CN113836480B - Heat exchanger efficiency prediction method based on Gaussian process regression - Google Patents
Heat exchanger efficiency prediction method based on Gaussian process regression Download PDFInfo
- Publication number
- CN113836480B CN113836480B CN202010581878.4A CN202010581878A CN113836480B CN 113836480 B CN113836480 B CN 113836480B CN 202010581878 A CN202010581878 A CN 202010581878A CN 113836480 B CN113836480 B CN 113836480B
- Authority
- CN
- China
- Prior art keywords
- heat exchanger
- gaussian process
- heat exchange
- exchange efficiency
- process regression
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000012549 training Methods 0.000 claims abstract description 14
- 230000000694 effects Effects 0.000 claims description 4
- 230000000737 periodic effect Effects 0.000 claims description 4
- 238000007476 Maximum Likelihood Methods 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000002939 conjugate gradient method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 3
- 238000012423 maintenance Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F28—HEAT EXCHANGE IN GENERAL
- F28F—DETAILS OF HEAT-EXCHANGE AND HEAT-TRANSFER APPARATUS, OF GENERAL APPLICATION
- F28F2200/00—Prediction; Simulation; Testing
Abstract
The invention relates to the field of nuclear power overhaul, in particular to a heat exchanger efficiency prediction method based on Gaussian process regression. At present, the heat exchange efficiency of a heat exchanger is not predicted on line in real time in a nuclear power plant. The invention comprises the following steps: step one: acquiring history and current operation data of the heat exchanger; step two: forming a training data set; step three: a regression model of the Gaussian process; step four: predicting heat exchange efficiency of the heat exchanger; step five: and carrying out sliding prediction on the heat exchange efficiency. The invention uses the Gaussian process regression method to realize the real-time online prediction of the heat exchange efficiency of the heat exchanger. Meanwhile, the accuracy of a prediction result is guaranteed by adopting a sliding prediction mode, and the method has certain guiding significance for heat exchanger performance monitoring.
Description
Technical Field
The invention relates to the field of nuclear power overhaul, in particular to a heat exchanger efficiency prediction method based on Gaussian process regression.
Background
The nuclear power plant has numerous devices, heavy maintenance tasks and high cost. The current practice of chinese nuclear power plants is to employ time-based periodic maintenance strategies rather than state-based maintenance strategies based on equipment operating conditions. Taking a plate heat exchanger of a nuclear power plant as an example, the maintenance strategy is to regularly wash, disassemble, replace parts and the like. The reason is mainly that the accurate operation state of the heat exchanger cannot be obtained. The heat exchange efficiency is one of the key performance parameters of the heat exchanger, and the parameters can reflect the performance of the heat exchanger to a great extent. The heat exchange efficiency in a period of time in the future is predicted according to the historical running state of the heat exchanger, and the method has a certain practical significance for realizing the maintenance of the heat exchanger based on the running state. At present, the heat exchange efficiency of a heat exchanger is not predicted on line in real time in a nuclear power plant.
The invention mainly aims at the problems, and develops a heat exchange efficiency prediction method of the heat exchanger based on Gaussian process regression, which can predict the heat exchange efficiency of the heat exchanger for a period of time in the future according to historical operation data of the heat exchanger to obtain a trend state of the future operation of the heat exchanger.
Disclosure of Invention
1. The purpose is as follows:
in order to fill the defect of online real-time prediction of heat exchange efficiency of a heat exchanger in nuclear power plant heat exchanger performance monitoring, the invention provides a heat exchange efficiency prediction method of the heat exchanger based on Gaussian process regression, which is simple in design and strong in applicability.
2. The technical scheme is as follows:
in order to achieve the above purpose, the invention adopts the following technical scheme:
a heat exchanger efficiency prediction method based on Gaussian process regression comprises the following steps:
step one: the method for acquiring the history and current operation data of the heat exchanger specifically comprises the following steps: and obtaining inlet and outlet temperatures and flow rates of the hot side and the cold side.
Step two: forming a training data set, specifically comprising: (1) According to a heat exchanger efficiency calculation formula, calculating to obtain historical heat exchange efficiency of the heat exchanger, wherein the historical heat exchange efficiency is calculated by taking hours and days as units respectively; (2) Constructing a heat exchange efficiency prediction training data set X= [ X ] 1 ,x 2 ,…x n ] T ,Y=[y 1 ,y 2 ,…y n ] T X is time and y is efficiency.
