CN107330251A - A kind of wind power prediction method based on Retrieval method - Google Patents
A kind of wind power prediction method based on Retrieval method Download PDFInfo
- Publication number
- CN107330251A CN107330251A CN201710434726.XA CN201710434726A CN107330251A CN 107330251 A CN107330251 A CN 107330251A CN 201710434726 A CN201710434726 A CN 201710434726A CN 107330251 A CN107330251 A CN 107330251A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- wind power
- layer
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
Abstract
The invention discloses a kind of short-term wind power prediction method based on Retrieval method.Due to the randomness and uncertainty of wind-resources, there is certain difficulty when carrying out wind power prediction according to the history wind power data of wind power plant and history meteorological data, especially during Mathematical Models.Retrieval method is applied to BP neural network by this method（Back propagation neural network, feedforward neural network）Forecast model is set up, using Retrieval method come the weights and threshold value of Optimizing BP Network parameter, accelerates the Optimization Learning speed of BP neural network, so as to accelerate the efficiency of wind power prediction processing；Experiment proves that the optimized algorithm is used in wind power plant wind power prediction and is favorably improved precision of prediction, accelerates network convergence speed.
Description
Technical field
The present invention relates to wind power plant wind power prediction field, more particularly to a kind of wind power prediction based on Retrieval method
Method.
Background technology
With going from bad to worse for ecological environment, the green energy resource such as wind-powered electricity generation is of increased attention.In recent years, China
Installed capacity of wind-driven power rapid development, the grid-connected problem that at the same time randomness of wind power, fluctuation are brought more is highlighted, and is needed
High-precision wind power prediction is carried out to the wind power plant in regional power grid, so as to which wind-powered electricity generation is included into dispatching of power netwoks in the works,
Improve digestion capability of the power network to wind-powered electricity generation.The continuous increase of current distribution wind power generator group installed capacity, although each wind
There is wind-resources assessment before electric field construction.But continuing to develop and progressive with society, stability of the people to power system
It is required that more and more higher, additionally, due to Current resource it is in short supply in the case of, the making full use of wind-power electricity generation of dynamics as maximum as possible and
Reduce conventional energy resource to generate electricity, reduce resource consumption and environmental pollution.
Due to the randomness and uncertainty of wind-resources, exist when carrying out power prediction according to wind power plant historical data certain
Difficulty, especially during Mathematical Models.Mathematical modeling typically uses certain statistical method, passes through a large amount of wind
Electric field historical data is trained to model, sets up input data and the mapping relations of wind power, so as to enter to wind power
Row prediction.The mathematical model establishing method of wind power prediction mainly has：The methods such as time series method, neutral net, fuzzy logic.
At present, BP neural network (back propagation neural network, feedforward neural network) at home and abroad some wind work(
Practical application has been obtained in rate forecasting software.However, there are some shortcomings and deficiencies in actual applications in BP neural network：1)BP
When neural network algorithm carries out modified weight, error function generally declines mode using gradient, is single-phase search, global optimizing energy
Power is not good enough；2) BP neural network parameter initialization is random, causes the possibility duplicated in study larger, causes network convergence
Speed is excessively slow, or even trains the state that paralyses.Retrieval method based on Clouds theory is applied to BP neural network by this patent
In algorithm, the global optimizing ability of Retrieval method can either be lifted, wind power prediction precision is improved, nerve net can be accelerated again
The global optimizing speed of network.
Find by prior art documents, a kind of short-term wind power prediction method (invention minimum based on variance
Patent：CN 201510768952.2), this application proposes to constitute combined prediction by Statistical Prediction Model and physical prediction model
Model, according to history wind power data and the corresponding relation of meteorological data, respectively with Statistical Prediction Model and physical prediction model
Wind power is predicted, the advantage of single model is combined, improves the precision of prediction.Although this method contributes to lifting prediction essence
Degree, but data volume used is big, it is more to be related to model, is readily incorporated accidental error, influences precision of prediction.
The content of the invention
It is an object of the invention to overcome deficiencies of the prior art, it is proposed that a kind of based on Retrieval method
Retrieval method is applied to BP neural network and sets up forecast model by wind power prediction method, institute's extracting method, is calculated using cloud heredity
Method carrys out the weights and threshold value of Optimizing BP Network parameter, accelerates the Optimization Learning speed of BP neural network, so that it is pre- to accelerate wind power
Survey the efficiency of processing.
