CN109539596B - GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method - Google Patents
GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method Download PDFInfo
- Publication number
- CN109539596B CN109539596B CN201811435232.4A CN201811435232A CN109539596B CN 109539596 B CN109539596 B CN 109539596B CN 201811435232 A CN201811435232 A CN 201811435232A CN 109539596 B CN109539596 B CN 109539596B
- Authority
- CN
- China
- Prior art keywords
- grnn
- fitness
- heat
- population
- photo
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/40—Solar thermal energy, e.g. solar towers
Abstract
The invention discloses a GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method, which comprises the following steps: 1) determining parameters of a solar heat collection system, 2) collecting and distributing training test data of a network, 3) constructing a GRNN structure, and 4) determining an optimal smoothing factor sigma through GA; 5) training the GA-GRNN to obtain the trained GA-GRNN; 6) testing the GA-GRNN, inputting the test data selected in the step 2 into the GA-GRNN trained in the step 5 for testing; 7) and (6) performing photo-thermal efficiency prediction by using the GA-GRNN trained in the step (6) to obtain a photo-thermal efficiency prediction result of the current solar heat collection system. The method for predicting the photo-thermal efficiency makes up the uncertainty of weather and climate factors, greatly reduces the difficulty of data collection, and can accurately predict the photo-thermal power of the solar thermal collector.
Description
Technical Field
The invention belongs to the technical field of prediction of solar power generation and heat collection power, and relates to a prediction method of solar heat collection system photo-thermal efficiency based on GA-GRNN.
Background
The energy crisis caused by the accelerated development of industry and economy has become a serious threat to global sustainable development. In order to relieve the pressure of environment and energy, solar energy is widely applied by virtue of no pollution and abundant resources. In addition to photovoltaic power generation and photothermal power generation, large-scale heating projects are also gradually beginning to utilize solar heat collection technology for liquid heating, heating and the like. In large-area residential areas in China and production processes of most of large-scale industrial enterprises, hot water needs to be supplied continuously, and in terms of traditional heat supply modes, the carbon dioxide emission of coal-fired boilers is too high, natural gas boilers need to build pipelines and gas sources in advance, and electric boilers are high in operation cost. Therefore, under the requirements of national energy conservation and emission reduction policies, solar heat supply becomes a promising solution.
The current mainstream solar heat collection technology mainly comprises a tower type, a groove type, a disc type, a linear Fresnel type and the like. The planar mirror field of the linear Fresnel type light condensing system collects heat to the heat collecting tube under the CPC by tracking the sun for subsequent heat utilization, has the advantages of simple structure, small wind resistance, low cost, high land utilization rate and the like, and is gradually applied in a large scale. However, due to the periodicity of solar irradiation and the sensitivity of the solar radiation due to weather influence, solar heat utilization is more complex and unstable than the traditional heat supply mode, and the technology related to solar heat utilization efficiency prediction is still in a starting stage, mostly needs to depend on manual measurement and calculation, so that the cost is higher, and the error is larger.
Disclosure of Invention
The invention aims to provide a GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method, and solves the problems that in the solar heat utilization efficiency prediction process in the prior art, most of the solar heat utilization efficiency prediction methods need to depend on manual measurement and calculation, the data cost is high, and the result error is large.
The technical scheme adopted by the invention is that a GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method is implemented according to the following steps:
taking the smoothing factor sigma optimized in the step 4 as a smoothing factor of GRNN, substituting the smoothing factor sigma into the GRNN structure constructed in the step 3, and determining a network transfer function; training by taking the photothermal efficiency corresponding to each integral point as a target quantity to obtain trained GA-GRNN;
inputting the test data selected in the step 2 into the GA-GRNN trained in the step 5 for testing;
and inputting the relative humidity, the surface temperature and the normal direct radiation intensity of the current day to obtain a photo-thermal efficiency prediction result of the current solar heat collection system.
The beneficial effects of the invention are that the invention comprises the following aspects:
1) and constructing and training a generalized neural network optimized by a genetic algorithm by taking three items of relative humidity, surface temperature and normal direct radiation intensity as input quantities and taking the photo-thermal power of a solar heat collecting system as output quantities. Because the fitting accuracy and the prediction effect of GRNN are good when the sample data is less, the difficulty of data collection is greatly reduced. After the network training is finished, the meteorological parameters of the day are input, and the prediction quantity of the photo-thermal efficiency can be obtained, so that guidance is provided for heat collection control scheduling in production and life.
2) The method utilizes the nonlinear mapping capability, the strong approximation capability, the high fault tolerance and the global optimization capability of the genetic algorithm of the generalized regression neural network, makes up the uncertainty of weather and climate factors based on the principle of probability statistics, has little influence on the fitting precision and the prediction effect of the GA-GRNN network under the condition of small sample data, reduces the data collection difficulty and the labor cost, and can accurately predict the photo-thermal power of the solar heat collector.
Drawings
FIG. 1 is a linear Fresnel solar energy collection system that is the subject of an embodiment of the method of the present invention;
FIG. 2 is a diagram of a generalized neural network architecture employed in the method of the present invention;
FIG. 3 is a flow chart of the GA-GRNN algorithm employed in the method of the present invention.
In the figure, 1 is a light and heat collecting subsystem, 2 is a heat exchange subsystem, 3 is a heat utilization device, 4 is a circulation subsystem, 5 is a plane reflector field, 6 is a Fresnel mirror, 7 is a light and heat collecting pipe, 8 is a superheater, 9 is a steam generator, 10 is a preheater, 11 is an oil circulator, 12 is a cooler, 13 is a heat conducting oil pipeline, and 14 is a steam pipeline.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
GA-GRNN is a Genetic Algorithm-generalized regression neural network, and is referred to as a generalized neural network optimized by a Genetic Algorithm.
Referring to fig. 1, the linear fresnel type solar heat collection system targeted by the embodiment of the photothermal efficiency prediction method of the present invention has a structure consisting of a light-gathering heat collection subsystem 1, a heat exchange subsystem 2, a heat utilization device 3 and a circulation subsystem 4,
the light and heat collecting subsystem 1 consists of a plane reflector field 5, a Fresnel mirror 6 and a light and heat collecting tube 7; the plane mirror field 5 is a plurality of groups of mirror arrays which are horizontally arranged in the east-west direction or obliquely arranged in the north-south direction, an automatic tracking component is fixedly installed below each group of mirrors to enable the mirror surface of the plane mirror array to track the sun at any time so as to obtain the direct light rays as much as possible, the light rays are reflected and gathered into a Fresnel mirror 6 installed on the upper portion of the plane mirror field 5 for one time, one part of the light rays are directly received by a light-gathering heat collecting tube 7, and the other part of the light rays are received by the light-gathering heat collecting tube 7 after being reflected secondarily by a built-in reflecting surface of. The photo-thermal conversion raises the temperature of the working medium in the light-gathering heat collecting tube 7, and converts solar energy into heat energy for power generation or heat supply at the rear end;
the heat exchange subsystem 2 consists of a superheater 8, a steam generator 9 and a preheater 10, heat conducting oil in the light-gathering heat collecting tube 7 is heated after photo-thermal conversion, enters the preheater 10 through a heat conducting oil pipeline 13, and further heats cold water to enter the steam generator 9 through oil-water heat exchange; subsequently, saturated steam is generated by the steam generator 9, sometimes the temperature of the general saturated steam cannot meet the production requirement, superheated steam with higher temperature can be further obtained by using the superheater 8, and high-temperature and high-pressure steam is conveyed to the heat utilization equipment 3 through the steam pipeline 14 for heat supply or power generation so as to be effectively utilized;
the circulation subsystem 4 consists of an oil circulator 11 and a cooler 12, when the heat conduction oil reaches the working temperature, the heat conduction oil enters the heat exchange subsystem 2 for heat exchange, and steam with a certain temperature is generated; the circulation flow of the heat conduction oil is adjusted to be adaptive to the radiation level through the oil circulator 11 during operation, so that the working temperature of the heat conduction oil is ensured.
Referring to fig. 3, based on the above structure, the method for predicting photothermal efficiency of the present invention is implemented according to the following steps:
in order to provide training data for the neural network, a part of photothermal efficiency needs to be calculated, and the photothermal power of the solar heat collection system is related to specific parameters of each subsystem.
Aiming at the working process of the linear Fresnel type solar heat collection system, in order to collect subsequent data, the calculation method of the photo-thermal efficiency is described herein, and the specific process is,
1.1) calculating the solar radiation illumination intensity,
solar radiation is an energy source of a solar heat collecting system, and the determination of the magnitude of solar irradiance projected on a lighting surface of the light-gathering heat collecting tube 7 is the primary step of system power calculation.
After solar radiation enters the atmosphere of the earth, absorption, reflection and scattering occur, in order to obtain more solar heat, an inclination angle theta is generally formed in the installation of the light-gathering heat-collecting tube 7, and the meteorological data shows the solar radiation amount on the horizontal plane, so that the solar radiation amount on the inclination angle inclined plane needs to be equivalently converted in advance.
The radiation that reaches the ground and can be utilized by the solar energy collection system mainly comprises direct solar radiation I that comes directly from the sun and is received without ever changing directiond·θSolar dispersion I redirected by atmospheric dispersions·θAnd the reflected illumination I reflected and received by the groundr·θThe total irradiation intensity of the sun IθExpressed as: i isθ＝Is·θ+Id·θ+Ir·θThe unit is W/m2； (1)
1.2) calculating the efficiency of the light-gathering heat-collecting tube 7,
the light-gathering heat-collecting tube 7 is a core component for heat absorption and heat transfer of solar energy, the light-gathering heat-collecting tube 7 of the existing linear Fresnel type solar heat-collecting system mainly adopts two forms of a single-tube CPC cavity heat collector and a compact multi-tubule cavity heat collector, and the single-tube form can reach higher heat-collecting temperature, so that the single-tube CPC cavity heat collector is often selected in engineering.
Set the specific heat capacity to be cpThe heat conducting oil flows through the light-gathering heat collecting pipe 7 at the mass flow rate m, and the inlet temperature of the heat conducting oil of the heat conducting working medium is TiOutlet temperature of ToDuring the period, the light-gathering heat-collecting tube 7 does useful work QuThe expression is as follows: qu＝m·cp(To-Ti)， (2)
Let the total area of the plane mirror field 5 be D, in m2The total irradiation intensity of the sun is IθIn the unit of W/m2The efficiency expression of the light-gathering heat-collecting tube 7 is as follows:
1.3) calculating the photo-thermal efficiency of the linear Fresnel type solar heat collection system,
the total photo-thermal efficiency of the linear Fresnel type solar heat collecting system is influenced by the efficiency of the reflecting mirror and the thermal efficiency of the light-gathering heat collecting tube 7, and the expression is as follows:
eta photothermal efficiency eta reflector eta heat collecting tube (4)
In the formula etaReflecting mirrorIn relation to the structural arrangement, mounting position and orientation of the flat mirror field 5 and the fresnel mirror 6, field measurements are required.
the main light-gathering structure of the linear Fresnel type solar heat collection system is a plane reflector field 5, and after the structure of the plane reflector field 5 is fixed, the main factors influencing the photo-thermal efficiency are meteorological conditions, such as air temperature, cloud amount, relative humidity, surface temperature, radiation intensity and the like.
In the step, three indexes of relative humidity, surface temperature and normal direct radiation intensity with high correlation are selected as input quantity of the neural network, in the embodiment, whole-point data of No. 1, No. 8, No. 15, No. 22 and No. 29 in each month in 2015, which are 8:00-18:00 every day for 60 days, is selected, and the table 1 is meteorological data of No. 15 noon 12 o' clock in each month.
TABLE 1 neural network partial input data collected in the examples
| Date | Humidity of air | Surface temperature | Intensity of normal direct radiation |
| 15 |
26 | 8.0 | 166.35 |
| 2015, 3 months and 15 |
6 | 20.3 | 590.19 |
| 15 |
37 | 36.3 | 613.48 |
| 15 |
69 | 14.7 | 3.04 |
| 9/15/2015 | 22 | 37.2 | 758.83 |
| 11/15/2015 | 37 | 9.1 | 639.01 |
Meanwhile, the photothermal efficiency measurement is calculated and recorded at the whole point of 60 days and 8:00-18:00 every day in total of No. 1, No. 8, No. 15, No. 22 and No. 29 every month of 12 months in the whole year of 2015, and the measurement data of 12 am at 15 days of odd months are shown in the table 2:
TABLE 2 measurement of photothermal efficiency
| Date | 2015.1.15 | 2015.3.15 | 2015.5.15 | 2015.7.15 | 2015.9.15 | 2015.11.15 |
| Efficiency of light and heat | 0.3312 | 0.4936 | 0.6133 | 0.3211 | 0.5250 | 0.3488 |
In the above embodiment, the data of 1, 8, 22 and 29 days per month are actually used as training data input networks (total 576 groups), and the data of 15 days per month are actually used as test data (total 144 groups).
referring to fig. 2, the GRNN is composed of an input layer, a mode layer, a summation layer, and an output layer, the input layer directly transfers input variables to the mode layer, each neuron of the mode layer corresponds to different types of sample inputs, and a transfer function is:
in the formula (5), X is a network input variable, XiAnd sigma is a smoothing factor for the learning sample corresponding to the ith neuron, and in order to improve the accuracy of prediction, the value of sigma is optimized and calculated through a genetic algorithm.
Performing arithmetic summation on the output of each mode layer neuron in a summation layer, wherein the connection weight of the mode layer and each neuron is 1, and the summation calculation formula is as follows:
the neuron transfer function is:
the number of the neurons in the output layer is equal to the dimension of the output vector in the learning sample, and the values are all 1;
each neuron divides the output of the summation layer, and the output of the neuron j corresponds to the jth element of the estimation result y (x), namely:
as can be seen from step 3, the essence of network training is to optimize the smoothing factor σ, which has a significant influence on the network prediction accuracy. The traditional test method makes the smoothing factor sigma increase progressively in a certain interval, counts the error sequence of the corresponding sample result, and determines the optimal parameter by taking the minimum root mean square error of the error sequence as a standard, but the method has complex calculation and is difficult to determine the set interval.
Referring to fig. 3, in order to make up for the selection problem of the smoothing factor in GRNN modeling, GRNN is regarded as a prediction function, and the global optimization capability of the genetic algorithm is used to optimize the smoothing factor σ,
4.1) setting basic parameters of the genetic algorithm,
setting the smoothing factor sigma as chromosome gene with value range of [0.05, 1%]The operation precision is set to 0.0001, and the bit length is ceil (log)2((1-0.05)./0.0001))＝14；
Generating 50 random numbers in a value range as a first generation population, generally, obtaining a more excellent solution through multiple evolutions, and setting the iteration number N to be 50; the crossover probability pc is 0.4; the variation probability pm is 0.2;
4.2) constructing a fitness function to evaluate the quality of the individual,
fitness (fitness) is a measure for measuring the goodness of each smooth factor individual in the population, which may reach or approach the optimal solution, and fitness function (fitness function) is used for evaluating the goodness of a new individual and improving the network accuracy through cross validation.
Dividing the training sample into a first subset and a second subset, adopting the first subset as the training sample, and taking the average error function of the predicted value and the actual value of the second subset as the fitness function, wherein the fitness function expression is as follows:
in the formula (9), E (σ) isThe pre-smoothing factor tests the mean value of the error of the sample,
for the prediction of the test sample, y (x)i) For the actual output of the test sample, n is the population number, the example chooses n-50,
calculating fitness values of 50 individuals according to the formula (9), sequencing in an ascending order to obtain an optimal individual, and recording the sample as a population P1;
4.3) carrying out genetic manipulation,
selecting individuals with good fitness from the population P1 to form a new population by adopting a proportional selection method in genetic operation, accumulating the individual fitness in the population P1 to obtain the fitness sum Sigma fitness, establishing a region corresponding relation with [0 ] and Sigma fitness ] according to the individual fitness value, generating a random number in the range of [0 ] and Sigma fitness ], and selecting the individuals corresponding to the region where the random number is located; the larger the individual fitness value is, the larger the probability of being selected is;
the cross operation enables the two selected individuals with high fitness to be obtained through advantage complementation, and more excellent individuals are obtained; arithmetic crossing is adopted in the step, new individuals are generated through linear combination, and the crossing probability pc is taken to be 0.4;
the variation increases the randomness of individual variation of the smooth factors, increases the individual diversity, enlarges the search range, and completes the variation operation by carrying out negation operation on the selected variation position.
4.4) carrying out genetic iteration,
after genetic selection, cross operation and mutation operation are carried out on the population P1, the fitness value of a new individual is calculated again, the individual with a large fitness value is reserved, a child population P2 is obtained, the P2 is used as a parent population again, the step 4.3) is repeated, iteration is carried out continuously and circularly, whether the genetic iteration frequency reaches 50 or not is judged, and if yes, the optimal individual output of the last calculation result is used as an optimized smooth factor sigma;
taking the smoothing factor sigma optimized in the step 4 as a smoothing factor of GRNN, substituting the smoothing factor sigma into the GRNN structure constructed in the step 3, and determining a network transfer function; inputting the relative humidity, the surface temperature and the normal direct radiation intensity data of the whole points of 576 groups of days 8:00-18:00 of each month in 2015 year and 29 days into GA-GRNN, and training by taking the photothermal efficiency corresponding to each whole point as a target amount to obtain the trained GA-GRNN;
and inputting 144 groups of data totaling 15 days 8:00-18:00 each month in 15 days of 15 days in the step 2 as test data into the GA-GRNN trained in the step 5 for testing. Tests show that the GA-GRNN has high prediction accuracy on the photo-thermal efficiency of the solar heat collection system, and the prediction value can provide guidance for a general production process.
and inputting the relative humidity, the surface temperature and the normal direct radiation intensity of the current day to obtain a photo-thermal efficiency prediction result of the current solar heat collection system.
Claims (1)
1. A prediction method for solar heat collection system photo-thermal efficiency based on GA-GRNN is characterized by comprising the following steps:
step 1, determining parameters of a solar heat collection system, specifically comprising the following steps of,
1.1) calculating the solar radiation illumination intensity,
the radiation reaching the ground utilized by the solar energy collection system mainly comprises direct solar radiation I received directly without ever changing direction from the sund·θSolar dispersion I redirected by atmospheric dispersions·θAnd the reflected illumination I reflected and received by the groundr·θThe total irradiation intensity of the sun IθExpressed as: i isθ＝Is·θ+Id·θ+Ir·θThe unit is W/m2；(1)
1.2) calculating the efficiency of the light-gathering heat-collecting tube,
the lower light-gathering heat-collecting pipes are all single-pipe CPC cavity type heat collectors,
set the specific heat capacity to be cpThe heat conducting oil flows through the light-gathering heat collecting pipe at mass flow rate m, and the inlet temperature of the heat conducting oil of the heat conducting working medium is TiOutlet temperature of ToDuring the period, the light-gathering heat-collecting tube does useful work QuThe expression is as follows:
Qu＝m·cp(To-Ti)， (2)
let the total area of the plane mirror field be D, in m2The total irradiation intensity of the sun is IθIn the unit of W/m2The efficiency expression of the light-gathering heat-collecting tube is as follows:
1.3) calculating the photo-thermal efficiency of the linear Fresnel type solar heat collection system,
the total photo-thermal efficiency of the linear Fresnel type solar heat collecting system is influenced by the efficiency of the reflecting mirror and the thermal efficiency of the light-gathering heat collecting tube, and the expression is as follows:
ηefficiency of light and heat＝ηReflecting mirror·ηHeat collecting pipe， (4)
In the formula etaReflecting mirrorThe structure setting, the installation position and the direction of the plane reflecting mirror field and the Fresnel mirror are related, and the field measurement is needed;
step 2, collecting and distributing the training test data of the network,
selecting three indexes of relative humidity, surface temperature and normal direct radiation intensity with high correlation degree as input quantity of the neural network by factors influencing the photothermal efficiency;
step 3, constructing a GRNN structure, specifically comprising the following steps,
the GRNN comprises an input layer, a mode layer, a summation layer and an output layer, wherein the input layer directly transmits input variables to the mode layer, each neuron of the mode layer corresponds to different types of sample inputs, and a transfer function is as follows:
in the formula (5), X is a network input variable, XiIs a learning sample corresponding to the ith neuron, sigma is a smoothing factor,
performing arithmetic summation on the output of each mode layer neuron in a summation layer, wherein the connection weight of the mode layer and each neuron is 1, and the summation calculation formula is as follows:
the neuron transfer function is:
the number of the neurons in the output layer is equal to the dimension of the output vector in the learning sample, and the values are all 1;
each neuron divides the output of the summation layer, and the output of the neuron j corresponds to the jth element of the estimation result y (x), namely:
step 4, determining an optimal smoothing factor sigma through GA, regarding GRNN as a prediction function, optimizing the smoothing factor sigma by using the global optimizing capability of the genetic algorithm,
4.1) setting basic parameters of the genetic algorithm,
setting the smoothing factor sigma as chromosome gene with value range of [0.05, 1%]The operation precision is set to 0.0001, and the bit length is ceil (log)2((1-0.05)./0.0001))＝14；
Generating 50 random numbers in a value range as a first generation population, obtaining a more excellent solution through multiple evolutions, and setting the iteration number N to be 50; the crossover probability pc is 0.4; the variation probability pm is 0.2;
4.2) constructing a fitness function to evaluate the quality of the individual,
fitness (fitness) is a measure for measuring the goodness of each smooth factor individual in the population, which may reach or approach the optimal solution, fitness function (fitness function) is used for evaluating the goodness of a new individual, network accuracy is improved through cross validation,
dividing the training sample into a first subset and a second subset, adopting the first subset as the training sample, and taking the average error function of the predicted value and the actual value of the second subset as the fitness function, wherein the fitness function expression is as follows:
in equation (9), E (σ) is the mean error value of the current smoothing factor test sample,
for the prediction of the test sample, y (x)i) For the actual output of the test sample, n is the population number, n is 50,
calculating fitness values of 50 individuals according to the formula (9), sequencing in an ascending order to obtain an optimal individual, and recording the sample as a population P1;
4.3) carrying out genetic manipulation,
selecting individuals with good fitness from the population P1 to form a new population by adopting a proportional selection method in genetic operation, accumulating the individual fitness in the population P1 to obtain the fitness sum Sigma fitness, establishing a region corresponding relation with [0 ] and Sigma fitness ] according to the individual fitness value, generating a random number in the range of [0 ] and Sigma fitness ], and selecting the individuals corresponding to the region where the random number is located;
the cross operation enables the two selected individuals with high fitness to be obtained through advantage complementation, and more excellent individuals are obtained; arithmetic crossing is adopted in the step, new individuals are generated through linear combination, and the crossing probability pc is taken to be 0.4;
the randomness of individual change of the smooth factors is increased through mutation, and mutation operation is completed through negation operation on the selected mutation positions;
4.4) carrying out genetic iteration,
after genetic selection, cross operation and mutation operation are carried out on the population P1, the fitness value of a new individual is calculated again, the individual with a large fitness value is reserved, a child population P2 is obtained, the P2 is used as a parent population again, the step 4.3) is repeated, iteration is carried out continuously and circularly, whether the genetic iteration frequency reaches 50 or not is judged, and if yes, the optimal individual output of the last calculation result is used as an optimized smooth factor sigma;
step 5, training the GA-GRNN,
taking the smoothing factor sigma optimized in the step 4 as a smoothing factor of GRNN, substituting the smoothing factor sigma into the GRNN structure constructed in the step 3, and determining a network transfer function; training by taking the photothermal efficiency corresponding to each integral point as a target quantity to obtain trained GA-GRNN;
step 6, testing the GA-GRNN,
inputting the test data selected in the step 2 into the GA-GRNN trained in the step 5 for testing;
step 7, using the GA-GRNN trained in the step 6 to predict the photothermal efficiency,
and inputting the relative humidity, the surface temperature and the normal direct radiation intensity of the current day to obtain a photo-thermal efficiency prediction result of the current solar heat collection system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811435232.4A CN109539596B (en) | 2018-11-28 | 2018-11-28 | GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811435232.4A CN109539596B (en) | 2018-11-28 | 2018-11-28 | GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109539596A CN109539596A (en) | 2019-03-29 |
| CN109539596B true CN109539596B (en) | 2020-10-23 |
Family
ID=65850865
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811435232.4A Active CN109539596B (en) | 2018-11-28 | 2018-11-28 | GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109539596B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110837932A (en) * | 2019-11-08 | 2020-02-25 | 陕西省水利电力勘测设计研究院 | Thermal power prediction method of solar heat collection system based on DBN-GA model |
| CN111625956B (en) * | 2020-05-25 | 2023-11-10 | 浙江大学 | Tower type solar energy mirror field optimization method based on self-adaptive differential evolution algorithm |
| CN111815051B (en) * | 2020-07-06 | 2022-09-23 | 安徽建筑大学 | GRNN photovoltaic power generation prediction method considering weather influence factors |
| CN112836333A (en) * | 2020-11-16 | 2021-05-25 | 苏州西热节能环保技术有限公司 | Method and device for calculating power generation efficiency of photo-thermal power station and readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1998020437A2 (en) * | 1996-11-04 | 1998-05-14 | 3-Dimensional Pharmaceuticals, Inc. | System, method and computer program product for identifying chemical compounds having desired properties |
| CN103810534A (en) * | 2013-12-11 | 2014-05-21 | 广西电网公司电力科学研究所 | Photovoltaic power output prediction method |
| CN103955768A (en) * | 2014-04-30 | 2014-07-30 | 河北省电力勘测设计研究院 | CSP radiation and heat energy predication method based on BP neural network model |
| CN106227915A (en) * | 2016-07-07 | 2016-12-14 | 南京工程学院 | Disc type solar energy heat collector exit temperature prediction method based on GA RBF |
| CN107577844A (en) * | 2017-08-04 | 2018-01-12 | 中国电力科学研究院 | The Forecasting Methodology and system of a kind of slot type photo-thermal power station mirror field outlet temperature |
| CN108131848A (en) * | 2017-12-15 | 2018-06-08 | 河海大学 | A kind of DSG slot types photo-thermal power station heat collecting field fluid properties Forecasting Methodology |
| CN108171363A (en) * | 2017-12-13 | 2018-06-15 | 北京金风慧能技术有限公司 | The Forecasting Methodology and device of photo-thermal power generation power |
-
2018
- 2018-11-28 CN CN201811435232.4A patent/CN109539596B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1998020437A2 (en) * | 1996-11-04 | 1998-05-14 | 3-Dimensional Pharmaceuticals, Inc. | System, method and computer program product for identifying chemical compounds having desired properties |
| CN103810534A (en) * | 2013-12-11 | 2014-05-21 | 广西电网公司电力科学研究所 | Photovoltaic power output prediction method |
| CN103955768A (en) * | 2014-04-30 | 2014-07-30 | 河北省电力勘测设计研究院 | CSP radiation and heat energy predication method based on BP neural network model |
| CN106227915A (en) * | 2016-07-07 | 2016-12-14 | 南京工程学院 | Disc type solar energy heat collector exit temperature prediction method based on GA RBF |
| CN107577844A (en) * | 2017-08-04 | 2018-01-12 | 中国电力科学研究院 | The Forecasting Methodology and system of a kind of slot type photo-thermal power station mirror field outlet temperature |
| CN108171363A (en) * | 2017-12-13 | 2018-06-15 | 北京金风慧能技术有限公司 | The Forecasting Methodology and device of photo-thermal power generation power |
| CN108131848A (en) * | 2017-12-15 | 2018-06-08 | 河海大学 | A kind of DSG slot types photo-thermal power station heat collecting field fluid properties Forecasting Methodology |
Non-Patent Citations (1)
| Title |
|---|
| 遗传算法的广义回归神经网络建模方法;孔国利等;《计算机工程与设计》;20170228;第38卷(第2期);第489页左栏第2段至第490页左栏第6段 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109539596A (en) | 2019-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109539596B (en) | GA-GRNN-based solar heat collection system photo-thermal efficiency prediction method | |
| Qiu et al. | A comprehensive model for analysis of real-time optical performance of a solar power tower with a multi-tube cavity receiver | |
| Liang et al. | Study on the thermal performance of a novel cavity receiver for parabolic trough solar collectors | |
| Xie et al. | Numerical and experimental analysis of a point focus solar collector using high concentration imaging PMMA Fresnel lens | |
| Khosravi et al. | Energy modeling of a solar dish/Stirling by artificial intelligence approach | |
| Zaaoumi et al. | Estimation of the energy production of a parabolic trough solar thermal power plant using analytical and artificial neural networks models | |
| Ghazouani et al. | Thermo-economic and exergy analysis and optimization of small PTC collectors for solar heat integration in industrial processes | |
| Wang et al. | Optical efficiency improvement of solar power tower by employing and optimizing novel fin-like receivers | |
| Cheng et al. | A novel optical optimization model for linear Fresnel reflector concentrators | |
| Gao et al. | Model building and optical performance analysis on a novel designed compound parabolic concentrator | |
| Hachicha | Numerical modelling of a parabolic trough solar collector | |
| Kuo et al. | A bifacial photovoltaic thermal system design with parameter optimization and performance beneficial validation | |
| Ahlgren et al. | A simplified model for linear correlation between annual yield and DNI for parabolic trough collectors | |
| Hu et al. | Design and performance analysis of a multi-reflection heliostat field in solar power tower system | |
| Chen et al. | Model construction and performance investigation of multi-section compound parabolic concentrator with solar vacuum tube | |
| Vakili et al. | A review of recent developments in the application of machine learning in solar thermal collector modelling | |
| Men et al. | Study on the annual optical comprehensive performance of linear Fresnel reflector concentrators with an effective multi-objective optimization model | |
| Wang et al. | Optical-thermal-mechanical analysis of high-temperature receiver integrated with gradually sparse biomimetic heliostat field layouts for the next-generation solar power tower | |
| Korzenszky | Linear model of DHW system using response surface method approach | |
| Mohanty et al. | Soft computing techniques for a solar collector using solar radiation data | |
| Giovannetti et al. | Comparison of process heat collectors with respect to technical and economic conditions | |
| Zima et al. | Mathematical model of an innovative double U-tube sun-tracked PTC and its experimental verification | |
| Deng et al. | Model verification and photo-thermal conversion assessment of a novel facade embedded compound parabolic concentrator | |
| Khosravi et al. | Solar power tower system | |
| CN105447325A (en) | Method for calculating cost and recovery period of solar concentrating photovoltaic and photothermal cogeneration system |
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 |