CN111861011B - Supercritical pressure fluid convection heat exchange performance prediction method and system - Google Patents
Supercritical pressure fluid convection heat exchange performance prediction method and system Download PDFInfo
- Publication number
- CN111861011B CN111861011B CN202010716914.3A CN202010716914A CN111861011B CN 111861011 B CN111861011 B CN 111861011B CN 202010716914 A CN202010716914 A CN 202010716914A CN 111861011 B CN111861011 B CN 111861011B
- Authority
- CN
- China
- Prior art keywords
- precision
- heat exchange
- exchange data
- data
- pressure fluid
- 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
- 239000012530 fluid Substances 0.000 title claims abstract description 172
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000007781 pre-processing Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims description 55
- 238000012360 testing method Methods 0.000 claims description 22
- 230000008569 process Effects 0.000 claims description 20
- 238000004088 simulation Methods 0.000 claims description 17
- 238000012546 transfer Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000000704 physical effect Effects 0.000 claims description 11
- 238000004140 cleaning Methods 0.000 claims description 10
- 238000002474 experimental method Methods 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000011160 research Methods 0.000 claims description 3
- 238000006467 substitution reaction Methods 0.000 description 6
- 238000013461 design Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000008570 general process Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 231100000956 nontoxicity Toxicity 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000000750 progressive effect Effects 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- 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
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Mathematical Physics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention relates to a method and a system for predicting the convection heat exchange performance of supercritical pressure fluid. The prediction method comprises the following steps: acquiring heat exchange data of the supercritical pressure fluid; preprocessing the supercritical pressure fluid heat exchange data, and determining the preprocessed supercritical pressure fluid heat exchange data; determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation; determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation; determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; and predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model. By adopting the prediction method and the prediction system, the prediction precision of the convection heat exchange performance can be improved, and the prediction cost of equipment can be reduced.
Description
Technical Field
The invention relates to the field of convective heat exchange performance prediction, in particular to a method and a system for predicting convective heat exchange performance of supercritical pressure fluid.
Background
The supercritical pressure fluid has the characteristics of large specific heat capacity, low viscosity, good fluidity, good cooling performance, no corrosiveness, non-inflammability and non-toxicity, is widely applied to the industrial fields of solar power generation, nuclear reactor cooling, aerospace thermal protection and the like at present, has wide application prospect, and the flow and heat exchange performance of the supercritical pressure fluid directly influences the efficiency and safety of a power system. However, the drastic change of physical properties of the supercritical pressure fluid near the quasi-critical temperature causes the convective heat exchange mechanism of the supercritical pressure fluid to be different from that of the conventional fluid, which brings great challenges to the design and heat exchange performance prediction of the system.
The conventional method for predicting the convection heat exchange performance of the supercritical pressure fluid mainly comprises two methods: 1) correlation prediction method: the method comprises the steps that a supercritical pressure fluid convection heat exchange rule correlation is used for quickly calculating convection heat exchange performance parameters (such as convection heat exchange coefficient and convection heat exchange Nussel number) of the supercritical pressure fluid, low-precision data are obtained through convection heat exchange rule correlation calculation, the supercritical pressure fluid convection heat exchange performance is difficult to accurately predict due to the lack of high-precision data and the possession of a large amount of low-precision data, and due to the coupling influence of multiple factors such as buoyancy force, flow acceleration and the like, the accuracy of the convection heat exchange rule correlation is poor and generally deviates from a true value by more than 30%; 2) numerical simulation method: with the development of a numerical simulation method, the flow of fluid can be simulated through computational fluid mechanics or direct numerical simulation, and the convection heat exchange performance is predicted, but the deviation between a calculated predicted value and a true value is more than 20% or even higher due to inaccurate turbulence models in the former, and although the heat exchange performance can be predicted more accurately in the latter, calculation and simulation are required on a super computer, the equipment cost is high, the cost of general hundreds of thousands of calculation resources is spent, the calculation period is also long, and the cost is generally several months or even longer. Therefore, although the convection heat exchange rule correlation can quickly predict the convection heat exchange performance of the supercritical pressure fluid, the accuracy is poor, and the correlation can be regarded as low-precision data; the numerical simulation method generally obtains high-precision data, and although the accuracy is higher, the cost of computing equipment and time is very high.
Disclosure of Invention
The invention aims to provide a method and a system for predicting the convection heat exchange performance of supercritical pressure fluid, which aim to solve the problems of low accuracy and high prediction cost of the conventional method for predicting the convection heat exchange performance of the supercritical pressure fluid.
In order to achieve the purpose, the invention provides the following scheme:
a method for predicting the convective heat exchange performance of supercritical pressure fluid comprises the following steps:
acquiring heat exchange data of the supercritical pressure fluid; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline;
preprocessing the supercritical pressure fluid heat exchange data, and determining the preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprise preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data;
Determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation;
determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation;
determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; the multi-precision model is a convective heat exchange performance prediction model which takes dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convective heat exchange performance as output;
and predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
Optionally, the preprocessing the supercritical pressure fluid heat exchange data to determine the preprocessed supercritical pressure fluid heat exchange data specifically includes:
cleaning the supercritical pressure fluid heat exchange data, deleting abnormal points and outliers in the supercritical pressure fluid heat exchange data, and determining the cleaned supercritical pressure fluid heat exchange data;
converting the heat exchange data of the cleaned supercritical pressure fluid into dimensionless form parameters; the dimensionless form parameter is the heat exchange data of the pretreated supercritical pressure fluid.
Optionally, the determining, by using the collaborative kriging method, a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data, and the high-precision model specifically includes:
dividing the high-precision heat exchange data into a training set and a test set;
by usingThe training set is based on a formula
Determining a multi-precision model; wherein,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process.
Optionally, the training set is used, according to a formula
Determining a multi-precision model, and then:
testing the multi-precision model according to the test set, and judging whether the prediction precision of the multi-precision model reaches a precision threshold value to obtain a first judgment result;
if the first judgment result indicates that the prediction precision of the multi-precision model reaches a precision threshold, determining that the current multi-precision model is an optimal multi-precision model;
and if the first judgment result shows that the prediction precision of the multi-precision model does not reach a precision threshold value, adding the test set into the training set to form a new training set, and optimizing the multi-precision model by using the new training set.
A system for predicting convective heat exchange performance of a supercritical pressure fluid, comprising:
the supercritical pressure fluid heat exchange data acquisition module is used for acquiring supercritical pressure fluid heat exchange data; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline;
the preprocessing module is used for preprocessing the supercritical pressure fluid heat exchange data and determining the preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprise preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data;
the low-precision model determining module is used for determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation;
the high-precision model determining module is used for determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation;
The multi-precision model determining module is used for determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; the multi-precision model is a convective heat exchange performance prediction model which takes dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convective heat exchange performance as output;
and the heat exchange performance prediction module is used for predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
Optionally, the preprocessing module specifically includes:
the cleaning unit is used for cleaning the supercritical pressure fluid heat exchange data, deleting abnormal points and outliers in the supercritical pressure fluid heat exchange data, and determining the cleaned supercritical pressure fluid heat exchange data;
the conversion module is used for converting the cleaned supercritical pressure fluid heat exchange data into dimensionless form parameters; the dimensionless form parameter is the heat exchange data of the pretreated supercritical pressure fluid.
Optionally, the multi-precision model determining module specifically includes:
the dividing unit is used for dividing the high-precision heat exchange data into a training set and a test set;
A multi-precision model determination unit for using the training set according to a formula
Determining a multi-precision model; wherein,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process.
Optionally, the method further includes:
the first judgment unit is used for testing the multi-precision model according to the test set, judging whether the prediction precision of the multi-precision model reaches a precision threshold value or not, and obtaining a first judgment result;
an optimal multi-precision model determining unit, configured to determine that the current multi-precision model is an optimal multi-precision model if the first determination result indicates that the prediction precision of the multi-precision model reaches a precision threshold;
and the optimization unit is used for adding the test set into the training set to form a new training set and optimizing the multi-precision model by using the new training set if the first judgment result indicates that the prediction precision of the multi-precision model does not reach a precision threshold value.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: the invention provides a method and a system for predicting the convection heat exchange performance of supercritical pressure fluid, wherein a substitution model capable of comprehensively utilizing high-precision and low-precision data sets is built and trained by means of a Gaussian process regression algorithm and a collaborative kriging method in machine learning, the prediction accuracy of the model is improved by introducing a large amount of low-precision data on the basis of using a small amount of high-precision data, and the problem that the convection heat exchange performance of the supercritical pressure fluid is difficult to accurately predict when the high-precision data is lacked and a large amount of low-precision data is possessed is solved.
The method for predicting the convection heat exchange performance of the supercritical pressure fluid, which is quick, cheap and high in precision, is established, and can further promote the development of relevant application of the supercritical pressure fluid; the principle of the invention has certain generalizability, and the invention can be suitable for other heat exchange prediction problems or thermal design after adjustment, and improves the design of a related heat exchange system by utilizing the convective heat exchange performance of the supercritical pressure fluid under different predicted conditions.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a method for predicting the convective heat exchange performance of a supercritical pressure fluid provided by the present invention;
FIG. 2 is a schematic diagram of a method for predicting the convective heat exchange performance of a supercritical pressure fluid according to the present invention;
fig. 3 is a structural diagram of a system for predicting the convective heat exchange performance of a supercritical pressure fluid provided by 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 only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for predicting the convection heat exchange performance of supercritical pressure fluid, which do not need a large amount of high-precision data, and establish a multi-precision model by using a small amount of high-precision data and a large amount of low-precision data, thereby improving the prediction precision of the convection heat exchange performance and reducing the prediction cost of equipment.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
The invention provides a method and a system for predicting the convection heat exchange performance of supercritical pressure fluid, wherein the general process comprises the following three steps: 1) arranging and cleaning the existing supercritical pressure fluid heat exchange data; 2) constructing a machine learning algorithm by combining a Gaussian regression process and a collaborative kriging method, and training a multi-precision surrogate model by using the existing data; 3) and substituting the parameters into a trained multi-precision substitution model according to dimensionless parameters of a given system to calculate the convective heat exchange performance (such as convective heat exchange coefficient, convective heat exchange Knudel number and the like).
Fig. 1 is a flowchart of a method for predicting a supercritical pressure fluid convection heat exchange performance provided in the present invention, and as shown in fig. 1, a method for predicting a supercritical pressure fluid convection heat exchange performance includes:
step 101: acquiring heat exchange data of the supercritical pressure fluid; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline.
Step 102: preprocessing the supercritical pressure fluid heat exchange data, and determining the preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprise preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data.
The data preprocessing process comprises the following steps:
and (4) collating and summarizing experimental data and numerical simulation data, wherein the data generally comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient, the characteristic length of a pipeline and the like. And cleaning the data set, and deleting abnormal points, outliers and the like contained in the data set. Preprocessing is performed by the following calculation method: inquiring a fluid physical property data table according to the wall temperature of the fluid along the flowing direction and numerical values of the main stream temperature and the pressure of the fluid, obtaining the ratio of the physical property of the fluid close to the wall and the physical property of the fluid under the main stream average temperature, and generally calculating the constant pressure specific heat c pAnd the density ρ; and obtaining a series of dimensionless parameters including Reynolds number Re, Printt number, Nu and Bo number of the fluid along the flowing direction according to the definition formula of the Reynolds number Re, the Printt number, the Nu and the Bo number in the heat transfer science. Each piece of data corresponds to a certain position in the pipe, so that a data set in a dimensionless form is obtained, and the influence of the dimension of the physical quantity on model training is avoided. The above preprocessing method is applicable to experimental data, numerical simulation data, and relational data.
The Knoop number is a common dimensionless number in heat transfer science, and in a physical sense, the Knoop number represents the intensity of convective heat transfer, and the larger the value of the Knoop number is, the stronger the intensity of convective heat transfer of fluid is, namely, the convective heat transfer performance is stronger. In general, it can be assumed that the knoop number is a function of a series of characteristic parameters, and assuming that a vector formed by the characteristic parameters is x, there are:
Nu＝f(x)
the form of the feature vector x may vary depending on the parameters contained in the data set. In the invention, the Reynolds number Re, the Prandtl number Pr, the Bo number, the ratio of physical properties and the corresponding dimensionless position of the parameter in the tube, which are obtained by calculation in the preprocessing step, are used as characteristic parameters in prediction.
The data set consists of a high-precision data set and a low-precision data set, which contain all the parameter values in the feature vector, as well as the values of the knoop Nu to be predicted finally. The high and low precision mainly refers to the precision of a target value Nu needing to be predicted finally, the high-precision data set refers to a data set obtained by experiment or numerical simulation data, and the low-precision data set refers to data obtained by non-dimensional rule correlation prediction obtained by existing research. A relatively common fluid is subjected to heat convection in a pipeline and has a dimensionless correlation Nu of 0.23Re0.8Pr0.3, and a Knudel number Nu predicted value corresponding to the corresponding characteristic parameter can be obtained through the correlation so as to obtain a corresponding data set with the same composition.
Step 103: and determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation.
Step 104: and determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation.
Step 105: determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; the multi-precision model is a convective heat exchange performance prediction model which takes dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convective heat exchange performance as output; the multi-precision model is expressed as a correspondence between the high-precision model and the low-precision model.
And (3) a multi-precision model training process:
as shown in fig. 2, in which,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process. Randomly sampling all data of a high-precision data set and a low-precision data set, splitting a training set and a testing set, and ensuring that the data of the high-precision data set and the data of the low-precision data set are not crossed with each other, wherein the training set is used for training a training machineAnd a machine learning model, wherein the test set is used for evaluating the predictive performance of the machine learning model.
In the training process, the input parameters of the model are all parameters (namely flow velocity, wall temperature, main flow temperature of fluid, fluid pressure and characteristic length) except the convective heat transfer coefficient in a training set, and the output parameters are the convective heat transfer coefficient. And comparing the output convective heat transfer coefficient with the convective heat transfer coefficient concentrated in training, and obtaining a substitution model for predicting the convective heat transfer coefficient of the supercritical fluid under the condition of continuous comparison and minimum deviation. In the training of the surrogate model, there are two methods: firstly, a Gaussian regression process; ② a multi-precision substitution model constructed by cooperating with the kriging method. The training process for these two surrogate models is given in fig. 1. Specifically, in the training process of the gaussian regression process, the training set in the high-precision data set is directly adopted for training to obtain the surrogate model. In the training process of the collaborative kriging method, a Gaussian process, namely a low-precision model, is obtained by training the training set data of the low-precision data 
Then, training by using the training set data in the high-precision data set to obtain a new Gaussian process delta; and finally, simultaneously using the low-precision data, the low-precision model, the high-precision data and the high-precision model for training the multi-precision model by using a collaborative kriging method, and finally storing relevant parameters in a training result.
Step 106: and predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
The same as the data preprocessing step, the thermophysical parameters of the supercritical pressure fluid system along the way are arranged into a dimensionless form, so that the thermophysical parameters are consistent with the input parameter format and the physical meaning in the training process.
And substituting the dimensionless parameters obtained after arrangement into the multi-precision substitution model obtained by training for calculation, thereby obtaining the convection heat transfer coefficient of the system.
Fig. 3 is a structural diagram of a supercritical pressure fluid convection heat exchange performance prediction system provided in the present invention, and as shown in fig. 3, a supercritical pressure fluid convection heat exchange performance prediction system includes:
a supercritical pressure fluid heat exchange data acquisition module 301, configured to acquire supercritical pressure fluid heat exchange data; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow speed of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline.
A preprocessing module 302, configured to preprocess the supercritical pressure fluid heat exchange data, and determine preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprises preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data.
The preprocessing module 302 specifically includes: the cleaning unit is used for cleaning the supercritical pressure fluid heat exchange data, deleting abnormal points and outliers in the supercritical pressure fluid heat exchange data, and determining the cleaned supercritical pressure fluid heat exchange data; the conversion module is used for converting the cleaned supercritical pressure fluid heat exchange data into dimensionless form parameters; the dimensionless form parameter is the heat exchange data of the pretreated supercritical pressure fluid.
And the low-precision model determining module 303 is configured to determine a low-precision model according to the preprocessed low-precision heat exchange data based on a gaussian regression equation.
And the high-precision model determining module 304 is used for determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation.
A multi-precision model determining module 305, configured to determine a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data, and the high-precision model by using a collaborative kriging method; the multi-precision model is a convection heat exchange performance prediction model which takes dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convection heat exchange performance as output.
The multi-precision model determining module 305 specifically includes: the dividing unit is used for dividing the high-precision heat exchange data into a training set and a test set; a multi-precision model determination unit for using the training set according to a formula
Determining a multi-precision model; wherein,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process.
And the heat exchange performance prediction module 306 is used for predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
The invention also includes: the first judgment unit is used for testing the multi-precision model according to the test set, judging whether the prediction precision of the multi-precision model reaches a precision threshold value or not, and obtaining a first judgment result; an optimal multi-precision model determining unit, configured to determine that the current multi-precision model is an optimal multi-precision model if the first determination result indicates that the prediction precision of the multi-precision model reaches a precision threshold; and the optimization unit is used for adding the test set into the training set to form a new training set and optimizing the multi-precision model by using the new training set if the first judgment result indicates that the prediction precision of the multi-precision model does not reach a precision threshold value.
By adopting the technical scheme, the invention can achieve the following technical effects: 1) the model training time and the prediction calculation time are reduced, and the efficiency of the system design and the heat exchange performance prediction process is improved; 2) the existing experimental and numerical simulation data are comprehensively utilized, high-precision and low-precision data are mixed, and a multi-precision substitution model capable of giving a high-precision prediction result is trained; 3) the trained surrogate model has good generalization performance, and once the training is completed, the model can be applied to any scene again.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A method for predicting the convective heat exchange performance of supercritical pressure fluid is characterized by comprising the following steps:
acquiring heat exchange data of the supercritical pressure fluid; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline;
preprocessing the supercritical pressure fluid heat exchange data, and determining the preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprise preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data;
preprocessing is performed by the following calculation method: inquiring a fluid physical property data table according to the wall temperature of the fluid along the flowing direction and numerical values of the main stream temperature and the pressure of the fluid, obtaining the ratio of the physical property of the fluid close to the wall and the physical property of the fluid under the main stream average temperature, and generally calculating the constant pressure specific heat c pAnd the density ρ; obtaining a series of dimensionless parameters of Reynolds number Re, Printt number, Nu and Bo number of the fluid along the flowing direction according to the definition formula of Reynolds number Re, Printt number, Nu and Bo number in the heat transfer science; each piece of data corresponds to a certain position in the pipe, so that a data set in a dimensionless form is obtained, and the influence of the dimension of the physical quantity on model training is avoided;
the data set consists of a high-precision data set and a low-precision data set, wherein all parameter values in the feature vector and the value of the Knudel number Nu to be predicted finally are contained; the high and low precision mainly refers to the precision of a target value Nu needing to be predicted finally, the high-precision data set refers to a data set obtained by experiment or numerical simulation data, and the low-precision data set refers to data obtained by using a dimensionless criterion correlation formula obtained by existing research; the more commonly used fluid has a dimensionless correlation of Nu ═ 0.23Re0.8Pr0.3The Nu predicted value corresponding to the corresponding characteristic parameter can be obtained through the correlation, so that a corresponding data set with the same composition is obtained, but the result of the correlation calculation is low in precision and usually has a certain deviation from a true value, so that the data set is called a low-precision data set;
Determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation;
determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation;
determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; the multi-precision model is a convective heat exchange performance prediction model which takes dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convective heat exchange performance as output;
and predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
2. The method for predicting the convection heat exchange performance of the supercritical pressure fluid according to claim 1, wherein the step of preprocessing the heat exchange data of the supercritical pressure fluid and determining the preprocessed heat exchange data of the supercritical pressure fluid specifically comprises:
cleaning the supercritical pressure fluid heat exchange data, deleting abnormal points and outliers in the supercritical pressure fluid heat exchange data, and determining the cleaned supercritical pressure fluid heat exchange data;
converting the heat exchange data of the cleaned supercritical pressure fluid into dimensionless form parameters; the dimensionless form parameter is the heat exchange data of the pretreated supercritical pressure fluid.
3. The method for predicting the convective heat exchange performance of the supercritical pressure fluid according to claim 1, wherein the determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by using the collaborative kriging method specifically comprises:
dividing the high-precision heat exchange data into a training set and a test set;
using the training set according to a formula
Determining a multi-precision model; wherein,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process.
4. The method of predicting convective heat exchange performance of supercritical pressure fluid according to claim 3, wherein the training set is utilized according to a formula
Determining a multi-precision model, and then:
testing the multi-precision model according to the test set, and judging whether the prediction precision of the multi-precision model reaches a precision threshold value to obtain a first judgment result;
if the first judgment result indicates that the prediction precision of the multi-precision model reaches a precision threshold, determining that the current multi-precision model is an optimal multi-precision model;
And if the first judgment result shows that the prediction precision of the multi-precision model does not reach a precision threshold value, adding the test set into the training set to form a new training set, and optimizing the multi-precision model by using the new training set.
5. A system for predicting convective heat exchange performance of a supercritical pressure fluid, comprising:
the supercritical pressure fluid heat exchange data acquisition module is used for acquiring supercritical pressure fluid heat exchange data; the supercritical pressure fluid heat exchange data comprises high-precision heat exchange data and low-precision heat exchange data; the high-precision heat exchange data is a data set obtained by experiment or numerical simulation data; the low-precision heat exchange data is a data set obtained by dimensionless rule correlation prediction; the supercritical pressure fluid heat exchange data comprise the flow velocity of the fluid along the way, the wall surface temperature, the main flow temperature of the fluid, the fluid pressure, the convection heat exchange coefficient and the characteristic length of the pipeline;
the preprocessing module is used for preprocessing the supercritical pressure fluid heat exchange data and determining the preprocessed supercritical pressure fluid heat exchange data; the preprocessed supercritical pressure fluid heat exchange data comprise preprocessed low-precision heat exchange data and preprocessed high-precision heat exchange data;
Preprocessing is performed by the following calculation method: inquiring a fluid physical property data table according to the wall surface temperature of the fluid along the flowing direction and the values of the main stream temperature and the pressure of the fluid, obtaining the ratio of the physical property of the fluid clinging to the wall surface to the physical property of the fluid at the main stream average temperature, and generally calculating the constant pressure specific heat cpAnd the density ρ; obtaining a series of dimensionless parameters including Reynolds number Re, Przert number Pr, Nu and Bo numbers of the fluid along the flow direction according to the definition formula of the Reynolds number Re, the Przert number, the Nussel number and the Bo number in the heat transfer science; each piece of data corresponds to a certain position in the pipe, so that a data set in a non-dimensional form is obtained, and the influence of the dimension of the physical quantity on model training is avoided;
the data set consists of a high-precision data set and a low-precision data set, wherein all parameter values in the feature vector and the value of the Knudel number Nu to be predicted finally are contained; the high and low precision mainly refers to the precision of a target value Nu needing to be predicted finally, the high-precision data set refers to a data set obtained by experiment or numerical simulation data, and the low-precision data set refers to data obtained by using a dimensionless criterion correlation formula obtained by existing research; the more commonly used fluid has a dimensionless correlation of Nu ═ 0.23Re 0.8Pr0.3The Nu predicted value corresponding to the corresponding characteristic parameter can be obtained through the correlation, so that a corresponding data set with the same composition is obtained, but the result of the correlation calculation is low in precision and usually has a certain deviation from a true value, so that the data set is called a low-precision data set;
the low-precision model determining module is used for determining a low-precision model according to the preprocessed low-precision heat exchange data based on a Gaussian regression equation;
the high-precision model determining module is used for determining a high-precision model according to the preprocessed high-precision heat exchange data based on a Gaussian regression equation;
the multi-precision model determining module is used for determining a multi-precision model according to the low-precision heat exchange data, the low-precision model, the high-precision heat exchange data and the high-precision model by utilizing a collaborative kriging method; the multi-precision model is a convective heat exchange performance prediction model which takes the dimensionless form parameters of the supercritical pressure fluid heat exchange data as input and takes convective heat exchange performance as output;
and the heat exchange performance prediction module is used for predicting the heat exchange performance of the supercritical pressure fluid according to the multi-precision model.
6. The system of claim 5, wherein the pre-processing module comprises:
The cleaning unit is used for cleaning the supercritical pressure fluid heat exchange data, deleting abnormal points and outliers in the supercritical pressure fluid heat exchange data, and determining the cleaned supercritical pressure fluid heat exchange data;
the conversion module is used for converting the cleaned supercritical pressure fluid heat exchange data into dimensionless form parameters; the dimensionless form parameter is the heat exchange data of the pretreated supercritical pressure fluid.
7. The system for predicting the convective heat exchange performance of a supercritical pressure fluid according to claim 5, wherein the multi-precision model determination module specifically comprises:
the dividing unit is used for dividing the high-precision heat exchange data into a training set and a test set;
a multi-precision model determination unit for using the training set according to a formula
Determining a multi-precision model; wherein,
a high-precision model; rho (X) is a scale factor used for quantifying the relationship between two precision model output values;
a low-precision model; δ (X) is a gaussian process.
8. The system of claim 7, further comprising:
the first judgment unit is used for testing the multi-precision model according to the test set, judging whether the prediction precision of the multi-precision model reaches a precision threshold value or not, and obtaining a first judgment result;
An optimal multi-precision model determining unit, configured to determine that the current multi-precision model is an optimal multi-precision model if the first determination result indicates that the prediction precision of the multi-precision model reaches a precision threshold;
and the optimization unit is used for adding the test set into the training set to form a new training set and optimizing the multi-precision model by using the new training set if the first judgment result indicates that the prediction precision of the multi-precision model does not reach a precision threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010716914.3A CN111861011B (en) | 2020-07-23 | 2020-07-23 | Supercritical pressure fluid convection heat exchange performance prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010716914.3A CN111861011B (en) | 2020-07-23 | 2020-07-23 | Supercritical pressure fluid convection heat exchange performance prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111861011A CN111861011A (en) | 2020-10-30 |
| CN111861011B true CN111861011B (en) | 2022-05-31 |
Family
ID=72950321
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010716914.3A Active CN111861011B (en) | 2020-07-23 | 2020-07-23 | Supercritical pressure fluid convection heat exchange performance prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111861011B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116955933B (en) * | 2023-09-21 | 2024-01-09 | 深圳市丰瑞德机电技术有限公司 | Variable flow microchannel heat exchanger performance detection method, device and equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170297941A1 (en) * | 2017-07-06 | 2017-10-19 | Xi'an Jiaotong University | Indirect heat transfer supercritical water oxidation system and control method thereof |
| CN108491668B (en) * | 2018-04-17 | 2021-06-08 | 北京理工大学 | Aircraft system optimization method based on dynamic multi-model fusion |
-
2020
- 2020-07-23 CN CN202010716914.3A patent/CN111861011B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111861011A (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3944125B1 (en) | Method and system for manufacturing a heat exchanger for supercritical pressure fluid | |
| CN110289111B (en) | Natural circulation loop flow determination method and device and server | |
| CN110414114B (en) | Multi-objective multi-parameter optimization design method for U-shaped ground heat exchanger | |
| Wang et al. | Aerodynamic optimization for low pressure turbine exhaust hood using Kriging surrogate model | |
| SaemI et al. | Computation of two-and three-dimensional water hammer flows | |
| Yu et al. | Improved treatment of the open boundary in the method of lattice boltzmann equation: general description of the method | |
| Chen et al. | Towards high-accuracy deep learning inference of compressible flows over aerofoils | |
| JP2023521581A (en) | Computer simulation method for boiling water reactor (BWR) | |
| CN113011111B (en) | Turbulence excitation evaluation method and device for rod-shaped structure of reactor internals | |
| CN114462336B (en) | Method for calculating average temperature of coolant of main pipeline of nuclear reactor | |
| Varun Kumar et al. | Artificial neural network modeling for predicting the transient thermal distribution in a stretching/shrinking longitudinal fin | |
| CN111861011B (en) | Supercritical pressure fluid convection heat exchange performance prediction method and system | |
| JP2014013424A (en) | Calculation method for gas-liquid two-phase flow simulation, program, and apparatus | |
| Ye et al. | Verification and validation of CFD simulations of the NTNU BT1 wind turbine | |
| CN114896891A (en) | Steam simulation calculation method based on error correction of nuclear extreme learning machine | |
| CN117371354A (en) | Natural gas pipe network transient steady-state simulation method, device, equipment and medium | |
| Pirozzoli et al. | Direct numerical simulation of one-sided forced thermal convection in plane channels | |
| Tong et al. | Investigating the thermal-hydraulic enhancement and ash deposition characteristics of leeward-cut spiral finned tube heat exchangers | |
| JP2023515756A (en) | Method and system for estimating hydraulic conditions in dynamic operation of steam heat supply networks | |
| CN113051846A (en) | Wall surface first layer grid thickness estimation method considering compressible and heat conduction effects | |
| Wang et al. | Effect of uncertain operating conditions on the aerodynamic performance of high-pressure axial turbomachinery blades | |
| CN105032951A (en) | Control method for improving precision of ultra-fast cooling temperature model and self-learning efficiency | |
| CN113536640A (en) | Optimal design method of internal flow channel structure of air distributor based on orthogonal test | |
| Vedam et al. | Evaluation of Gradient and Curvature-Based Adaptive Mesh Refinement for Viscous Transonic Flows | |
| Mahood et al. | Analytical and numerical investigation of transient gas blow down |
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 |