A kind of short-term wind power prediction method based on Retrieval method, Retrieval method is applied to BP neural network by it
In algorithm, using the weights and threshold value of Retrieval method Optimizing BP Network parameter, accelerate the Optimization Learning speed of BP neural network,
So as to accelerate its power prediction treatment effeciency.
Further, it is comprised the following steps that：
Step 1：BP neural network structure is determined, selection sample data used sets up BP neural network structure.BP nerve nets
It is a kind of nonlinear relation between the input and output of network, if input number of nodes is N, output node number is M, then
Network is the mapping that theorem in Euclid space is tieed up from N-dimensional theorem in Euclid space to M.Three layers of BP networks of this paper we selected typicals are used as reference, difference
For input layer, hidden layer, output layer, the node for making the input layer, hidden layer and output layer of network is respectively m, p, q, inputs sample
This sum is n, if input layer xni, hidden layer node snj, output node layer ynt, input layer and hidden layer node
Network weight is wij, the network weight between hidden layer node and output node layer is vjt。；
Step 2：Network weight is initialized, the weights and threshold value of BP network parameters are initialized, several inputs are generated at random
Weight w between layer and hidden layerij, the weight w between hidden layer and output layerjk；
During forward-propagating, each node layer output：
snj=f (net) j=1,2,3...p
ynt=f (netnt) t=1,2 ..., p
In various above, it is all unipolarity function generally to choose excitation function
Step 3：Initialize the coding in population, i.e. genetic algorithm.N number of original string structured data is randomly generated, each string
Structured data is referred to as an individual.Individual constitutes a colony.Genetic algorithm is using this N number of original string structured data as first
Initial point starts iteration.
Step 4：Object function is determined, optimal search fitness letter is used as using the inverse of the error function of BP neural network
Number,
In formula, cnlFor the output node desired value of setting, ynlFor output node.When neutral net error reaches minimum value
When, it is optimal network weight that fitness function, which is obtained under maximum, this state,.；
Step 5：Individual using the population of initialization substitutes into as input quantity and calculates every in each population in fitness function
Individual fitness；
Step 6：If fitness meets the fitness value requirement or iterations requirement set by system, system is redirected
To step 10, next step is otherwise carried out；
Step 7：Fitness highest individual in population is preserved, for carrying out the operation such as intersection, variation in genetic algorithm；
Step 8：Progeny population is generated using the Y condition generators in Clouds theory, crossover operation is realized.
Normal cloud model is a followed normal distribution regularity of distribution, the random manifold with steady tendency, with desired value Ex, entropy
En, tri- numerical value of super entropy He are characterized.Desired value Ex：It is best able to represent the point of this qualitativing concept in number field space, reflects
The position of centre of gravity of cloud.Entropy En：On the one hand the scope that can be received in number field space by Linguistic Value is reflected；On the other hand also reflect
Point in number field space can represent the probability of this Linguistic Value, represent the randomness that the water dust of qualitativing concept occurs.It is disclosed
The relevance of ambiguity and randomness.Super entropy He：It is the uncertain measurement of entropy, reflects and represent the Linguistic Value in number field space
Uncertain coherency a little.
Crossover operation：
(1) random generation or artificial formulation degree of certainty μ；
(2)
(3) En=variables hunting zone/c1, c1 take 3*p, p to be Population Size；
(4) He=En/c2, c2 take the value between 5~15；
(5) progeny population is produced by Y condition generators；
In formula：xfAnd xmFather's individual and mother's individual respectively in crossover operation；FfAnd FmTheir adaptation is then corresponded to respectively
Degree.
Step 9：Mutation operation in gene mutation, i.e. genetic algorithm is realized using basic Normal Cloud cloud generator；
Mutation operation：
(1) Ex takes former individual；
(2) En=variables hunting zone/c3, c3 take 5；
(3) He=En/c4, c4 take the value between 5~15；
(4) basic Normal Cloud cloud generator is performed, random number Temp is generated, works as μ>It is individual in Population Regeneration during Temp.
Jump to step 5.
