CN111242345A - Nuclear power unit electric power prediction method based on cluster analysis and random forest regression - Google Patents
Nuclear power unit electric power prediction method based on cluster analysis and random forest regression Download PDFInfo
- Publication number
- CN111242345A CN111242345A CN201911364412.2A CN201911364412A CN111242345A CN 111242345 A CN111242345 A CN 111242345A CN 201911364412 A CN201911364412 A CN 201911364412A CN 111242345 A CN111242345 A CN 111242345A
- Authority
- CN
- China
- Prior art keywords
- random forest
- forest regression
- electric power
- regression model
- unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000007621 cluster analysis Methods 0.000 title claims abstract description 16
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 13
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 230000000694 effects Effects 0.000 claims description 8
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 6
- 238000003066 decision tree Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 claims description 2
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 abstract description 4
- 230000009467 reduction Effects 0.000 abstract description 3
- 230000008859 change Effects 0.000 description 4
- 230000000630 rising effect Effects 0.000 description 3
- 239000013535 sea water Substances 0.000 description 3
- 238000010977 unit operation Methods 0.000 description 2
- 230000003749 cleanliness Effects 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Primary Health Care (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the field of power systems, and relates to a nuclear power unit electric power prediction method based on cluster analysis and random forest regression. The method carries out feature extraction on the historical operation data of the unit through cluster analysis to train a random forest regression model, so that the electric power of the unit can be predicted according to the real-time operation data of the unit. Because the data dimensionality reduction is realized through the cluster analysis, and the establishment of a mechanism model is avoided through the establishment of a random forest regression model, the method has the advantages of high prediction precision, high prediction speed and strong generalization capability.
Description
Technical Field
The invention belongs to the field of power systems, and relates to a nuclear power unit electric power prediction method based on cluster analysis and random forest regression.
Background
Nuclear energy plays an important role in the world energy structure due to its characteristics of cleanliness and stable power generation. A number of developed countries, including france, the united states, and germany, have long developed nuclear power technologies. In the face of increasingly severe resource and environmental pressure, the energy situation in China is increasingly severe, and the adjustment of the energy structure is not easy. In 2020, the installed capacity of the nuclear power is expected to reach 5800 ten thousand kilowatts, and the continuously rising installed scale prompts people to pay attention to the problems generated in the operation of the nuclear power unit.
In recent years, due to the arrangement problem of water taking and discharging ports of a certain domestic nuclear power station, the average temperature rise of a whole tide is larger than the designed temperature rise when a unit runs in summer, the change of seawater temperature causes the change of unit backpressure, the phenomenon of lower output electric power of the unit under working conditions in summer occurs, and the phenomenon is particularly represented as the rapid change of the temperature of a seawater inlet of a condenser during the rising tide and falling tide. In order to avoid the thermal power of the reactor of the unit exceeding 100% during the highest rising period of the seawater temperature, the output of the unit is limited during daily operation. Aiming at the problem, the operation data of the unit is monitored, and the influence relation of the operation parameters of the unit on the electric power of the unit is established by using the historical data of the operation of the unit, so that the change of the electric power of the unit is accurately predicted, and the operation guidance is provided for the adjustment of the opening of the high-pressure regulating valve of the steam turbine.
When a nuclear power unit actually operates, parameters influencing the electric power of the unit are various, and the parameters have strong relativity and collinearity, so that the mechanism that each parameter acts on the electric power of the unit independently is difficult to analyze theoretically, and a traditional multiple linear regression model is difficult to achieve high fitting accuracy; compared with the traditional regression model, the random forest regression model has the advantages of strong generalization capability, high training speed, simplicity in implementation, accuracy in prediction and the like, is applied to prediction in the field of wind power at present, but is not applied to nuclear power units, so that the random forest regression model has important significance in application research of the random forest regression model in conventional islands of nuclear power stations.
Disclosure of Invention
The invention adopts a nuclear power unit electric power prediction method based on cluster analysis and random forest regression. The characteristic extraction is carried out on the historical operation data of the unit through cluster analysis, and the characteristic extraction is used for training a random forest regression model, so that the electric power of the unit can be predicted according to the real-time operation data of the unit. Because the data dimensionality reduction is realized through the cluster analysis, and the establishment of a mechanism model is avoided through the establishment of a random forest regression model, the method has the advantages of high prediction precision, high prediction speed and strong generalization capability.
The invention adopts the following technical scheme:
the nuclear power unit electric power prediction method based on cluster analysis and random forest regression comprises the following steps:
(1) and acquiring historical operating data of the unit.
(2) And cleaning historical operating data of the unit and removing abnormal data.
(3) And performing feature extraction on the washed historical operating data by adopting clustering analysis to obtain a feature value.
