CN106372745A - Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network - Google Patents
Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network Download PDFInfo
- Publication number
- CN106372745A CN106372745A CN201610739595.1A CN201610739595A CN106372745A CN 106372745 A CN106372745 A CN 106372745A CN 201610739595 A CN201610739595 A CN 201610739595A CN 106372745 A CN106372745 A CN 106372745A
- Authority
- CN
- China
- Prior art keywords
- power generation
- fuzzy
- ebf
- photovoltaic power
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The invention discloses a photovoltaic generating capacity prediction method based on a fuzzy EBF (Elliptical Basis Function) network. The photovoltaic generating capacity prediction method comprises the following steps of: selecting the influence factor of a photovoltaic generating capacity, collecting the historical data of the influence factor of the photovoltaic generating capacity and photovoltaic generating capacity historical data corresponding to the historical data of the influence factor of the photovoltaic generating capacity, and determining a sample set; generating the historical data of the influence factor of the photovoltaic generating capacity in a training sample set into an input vector, taking the photovoltaic generating capacity historical data corresponding to the historical data of the influence factor of the photovoltaic generating capacity as an output vector, carrying out normalization processing, and determining a training sample; utilizing the training sample to train the fuzzy EBF network through a Levenberg-Marquardt algorithm, collecting the influence factor data of the photovoltaic generating capacity of a day to be predicted, generating a prediction input vector, carrying out the normalization processing, inputting the prediction input vector into the trained fuzzy EBF network to obtain a photovoltaic generating capacity prediction output vector, and carrying out reverse normalization processing on the prediction output vector to obtain a prediction photovoltaic generating capacity vector of the day to be predicted. By use of the photovoltaic generating capacity prediction method, the prediction accuracy of the photovoltaic generating capacity is improved.
Description
Technical field
The present invention relates to a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network.
Background technology
Develop rapidly with social, a large amount of consumption of traditional energy make people face in industrial development and daily life
The problems such as exhaust with serious environmental pollution with regard to non-renewable energy resources.Solar energy as a kind of regenerative resource, due to clean,
The features such as environmental protection, safety, low cost, it has also become the mankind use the important component part of the energy, and are constantly developed.
It is a kind of effective means of utilization solar energy that large-scale photovoltaic generates electricity, but solar radiation, weather temperature, weather pattern
Easily impact is produced on photovoltaic generation with factors such as photovoltaic electroplax conversion ratios, and in non-linear.Therefore, solar energy is sent out
The prediction of electricity is had to greatest extent important to reasonable arrangement electric appliance usage time and using solar energy resources, reduction electric cost
Meaning.
Abroad, the theCourse of PV Industry impetus is swift and violent, and market at home, because engineering and manufacturing technology, cost are too high
Reason solar energy industry slower development.In recent years, under relevant policies are constantly helped and government supports to encourage energetically, solar energy is made
For the emerging infinite regeneration energy, its prospect is very wide.
During solar panel generates electricity, with the change of intensity of solar radiation and atmospheric temperature, its conversion ratio
All change continuous with output.Therefore, the temperature of photovoltaic electroplax also becomes impact solar photovoltaic generation system
A kind of key factor of generated energy.
Fuzzy ebf neural network forecast can be not only used for system modelling, identification and control, and can be also used for fuzzy rule
Automatically generate and extract, but traditional algorithm parameter initialization selects too extensive, is easily caused local minimum, fixing
The width of every rule input variable easily paralysis in learning process and leads to learning time long.
Content of the invention
The present invention in order to solve the above problems it is proposed that a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network,
Present method solves algorithm holds is absorbed in local extremum, parameter initialization selects too extensive, the low problem of solving precision.
To achieve these goals, the present invention adopts the following technical scheme that
A kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network, comprises the following steps:
(1) choose photovoltaic power generation quantity influence factor, collection photovoltaic power generation quantity influence factor's historical data and its corresponding to
Photovoltaic power generation quantity historical data, determine sample set;
(2) historical data concentrating photovoltaic power generation quantity influence factor with training sample generates input vector, with corresponding to it
The historical data of photovoltaic power generation quantity be output vector, and be normalized, determine training sample;
(3) utilize training sample, ebf network is obscured by levenberg-marquardt Algorithm for Training;
(4) collection day to be predicted photovoltaic power generation quantity influence factor data genaration prediction input vector, is normalized
Input the fuzzy ebf network training afterwards, obtain photovoltaic power generation quantity prediction output vector, prediction output vector through renormalization at
Reason obtains the prediction photovoltaic power generation quantity vector of day to be predicted.
