is used for deleting repeated variables and simultaneously reducing the dimension of the multiple variables, and a pearson correlation coefficient r for measuring the linear correlation strength between continuous variables is adopted, and the value of the pearson correlation coefficient r is between-1 and 1; 0 ≦ r ≦ 0.3, low correlation; 0.3 ≦ r ≦ 0.8, moderately relevant; 0.8 ≦ r ≦ 1, highly relevant.
4. The intelligent main steam temperature prediction method based on big data as claimed in claim 2, wherein the improved grey correlation analysis method is specifically:
the reference data column is often denoted x0The value at the 1 st time is recorded as x0(1) The value at the 2 nd time is x0(2) The value at the k-th time is x0(k) (ii) a Reference sequence x0Can be expressed as x0＝(x0(1)，x0(2)，…x0(n)), the compared series in the correlation analysis is often denoted as x1，x2，…，xkAnalogous reference sequence x0Is represented by the formula (I) having x1＝(x1(1)，x1(2)，…x1(n))…xk＝(xk(1)，xk(2)，…xk(n))；
For a reference data column x0There are several comparison series x1，x2，…，xnIn the case of (1), data is subjected to non-dimensionalization by using an averaging method, and the difference between each comparison curve and a reference curve at each point or moment is represented by the following relationship;
in the formula, xii(k) Is the comparison curve x at the kth momentiWith reference curve x0Is called xiFor x0A correlation coefficient at the time k, wherein 0.5 is a resolution coefficient, is marked as zeta, and is selected between 0 and 1;
in the improved grey correlation analysis method, the grey correlation coefficients of two data columns involved in comparison are weighted.
5. The big data-based intelligent prediction method for main steam temperature as claimed in claim 2, wherein the final input variables of the main steam temperature prediction model comprise: the system comprises overheating desuperheating water flow, steam temperature at the front opening of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal feeding quantity of each coal mill and unit power.
6. The big data-based intelligent prediction method for main steam temperature as claimed in claim 2, wherein the format of the raw data in step 2) is csv or database form.
7. The intelligent main steam temperature prediction method based on big data as claimed in claim 2, wherein the raw data in step 2) includes superheat desuperheating water flow, steam temperature at the front inlet of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air damper, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill, and unit power.
8. The big data-based intelligent prediction method for main steam temperature as claimed in claim 2, wherein the preprocessing in step 2) comprises normalization, removing outliers and supplementing missing values by interpolation.
9. The big-data-based intelligent main steam temperature prediction method according to claim 2, wherein the LSTM sub-neural network model in the step 2) is specifically:
3 gates, namely an input gate, a forgetting gate and an output gate, and memory cells with the same shape as the hidden state are introduced;
assuming that the number of hidden units is h, inputting the small batch X with given time step tt∈Rn*dAnd previous time step hidden state Ht-1∈Rn*hN is the number of samples, d is the number of inputs; the formula for the three gates and the candidate memory cells is as follows:
an input gate: i ist＝σ(XtWxi+Ht-1Whi+bi)
Forget the door: ft＝σ(XtWxf+Ht-1Whf+bf)
An output gate: o ist＝σ(XtWxo+Ht-1Who+bo)
Alternative memory cells:
memory cell H at the current time stept∈Rn*hThe calculation combines the information of the memory cells of the last time step and the alternative memory cells of the current time step, and updates the memory cells of the current time step:
wherein XtIs the time stepSmall batch input of t, Ht-1Is the hidden state of the last time step t-1; wxiAnd WhiIs inputting XtAnd hidden state Ht-1Corresponding input gate weight matrix, biIs the input gate offset; wxfAnd WhfIs inputting XtAnd hidden state Ht-1Corresponding forgetting gate weight matrix, bfIs a forgetting gate bias; wxoAnd WhoIs inputting XtAnd hidden state Ht-1Corresponding output gate weight matrix, boIs the output gate offset; wxcAnd WhcIs inputting XtAnd hidden state Ht-1Corresponding candidate memory cell weight matrix, bcIs an alternative memory cell bias; ct-1Is the memory cell of the previous time step t-1;
then, the output gate controls the memory cell to the hidden state HtFlow of information of (1): ht＝Ot⊙tanh(Ct)；
In the LSTM, selecting an activation function, and using a sigma function as the activation function by a forgetting gate, an input gate and an output gate; in generating the alternative memory, a hyperbolic tangent function tanh is used as the activation function.
