CN109539596A - Solar thermal collection system light thermal efficiency forecast method based on GA-GRNN - Google Patents
Solar thermal collection system light thermal efficiency forecast method based on GA-GRNN Download PDFInfo
- Publication number
- CN109539596A CN109539596A CN201811435232.4A CN201811435232A CN109539596A CN 109539596 A CN109539596 A CN 109539596A CN 201811435232 A CN201811435232 A CN 201811435232A CN 109539596 A CN109539596 A CN 109539596A
- Authority
- CN
- China
- Prior art keywords
- grnn
- thermal efficiency
- fitness
- collection system
- individual
- 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.)
- Granted
Links
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 solar thermal collection system light thermal efficiency forecast method based on GA-GRNN that the invention discloses a kind of, step includes: 1) to determine the parameter of solar thermal collection system, 2) collect, distribute the training test data of network, 3) GRNN structure is constructed, 4) optimal smoothing factor sigma determined by GA；5) training GA-GRNN, obtains trained GA-GRNN；6) GA-GRNN is tested, the test data selected in step 2 is input in the trained GA-GRNN of step 5 and is tested；7) photo-thermal EFFICIENCY PREDICTION is carried out with the trained GA-GRNN of step 6, obtains the photo-thermal EFFICIENCY PREDICTION result of current solar thermal collection system.Light thermal efficiency forecast method of the invention, compensates for the uncertainty of weather, climatic factor, greatly reduces the difficulty of data collection, is capable of the photo-thermal power of accurate prediction solar thermal collector.
Description
Technical field
The invention belongs to solar generating heat collecting power prediction technical fields, are related to a kind of solar energy based on GA-GRNN
Collecting system light thermal efficiency forecast method.
Background technique
Industry has become seriously threatening for Global Sustainable Development with energy crisis caused by economic accelerated development.For
Alleviation environment and Pressure on Energy, the advantage that solar energy is pollution-free with its, resourceful are widely applied.In addition to photovoltaic is sent out
With outside photo-thermal power generation, big thermoelecrtic project also gradually starts with solar energy heating technology and carries out liquid heating, heating etc. electricity.I
Constant GEOTHERMAL WATER supply is required in the production procedure of state's large area residential block and most large size industrial enterprises, just
For traditional heating mode, the CO2 emissions of coal-burning boiler are excessively high, and natural gas boiler needs to build pipeline and gas in advance
Source, electric boiler operating cost are high.Therefore under national energy conservation and emission reduction policy requirements, solar-heating, which becomes one, extremely prospect
Solution.
The solar energy heating technology of mainstream mainly has tower, slot type, dish-style and linear Fresnel formula etc. at present.Wherein, line
Property Fresnel condenser system plane mirror field heat gathered in the thermal-collecting tube under CPC by the tracking sun and carries out subsequent thermal
Amount utilizes, and has many advantages, such as that structure is simple, windage is small, at low cost, land utilization ratio is high, gradually obtains large-scale application.But by
In the periodicity shined upon and the sensibility influenced by weather, solar thermal utilization is more more complicated than traditional heating mode, no
To stablize, and the technology about solar thermal utilization EFFICIENCY PREDICTION is still at an early stage, majority needs to calculate by manual measurement,
Higher cost, error are larger.
Summary of the invention
The solar thermal collection system light thermal efficiency forecast method based on GA-GRNN that the object of the present invention is to provide a kind of, solution
The prior art of having determined during solar thermal utilization EFFICIENCY PREDICTION, majority need by manual measurement calculate, data cost compared with
Height, the larger problem of resultant error.
The technical scheme adopted by the invention is that a kind of solar thermal collection system photo-thermal EFFICIENCY PREDICTION based on GA-GRNN
Method follows the steps below to implement:
Step 1, the parameter for determining solar thermal collection system,
Step 2, collection, the training test data for distributing network,
Step 3, building GRNN structure,
Step 4 determines optimal smoothing factor sigma by GA, regards GRNN as an anticipation function, utilizes the complete of genetic algorithm
Office's optimizing ability optimizes smoothing factor σ；
Step 5, training GA-GRNN,
Using the smoothing factor σ after optimizing in step 4 as the smoothing factor of GRNN, the GRNN knot constructed in step 3 is substituted into
In structure, network transfer function is determined；It is trained, obtains trained using the corresponding photo-thermal efficiency of each integral point as aim parameter
GA-GRNN；
Step 6, test GA-GRNN,
The test data selected in step 2 is input in the trained GA-GRNN of step 5 and is tested；
Step 7 carries out photo-thermal EFFICIENCY PREDICTION with the trained GA-GRNN of step 6,
Relative humidity, surface temperature and the normal direction direct solar radiation intensity for inputting the same day, obtain current solar thermal collection system
Photo-thermal EFFICIENCY PREDICTION result.