In described step (1), the influence factor of photovoltaic power generation quantity includes: actual total in prediction the previous day day setting time
Generated energy, the average intensity of solar radiation of day to be predicted, the average weather temperature of day to be predicted, day to be predicted solar photovoltaic
The disturbance degree of plate;In described step (1), its corresponding photovoltaic power generation quantity is that actual total in day to be predicted setting time generates electricity
Amount.
In described step (2), concrete steps include:
(2-1) construct input vector using the historical data of gained photovoltaic power generation quantity influence factor respectively by season, with it
The historical data construction output vector of corresponding photovoltaic power generation quantity;
(2-2) confirm and store input vector, the minima of each component and maximum in output vector, to input vector and
Output vector is normalized, and obtains normalization input vector and normalized output vector.
In described step (3), fuzzy ebf network structure includes input layer, obfuscation layer, fuzzy reasoning layer and output layer;
Described input layer number is 4, includes the gross actual power generation, to be predicted of setting period the previous day photovoltaic generation day to be predicted
Day average intensity of solar radiation, the average weather temperature of day to be predicted, the disturbance degree of day to be predicted photovoltaic electroplax;Institute
State output layer nodes and be 1, set the gross generation of period including day to be predicted;Described obfuscation node layer number is 12,
It is divided into 3 groups, every group node number is 4, and its membership function is determined by below equation:
In formula, f1(i, j) represents i-th input variable x of input layeriIt is subordinate to the degree of membership of j-th fuzzy subset, cijAnd bij
Represent average in obfuscation layer jth group for i-th input variable of input layer and variance, c respectivelyijInitial span preferred
For [- 1,1], bijInitial span be preferably [- 2,1], exp () is the exponential function with natural constant e as the truth of a matter.
In described step (3), concrete steps include:
(3-1) error signal of the number according to training sample and each sample corresponding, determines fuzzy ebf neutral net
Performance indications;
(3-2) it is based on improvement levenberg-marquardt algorithm weights are revised;
(3-3) i-th input variable of iterative input layer and the membership function of obfuscation j-th fuzzy set of layer is equal
It is known that meeting iterated conditional, fuzzy ebf network training terminates value.
In described step (3-2), improving levenberg-marquardt weights revisal formulas is:
W (n+1)=w (n)-[jt(n)j(n)+cij(n)i]-1jt(n)en(11)
In formula, w (n) is the weight matrix of fuzzy ebf network nth iteration, and i is unit matrix, and j is Jacobi
(jacobian) matrix, is determined by below equation:
In formula, enFor the error of n-th sample, wpP-th weights for fuzzy ebf network model;
jtFor the transposition of matrix j, enVectorial for the output error of n training sample, particularly as follows: en=[e1e2...en].
In described step (3-3), calculateIf the norm of j is less than value a of a setting0When, assert it
Meet iterated conditional.
In described step (3-3), calculate cij(n+1), its formula is as follows:
cij(n+1)=cij(n)/β(n+1) (13)
In formula, cijN () is the person in servitude of i-th input variable of input layer and obfuscation j-th fuzzy set of layer during nth iteration
The average of membership fuction, β (n+1) is the dynamic convergence factor, and computing formula is:
In formula, n is number of training, emaxFor maximum training error, eminFor fuzzy ebf network desired output essence
Degree, k is constant.
In described step (4), collection day to be predicted photovoltaic power generation quantity influence factor data genaration prediction input vector, carry out
Normalized, obtains the prediction input vector after normalized, normalization formula is:
Wherein, x* iFor i-th component in the prediction input vector before normalized,Defeated for predicting after normalized
I-th component in incoming vector, xi,min, xi,maxIt is originally inputted the minima of i-th component in vector before being respectively normalized
And maximum.