10. The intelligent main steam temperature prediction method based on big data as claimed in claim 2, wherein the online SIS real-time input data of step 4) comprises superheat desuperheating water flow, steam temperature at the front inlet of each stage of superheater desuperheater, water supply temperature, main steam pressure, position feedback of each secondary air baffle, water-coal ratio, primary air quantity of each coal mill, coal supply quantity of each coal mill and unit power.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011240988.0A CN112365045A (en) | 2020-11-09 | 2020-11-09 | Main steam temperature intelligent prediction method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011240988.0A CN112365045A (en) | 2020-11-09 | 2020-11-09 | Main steam temperature intelligent prediction method based on big data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112365045A true CN112365045A (en) | 2021-02-12 |
Family
ID=74510340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011240988.0A Pending CN112365045A (en) | 2020-11-09 | 2020-11-09 | Main steam temperature intelligent prediction method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112365045A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113359425A (en) * | 2021-07-06 | 2021-09-07 | 浙江浙能技术研究院有限公司 | Thermal power plant boiler main steam temperature intelligent control system based on LSTM neural network PID optimization |
| CN113378477A (en) * | 2021-06-29 | 2021-09-10 | 西北师范大学 | Boiler superheater area high and low temperature prediction method based on deep learning method |
| CN113467237A (en) * | 2021-06-22 | 2021-10-01 | 东南大学 | Dynamic modeling method for main steam temperature based on deep learning |
| CN114383130A (en) * | 2022-01-12 | 2022-04-22 | 浙江中智达科技有限公司 | Method, device and equipment for controlling temperature of superheated steam of boiler and storage medium |
| CN115079572A (en) * | 2022-06-30 | 2022-09-20 | 福建省龙德新能源有限公司 | Energy management control system for preparing lithium hexafluorophosphate and control method thereof |
| CN115440011A (en) * | 2022-08-23 | 2022-12-06 | 华能太仓发电有限责任公司 | Multi-parameter coupling main reheat steam temperature abnormity early warning system and method |
| CN115525076A (en) * | 2022-10-08 | 2022-12-27 | 北京航空航天大学 | Atomic gas chamber temperature active disturbance rejection control method based on LSTM neural network |
| WO2024045243A1 (en) * | 2022-08-31 | 2024-03-07 | 福建省龙德新能源有限公司 | Automatic batching system for preparation of lithium hexafluorophosphate and batching method of automatic batching system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101825869A (en) * | 2010-05-13 | 2010-09-08 | 上海交通大学 | Method for identifying superheater model parameters based on data drive |
| CN105242155A (en) * | 2015-11-18 | 2016-01-13 | 南京工程学院 | Transformer fault diagnosis method based on entropy weight method and grey correlation analysis |
| CN105550700A (en) * | 2015-12-08 | 2016-05-04 | 国网山东省电力公司电力科学研究院 | Time series data cleaning method based on correlation analysis and principal component analysis |
| CN107844865A (en) * | 2017-11-20 | 2018-03-27 | 天津科技大学 | Feature based parameter chooses the stock index prediction method with LSTM models |
| CN109063939A (en) * | 2018-11-01 | 2018-12-21 | 华中科技大学 | A kind of wind speed forecasting method and system based on neighborhood door shot and long term memory network |
| CN110285403A (en) * | 2019-06-10 | 2019-09-27 | 华北电力大学 | Main Steam Temperature Control method based on controlled parameter prediction |
| CN110414788A (en) * | 2019-06-25 | 2019-11-05 | 国网上海市电力公司 | A kind of power quality prediction technique based on similar day and improvement LSTM |
-
2020
- 2020-11-09 CN CN202011240988.0A patent/CN112365045A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101825869A (en) * | 2010-05-13 | 2010-09-08 | 上海交通大学 | Method for identifying superheater model parameters based on data drive |
| CN105242155A (en) * | 2015-11-18 | 2016-01-13 | 南京工程学院 | Transformer fault diagnosis method based on entropy weight method and grey correlation analysis |
| CN105550700A (en) * | 2015-12-08 | 2016-05-04 | 国网山东省电力公司电力科学研究院 | Time series data cleaning method based on correlation analysis and principal component analysis |
| CN107844865A (en) * | 2017-11-20 | 2018-03-27 | 天津科技大学 | Feature based parameter chooses the stock index prediction method with LSTM models |
| CN109063939A (en) * | 2018-11-01 | 2018-12-21 | 华中科技大学 | A kind of wind speed forecasting method and system based on neighborhood door shot and long term memory network |
| CN110285403A (en) * | 2019-06-10 | 2019-09-27 | 华北电力大学 | Main Steam Temperature Control method based on controlled parameter prediction |