The invention has the advantages that including the following aspects:
1) using relative humidity, surface temperature, method phase direct solar radiation intensity three as input quantity, with solar thermal collection system
Photo-thermal power be output quantity, construct simultaneously training the General Neural Network through genetic algorithm optimization.Since GRNN is in sample data
Fitting precision and prediction effect are also fine when less, therefore greatly reduce the difficulty of data collection.After network training is good, input
Same day meteorologic parameter can be obtained the premeasuring of photo-thermal efficiency, to make guidance for the collection thermal control scheduling in production and living.
2) appearance of the non-linear mapping capability of the invention for utilizing generalized regression nerve networks, stronger approximation capability, height
The global optimizing ability of mistake and genetic algorithm, the principle based on probability statistics compensate for weather, climatic factor not really
It is qualitative, and GA-GRNN network influences very little to its fitting precision and prediction effect in Small Sample Database, reduces data and receives
Collect difficulty and human cost, is capable of the photo-thermal power of accurate prediction solar thermal collector.
Detailed description of the invention
Fig. 1 is the linear Fresnel formula solar thermal collection system of embodiment of the present invention method object；
Fig. 2 is the General Neural Network structure chart that the method for the present invention uses；
Fig. 3 is the GA-GRNN algorithm flow chart that the method for the present invention uses.
In figure, 1. light and heat collection subsystems, 2. heat exchange subsystems, 3. heat utilization equipment, 4. cycle subsystems, 5. planes are anti-
Penetrate Jing Chang, 6. Fresnel mirrors, 7. light-gathering heat collection pipes, 8. superheaters, 9. steam generators, 10. preheaters, 11. oily circulators,
12. cooler, 13. thermal conductive oil pipelines, 14. steam pipeworks.
Specific embodiment
The following describes the present invention in detail with reference to the accompanying drawings and specific embodiments.
GA-GRNN full name is Genetic Algorithm-generalized regression neural network,
Connotation is the General Neural Network of genetic algorithm optimization.
Referring to Fig.1, the linear Fresnel formula solar thermal collection system of photo-thermal EFFICIENCY PREDICTION embodiment of the method object of the present invention
Structure be to be made of light and heat collection subsystem 1, heat exchange subsystem 2, heat utilization equipment 3 and cycle subsystem 4,
Light and heat collection subsystem 1 is made of plane mirror field 5, Fresnel mirror 6 and light-gathering heat collection pipe 7；Plane mirror
Field 5 is the reflection mirror array of the horizontal east-west direction arrangement of multiple groups or the arrangement of inclination north-south, and peace is fixed below every group of reflecting mirror
The mirror surface moment of planar mirror array is made to track the sun equipped with component is automatically tracked, to obtain method phase direct lines more as far as possible,
Light primary event is gathered and is mounted in the Fresnel mirror 6 on 5 top of plane mirror field, a part of light is directly by optically focused
Thermal-collecting tube 7 receives, and another part light is received by light-gathering heat collection pipe 7 again after 6 built-in reflective face secondary reflection of Fresnel mirror.Light
Thermal transition increases the temperature of working medium in light-gathering heat collection pipe 7, converts solar energy into thermal energy and uses for rear end power generation or heat supply；
Heat exchange subsystem 2 is made of superheater 8, steam generator 9, preheater 10, the conduction oil warp in light-gathering heat collection pipe 7
It heats up after crossing photothermal conversion, preheater 10 is entered by thermal conductive oil pipeline 13, further heats up cold water by grease heat exchange
Heating enters steam generator 9；Subsequent to generate saturated vapor by steam generator 9, the temperature of general saturated vapor can not sometimes
Meet production requirement, the superheated steam that superheater 8 further obtains higher temperature can be used, high temperature and high pressure steam passes through steam
Pipeline 14 is transported to heat utilization equipment 3, is used for heat supply or power generation, is utilized effectively；
Cycle subsystem 4 is made of oily circulator 11, cooler 12, when conduction oil reaches operating temperature, conduction oil into
The subsystem 2 that enters to exchange heat carries out heat exchange, the steam of output certain temperature；Following for conduction oil is adjusted by oily circulator 11 in operation
Circulation and radiation level are adapted, to ensure the operating temperature of conduction oil.