In described step (4), at the photovoltaic power generation quantity that fuzzy ebf network obtains predicts output vector through renormalization
Reason obtains the prediction photovoltaic power generation quantity vector of day to be predicted, and the formula that renormalization is processed is:
It is in the photovoltaic power generation quantity prediction output vector of the renormalization before processing that fuzzy ebf neural network forecast obtains
I-th component, y* iI-th component in prediction photovoltaic power generation quantity vector after processing for renormalization, yi,min、yi,maxIt is respectively
The minima of the i-th component and maximum in original output vector before normalized.
The invention has the benefit that
The present invention passes through solar radiation intensity, and weather temperature and photovoltaic electroplax conversion ratio are it is achieved that photovoltaic is sent out
The prediction of electricity, is easily absorbed in the defect of local extremum for fuzzy ebf neutral net, using improvement levenberg-
Marquardt Algorithm for Training obscures ebf network, preferably solves algorithm and is absorbed in local minimum points, parameter initialization selects too
Cross extensively, the width of fixing every rule input variable easily paralysis in learning process and leads to learning time
Long problem, improves the precision of prediction of photovoltaic power generation quantity.
Brief description
Fig. 1 is fuzzy ebf network architecture figure.
Specific embodiment:
The invention will be further described with embodiment below in conjunction with the accompanying drawings.
As shown in figure 1, a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network it is characterised in that: include following
Step:
(1a) choose photovoltaic power generation quantity influence factor, collection photovoltaic power generation quantity influence factor's historical data and its corresponding to
Photovoltaic power generation quantity historical data, according to season division be the spring, the summer, the autumn, four groups of training sample sets of winter;
(1b) historical data according to step (1a) gained photovoltaic power generation quantity influence factor by season respectively generate input to
Amount, using the historical data of its corresponding photovoltaic power generation quantity as output vector, and is normalized, and obtains training sample
This;
(1c) utilize step (1b) gained training sample, obscured using improving levenberg-marquardt Algorithm for Training
Ebf network, the fuzzy ebf network after being trained;
(1d) gather day photovoltaic power generation quantity influence factor data genaration prediction input vector to be predicted, be normalized place
Reason, obtains the prediction input vector after normalized;
(1e) the prediction input vector after described for step (1d) normalized is pressed the corresponding fuzzy ebf net of season input
Network, obtains photovoltaic power generation quantity prediction output vector, and prediction output vector processes the pre- light-metering obtaining day to be predicted through renormalization
Volt generated energy vector.
Aforesaid one kind, is characterized in that: in described step (1a), the influence factor of photovoltaic power generation quantity includes: to be predicted a few days ago
The gross actual power generation of one day 6:00-19:00, the average intensity of solar radiation of day to be predicted, the average sky temperature of day to be predicted
Degree, the disturbance degree of day to be predicted photovoltaic electroplax;In described step (1a), its corresponding photovoltaic power generation quantity is to be predicted
The gross actual power generation of day 6:00-19:00.
Aforesaid one kind, is characterized in that: the average intensity of solar radiation of described day to be predictedFor:
In formula,For predicting the average intensity of solar radiation of day;giIntensity of solar radiation for the i-th moment;
The average weather temperature of described day to be predictedFor:
In formula,For predicting the average weather temperature of day;tnWeather temperature for the i-th moment;
Described day to be predicted photovoltaic electroplax disturbance degreeFor:
In formula, a1Be more than 0 and less than 0.8 natural constant, a2Be more than 0.4 and less than 1 natural constant, rf
Basic conversion ratio for photovoltaic electroplax.
Aforesaid one kind, is characterized in that: described step (1b) specifically includes step:
(4a) construct input vector using the historical data of gained photovoltaic power generation quantity influence factor respectively by season, with its institute
The historical data construction output vector of corresponding photovoltaic power generation quantity;
(4b) step (4a) gained input vector and output vector are normalized, obtain normalization input vector
With normalized output vector, the formula of wherein normalized is:
Wherein, niFor input layer number, l is output layer nodes, xi, yiIt is respectively photovoltaic generation before normalized
Amount is originally inputted the i-th component in the original output vector of vector sum, xi,min, xi,maxIt is respectively photovoltaic power generation quantity before normalized
It is originally inputted the minima of the i-th component and maximum, y in vectori,min, yi,maxIt is respectively photovoltaic power generation quantity before normalized
The minima of the i-th component and maximum in original output vector,Be respectively normalized after be originally inputted vector
With the i-th component in original output vector;
(4c) before preserving normalized, photovoltaic power generation quantity is originally inputted minima x of each component in vectori,minAnd maximum
Value xi,maxMinima y with component each in original output vectori,minWith maximum yi,max.