Step 10：Obtain best initial weights wij、wjk, BP neural network forward-propagating calculating is substituted into, each layer sample is calculated
Square-error between training output valve and desired value；
Step 11：Export weights；
Further, it is characterised in that
Basic Normal Cloud cloud generator：Three numerical characteristics { Ex, En, He } of given cloud and n water dust number, generation n
Desired value is En, and variance is He normal random number
En'=RANDN (En, He)
xi=RANDN (Ex, En')
N water dust (x is produced into by above formulai,μi), wherein, xiIt is the normal random number that desired value is Ex, μiIt is desired value
For Ex Normal Cloud expectation curve in xiThe degree of membership at place.
Y condition cloud generators：Three numerical characteristics { Ex, En, He } of given cloud and specific degree of certainty μ0, generation n
Desired value is En, and variance is He normal random number
En'=RANDN (En, He)
N water dust (x is produced by above formulai,μ0)。
Compared with prior art, the invention has the advantages that and technique effect：The present invention is proposed to be calculated based on cloud heredity
Retrieval method is applied to BP neural network and sets up forecast model by the wind power prediction method of method, this method, utilizes cloud heredity
Algorithm carrys out the weights and threshold value of Optimizing BP Network parameter, accelerates the Optimization Learning speed of BP neural network, so as to accelerate wind power
Predict the efficiency of processing.
Brief description of the drawings
Fig. 1 is the BP neural network flow chart based on Retrieval method.
Fig. 2 is three numerical characteristic schematic diagrames of normal cloud model.
Fig. 3 is the BP neural network wind power prediction result figure based on Retrieval method.
Embodiment
Below in conjunction with the accompanying drawings, the present invention is done and further described in detail, but the implementation of the present invention is not limited to this, needs
, it is noted that if following have not the especially process or parameter of detailed description, it is existing to be that those skilled in the art are referred to
Technology understand or realize.
Fig. 1 is the BP neural network flow chart based on Retrieval method, and it is comprised the following steps that：
The present invention proposes a kind of wind power prediction method based on Retrieval method, and institute's extracting method transports Retrieval method
Forecast model is set up for BP neural network, using Retrieval method come the weights and threshold value of Optimizing BP Network parameter, accelerates BP
The Optimization Learning speed of neutral net, so as to accelerate the efficiency of wind power prediction processing.
A kind of short-term wind power prediction method based on Retrieval method, it is characterised in that：Retrieval method is applied to
In BP neural network algorithm, using the weights and threshold value of Retrieval method Optimizing BP Network parameter, accelerate the excellent of BP neural network
Change pace of learning, so as to accelerate its power prediction treatment effeciency.
Further, it is comprised the following steps that：
Step 1：BP neural network structure is determined, selection sample data used sets up BP neural network structure.BP nerve nets
It is a kind of nonlinear relation between the input and output of network, if input number of nodes is N, output node number is M, then
Network is the mapping that theorem in Euclid space is tieed up from N-dimensional theorem in Euclid space to M.Three layers of BP networks of this paper we selected typicals are used as reference, difference
For input layer, hidden layer, output layer, the node for making the input layer, hidden layer and output layer of network is respectively m, p, q, inputs sample
This sum is n, if input layer xni, hidden layer node snj, output node layer ynt, input layer and hidden layer node
Network weight is wij, the network weight between hidden layer node and output node layer is vjt.；
Step 2：Network weight is initialized, the weights and threshold value of BP network parameters are initialized, several inputs are generated at random
Weight w between layer and hidden layerij, the weight w between hidden layer and output layerjk；
During forward-propagating, each node layer output：
snj=f (net) j=1,2,3...p
ynt=f (netnt) t=1,2 ..., p
In various above, it is all unipolarity function generally to choose excitation function
Step 3：Initialize the coding step in population, i.e. genetic algorithm.N number of original string structured data is randomly generated, often
Individual string structure data are referred to as an individual.Individual constitutes a colony.Genetic algorithm is made with this N number of original string structured data
Start iteration for initial point.
Step 4：Object function is determined, optimal search fitness letter is used as using the inverse of the error function of BP neural network
Number,
In formula, cnlFor the output node desired value of setting, ynlFor output node.When neutral net error reaches minimum value
When, it is optimal network weight that fitness function, which is obtained under maximum, this state,.；
Step 5：Individual using the population of initialization substitutes into as input quantity and calculates every in each population in fitness function
Individual fitness；
Step 6：If fitness meets the fitness value requirement or iterations requirement set by system, system is redirected
To step 10, next step is otherwise carried out；
Step 7：Fitness highest individual in population is preserved, for carrying out the operation such as intersection, variation in genetic algorithm；
Step 8：Progeny population is generated using the Y condition generators in Clouds theory, crossover operation is realized.
Normal cloud model is a followed normal distribution regularity of distribution, the random manifold with steady tendency, with desired value Ex, entropy
En, tri- numerical value of super entropy He are characterized.Desired value Ex：It is best able to represent the point of this qualitativing concept in number field space, reflects
The position of centre of gravity of cloud.
Fig. 2 is three numerical characteristic schematic diagrames of normal cloud model.In figure abscissa represent desired value be Ex normal state with
Machine number, ordinate represents the degree of membership for the normal random number x that desired value is Ex.Entropy En：On the one hand reflecting can in number field space
The scope received by Linguistic Value；On the other hand the point also reflected in number field space can represent the probability of this Linguistic Value, table
Show the randomness that the water dust of qualitativing concept occurs.It discloses the relevance of ambiguity and randomness.Super entropy He：It is the not true of entropy
Fixed measurement, reflect represented in number field space the Linguistic Value uncertain coherency a little.
Crossover operation：
(1) random generation or artificial formulation degree of certainty μ；
(2)
(3) En=variables hunting zone/c1, c1 take 3*p, p to be Population Size；
(4) He=En/c2, c2 take the value between 5~15；
(5) progeny population is produced by Y condition generators；
In formula：xfAnd xmFather's individual and mother's individual respectively in crossover operation；FfAnd FmTheir adaptation is then corresponded to respectively
Degree.
Step 9：Mutation operation in gene mutation, i.e. genetic algorithm is realized using basic Normal Cloud cloud generator；
Mutation operation：
(1) Ex takes former individual；
(2) En=variables hunting zone/c3, c3 take 5；
(3) He=En/c4, c4 take the value between 5~15；
(4) basic Normal Cloud cloud generator is performed, random number Temp is generated, works as μ>It is individual in Population Regeneration during Temp.
Jump to step 5.
Step 10：Obtain best initial weights wij、wjk, BP neural network forward-propagating calculating is substituted into, each layer sample is calculated
Square-error between training output valve and desired value；
Step 11：Export weights；
In the basic Normal Cloud cloud generator：Three numerical characteristics { Ex, En, He } of given cloud and n water dust number, it is raw
It is En into n desired value, variance is He normal random number
En'=RANDN (En, He)
xi=RANDN (Ex, En')
N water dust (x is produced into by above formulai,μi)。
The Y conditions cloud generator：Three numerical characteristics { Ex, En, He } of given cloud and specific degree of certainty μ0, generation
N desired value is En, and variance is He normal random number
En'=RANDN (En, He)
N water dust (x is produced by above formulai,μ0)。
Using the annual data of the fluffy wind power plant in the Zhanjiang pool as training sample；Parameter c2 takes 10, c4 to take in Retrieval method
10；The BP neural network wind power prediction based on Retrieval method is carried out, as a result as shown in Figure 3.Experiment proves the optimized algorithm
It is used in wind power plant wind power prediction and is favorably improved precision of prediction, accelerate network convergence speed, this method calculates cloud heredity
Method applies to BP neural network and sets up forecast model, using Retrieval method come the weights and threshold value of Optimizing BP Network parameter, plus
The Optimization Learning speed of fast BP neural network, so as to accelerate the efficiency of wind power prediction processing.
Claims (4)
1. a kind of short-term wind power prediction method based on Retrieval method, it is characterised in that：Retrieval method is applied into BP
In neural network algorithm, using the weights and threshold value of Retrieval method Optimizing BP Network parameter, accelerate the optimization of BP neural network
Pace of learning, so as to accelerate power prediction treatment effeciency.