Referring to Fig. 3, based on above-mentioned structure, light thermal efficiency forecast method of the invention is followed the steps below to implement:
Step 1, the parameter for determining solar thermal collection system,
To provide training data to neural network, need to calculate a part of photo-thermal efficiency, the photo-thermal of solar thermal collection system
Power is related with the design parameter of subsystems.
For the workflow of the above linear Fresnel formula solar thermal collection system, to carry out subsequent data collection,
This illustrates that the calculation method of photo-thermal efficiency, detailed process are,
1.1) solar radiation illumination is calculated,
Solar radiation is the energy source of solar thermal collection system, and determination is incident upon on 7 lighting surface of light-gathering heat collection pipe
The magnitude of solar irradiance is the first step for carrying out system power calculating.
Solar radiation enters after Earth'S atmosphere, it may occur that absorbs, reflects and scatters, in order to obtain more sun
Heat, generally there are an inclination angle thetas for the installation of light-gathering heat collection pipe 7, and it is solar irradiation on horizontal plane that meteorological data, which provides,
Amount, therefore equivalency transform need to be shifted to an earlier date into the solar irradiation on mounted angle inclined-plane.
The radiation that reaching ground can be utilized by solar thermal collection system mainly includes never changing directly from the sun
Direction and the direct sunlight I that receivesd·θ, by atmosphere scattering influenced and the sun that is changing direction and receiving scatter
Is·θAnd through ground return and the illumination by reflected light I that receivesr·θ, then total irradiation intensity I of the sunθIt indicates are as follows: Iθ=Is·θ+
Id·θ+Ir·θ, unit is W/m2；(1)
1.2) efficiency of light-gathering heat collection pipe 7 is calculated,
Light-gathering heat collection pipe 7 is the core component of solar energy heat absorbing heat transfer, existing linear Fresnel formula solar energy heating system
Light-gathering heat collection pipe 7 of uniting mainly uses single tube CPC lumen type heat collector and compact two kinds of forms of more tubule cavity heat collectors, due to list
Form of tubes can then reach higher heat-collecting temperature, and single tube CPC lumen type heat collector is often selected in engineering.
If specific heat capacity is cpConduction oil light-gathering heat collection pipe 7, the import temperature of heat-transfer working medium conduction oil are flowed through with mass flow m
Degree is Ti, outlet temperature To, then during the useful work Q that does of light-gathering heat collection pipe 7uExpression formula are as follows:
Qu=mcp(To-Ti), (2)
If the gross area of plane mirror field 5 is D, unit m2, total irradiation intensity of the sun is Iθ, unit W/m2, gather
The efficiency expression formula of light thermal-collecting tube 7 are as follows:
1.3) the photo-thermal efficiency of linear Fresnel formula solar thermal collection system is calculated,
The total photo-thermal efficiency of linear Fresnel formula solar thermal collection system is by 7 thermal efficiency of reflector efficiency and light-gathering heat collection pipe
Influence, expression formula is as follows:
ηPhoto-thermal efficiency=ηReflecting mirror·ηThermal-collecting tube, (4)
η in formulaReflecting mirrorIt is related with the structure setting of plane mirror field 5 and Fresnel mirror 6, installation site and direction, it needs
Carry out on-site measurement.
Step 2, collection, the training test data for distributing network,
The main concentration structure of linear Fresnel formula solar thermal collection system is plane mirror field 5, plane mirror field 5
Structure it is fixed after, the principal element for influencing photo-thermal efficiency is meteorological condition, as temperature, cloud amount, relative humidity, surface temperature,
Radiation intensity etc..
This step chooses three the higher relative humidity of the degree of correlation, surface temperature and normal direction direct solar radiation intensity index conducts
Neural network input quantity, choose in embodiment No. 1 of in every month, 2015, No. 8, No. 15, No. 22, No. 29 it is 60 days daily 8 total:
The integral point data of 00-18:00, table 1 are monthly 12 points of meteorological data of No. 15 high noons therein.