Aforesaid one kind, is characterized in that: the fuzzy ebf network structure adopting in described step (1c) includes input layer, mould
Gelatinizing layer, fuzzy reasoning layer and output layer;Described input layer number is 4, including the previous day photovoltaic generation day to be predicted 6:
The gross actual power generation of 00-19:00 period, the average intensity of solar radiation of day to be predicted, the average weather temperature of day to be predicted,
Day to be predicted photovoltaic electroplax disturbance degree;Described output layer nodes are 1, including day 06:00-19:00 to be predicted
The gross generation of period;Described obfuscation node layer number be 12, be divided into 3 groups, every group node number be 4, its membership function by
Below equation determines:
In formula, f1(i, j) represents i-th input variable x of input layeriIt is subordinate to the degree of membership of j-th fuzzy subset, cijAnd bij
Represent average in obfuscation layer jth group for i-th input variable of input layer and variance, c respectivelyijInitial span preferred
For [- 1,1], bijInitial span be preferably [- 2,1].Exp () is the exponential function with natural constant e as the truth of a matter.
The nodes of described fuzzy reasoning layer are 3, and output is determined by below equation:
In formula, f2J () represents the fitness to j-th strip fuzzy rule for the input layer input variable.
Special instruction, the t norm of this layer of j-th strip fuzzy rule to be represented by below equation:
φj=exp [- md2(j)] (j=1,2,3) (8)
In formula, md (j) is the mahalanobis distance of j-th strip rule, is determined by below equation:
In formula, x=(x1,…,xr)t, c=(c1j,…,crj)t,
Can be obtained from above, the reception domain of this model is the suprasphere in super ellipsoids body rather than rbf unit.
The output of described output layer is determined by below equation:
In formula, l is the number of output node layer, and w (l, j) is l node of output layer and fuzzy reasoning j-th node of layer
Connection weight matrix.
Aforesaid one kind, is characterized in that: being instructed using improvement levenberg-marquardt algorithm in described step (1c)
Practice the training of fuzzy ebf network, the fuzzy ebf network after being trained, concretely comprise the following steps:
(6a) performance indications setting fuzzy ebf neutral net are:
In formula, n is the number of training sample, enIt is the error signal of corresponding n-th sample, solution formula is:
en=dn-yn(12)
In formula, dnFor the desired output of corresponding n-th sample of output layer, ynDefeated for the reality of corresponding n-th sample of output layer
Go out.
(6b) improving levenberg-marquardt weights revisal formulas is:
W (n+1)=w (n)-[jt(n)j(n)+cij(n)i]-1jt(n)en(13)
In formula, w (n) is the weight matrix of fuzzy ebf network nth iteration, and i is unit matrix, and j is Jacobi
(jacobian) matrix, is determined by below equation:
In formula, enFor the error of n-th sample, wmP-th weights for fuzzy ebf network model.
jtFor the transposition of matrix j, enVectorial for the output error of n training sample, particularly as follows: en=[e1e2...en].
(6c) calculateIf the norm of j is less than value a of a setting0When, turn (6e).
(6d) calculate cij(n+1), its formula is as follows:
cij(n+1)=cij(n)/β(n+1) (15)
In formula, cijN () is the person in servitude of i-th input variable of input layer and obfuscation j-th fuzzy set of layer during nth iteration
The average of membership fuction, β (n+1) is the dynamic convergence factor, and computing formula is:
In formula, n is number of training, emaxFor maximum training error, eminFor fuzzy ebf network desired output essence
Degree, k is the constant more than 4 less than 9.
It is then back to step (6b).
(6e) obscure ebf network training to terminate.