2. a kind of short-term wind power prediction method based on Retrieval method according to claim 1, it is characterised in that tool
Body step is as follows：
Step 1：BP neural network structure is determined, selection sample data used sets up BP neural network structure；The BP nerve nets
There is nonlinear relation between the input and output of network, if input number of nodes is N, output node number is M, then network
It is the mapping that theorem in Euclid space is tieed up from N-dimensional theorem in Euclid space to M；For three layers of BP networks, there are input layer, hidden layer, output layer, make net
The node of the input layer of network, hidden layer and output layer is respectively m, p, q, and input sample sum is n, if input layer xni, it is hidden
S containing node layernj, output node layer ynt, the network weight of input layer and hidden layer node is wij, hidden layer node with it is defeated
The network weight gone out between node layer is vjt；
Step 2：Initialize network weight, initialize BP network parameters weights and threshold value, generate at random several input layers and
Weight w between hidden layerij, the weight w between hidden layer and output layerjk；
During forward-propagating, each node layer output：
Snj=f (net) j=1,2 ... p
<mrow>
<mi>n</mi>
<mi>e</mi>
<mi>t</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<msub>
<mi>w</mi>
<mrow>
<mi>i</mi>
<mi>j</mi>
</mrow>
</msub>
<msub>
<mi>x</mi>
<mrow>
<mi>n</mi>
<mi>i</mi>
</mrow>
</msub>
<mo>,</mo>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mi>p</mi>
</mrow>
ynt=f (netnt) t=1,2 ..., p
<mrow>
<msub>
<mi>net</mi>
<mrow>
<mi>n</mi>
<mi>t</mi>
</mrow>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mi>p</mi>
</munderover>
<msub>
<mi>v</mi>
<mrow>
<mi>j</mi>
<mi>t</mi>
</mrow>
</msub>
<msub>
<mi>s</mi>
<mrow>
<mi>n</mi>
<mi>j</mi>
</mrow>
</msub>
<munder>
<mo>&Sigma;</mo>
<mi>j</mi>
</munder>
<msub>
<mi>ni</mi>
<mi>j</mi>
</msub>
<mo>-</mo>
<msub>
<mi>&theta;</mi>
<mi>t</mi>
</msub>
<mo>,</mo>
<mi>t</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>q</mi>
</mrow>
In various above, it is all unipolarity function generally to choose excitation function
X represent it is above-mentioned it is various in respective amount；
Step 3：The coding in population, i.e. genetic algorithm is initialized, N number of original string structured data, each string structure is randomly generated
Data are referred to as an individual, and individual constitutes a colony, and genetic algorithm is used as initial point using this N number of original string structured data
Start iteration；
Step 4：Object function is determined, using the reciprocal as optimal search fitness function of the error function of BP neural network,
In formula, cnlFor the output node desired value of setting, ynlFor output node；When neutral net error reaches minimum value, fit
It is optimal network weight that response function, which is obtained under maximum, this state,；
Step 5：Individual using the population of initialization is substituted into and calculated in fitness function in each population per each and every one as input quantity
The fitness of body；
Step 6：If fitness meets set fitness value requirement or iterations requirement, step 10 is jumped to, it is no
Then carry out next step；
Step 7：Fitness highest individual in population is preserved, for carrying out the intersection in genetic algorithm, mutation operation；
Step 8：Progeny population is generated using the Y condition generators in Clouds theory, crossover operation is realized；
Normal cloud model is a followed normal distribution regularity of distribution, the random manifold with steady tendency, with desired value Ex, entropy En,
Super tri- numerical value of entropy He are characterized；Desired value Ex：It is best able to represent the point of this qualitativing concept in number field space, reflects cloud
Position of centre of gravity；Entropy En：On the one hand the scope that can be received in number field space by Linguistic Value is reflected, is on the other hand also reflected in number
The point of domain space can represent the probability of this Linguistic Value, represent the randomness that the water dust of qualitativing concept occurs；Super entropy He：It is entropy
Uncertain measurement, reflect represented in number field space the Linguistic Value uncertain coherency a little；
The crossover operation is：
(1) random generation or artificial formulation degree of certainty μ；
(2)
(3) En=variables hunting zone/c1, c1 take 3*p, p to be Population Size；
(4) He=En/c2, c2 take the value between 5~15；
(5) progeny population is produced by Y condition generators；
In formula：xfAnd xmFather's individual and mother's individual respectively in crossover operation；FfAnd FmTheir fitness is then corresponded to respectively；
Step 9：Mutation operation in gene mutation, i.e. genetic algorithm is realized using basic Normal Cloud cloud generator；
The mutation operation is：
(1) Ex takes former individual；
(2) En=variables hunting zone/c3, c3 take 5；
(3) He=En/c4, c4 take the value between 5~15；
(4) basic Normal Cloud cloud generator is performed, random number Temp is generated, works as μ>It is individual in Population Regeneration during Temp；Jump to
Step 5；
Step 10：Obtain best initial weights wij、wjk, BP neural network forward-propagating calculating is substituted into, each layer sample training is calculated
Square-error between output valve and desired value；
Step 11：Export weights.
3. a kind of short-term wind power prediction method based on Retrieval method according to claim 2, it is characterised in that：
In the basic Normal Cloud cloud generator：Three numerical characteristics { Ex, En, He } of given cloud and n water dust number, generate n
Individual desired value is En, and variance is He normal random number
En'=RANDN (En, He)
xi=RANDN (Ex, En')
<mrow>
<msub>
<mi>&mu;</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<msup>
<mi>e</mi>
<mfrac>
<mrow>
<mo>-</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mi>E</mi>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mn>2</mn>
<msup>
<mrow>
<mo>(</mo>
<msup>
<mi>En</mi>
<mo>&prime;</mo>
</msup>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
</msup>
</mrow>
N water dust (x is produced into by above formulai,μi), wherein, xiIt is the normal random number that desired value is Ex, μiIt is that desired value is Ex
The expectation curve of Normal Cloud is in xiThe degree of membership at place.
4. a kind of short-term wind power prediction method based on Retrieval method according to claim 2, it is characterised in that：
In the Y conditions life device：Three numerical characteristics { Ex, En, He } of given cloud and specific degree of certainty μ0, generate n expectation
It is worth for En, variance is He normal random number
En'=RANDN (En, He)
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mi>E</mi>
<mi>x</mi>
<mo>&PlusMinus;</mo>
<msup>
<mi>En</mi>
<mo>&prime;</mo>
</msup>
<msqrt>
<mrow>
<mo>-</mo>
<mn>2</mn>
<mi>l</mi>
<mi>n</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>&mu;</mi>
<mn>0</mn>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</msqrt>
</mrow>
N water dust (x is produced by above formulai,μ0).Wherein, xiIt is the normal random number that desired value is Ex, μ0It is that desired value is Ex
The expectation curve of Normal Cloud is in xiThe degree of membership at place.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434726.XA CN107330251A (en) | 2017-06-10 | 2017-06-10 | A kind of wind power prediction method based on Retrieval method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710434726.XA CN107330251A (en) | 2017-06-10 | 2017-06-10 | A kind of wind power prediction method based on Retrieval method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107330251A true CN107330251A (en) | 2017-11-07 |
Family
ID=60194918
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710434726.XA Pending CN107330251A (en) | 2017-06-10 | 2017-06-10 | A kind of wind power prediction method based on Retrieval method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107330251A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109117491A (en) * | 2018-06-15 | 2019-01-01 | 北京理工大学 | A kind of agent model construction method for the higher-dimension small data merging expertise |
| CN109146192A (en) * | 2018-09-03 | 2019-01-04 | 贵州电网有限责任公司 | A kind of wind power forecasting method considering running of wind generating set operating condition |
| CN109409303A (en) * | 2018-10-31 | 2019-03-01 | 南京信息工程大学 | A kind of cascade multitask Face datection and method for registering based on depth |
| CN109635919A (en) * | 2018-11-05 | 2019-04-16 | 新乡航空工业(集团)有限公司 | A kind of air class check valve valid circulation area measuring method and its device |
| CN111240318A (en) * | 2019-12-24 | 2020-06-05 | 华中农业大学 | Robot personnel discovery algorithm |
| CN111626472A (en) * | 2020-04-17 | 2020-09-04 | 国网甘肃省电力公司信息通信公司 | Scene trend judgment index computing system and method based on deep hybrid cloud model |
| CN112052939A (en) * | 2020-08-19 | 2020-12-08 | 国网山西省电力公司 | Active early warning system based on neural network algorithm |
-
2017
- 2017-06-10 CN CN201710434726.XA patent/CN107330251A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109117491A (en) * | 2018-06-15 | 2019-01-01 | 北京理工大学 | A kind of agent model construction method for the higher-dimension small data merging expertise |
| CN109117491B (en) * | 2018-06-15 | 2023-04-07 | 北京理工大学 | Agent model construction method of high-dimensional small data fusing expert experience |
| CN109146192A (en) * | 2018-09-03 | 2019-01-04 | 贵州电网有限责任公司 | A kind of wind power forecasting method considering running of wind generating set operating condition |
| CN109146192B (en) * | 2018-09-03 | 2021-09-03 | 贵州电网有限责任公司 | Wind power prediction method considering operation condition of wind turbine generator |
| CN109409303A (en) * | 2018-10-31 | 2019-03-01 | 南京信息工程大学 | A kind of cascade multitask Face datection and method for registering based on depth |
| CN109635919A (en) * | 2018-11-05 | 2019-04-16 | 新乡航空工业(集团)有限公司 | A kind of air class check valve valid circulation area measuring method and its device |
| CN109635919B (en) * | 2018-11-05 | 2023-01-31 | 新乡航空工业(集团)有限公司 | Method and device for measuring and calculating effective circulation area of air check valve |
| CN111240318A (en) * | 2019-12-24 | 2020-06-05 | 华中农业大学 | Robot personnel discovery algorithm |
| CN111626472A (en) * | 2020-04-17 | 2020-09-04 | 国网甘肃省电力公司信息通信公司 | Scene trend judgment index computing system and method based on deep hybrid cloud model |
| CN111626472B (en) * | 2020-04-17 | 2023-10-27 | 国网甘肃省电力公司信息通信公司 | Scene trend judgment index computing system and method based on depth hybrid cloud model |
| CN112052939A (en) * | 2020-08-19 | 2020-12-08 | 国网山西省电力公司 | Active early warning system based on neural network algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107330251A (en) | A kind of wind power prediction method based on Retrieval method | |
| Lee et al. | Quantum genetic algorithm for dynamic economic dispatch with valve-point effects and including wind power system | |
| Wan et al. | Optimal micro-siting of wind turbines by genetic algorithms based on improved wind and turbine models | |
| CN103606014B (en) | A kind of island distributed power source optimization method based on multiple target | |
| CN103049798B (en) | A kind of short-term power generation power Forecasting Methodology being applied to photovoltaic generating system | |
| Kassa et al. | Short term wind power prediction using ANFIS | |
| Kumar Khadanga et al. | Frequency control in hybrid distributed power systems via type‐2 fuzzy PID controller | |
| CN106130079A (en) | A kind of edema due to wind pathogen fire short-term joint optimal operation method | |
| CN105809279A (en) | Multi-objective quantum Shuffled Frog Leaping Algorithm (SFLA) based water resource optimization and diversion method | |
| CN107256436B (en) | Predictive matching and consumption control method for heat storage electric boiler and clean energy | |
| CN107482618A (en) | Electricity pricing method with the generating of wind fire and the power network of flexible load containing polymorphic type | |
| CN105186499A (en) | Multi-target probabilistically optimal power flow fuzzy modelling and solving method for power distribution network | |
| CN110175707A (en) | Electrical power system transient frequency predication model update method based on data inheritance thought | |
| CN103065202A (en) | Wind power plant ultrashort term wind speed prediction method based on combination kernel function | |
| CN110009141B (en) | Climbing event prediction method and system based on SDAE feature extraction and SVM classification model | |
| CN106130066A (en) | A kind of Multi-objective Robust control method for frequency for independent micro-grid system | |
| CN111555368B (en) | Deep generation type countermeasure network scheduling and control method of comprehensive energy system | |
| CN106602593A (en) | Micro-grid multi-objective-to-single-objective conversion method | |
| CN105956252A (en) | Generative deep belief network-based multi-scale forecast modeling method for ultra-short-term wind speed | |
| CN108665068A (en) | The improved adaptive GA-IAGA of water distribution hydraulic model automatic Check problem | |
| CN105633971B (en) | A kind of intelligent distribution network probabilistic loadflow method based on imperial competition algorithm | |
| CN104216383A (en) | Operating efficiency optimizing method of small hydropower station unit | |
| CN107330248A (en) | A kind of short-term wind power prediction method based on modified neutral net | |
| Chen et al. | Modified antipredatory particle swarm optimization for dynamic economic dispatch with wind power | |
| Sun et al. | An on-line generator start-up strategy based on deep learning and tree search |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171107 |