Part of neural network input data collected by table 1, embodiment
| Date | Air humidity | Surface temperature | Normal direction direct solar radiation intensity |
| On January 15th, 2015 | 26 | 8.0 | 166.35 |
| On March 15th, 2015 | 6 | 20.3 | 590.19 |
| On May 15th, 2015 | 37 | 36.3 | 613.48 |
| On July 15th, 2015 | 69 | 14.7 | 3.04 |
| On September 15th, 2015 | 22 | 37.2 | 758.83 |
| On November 15th, 2015 | 37 | 9.1 | 639.01 |
No. 1 at annual 12 months 2015 monthly, No. 8, No. 15, No. 22, No. 29 total 60 days daily 8:00- simultaneously
The integral point of 18:00 carries out the calculating of photo-thermal efficiency measurement and records, and table 2 is 12 points of the measurement data of high noon of the odd number moon 15:
Table 2, photo-thermal efficiency measure
| Date | 2015.1.15 | 2015.3.15 | 2015.5.15 | 2015.7.15 | 2015.9.15 | 2015.11.15 |
| Photo-thermal efficiency | 0.3312 | 0.4936 | 0.6133 | 0.3211 | 0.5250 | 0.3488 |
Then in the above embodiments, it is practical using monthly 1 day, 8 days, 22 days, data on the 29th it is defeated as training data
Enter network (576 groups total), it is practical to use monthly 15 days data as test data (144 groups total).
Step 3, building GRNN structure (generalized regression nerve networks),
Referring to Fig. 2, GRNN is made of input layer, mode layer, summation layer and output layer, and input layer directly passes input variable
Mode layer is passed, each neuron of mode layer corresponds to different types of sample input, transmission function are as follows:
In formula (5), X is network inputs variable, XiFor the corresponding learning sample of i-th of neuron, σ is smoothing factor, is
The accuracy of prediction is improved, value will optimize calculating by genetic algorithm.
The summation that counts, the connection of mode layer and each neuron are carried out to the output of each mode layer neuron in summation layer
Weight is 1, read group total formula are as follows:
Neural transferring function are as follows:
Neuron number is equal to the dimension of output vector in learning sample in output layer, and value is 1；
The output for layer of summing is divided by by each neuron, and the output of neuron j corresponds to j-th of element of estimated result Y (X),
That is:
Step 4 determines the optimal smoothing factor by GA,
The essence of network training is exactly to optimize smoothing factor σ it can be seen from step 3, and smoothing factor σ is to neural network forecast essence
Degree has significant impact.Conventional test methods enable smoothing factor σ be incremented by a certain section, count the error sequence of corresponding sample results
Column determine optimal parameter with the minimum standard of the root-mean-square error of error sequence, but this method calculates complicated, setting interval
Determination it is more difficult.
Referring to Fig. 3, this step regards GRNN as a prediction to make up the select permeability of smoothing factor in GRNN modeling
Function optimizes smoothing factor σ using the global optimizing ability of genetic algorithm, and detailed process is,
4.1) genetic algorithm basic parameter is set,
Smoothing factor σ is set as chromosomal gene, value range is [0.05,1], and operational precision is set as 0.0001, into
And bit length=ceil (log2(/ 0.0001 (1-0.05)))=14；
In value range generate 50 random numbers as first generation population, in general, repeatedly evolution can obtain it is more excellent
Solution sets the number of iterations N=50；Crossover probability pc=0.4；Mutation probability pm=0.2；
4.2) building fitness function assesses individual superiority and inferiority,
Fitness (fitness) is measured each smoothing factor individual in population and is likely to be breached or close to the excellent of optimal solution
One kind of good degree is estimated, and fitness function (fitness function) is used to assess the superiority and inferiority of new individual, is tested by intersecting
Card improves neural network accuracy.