(4) The characteristic values and the target values (electric power) are subjected to non-dimensionalization processing by adopting a standardized method, so that a data set for random forest regression is obtained.
The normalized formula can be expressed as:
in the formula, X′For normalized sample values, X is the original sample value,
is the mean value of the original sample, and s is the standard deviation of the original sample.
(5) And (4) splitting the data set obtained in the step (4) to obtain a training set and a test set. The training set is used for training the random forest regression model, and the testing set is used for evaluating the prediction effect of the random forest regression model.
(6) And establishing a random forest regression model, and training the random forest regression model by using a training set.
(7) And testing the random forest regression model by using a test set, and evaluating the prediction effect of the random forest regression model by using three evaluation indexes, namely Mean Absolute Error (MAE), Mean Square Error (MSE) and R square value.
(8) And (3) optimizing a random forest regression model, and adjusting the minimum leaf number and the decision tree number of the random forest (the smaller minimum leaf number enables the model to capture noise in a training set more easily, the larger decision tree number enables the model to have better performance, and meanwhile, the training speed of the model is slowed down), and repeating the steps (6) to (8) until the R square value of the random forest regression model is larger than 0.999.
(9) And predicting the electric power of the unit by using a random forest regression model according to the real-time operation data of the unit.
Further, in the step (1), the historical operation data of the unit is historical operation data of a conventional island of the nuclear power plant, and the historical operation data includes main steam parameters, condenser operation parameters, regenerator reheater operation parameters and the like.
Further, in the step (2), the abnormal data are measured point abnormal data and operation process data, such as data generated when the unit is started and stopped.
Further, in the step (5), before splitting the data set, the order of the data set should be disordered to randomly sort the data set, and then splitting is performed.
Further, in step (7), the R-squared value is also called a decision coefficient, which reflects the proportion of the total variation of the dependent variable that can be explained by the independent variable through the regression relationship.
Due to the adoption of the technical scheme, compared with the prior art, the invention has the following advantages:
in the invention, a statistical learning method is introduced, the characteristic extraction is carried out on the historical operating data of the unit by adopting cluster analysis, the data dimension reduction is realized, and then the characteristic value and the target value are trained by random forest regression, thereby avoiding the establishment of a mechanism model. The unit power prediction is carried out by adopting a random forest regression model, so that the prediction precision is greatly improved, and the prediction time is shortened. The random forest regression model has strong generalization capability and can adapt to different nuclear power generating units.
Drawings
FIG. 1 is a flow chart of a nuclear power generating unit electric power prediction method based on cluster analysis and random forest regression.
FIG. 2 is a local graph of cluster analysis results in an embodiment of the present invention.
FIG. 3 is a local graph comparing the predicted value of the random forest regression model with the actual value of the unit operation in the embodiment of the invention.
Detailed Description
The invention is further explained by an embodiment of predicting the electric power of a 1000MW nuclear power plant unit in combination with the attached drawings.
As shown in FIG. 1, the invention provides a nuclear power unit electric power prediction method based on cluster analysis and random forest regression, which specifically comprises the following steps:
(1) and acquiring historical operating data of the unit. In the embodiment, historical operating data of a conventional island of a 1000MW nuclear power station from 2016 to 8 months in 2019 are obtained, 200 operating data measuring points are obtained, and the sampling interval is 10 min.
(2) And cleaning historical operating data of the unit and removing abnormal data. In the present embodiment, there are 174410 samples in total after the abnormal data is eliminated.
(3) And performing feature extraction on the washed historical operating data by adopting clustering analysis to obtain a feature value. FIG. 2 is a local graph of cluster analysis results in an embodiment of the present invention. According to the graph, the inlet temperature and the outlet temperature of the circulating water of the condenser are similar, and the average values of the measuring points of the inlet temperature and the outlet temperature of the circulating water of the condenser are respectively taken as characteristic values. For the same reason, the average of the measured points of the condenser inlet pressure was taken as the characteristic value.
The characteristic values of the embodiment are finally selected as shown in table 1 by combining the theory and the actual experience of the power plant operation.
TABLE 1 list of feature values (average of multiple points in the same feature)
(4) The characteristic values and the target values (electric power) are subjected to non-dimensionalization processing by adopting a standardized method, so that a data set for random forest regression is obtained. The normalized formula can be expressed as:
in the formula, X′For normalized sample values, X is the original sample value,
is the mean value of the original sample, and s is the standard deviation of the original sample.
(5) And randomly splitting the data set obtained in the last step to obtain a training set and a testing set. The training set is used for training the random forest regression model, and the testing set is used for evaluating the prediction effect of the random forest regression model. In this embodiment, the number of test set samples is 30% of the number of data lump samples.