Aforesaid one kind, is characterized in that: described step (1d) gathers day photovoltaic power generation quantity influence factor data life to be predicted
Become prediction input vector, be normalized, obtain the prediction input vector after normalized, normalization formula is:
Wherein, x* iFor i-th component in the prediction input vector before normalized,Defeated for predicting after normalized
I-th component in incoming vector, xi,min, xi,maxIt is originally inputted the minima of i-th component in vector before being respectively normalized
And maximum.
Aforesaid one kind, is characterized in that: the photovoltaic power generation quantity prediction obtaining through fuzzy ebf network in described step (1e)
Output vector processes the prediction photovoltaic power generation quantity vector obtaining day to be predicted through renormalization, and the formula that renormalization is processed is:
It is in the photovoltaic power generation quantity prediction output vector of the renormalization before processing that fuzzy ebf neural network forecast obtains
I-th component, y* iI-th component in prediction photovoltaic power generation quantity vector after processing for renormalization, yi,min、yi,maxIt is respectively
The minima of the i-th component and maximum in original output vector before normalized.
Although the above-mentioned accompanying drawing that combines is described to the specific embodiment of the present invention, not model is protected to the present invention
The restriction enclosed, one of ordinary skill in the art should be understood that on the basis of technical scheme, and those skilled in the art are not
Need to pay the various modifications that creative work can make or deformation still within protection scope of the present invention.
Claims (10)
1. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network, is characterized in that: comprise the following steps:
(1) influence factor of photovoltaic power generation quantity, collection photovoltaic power generation quantity influence factor's historical data and its corresponding light are chosen
Volt generated energy historical data, determines sample set;
(2) historical data concentrating photovoltaic power generation quantity influence factor with training sample generates input vector, with its corresponding light
The historical data of volt generated energy is output vector, and is normalized, and determines training sample;
(3) utilize training sample, ebf network is obscured by levenberg-marquardt Algorithm for Training;
(4) collection day to be predicted photovoltaic power generation quantity influence factor data genaration prediction input vector, defeated after being normalized
Enter the fuzzy ebf network training, obtain photovoltaic power generation quantity prediction output vector, prediction output vector is processed through renormalization
Prediction photovoltaic power generation quantity vector to day to be predicted.
2. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (1), the influence factor of photovoltaic power generation quantity includes: prediction the previous day day setting time in gross actual power generation, treat pre-
Survey the average intensity of solar radiation of day, the average weather temperature of day to be predicted, the disturbance degree of day to be predicted photovoltaic electroplax;
In described step (1), its corresponding photovoltaic power generation quantity is the gross actual power generation in day to be predicted setting time.
3. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (2), concrete steps include:
(2-1) construct input vector using the historical data of gained photovoltaic power generation quantity influence factor respectively by season, right with its institute
The historical data construction output vector of the photovoltaic power generation quantity answered;
(2-2) input vector, the minima of each component and maximum in output vector are confirmed and store, to input vector and output
Vector is normalized, and obtains normalization input vector and normalized output vector.
4. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (3), fuzzy ebf network structure includes input layer, obfuscation layer, fuzzy reasoning layer and output layer;Described input layer section
Count as 4, set gross actual power generation, the average sun of day to be predicted of period including the previous day photovoltaic generation day to be predicted
Radiant intensity, the average weather temperature of day to be predicted, the disturbance degree of day to be predicted photovoltaic electroplax;Described output node layer
Number is 1, sets the gross generation of period including day to be predicted;Described obfuscation node layer number is 12, is divided into 3 groups, every group
Nodes are 4, and its membership function is determined by below equation:
In formula, f1(i, j) represents i-th input variable x of input layeriIt is subordinate to the degree of membership of j-th fuzzy subset, cijAnd bijRespectively
Represent average in obfuscation layer jth group for i-th input variable of input layer and variance, cijInitial span be preferably [-
1,1],bijInitial span be preferably [- 2,1], exp () is the exponential function with natural constant e as the truth of a matter.
5. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (3), concrete steps include:
(3-1) error signal of the number according to training sample and each sample corresponding, determines the performance of fuzzy ebf neutral net
Index;
(3-2) it is based on improvement levenberg-marquardt algorithm weights are revised;
(3-3) average of the membership function of i-th input variable of iterative input layer and obfuscation j-th fuzzy set of layer,
Know and meet iterated conditional, fuzzy ebf network training terminates.
6. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 5, is characterized in that: described
In step (3-2), improving levenberg-marquardt weights revisal formulas is:
W (n+1)=w (n)-[jt(n)j(n)+cij(n)i]-1jt(n)en(11)
In formula, w (n) is the weight matrix of fuzzy ebf network nth iteration, and i is unit matrix, and j is Jacobi (jacobian)
Matrix, cijN () is the membership function of i-th input variable of input layer and obfuscation j-th fuzzy set of layer during nth iteration
Average, is determined by below equation:
In formula, enFor the error of n-th sample, wpP-th weights for fuzzy ebf network model;
jtFor the transposition of matrix j, enVectorial for the output error of n training sample, particularly as follows: en=[e1e2...en].
7. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 5, is characterized in that: described
In step (3-3), calculateIf the norm of j is less than value a of a setting0When, assert that it meets iterated conditional.
8. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 5, is characterized in that: described
In step (3-3), calculate cij(n+1), its formula is as follows:
cij(n+1)=cij(n)/β(n+1) (13)
In formula, cij(n) be during nth iteration i-th input variable of input layer and obfuscation j-th fuzzy set of layer be subordinate to letter
The average of number, β (n+1) is the dynamic convergence factor, and computing formula is:
In formula, n is number of training, emaxFor maximum training error, eminFor obscuring the desired output accuracy of ebf network, k is
Constant.
9. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (4), collection day to be predicted photovoltaic power generation quantity influence factor data genaration prediction input vector, it is normalized,
Obtain the prediction input vector after normalized, normalization formula is:
Wherein, x* iFor i-th component in the prediction input vector before normalized,For after normalized prediction input to
I-th component in amount, xi,min, xi,maxIt is originally inputted in vector the minima of i-th component and before being respectively normalized
Big value.
10. a kind of method for forecasting photovoltaic power generation quantity based on fuzzy ebf network as claimed in claim 1, is characterized in that: described
In step (4), the photovoltaic power generation quantity prediction output vector that obtains through fuzzy ebf network process through renormalization obtain to be predicted
The prediction photovoltaic power generation quantity vector of day, the formula that renormalization is processed is:
It is i-th in the photovoltaic power generation quantity prediction output vector of the renormalization before processing that fuzzy ebf neural network forecast obtains
Component, y* iI-th component in prediction photovoltaic power generation quantity vector after processing for renormalization, yi,min、yi,maxIt is respectively normalization
The minima of the i-th component and maximum in the original output vector of before processing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610739595.1A CN106372745A (en) | 2016-08-27 | 2016-08-27 | Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610739595.1A CN106372745A (en) | 2016-08-27 | 2016-08-27 | Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106372745A true CN106372745A (en) | 2017-02-01 |
Family
ID=57904259
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610739595.1A Pending CN106372745A (en) | 2016-08-27 | 2016-08-27 | Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106372745A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107958307A (en) * | 2017-11-28 | 2018-04-24 | 珠海格力电器股份有限公司 | Electricity charge Forecasting Methodology and device |
TWI623890B (en) * | 2017-07-04 | 2018-05-11 | 台灣電力股份有限公司 | System for predicating power generation by utilizing multiple neural networks and method thereof |
CN109284874A (en) * | 2018-10-26 | 2019-01-29 | 昆明电力交易中心有限责任公司 | Daily generation prediction technique, device, equipment and the storage medium of photovoltaic plant |
CN110675278A (en) * | 2019-09-18 | 2020-01-10 | 上海电机学院 | Photovoltaic power short-term prediction method based on RBF neural network |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218674A (en) * | 2013-04-07 | 2013-07-24 | 国家电网公司 | Method for predicating output power of photovoltaic power generation system based on BP (Back Propagation) neural network model |
CN104616085A (en) * | 2015-02-16 | 2015-05-13 | 河海大学常州校区 | Photovoltaic generating capacity predicting method based on BP neural network |
-
2016
- 2016-08-27 CN CN201610739595.1A patent/CN106372745A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103218674A (en) * | 2013-04-07 | 2013-07-24 | 国家电网公司 | Method for predicating output power of photovoltaic power generation system based on BP (Back Propagation) neural network model |
CN104616085A (en) * | 2015-02-16 | 2015-05-13 | 河海大学常州校区 | Photovoltaic generating capacity predicting method based on BP neural network |
Non-Patent Citations (2)
Title |
---|
张维杰: "基于分季建模的风电功率单一预测模型的建立", 《中国优秀硕士学位论文全文数据库》 * |
徐奉友等: "Levenberg-Marquardt算法在T-S型模糊RBF神经网络训练中的应用", 《计算机系统应用》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI623890B (en) * | 2017-07-04 | 2018-05-11 | 台灣電力股份有限公司 | System for predicating power generation by utilizing multiple neural networks and method thereof |
CN107958307A (en) * | 2017-11-28 | 2018-04-24 | 珠海格力电器股份有限公司 | Electricity charge Forecasting Methodology and device |
CN109284874A (en) * | 2018-10-26 | 2019-01-29 | 昆明电力交易中心有限责任公司 | Daily generation prediction technique, device, equipment and the storage medium of photovoltaic plant |
CN109284874B (en) * | 2018-10-26 | 2021-08-17 | 昆明电力交易中心有限责任公司 | Method, device and equipment for predicting daily generated energy of photovoltaic power station and storage medium |
CN110675278A (en) * | 2019-09-18 | 2020-01-10 | 上海电机学院 | Photovoltaic power short-term prediction method based on RBF neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhou et al. | Prediction of photovoltaic power output based on similar day analysis, genetic algorithm and extreme learning machine | |
Tang et al. | Entropy method combined with extreme learning machine method for the short-term photovoltaic power generation forecasting | |
CN104616085A (en) | Photovoltaic generating capacity predicting method based on BP neural network | |
CN103106544B (en) | A kind of photovoltaic generation prognoses system based on T-S Fuzzy neutral net | |
CN106372745A (en) | Photovoltaic generating capacity prediction method based on fuzzy EBF (Elliptical Basis Function) network | |
CN104050517A (en) | Photovoltaic power generation forecasting method based on GRNN | |
CN103400204A (en) | Forecasting method for solar photovoltaic electricity generation amount based on SVM (support vector machine) - Markov combination method | |
CN104463349A (en) | Photovoltaic generated power prediction method based on multi-period comprehensive similar days | |
CN103218674A (en) | Method for predicating output power of photovoltaic power generation system based on BP (Back Propagation) neural network model | |
CN103489038A (en) | Photovoltaic ultra-short-term power prediction method based on LM-BP neural network | |
CN103823504B (en) | A kind of maximum power tracking and controlling method based on least square method supporting vector machine | |
CN103218673A (en) | Method for predicating short-period output power of photovoltaic power generation based on BP (Back Propagation) neural network | |
CN105186499A (en) | Multi-target probabilistically optimal power flow fuzzy modelling and solving method for power distribution network | |
CN105678397A (en) | Short-term photovoltaic power prediction method based on improved EMD algorithm and Elman algorithm | |
CN106803135A (en) | The Forecasting Methodology and device of a kind of photovoltaic power generation system output power | |
CN109376951A (en) | A kind of photovoltaic probability forecasting method | |
CN104598765A (en) | Building energy consumption prediction method based on elastic adaptive neural network | |
Kumar et al. | Three years ahead solar irradiance forecasting to quantify degradation influenced energy potentials from thin film (a-Si) photovoltaic system | |
CN115238993A (en) | Reservoir group multi-objective optimization scheduling scheme decision system based on NSGA-II algorithm | |
CN103293950B (en) | The control method that a kind of maximum photovoltaic power point based on LSSVM is followed the tracks of | |
Kallio et al. | Photovoltaic power prediction for solar micro-grid optimal control | |
CN108964037A (en) | Based on the reconstitution model of high voltage distribution network | |
CN106202700A (en) | A kind of photovoltaic generation exert oneself prediction data analysing method | |
Xu et al. | Influence of different optimization aalgorithms on prediction accuracy of photovoltaic output power based on BP neural network | |
Huang et al. | Wind prediction based on improved bp artificial neural network in wind farm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170201 |
|
RJ01 | Rejection of invention patent application after publication |