Training sample is divided into subset one and subset two, training sample is used as using subset one, with the predicted value of subset two
With actual value mean error function as fitness function, fitness function expression formula is as follows:
In formula (9), E (σ) is the error mean of current smooth factor test sample,For the predicted value of test sample,
y(xi) be test sample reality output, n is population number, and embodiment chooses n=50,
50 individual fitness values are calculated according to formula (9), ascending sort obtains optimum individual, and sample is record at this time
Population P1；
4.3) genetic manipulation is carried out,
Genetic manipulation uses ratio back-and-forth method, the individual composition new population that fitness is good is selected from population P1, to population
Individual adaptation degree is cumulative in P1 obtains the sum of fitness ∑ fitness, is established and [0, ∑ fitness] according to ideal adaptation angle value
On region corresponding relationship, a random number is generated in [0, ∑ fitness] range, corresponding of the region where the random number
Body is selected；Ideal adaptation angle value is bigger, and the probability selected is bigger；
Crossover operation make through selecting be inherently high fitness two individuals, by have complementary advantages, obtain
More outstanding individual；This step generates new individual using the intersection that counts, by linear combination, takes crossover probability pc=0.4；
Variation increases the randomness of smoothing factor individual variation, and increases diversity of individuals, expands search model
It encloses, negates operation completion mutation operation by carrying out to selected variable position.
4.4) genetic iteration is carried out,
After carrying out hereditary selection, crossover operation, mutation operation to population P1, the fitness value of new individual is calculated again, is protected
The individual that fitness value is big is stayed, progeny population P2 is obtained, P2 is used as parent population repeat step 4.3) operation again, constantly followed
Ring iterative, and determine whether genetic iteration number reaches 50, if so, the optimum individual of last time calculated result is exported conduct
Smoothing factor σ after optimization；
Step 5, training GA-GRNN,
Using the smoothing factor σ after optimizing in step 4 as the smoothing factor of GRNN, the GRNN knot constructed in step 3 is substituted into
In structure, network transfer function is determined；By 1 day of in every month, 2015 in step 2,8 days, 22 days, daily 8:00-18:00 on the 29th
Relative humidity, surface temperature and the normal direction direct solar radiation intensity data of total 576 groups of integral point are input in GA-GRNN, with every
The corresponding photo-thermal efficiency of a integral point is trained for aim parameter, obtains trained GA-GRNN；
Step 6, test GA-GRNN,
Using 2015 in step 2 monthly 8:00-18:00 on the 15th amount to 144 groups of data as test data and be input to step 5
It is tested in trained GA-GRNN.After tested, predictablity rate of the GA-GRNN to solar thermal collection system photo-thermal efficiency
Higher, prediction numerical value can provide guidance for general production process.
Step 7 carries out photo-thermal EFFICIENCY PREDICTION with the trained GA-GRNN of step 6,
Relative humidity, surface temperature and the normal direction direct solar radiation intensity for inputting the same day, obtain current solar thermal collection system
Photo-thermal EFFICIENCY PREDICTION result.
Claims (5)
1. a kind of solar thermal collection system light thermal efficiency forecast method based on GA-GRNN, which is characterized in that according to the following steps
Implement:
Step 1, the parameter for determining solar thermal collection system,
Step 2, collection, the training test data for distributing network,
Step 3, building GRNN structure,
Step 4 determines optimal smoothing factor sigma by GA, regards GRNN as an anticipation function, is sought using the overall situation of genetic algorithm
Excellent ability optimizes smoothing factor σ；
Step 5, training GA-GRNN,
Using the smoothing factor σ after optimizing in step 4 as the smoothing factor of GRNN, substitute into the GRNN structure constructed in step 3,
Determine network transfer function；It is trained using the corresponding photo-thermal efficiency of each integral point as aim parameter, obtains trained GA-
GRNN；
Step 6, test GA-GRNN,
The test data selected in step 2 is input in the trained GA-GRNN of step 5 and is tested；
Step 7 carries out photo-thermal EFFICIENCY PREDICTION with the trained GA-GRNN of step 6,
Relative humidity, surface temperature and the normal direction direct solar radiation intensity for inputting the same day, obtain the light of current solar thermal collection system
Thermal efficiency prediction result.