| CN110414788A (en) * | 2019-06-25 | 2019-11-05 | 国网上海市电力公司 | A kind of power quality prediction technique based on similar day and improvement LSTM |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113467237A (en) * | 2021-06-22 | 2021-10-01 | 东南大学 | Dynamic modeling method for main steam temperature based on deep learning |
| CN113378477A (en) * | 2021-06-29 | 2021-09-10 | 西北师范大学 | Boiler superheater area high and low temperature prediction method based on deep learning method |
| CN113378477B (en) * | 2021-06-29 | 2023-05-26 | 西北师范大学 | Deep learning method-based boiler superheater area high-low temperature prediction method |
| CN113359425A (en) * | 2021-07-06 | 2021-09-07 | 浙江浙能技术研究院有限公司 | Thermal power plant boiler main steam temperature intelligent control system based on LSTM neural network PID optimization |
| CN114383130A (en) * | 2022-01-12 | 2022-04-22 | 浙江中智达科技有限公司 | Method, device and equipment for controlling temperature of superheated steam of boiler and storage medium |
| CN114383130B (en) * | 2022-01-12 | 2022-08-05 | 浙江中智达科技有限公司 | Method, device and equipment for controlling temperature of superheated steam of boiler and storage medium |
| CN115079572A (en) * | 2022-06-30 | 2022-09-20 | 福建省龙德新能源有限公司 | Energy management control system for preparing lithium hexafluorophosphate and control method thereof |
| CN115440011A (en) * | 2022-08-23 | 2022-12-06 | 华能太仓发电有限责任公司 | Multi-parameter coupling main reheat steam temperature abnormity early warning system and method |
| WO2024045243A1 (en) * | 2022-08-31 | 2024-03-07 | 福建省龙德新能源有限公司 | Automatic batching system for preparation of lithium hexafluorophosphate and batching method of automatic batching system |
| CN115525076A (en) * | 2022-10-08 | 2022-12-27 | 北京航空航天大学 | Atomic gas chamber temperature active disturbance rejection control method based on LSTM neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112365045A (en) | Main steam temperature intelligent prediction method based on big data | |
| CN110285403B (en) | Main steam temperature control method based on controlled parameter estimation | |
| Zhang et al. | Cascade control of superheated steam temperature with neuro-PID controller | |
| CN110413601B (en) | Generator data screening method based on Gauss naive Bayes and prediction error method | |
| CN104776446A (en) | Combustion optimization control method for boiler | |
| CN107609667B (en) | Heat supply load prediction method and system based on Box _ cox transformation and UFCNN | |
| Hou et al. | Novel fuzzy modeling and energy-saving predictive control of coordinated control system in 1000 MW ultra-supercritical unit | |
| CN110598929B (en) | Wind power nonparametric probability interval ultrashort term prediction method | |
| CN113359425A (en) | Thermal power plant boiler main steam temperature intelligent control system based on LSTM neural network PID optimization | |
| CN115077640A (en) | Real-time prediction method for main steam flow in urban solid waste incineration process | |
| Wang et al. | Deep-learning modeling and control optimization framework for intelligent thermal power plants: A practice on superheated steam temperature | |
| CN113515891B (en) | Emulsion explosive quality prediction and optimization method | |
| CN114818847A (en) | Steam turbine backpressure trend prediction method based on catboost algorithm | |
| Hu et al. | Feedforward DMC-PID cascade strategy for main steam temperature control system in fossil-fired power plant | |
| CN113836819B (en) | Bed temperature prediction method based on time sequence attention | |
| CN109270917B (en) | Intelligent power plant steam turbine bearing-oriented closed-loop control system fault degradation state prediction method | |
| CN116029433A (en) | Energy efficiency reference value judging method, system, equipment and medium based on grey prediction | |
| CN113158536B (en) | Primary frequency modulation capability calculation monitoring method and system | |
| CN111624979B (en) | Industrial closed-loop control loop multi-oscillation detection and tracing method based on slow characteristic analysis | |
| Lusenko et al. | Development of a waste-heat boiler model based on recurrent neural networks | |
| Ma et al. | ANN and PSO based intelligent model predictive optimal control for large-scale supercritical power unit | |
| CN113885326B (en) | Closed loop system identification method based on set value effective excitation by considering feedforward | |
| Alitasb et al. | Multiple-input multiple-output Radial Basis Function Neural Network modeling and model predictive control of a biomass boiler | |
| CN114383130B (en) | Method, device and equipment for controlling temperature of superheated steam of boiler and storage medium | |
| Nazaruddin et al. | Improving the performance of industrial boiler using artificial neural network modeling and advanced combustion control |
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 |