Step three: establishing a Gaussian process regression model, which specifically comprises the following steps:
(1) Determining a kernel function and a hyper-parameter of the kernel function;
(2) And comparing and selecting different kernel function implementation effects, and selecting a square index covariance function:
wherein m=diag (λ 1 ,λ 2 ,…,λ d ) λ represents a feature length scale, d is the input vector dimension,for outputting the scale parameter>Variance of noise, delta ij Is a kronecker subfunction. When i=j, δ ij When i+.j, =1, δ ij =0;
(3) Let θ be the set of hyper-parameters taking logarithms, i.e., θ= { lnλ 1 ,lnλ 2 ,…lnσ f ,lnσ n -a }; and calculating the hyper-parameters of the Gaussian process regression by adopting a maximum likelihood estimation method, wherein the hyper-parameters are as follows:
firstly initializing a super parameter theta, then solving an upper expression by using a conjugate gradient method, and when the upper expression is maximum, obtaining the super parameter as an optimal solution.
The kernel functions described above may be other kernel functions, including: a rational quadratic covariance function, a linear covariance function, a Matem covariance function, a periodic covariance function, a neural network covariance function, a Bn spline curve kernel function, a Fisher kernel function and a String kernel function.
The comparison selects different kernel functions, namely, comparing residual errors of the predicted value and the actual value of the heat exchange efficiency.
Step four: predicting heat exchange efficiency of a heat exchanger specifically comprises: and according to the history and the current operation data of the heat exchanger, a Gaussian process regression model is used.
Step five: the sliding prediction of the heat exchange efficiency specifically comprises the following steps: and carrying out sliding prediction on the heat exchange efficiency by using an iterative mode, and selecting the number and time of historical data for prediction and the length of a predicted time period by changing parameters.
3. The effect is as follows:
the invention uses the Gaussian process regression method to realize the real-time online prediction of the heat exchange efficiency of the heat exchanger. Meanwhile, the accuracy of a prediction result is guaranteed by adopting a sliding prediction mode, and the method has certain guiding significance for heat exchanger performance monitoring.
Drawings
FIG. 1 is a flow chart for predicting heat exchange efficiency of a heat exchanger based on Gaussian process regression
FIG. 2 is a heat exchanger efficiency prediction FIG. 1
FIG. 3 is a heat exchanger efficiency prediction FIG. 2
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, the idea of the present invention is: acquiring historical and current operation data of the heat exchanger, wherein the historical and current operation data must comprise inlet and outlet temperatures and flow rates of hot side and cold side measurement; and calculating the heat exchange efficiency of the heat exchanger by using a basic formula of heat transfer science to form a training data set. According to actual requirements, the heat exchange efficiency can be calculated once every second, every hour or every day and the like; constructing a regression model of the heat exchanger efficiency Gaussian process by using the training data set; according to the history and current operation data of the heat exchanger, predicting the heat exchange efficiency of the heat exchanger by using a Gaussian process regression model; and carrying out sliding prediction on the heat exchange efficiency by using an iterative mode. The number and time of the historical data used for prediction can be selected by changing parameters, and the length of the predicted time period can also be selected;
in this embodiment, a plate heat exchanger of a certain nuclear power plant is specifically selected as an object, and 42 days of operation data of the plate heat exchanger is collected, which specifically includes: hot side inlet temperature, outlet temperature, flow and cold side inlet temperature, outlet temperature, flow.
And calculating according to a heat exchanger efficiency calculation formula to obtain the historical heat exchange efficiency of the heat exchanger. As shown in fig. 2 and 3, the heat exchange efficiency is calculated in units of hours and days, respectively.
Constructing a heat exchange efficiency prediction training data set X= [ X ] 1 ,x 2 ,…x n ] T ,Y=[y 1 ,y 2 ,…y n ] T . x is timeY is the efficiency. And constructing a Gaussian process regression model of the heat exchanger efficiency by using the training data set, wherein the Gaussian process regression model is critical in determining the kernel function and the hyper-parameters of the kernel function.
By comparing the effects of selecting different kernel functions in this embodiment (comparing the residual error between the predicted value and the actual value of the heat exchange efficiency), the square index covariance function is selected as the kernel function of this embodiment, as follows:
wherein m=diag (λ 1 ,λ 2 ,…,λ d ) λ represents a feature length scale, d is the input vector dimension,for outputting the scale parameter>Variance of noise, delta ij Is a kronecker subfunction. When i=j, δ ij When i+.j, =1, δ ij =0。
In the present invention, the kernel functions of other embodiments may also be other kernel functions: a rational quadratic covariance function, a linear covariance function, a Matem covariance function, a periodic covariance function, a neural network covariance function, a Bn spline curve kernel function, a Fisher kernel function, a String kernel function, and the like.
Let θ be the set of hyper-parameters taking logarithms, i.e., θ= { lnλ 1 ,lnλ 2 ,…lnσ f ,lnσ n }. The super-parameters of the gaussian process regression are typically calculated using Maximum Likelihood Estimation (MLE) methods. The method comprises the following steps:
firstly initializing a super parameter theta, then solving an upper expression by using a conjugate gradient method, and when the upper expression is maximum, obtaining the super parameter as an optimal solution.