(6) And establishing a random forest regression model, and training the random forest regression model by using a training set. In this embodiment, the minimum number of leaves is 1, and the number of decision trees is 100.
(7) And testing the random forest regression model by using the test set, and evaluating the prediction effect of the random forest regression model. In this embodiment, the Mean Absolute Error (MAE) of the random forest regression model is 0.01066, the Mean Square Error (MSE) is 0.00043, and the R-square value is 0.99957.
(8) Optimizing a random forest regression model, and adjusting the minimum leaf number and the decision tree number of the random forest. In this embodiment, since the R-square value of the random forest regression model is greater than 0.999, the parameters of the random forest regression model are not adjusted.
(9) And predicting the electric power of the unit by using a random forest regression model according to the real-time operation data of the unit. FIG. 3 is a local graph comparing the predicted value of the random forest regression model with the actual value of the unit operation in the embodiment of the invention. In the graph, the actual electric power value is a scattered point, the predicted electric power value is a solid line, and the predicted electric power value and the actual electric power value are almost overlapped, which shows that the random forest regression model can realize accurate prediction of the electric power of the unit.
In addition to the random forest regression model, other regression models were tried, and the regression models and their predicted effects of this example are listed in table 2.
TABLE 2 regression model and its predicted effect
As can be seen from the table, the random forest regression model has the minimum MAE and MSE and the R square value closest to 1, so that the random forest regression model has the highest accuracy on the prediction of the electric power of the unit in the regression model, and can better provide guidance for the operation of the unit.
Claims (7)
1. The nuclear power unit electric power prediction method based on cluster analysis and random forest regression comprises the following steps:
(1) acquiring historical operating data of the unit;
(2) cleaning historical operating data of the unit, and eliminating abnormal data;
(3) performing feature extraction on the washed historical operating data by adopting clustering analysis to obtain a feature value;
(4) carrying out non-dimensionalization processing on the characteristic value and the target value by adopting a standardized method, thereby obtaining a data set for random forest regression;
the normalized formula is expressed as:
wherein X' is a normalized sample value, and X is an original sample value，
Is the mean value of the original sample, and s is the standard deviation of the original sample;
(5) splitting the data set obtained in the step (4) to obtain a training set and a test set; the training set is used for training a random forest regression model, and the testing set is used for evaluating the prediction effect of the random forest regression model;
(6) establishing a random forest regression model, and training the random forest regression model by using a training set;
(7) testing the random forest regression model by using a test set, and evaluating the prediction effect of the random forest regression model by using three evaluation indexes, namely Mean Absolute Error (MAE), Mean Square Error (MSE) and R square value;
(8) optimizing a random forest regression model, and adjusting the minimum leaf number and the decision tree number of a random forest; repeating the steps (6) to (8) until the R square value of the random forest regression model is more than 0.999;
(9) and predicting the electric power of the unit by using a random forest regression model according to the real-time operation data of the unit.
2. The nuclear power plant electric power prediction method of claim 1, characterized in that: in the step (1), the historical operating data of the unit is historical operating data of a conventional island of the nuclear power station, and the historical operating data comprises main steam parameters, condenser operating parameters and regenerator reheater operating parameters.
3. The nuclear power plant electric power prediction method of claim 1, characterized in that: and (2) the abnormal data in the step (1) are measured point abnormal data and operation process data.
4. The nuclear power plant electric power prediction method of claim 1, characterized in that: in the step (3), the characteristic value comprises one or more of main feed water flow, high-pressure regulating valve opening, main steam main pipe pressure, main steam main pipe temperature, condenser vacuum degree, low-pressure cylinder exhaust steam temperature, circulating water inlet temperature, circulating water outlet temperature, condenser inlet pressure and condenser outlet pressure.
5. The nuclear power plant electric power prediction method of claim 1, characterized in that: and (3) taking the average value of a plurality of measuring points in the same characteristic as a characteristic value.
6. The nuclear power plant electric power prediction method of claim 1, characterized in that: before splitting the data set in the step (5), the order of the data set is disordered to randomly sort the data set, and then splitting is carried out.
7. The nuclear power plant electric power prediction method of claim 1, characterized in that: in step (7), the R-squared value is also referred to as a coefficient of determination, which reflects the proportion of the total variation of the dependent variable that can be explained by the independent variable through a regression relationship.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911364412.2A CN111242345A (en) | 2019-12-26 | 2019-12-26 | Nuclear power unit electric power prediction method based on cluster analysis and random forest regression |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911364412.2A CN111242345A (en) | 2019-12-26 | 2019-12-26 | Nuclear power unit electric power prediction method based on cluster analysis and random forest regression |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111242345A true CN111242345A (en) | 2020-06-05 |
Family
ID=70863926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911364412.2A Pending CN111242345A (en) | 2019-12-26 | 2019-12-26 | Nuclear power unit electric power prediction method based on cluster analysis and random forest regression |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111242345A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112091A (en) * | 2021-05-06 | 2021-07-13 | 云南电力技术有限责任公司 | Nuclear power unit power prediction method based on PCA and LSTM |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180175790A1 (en) * | 2015-06-23 | 2018-06-21 | Qatar Foundation For Education, Science And Community Development | Method of forecasting for solar-based power systems |
| CN108197752A (en) * | 2018-01-25 | 2018-06-22 | 国网福建省电力有限公司 | Wind turbine output power short term prediction method based on random forest |
| CN108830411A (en) * | 2018-06-07 | 2018-11-16 | 苏州工业职业技术学院 | A kind of wind power forecasting method based on data processing |
| CN110363354A (en) * | 2019-07-16 | 2019-10-22 | 上海交通大学 | Wind field wind power prediction method, electronic device and storage medium |
-
2019
- 2019-12-26 CN CN201911364412.2A patent/CN111242345A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180175790A1 (en) * | 2015-06-23 | 2018-06-21 | Qatar Foundation For Education, Science And Community Development | Method of forecasting for solar-based power systems |
| CN108197752A (en) * | 2018-01-25 | 2018-06-22 | 国网福建省电力有限公司 | Wind turbine output power short term prediction method based on random forest |
| CN108830411A (en) * | 2018-06-07 | 2018-11-16 | 苏州工业职业技术学院 | A kind of wind power forecasting method based on data processing |
| CN110363354A (en) * | 2019-07-16 | 2019-10-22 | 上海交通大学 | Wind field wind power prediction method, electronic device and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| LIU D等: "Random forest solar power forecast based on classification optimization", 《ENERGY》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113112091A (en) * | 2021-05-06 | 2021-07-13 | 云南电力技术有限责任公司 | Nuclear power unit power prediction method based on PCA and LSTM |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111159844B (en) | Abnormity detection method for exhaust temperature of gas turbine of power station | |
| CN108446529B (en) | Organic Rankine cycle system fault detection method based on generalized mutual entropy-DPCA algorithm | |
| CN103850726B (en) | Method for quickly determining stationary sliding pressing optimization curve of steam turbine | |
| CN111581597A (en) | Wind turbine generator gearbox bearing temperature state monitoring method based on self-organizing kernel regression model | |
| CN109538311B (en) | Real-time monitoring method for control performance of steam turbine in high-end power generation equipment | |
| CN110987494A (en) | Method for monitoring cavitation state of water turbine based on acoustic emission | |
| CN115294671A (en) | Air compressor outlet pressure prediction method and prediction system | |
| CN108182553B (en) | Coal-fired boiler combustion efficiency online measurement method | |
| CN111242345A (en) | Nuclear power unit electric power prediction method based on cluster analysis and random forest regression | |
| CN110646193B (en) | Test method for obtaining flow characteristic of high-pressure regulating valve of steam turbine | |
| CN110928248A (en) | Method for determining performance degradation degree of gas turbine | |
| CN112348696B (en) | BP neural network-based heating unit peak regulation upper limit evaluation method and system | |
| CN115541227A (en) | Wind power gear box fault diagnosis method based on time-shifted cosine similar entropy | |
| CN111624979B (en) | Industrial closed-loop control loop multi-oscillation detection and tracing method based on slow characteristic analysis | |
| Qiao et al. | Research on SCADA data preprocessing method of Wind Turbine | |
| CN111581787B (en) | Method and system for screening heat rate analysis data of steam turbine in real time | |
| CN110032791B (en) | Turbine low-pressure cylinder efficiency real-time calculation method based on generalized regression neural network | |
| Peng et al. | Accuracy research on the modeling methods of the gas turbine components characteristics | |
| Tang et al. | Computer Prediction Model of Heat Consumption in Thermal System of Coal-Fired Power Station Based on Big Data Analysis and Information Sorting | |
| CN111307493B (en) | Knowledge-based fault diagnosis method for tower type solar molten salt heat storage system | |
| CN109241573B (en) | Steam turbine last stage blade model selection method | |
| CN113112091A (en) | Nuclear power unit power prediction method based on PCA and LSTM | |
| Chi et al. | Turbine blade fault detection based on feature extraction | |
| Yang et al. | Running State Assessment for Induced Draft Fans Using Auto Encoder Model Combined With Fuzzy Synthetic | |
| CN116242613A (en) | Method for measuring and calculating turbine regulating stage through-flow efficiency characteristics and terminal equipment |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200605 |
|
| WD01 | Invention patent application deemed withdrawn after publication |