2. the solar thermal collection system light thermal efficiency forecast method according to claim 1 based on GA-GRNN, feature exist
In: in the step 1, detailed process is,
1.1) solar radiation illumination is calculated,
The radiation that reaching ground can be utilized by solar thermal collection system mainly includes never changing direction directly from the sun
And the direct sunlight I that receivesd·θ, by atmosphere scattering influenced and the sun that is changing direction and receiving scatter Is·θ、
And through ground return and the illumination by reflected light I that receivesr·θ, then total irradiation intensity I of the sunθIt indicates are as follows: Iθ=Is·θ+Id·θ+
Ir·θ, unit is W/m2； (1)
1.2) efficiency of light-gathering heat collection pipe 7 is calculated,
Following light-gathering heat collection pipe 7 selects single tube CPC lumen type heat collector,
If specific heat capacity is cpConduction oil light-gathering heat collection pipe 7 is flowed through with mass flow m, the inlet temperature of heat-transfer working medium conduction oil is
Ti, outlet temperature To, then during the useful work Q that does of light-gathering heat collection pipe 7uExpression formula are as follows:
Qu=mcp(To-Ti), (2)
If the gross area of plane mirror field 5 is D, unit m2, total irradiation intensity of the sun is Iθ, unit W/m2, optically focused collection
The efficiency expression formula of heat pipe 7 are as follows:
1.3) the photo-thermal efficiency of linear Fresnel formula solar thermal collection system is calculated,
Shadow of the total photo-thermal efficiency of linear Fresnel formula solar thermal collection system by 7 thermal efficiency of reflector efficiency and light-gathering heat collection pipe
It rings, expression formula is as follows:
η photo-thermal efficiency=η reflecting mirror η thermal-collecting tube, (4)
η in formulaReflecting mirrorIt is related with the structure setting of plane mirror field 5 and Fresnel mirror 6, installation site and direction, it needs to carry out
On-site measurement.
3. the solar thermal collection system light thermal efficiency forecast method according to claim 2 based on GA-GRNN, feature exist
In: in the step 2, the higher relative humidity of the factor selection degree of correlation, surface temperature and the normal direction for influencing photo-thermal efficiency are straight
Three Xiang Zhibiao of radiation intensity is penetrated, as neural network input quantity.
4. the solar thermal collection system light thermal efficiency forecast method according to claim 3 based on GA-GRNN, feature exist
In: in the step 3, GRNN is made of input layer, mode layer, summation layer and output layer, and input layer is directly by input variable
Mode layer is passed to, each neuron of mode layer corresponds to different types of sample input, transmission function are as follows:
In formula (5), X is network inputs variable, XiFor the corresponding learning sample of i-th of neuron, σ is smoothing factor,
The summation that counts, the connection weight of mode layer and each neuron are carried out to the output of each mode layer neuron in summation layer
It is 1, read group total formula are as follows:
Neural transferring function are as follows:
Neuron number is equal to the dimension of output vector in learning sample in output layer, and value is 1；
The output for layer of summing is divided by by each neuron, and the output of neuron j corresponds to j-th of element of estimated result Y (X), it may be assumed that
5. the solar thermal collection system light thermal efficiency forecast method according to claim 4 based on GA-GRNN, feature exist
In: in the step 4, regard GRNN as an anticipation function, using the global optimizing ability of genetic algorithm to smoothing factor σ
It optimizing, detailed process is,
4.1) genetic algorithm basic parameter is set,
Smoothing factor σ is set as chromosomal gene, value range is [0.05,1], and operational precision is set as 0.0001, Jin Erwei
Length=ceil (log2(/ 0.0001 (1-0.05)))=14；
50 random numbers are generated in value range as first generation population, in general, more excellent solution can be obtained by repeatedly evolving, if
Determine the number of iterations N=50；Crossover probability pc=0.4；Mutation probability pm=0.2；
4.2) building fitness function assesses individual superiority and inferiority,
Fitness fitness be measure population in each smoothing factor individual be likely to be breached or close to optimal solution excellent degree
One kind estimate, fitness function fitness function is used to assess the superiority and inferiority of new individual, improves net by cross validation
Network precision,
Training sample is divided into subset one and subset two, training sample is used as using subset one, with the predicted value and reality of subset two
For actual value mean error function as fitness function, fitness function expression formula is as follows:
In formula (9), E (σ) is the error mean of current smooth factor test sample,For the predicted value of test sample, y (xi)
For the reality output of test sample, n is population number, and embodiment chooses n=50,
50 individual fitness values are calculated according to formula (9), ascending sort obtains optimum individual, and sample is population to record at this time
P1；
4.3) genetic manipulation is carried out,
Genetic manipulation uses ratio back-and-forth method, the individual composition new population that fitness is good is selected from population P1, in population P1
Individual adaptation degree is cumulative to obtain the sum of fitness ∑ fitness, according on the foundation of ideal adaptation angle value and [0, ∑ fitness]
Region corresponding relationship generates a random number in [0, ∑ fitness] range, the corresponding individual quilt in the region where the random number
Selection；
Crossover operation make through selecting be inherently high fitness two individuals, by have complementary advantages, obtain more
Outstanding individual；This step generates new individual using the intersection that counts, by linear combination, takes crossover probability pc=0.4；
Variation increases the randomness of smoothing factor individual variation, negates operation completion change by carrying out to selected variable position
ETTHER-OR operation；
4.4) genetic iteration is carried out,
After carrying out hereditary selection, crossover operation, mutation operation to population P1, the fitness value of new individual is calculated again, is retained suitable
The individual that angle value is big is answered, progeny population P2 is obtained, P2 is used as again parent population repeat step 4.3) operation, constantly circulation changes
Generation, and determine whether genetic iteration number reaches 50, if so, by the optimum individual output of last time calculated result as optimization
Smoothing factor σ afterwards.
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 true CN109539596A (en) | 2019-03-29 |
| CN109539596B 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) |
Cited By (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 |
| CN111625956A (en) * | 2020-05-25 | 2020-09-04 | 浙江大学 | Tower type solar mirror field optimization method based on self-adaptive differential evolution algorithm |
| CN111815051A (en) * | 2020-07-06 | 2020-10-23 | 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 |
|---|
| 孔国利等: "遗传算法的广义回归神经网络建模方法", 《计算机工程与设计》 * |
Cited By (5)
| 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 |
| CN111625956A (en) * | 2020-05-25 | 2020-09-04 | 浙江大学 | Tower type solar mirror field optimization method based on self-adaptive differential evolution algorithm |
| 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 |
| CN111815051A (en) * | 2020-07-06 | 2020-10-23 | 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 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109539596B (en) | 2020-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tian et al. | Thermo-economic optimization of a hybrid solar district heating plant with flat plate collectors and parabolic trough collectors in series | |
| Yao et al. | Modeling and simulation of the pioneer 1 MW solar thermal central receiver system in China | |
| CN109539596A (en) | Solar thermal collection system light thermal efficiency forecast method based on GA-GRNN | |
| Qiu et al. | A comprehensive model for analysis of real-time optical performance of a solar power tower with a multi-tube cavity receiver | |
| Gupta et al. | Thermodynamic performance evaluation of solar and other thermal power generation systems: A review | |
| Ghazouani et al. | Thermo-economic and exergy analysis and optimization of small PTC collectors for solar heat integration in industrial processes | |
| Afzal et al. | Optimizing the thermal performance of solar energy devices using meta-heuristic algorithms: A critical review | |
| Yağli | Examining the receiver heat loss, parametric optimization and exergy analysis of a solar power tower (SPT) system | |
| Elbeh et al. | Analysis and optimization of concentrated solar power plant for application in arid climate | |
| Liu et al. | Analysis of the Complementary Property of Solar Energy and Thermal Power Based on Coupling Model. | |
| Yan et al. | Dynamic modeling and simulation of a solar direct steam‐generating system | |
| Abid et al. | Energetic and exegetic analysis of a novel multi-generation system using solar power tower | |
| Bhattacharjee et al. | Viability of a concentrated solar power system in a low sun belt prefecture | |
| Ashour et al. | Developing and improving a prototype scale concentrating solar power tower-system | |
| Zaaraoui et al. | Technical and economical performance of parabolic trough collector power plant under Algerian climate | |
| Vakili et al. | A review of recent developments in the application of machine learning in solar thermal collector modelling | |
| Dehghanimadvar et al. | Mapping the development of various solar thermal technologies with hype cycle analysis | |
| Verma et al. | Ground and solar assisted heat pump systems for space heating and cooling applications in the northern region of India–A study on energy and CO2 saving potential | |
| Kong et al. | Operation optimization of a solar collector integrated with phase change material storage heating system | |
| Buonomano et al. | Solar-assisted district heating networks: Development and experimental validation of a novel simulation tool for the energy optimization | |
| Korzenszky | Linear model of DHW system using response surface method approach | |
| Kee et al. | Small-scale concentrated solar power system with thermal energy storage: system-level modelling and techno-economic optimisation | |
| Wang et al. | Rolling optimization based on holism for the operation strategy of solar tower power plant | |
| Xu et al. | The Badaling 1MW parabolic trough solar thermal power pilot plant | |
| Zhang et al. | Techno‐economic and sensitivity analysis of a hybrid concentrated photovoltaic/thermal system and an organic Rankine cycle to supply energy to sports stadiums |
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 |