Fig. 2 and 3 are heat exchange efficiency predictions of the heat exchanger obtained using the gaussian process regression model described above. Wherein, in the unit of hours, fig. 2 selects 100 data as training sets, and predicts heat exchange efficiency for 100 hours in the future; fig. 3 is a graph of heat exchange efficiency prediction for 7 days in the future, taking the 7 data as training set. The new heat exchange efficiency can replace the heat exchange efficiency in the training set in real time, and the heat exchange efficiency is calculated and updated on line in real time by using a sliding method for prediction. The actual heat exchange efficiency is more consistent with the predicted value, and the effectiveness and the accuracy of the method and the model are proved.
Claims (7)
1. A heat exchanger efficiency prediction method based on Gaussian process regression is characterized by comprising the following steps of: step one: acquiring history and current operation data of the heat exchanger; step two: forming a training data set; step three: establishing a Gaussian process regression model; step four: predicting heat exchange efficiency of the heat exchanger; step five: sliding prediction is carried out on heat exchange efficiency; wherein, step three: the regression model for the Gaussian process specifically comprises the following steps: (1) determining a kernel function and a hyper-parameter of the kernel function; (2) And comparing and selecting different kernel function implementation effects, and selecting a square index covariance function:
wherein m=diag (λ 1 ,λ 2 ,…,λ d ) λ represents a feature length scale, d is the input vector dimension,for outputting the scale parameter>Variance of noise, delta ij Is a kronecker subfunction; when i=j, δ ij When i+.j, =1, δ ij =0;
(3) Let θ be the set of hyper-parameters taking logarithms, i.e., θ= { lnλ 1 ,lnλ 2 ,…lnσ f ,lnσ n -a }; and calculating the hyper-parameters of the Gaussian process regression by adopting a maximum likelihood estimation method, wherein the hyper-parameters are as follows:
wherein n is the number of data points, and K is the covariance matrix of the training data set X;
firstly initializing a super parameter theta, then solving an upper expression by using a conjugate gradient method, and when the upper expression is maximum, obtaining the super parameter as an optimal solution.
2. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: the first step is as follows: the method for acquiring the history and current operation data of the heat exchanger specifically comprises the following steps: and obtaining inlet and outlet temperatures and flow rates of the hot side and the cold side.
3. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: the second step is as follows: forming a training data set, specifically comprising: (1) According to a heat exchanger efficiency calculation formula, calculating to obtain historical heat exchange efficiency of the heat exchanger, wherein the historical heat exchange efficiency is calculated by taking hours and days as units respectively; (2) Constructing a heat exchange efficiency prediction training data set X= [ X ] 1 ,x 2 ,…x n ] T ,Y=[y 1 ,y 2 ,…y n ] T X is time and y is efficiency.
4. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: and step four: predicting heat exchange efficiency of a heat exchanger specifically comprises: and according to the history and the current operation data of the heat exchanger, a Gaussian process regression model is used.
5. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: step five, the said step: the sliding prediction of the heat exchange efficiency specifically comprises the following steps: and carrying out sliding prediction on the heat exchange efficiency by using an iterative mode, and selecting the number and time of historical data for prediction and the length of a predicted time period by changing parameters.
6. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: the first step is as follows: acquiring historical and current operation data of the heat exchanger, wherein the kernel function can be other kernel functions, including: a rational quadratic covariance function, a linear covariance function, a Matem covariance function, a periodic covariance function, a neural network covariance function, a Bn spline curve kernel function, a Fisher kernel function and a String kernel function.
7. The heat exchanger efficiency prediction method based on gaussian process regression according to claim 1, wherein: the comparison selects different kernel functions, namely, comparing residual errors of a predicted value and an actual value of heat exchange efficiency.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010581878.4A CN113836480B (en) | 2020-06-23 | 2020-06-23 | Heat exchanger efficiency prediction method based on Gaussian process regression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010581878.4A CN113836480B (en) | 2020-06-23 | 2020-06-23 | Heat exchanger efficiency prediction method based on Gaussian process regression |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113836480A CN113836480A (en) | 2021-12-24 |
CN113836480B true CN113836480B (en) | 2024-01-12 |
Family
ID=78964154
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010581878.4A Active CN113836480B (en) | 2020-06-23 | 2020-06-23 | Heat exchanger efficiency prediction method based on Gaussian process regression |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113836480B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105938655A (en) * | 2016-06-16 | 2016-09-14 | 上海交通大学 | Real-time traffic state evaluation method based on Gaussian mixture model |
CN109376858A (en) * | 2018-09-12 | 2019-02-22 | 天津大学 | Method for predicting performance of condensing heat exchanger based on partial load rate |
CN109871602A (en) * | 2019-01-30 | 2019-06-11 | 西安工程大学 | A kind of critical heat flux density prediction technique returned based on Gaussian process |
CN109933942A (en) * | 2019-03-26 | 2019-06-25 | 中冶华天南京电气工程技术有限公司 | A kind of heat exchange station Temperature Control Model modeling method based on support vector machines |
CN109978201A (en) * | 2017-12-27 | 2019-07-05 | 深圳市景程信息科技有限公司 | Probability load prediction system and method based on Gaussian process quantile estimate model |
WO2020087845A1 (en) * | 2018-10-30 | 2020-05-07 | 东南大学 | Initial alignment method for sins based on gpr and improved srckf |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9841184B2 (en) * | 2010-02-26 | 2017-12-12 | Dominion Engineering, Inc. | Method and apparatus for evaluating repair and remediation alternatives for heat exchangers |
-
2020
- 2020-06-23 CN CN202010581878.4A patent/CN113836480B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105938655A (en) * | 2016-06-16 | 2016-09-14 | 上海交通大学 | Real-time traffic state evaluation method based on Gaussian mixture model |
CN109978201A (en) * | 2017-12-27 | 2019-07-05 | 深圳市景程信息科技有限公司 | Probability load prediction system and method based on Gaussian process quantile estimate model |
CN109376858A (en) * | 2018-09-12 | 2019-02-22 | 天津大学 | Method for predicting performance of condensing heat exchanger based on partial load rate |
WO2020087845A1 (en) * | 2018-10-30 | 2020-05-07 | 东南大学 | Initial alignment method for sins based on gpr and improved srckf |
CN109871602A (en) * | 2019-01-30 | 2019-06-11 | 西安工程大学 | A kind of critical heat flux density prediction technique returned based on Gaussian process |
CN109933942A (en) * | 2019-03-26 | 2019-06-25 | 中冶华天南京电气工程技术有限公司 | A kind of heat exchange station Temperature Control Model modeling method based on support vector machines |
Non-Patent Citations (1)
Title |
---|
垂直U型管地下换热器模型比较;张燕;苏华;;技术与市场(05);第110-112页 * |
Also Published As
Publication number | Publication date |
---|---|
CN113836480A (en) | 2021-12-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112926273B (en) | Method for predicting residual life of multivariate degradation equipment | |
Yang et al. | Local identification of LPV dual-rate system with random measurement delays | |
CN108121215B (en) | Process control loops method of evaluating performance and device based on full loop reconstruct emulation | |
Guo et al. | Modeling and characteristic analysis of fouling in a wet cooling tower based on wavelet neural networks | |
CA2689246C (en) | Monitoring methods and apparatus | |
CN104463347A (en) | Method for electronic product degenerate state trend prediction with singular signals | |
CN113836480B (en) | Heat exchanger efficiency prediction method based on Gaussian process regression | |
US10429828B2 (en) | Plant simulation device and plant simulation method with first parameter adjustable at start and second parameter adjustable during operation of the plant | |
Yang et al. | An improved mixture density network via wasserstein distance based adversarial learning for probabilistic wind speed predictions | |
CN113743022B (en) | Storage and visualization method for high-precision climate change data | |
Sheriff et al. | An operating economics-driven perspective on monitoring and maintenance in multiple operating regimes: Application to monitor fouling in heat exchangers | |
CN114429003A (en) | System of boiler four-tube service life prediction method | |
Cheng et al. | Deep reinforcement learning for cost-optimal condition-based maintenance policy of offshore wind turbine components | |
Jiang et al. | A hybrid degradation tendency measurement method for mechanical equipment based on moving window and Grey–Markov model | |
CN112016241A (en) | Soft measurement modeling method based on mLASSO-MLP model | |
CN111998423A (en) | Electric heat storage circulating air quantity-water temperature control system and prediction control method thereof | |
Na et al. | A soft-sensing model for feedwater flow rate using fuzzy support vector regression | |
Zhao et al. | Probabilistic fault prediction of incipient fault | |
Humberstone et al. | An adaptive model for expanded process monitoring | |
Rudenko et al. | Using a Neural Network Approach to Predict Deposits on the Surfaces of Heat Exchange Equipment | |
CN116702630B (en) | Error iterative evaluation method for capacitive voltage transformer CVT | |
AbouOmar et al. | Data-Driven Soft Sensors Based on Support Vector Regression and Gray Wolf Optimizer | |
Ping et al. | Sparse Bayesian Learning Based Hybrid Method for Load Forecasting in Integrated Energy Systems | |
CN117709198A (en) | Comprehensive energy system modeling method based on meta-learning and physical information neural network | |
Perov et al. | Modeling of De-Icing Heating System by Machine Learning Methods |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |