CN107516150A - A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system - Google Patents

A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system Download PDF

Info

Publication number
CN107516150A
CN107516150A CN201710742259.7A CN201710742259A CN107516150A CN 107516150 A CN107516150 A CN 107516150A CN 201710742259 A CN201710742259 A CN 201710742259A CN 107516150 A CN107516150 A CN 107516150A
Authority
CN
China
Prior art keywords
cso
particles
pso
particle
learning machine
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
Application number
CN201710742259.7A
Other languages
Chinese (zh)
Inventor
殷豪
董朕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Technology
Original Assignee
Guangdong University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN201710742259.7A priority Critical patent/CN107516150A/en
Publication of CN107516150A publication Critical patent/CN107516150A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • Tourism & Hospitality (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Water Supply & Treatment (AREA)
  • Biophysics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Primary Health Care (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The embodiment of the invention discloses a kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system, including wind power historical data is obtained, and wind power historical data is pre-processed to obtain training sample data and test sample data；Test sample data are predicted using the extreme learning machine Optimized model pre-established, obtain wind power prediction result；Wherein, the process of establishing of extreme learning machine Optimized model is that training sample data are added in extreme learning machine；Optimizing processing, the extreme learning machine Optimized model after being trained are carried out to the parameter of extreme learning machine using the particle cluster algorithm for combining chaos crossover algorithm in length and breadth；Parameter includes input weights and hidden layer biasing.The embodiment of the present invention improves the local search ability and global convergence precision of extreme learning machine Optimized model, and test sample data are predicted using the extreme learning machine Optimized model after optimization, so as to get prediction result it is more accurate.

Description

A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system
Technical field
The present embodiments relate to technical field of wind power generation, more particularly to a kind of prediction side of short-term wind-electricity power Method, apparatus and system.
Background technology
With the continuous growth of global power demand, regenerative resource has been widely used, and wind energy is as a kind of new Emerging regenerative resource, promote the development of wind-powered electricity generation science and technology.But the unstability of wind-powered electricity generation makes wind power system be combined with main power network One of obstacle, effectively utilize the wind energy of sustainable growth for greater safety, high-precision wind power forecasting method is to power network Run important in inhibiting.
At present, mainly short-term wind-electricity power is predicted using single model in the prior art, such as time series Method, Grey Modelss, artificial neural network, SVMs and extreme learning machine etc., still, due to wind power time sequence Row have the characteristics of non-linear strong and non-stationary high, so being difficult to using single model accurate to short-term wind-electricity power progress pre- Survey.
Therefore, how a kind of Forecasting Methodology for the short-term wind-electricity power for solving above-mentioned technical problem, apparatus and system are provided Need to solve the problems, such as those skilled in the art.
The content of the invention
The purpose of the embodiment of the present invention is to provide a kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system, improves The local search ability of model, global convergence precision is further increased, so that prediction result is more accurate.
In order to solve the above technical problems, the embodiments of the invention provide a kind of Forecasting Methodology of short-term wind-electricity power, including：
S11：Wind power historical data is obtained, and the wind power historical data is pre-processed to obtain training sample Notebook data and test sample data；
S12：The test sample data are predicted using the extreme learning machine Optimized model pre-established, obtain wind Electrical power prediction result；
Wherein, the process of establishing of the extreme learning machine Optimized model is：
S21：The training sample data are added in extreme learning machine；
S22：The parameter of the extreme learning machine is sought using the particle cluster algorithm with reference to chaos crossover algorithm in length and breadth Excellent processing, the extreme learning machine Optimized model after being trained；The parameter includes input weights and hidden layer biasing.
Optionally, the particle cluster algorithm using with reference to chaos crossover algorithm in length and breadth is to the parameter of the extreme learning machine Carry out optimizing processing process be specially：
S220：In advance to the PSO Population Sizes of particle cluster algorithm, the first maximum iteration, initial weight and accelerate because Son is configured, and in advance to the CSO Population Sizes of crossover algorithm in length and breadth, the second maximum iteration and crossed longitudinally probability It is configured；
S221：Input weights and hidden layer biasing to extreme learning machine initialize, and randomly generate PSO initialization Population；
S222：Calculated by the training sample data and fitness value calculation relational expression in the PSO initial populations Each PSO particles fitness value, and determine global optimum position P according to each fitness valuegbest；
The fitness value calculation relational expression isWherein, ptWithRespectively actual wind Electrical power exports and the output of target wind power, T are the quantity of the training sample data；
S223：The position for initializing each PSO particles in population to the PSO according to population operator is updated, And the fitness value of each PSO particles after renewal is calculated, according to described in the fitness value renewal of each PSO particles after renewal Global optimum position Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
S224：Judge whether current iteration number k is more than preset times k', if it is, into S225, otherwise, return S223, to carry out next iteration；
S225：Judge the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep constant, if it is, entering Enter S226；Otherwise, into S228；
S226：To the global optimum PSO particles XbestInitialized to obtain CSO populations, and calculated using intersecting in length and breadth Method carries out local optimal searching to each CSO particles in the CSO populations, finds optimal CSO particles；
S227：Judge whether the fitness of the optimal CSO particles is better than the fitness of the global optimum PSO particles, If it is, by the global optimum PSO particles XbestIt is updated to the optimal CSO particles；Otherwise, using the optimal CSO Particle replaces the PSO particles that fitness is worst in PSO populations, and returns to S223, to carry out next iteration；
S228：Judge whether current iteration number reaches the first maximum iteration, if it is, by global optimum's particle XbestOutput, and according to global optimum's particle XbestOptimized parameter is obtained, otherwise, S223 is returned to, to be changed next time Generation.
Optionally, it is described to the global optimum PSO particles XbestInitialized to obtain CSO populations, and using in length and breadth Crossover algorithm carries out local optimal searching to each CSO particles in the CSO populations, and the process for finding optimal CSO particles is：
S2261：Two are carried out to each CSO particles in CSO populations neither to repeat to be grouped at random, and according to lateral cross meter Calculate relational expression to be updated the CSO particles in every group, and calculate the fitness value of each filial generation CSO particles after renewal, will The fitness value of each filial generation CSO particles compared with the fitness value of each self-corresponding parent CSO particles, preferentially preserve to In CSO populations；
Wherein, the lateral cross calculation relational expression is：
MShc(i, d)=e1×F(i,d)+(1-e1)×F(j,d)+f1×(F(i,d)-F(j,d))
MShc(j, d)=e2×F(j,d)+(1-e2)×F(i,d)+f2×(F(j,d)-F(i,d))
i,j∈N(1,M0), d ∈ N (1, D)
Wherein, e1And e2For the random number between [0,1]；f1And f2Random number between [- 1,1]；M0It is big for CSO populations It is small；D is the dimension of CSO populations；The d that F (i, d) and F (j, d) is respectively parent CSO particle F (i) and F (j) is tieed up；MShc (i, d) and MShc(j, d) is respectively that F (i, d) and F (j, d) tie up filial generation CSO particles by d caused by lateral cross；
S2262：Two are carried out to all dimensions of each CSO particles in the CSO populations that obtain by lateral cross neither to weigh Multiple random packet, is updated by crossed longitudinally calculation relational expression to each CSO particles of CSO populations, and after calculating renewal Each current filial generation CSO particles fitness value, by the fitness value of each current filial generation CSO particles with it is respective corresponding respectively The fitness values of parent CSO particles be compared, preferentially preserve into CSO populations；
Wherein, the crossed longitudinally calculation relational expression is：
MSvc(i,d1)=eF (i, d1)+(1-e)·F(i,d2)
i∈N(1,M0), d1,d2∈ N (1, D), r ∈ [0,1]
Wherein, MSvc(i,d1) be parent CSO particle F (i) d1Peacekeeping d2Dimension passes through crossed longitudinally caused filial generation CSO particles, random numbers of the e between [0,1]；
S2262：Judge whether current iteration number reaches the second maximum iteration, if it is, according in CSO populations The fitness value of each current CSO particles determines optimal CSO particles；Otherwise, S2261 is returned to, to carry out next iteration.
Optionally, the mistake being updated according to population operator to each particle in the described first initialization population Cheng Wei：
Relational expression and location updating relational expression are updated by speed to enter each particle in the described first initialization population Row location updating；
The speed updates relational expressionIts In, r1、r2And r3Random number between respectively 0 and 1, w are inertia weight coefficient, c1、c2And c3It is accelerated factor, PgbestFor Particle global optimum position,WithPosition and speed of i-th of particle in kth time iteration respectively in population,For the optimal location of i-th of particle in population,For speed of i-th of the particle in population in+1 iteration of kth；
The location updating relational expression isWherein,It is i-th of particle in population in kth+1 Position in secondary iteration.
Optionally, the calculating process of the inertia weight coefficient w is：
According to chaos Inertia Weight calculation relational expression w=wmin+(wmax-wmin) × z (k) obtains the Inertia Weight coefficient w；Wherein, z (k)=μ z (k-1) (1-z (k-1)), μ=4, z (1)=0.8, wmaxFor 0.9, wminFor 0.4.
The embodiment of the present invention provides a kind of prediction meanss of short-term wind-electricity power accordingly, including：
Pretreatment module, located in advance for obtaining wind power historical data, and to the wind power historical data Reason obtains training sample data and test sample data；
Prediction module, it is pre- for being carried out using the extreme learning machine Optimized model pre-established to the test sample data Survey, obtain wind power prediction result；
Wherein, the extreme learning machine Optimized model includes：
Add module, for the training sample data to be added in extreme learning machine；
Training module, for using the particle cluster algorithm with reference to chaos crossover algorithm in length and breadth to the ginseng of the extreme learning machine Number carries out optimizing processing, the extreme learning machine Optimized model after being trained；The parameter includes input weights and implied Layer biasing.
Optionally, the training module includes：
Default unit, in advance to the PSO Population Sizes of particle cluster algorithm, the first maximum iteration, initial weight It is configured with accelerated factor, and in advance to the CSO Population Sizes of crossover algorithm in length and breadth, the second maximum iteration and longitudinal direction Crossover probability is configured；
Initialization unit, initialize for the input weights to extreme learning machine and hidden layer biasing, and produce at random Raw PSO initialization population；
Fitness value calculation unit, for calculating institute by the training sample data and fitness value calculation relational expression The fitness value of each PSO particles in PSO initial populations is stated, and global optimum position is determined according to each fitness value Put Pgbest；
The fitness value calculation relational expression isWherein, ptWithRespectively actual wind Electrical power exports and the output of target wind power, T are the quantity of the training sample data；
Updating block, the position for initializing each PSO particles in population to the PSO according to population operator are entered Row renewal, and the fitness value of each PSO particles after renewal is calculated, the fitness value according to each PSO particles after renewal Update the global optimum position Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
First judging unit, for judging whether current iteration number k is more than preset times k', if it is, triggering the Two judging units, otherwise, the updating block is triggered, to carry out next iteration；
Second judging unit, for judging the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep It is constant, if it is, crossover algorithm optimizing unit in length and breadth described in entering；Otherwise, the 4th judging unit is triggered；
Crossover algorithm optimizing unit in length and breadth, for the global optimum PSO particles XbestInitialized to obtain CSO kinds Group, and local optimal searching is carried out to each CSO particles in the CSO populations using crossover algorithm in length and breadth, find optimal CSO grains Son；
3rd judging unit, for judging whether the fitness of the optimal CSO particles is better than the global optimum PSO grains The fitness of son, if it is, by the global optimum PSO particles XbestIt is updated to the optimal CSO particles；Otherwise, use The optimal CSO particles replace the PSO particles that fitness is worst in PSO populations, and trigger the updating block, next to carry out Secondary iteration；
4th judging unit, for judging whether current iteration number reaches the first maximum iteration, if it is, Then by global optimum particle XbestOutput, and according to global optimum's particle XbestOptimized parameter is obtained, otherwise, described in triggering Updating block, to carry out next iteration.
The embodiment of the present invention additionally provides a kind of forecasting system of short-term wind-electricity power, including short-term wind as described above The prediction meanss of electrical power.
The embodiments of the invention provide a kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system, including：Obtain wind-powered electricity generation Power historical data, and wind power historical data is pre-processed to obtain training sample data and test sample data；Adopt Test sample data are predicted with the extreme learning machine Optimized model pre-established, obtain wind power prediction result；Its In, the process of establishing of extreme learning machine Optimized model is that training sample data are added in extreme learning machine；It is mixed using combining The particle cluster algorithm of ignorant crossover algorithm in length and breadth carries out optimizing processing to the parameter of extreme learning machine, the limit study after being trained Machine Optimized model；Parameter includes input weights and hidden layer biasing.
The embodiment of the present invention is optimized by the particle cluster algorithm of chaos crossover algorithm in length and breadth to extreme learning machine model, The local search ability and global convergence precision of model are improved, and using the extreme learning machine Optimized model after optimization to through pre- Test sample data after processing are predicted, and obtain short-term wind-electricity power prediction result, make prediction result more accurate.
Brief description of the drawings
Technical scheme in order to illustrate the embodiments of the present invention more clearly, below will be to institute in prior art and embodiment The accompanying drawing needed to use is briefly described, it should be apparent that, drawings in the following description are only some implementations of the present invention Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also be obtained according to these accompanying drawings Obtain other accompanying drawings.
Fig. 1 is a kind of schematic flow sheet of the Forecasting Methodology of short-term wind-electricity power provided in an embodiment of the present invention；
Fig. 2 is the schematic flow sheet that a kind of extreme learning machine Optimized model provided in an embodiment of the present invention is established；
Fig. 3 is a kind of structural representation of the prediction meanss of short-term wind-electricity power provided in an embodiment of the present invention；
Fig. 4 is a kind of structural representation of extreme learning machine Optimized model provided in an embodiment of the present invention.
Embodiment
The embodiments of the invention provide a kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system, model is improved Local search ability, global convergence precision is further increased, so that prediction result is more accurate.
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is Part of the embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
Fig. 1 is refer to, Fig. 1 is that a kind of flow of Forecasting Methodology of short-term wind-electricity power provided in an embodiment of the present invention is illustrated Figure.
This method includes：
S11：Wind power historical data is obtained, and wind power historical data is pre-processed to obtain number of training According to and test sample data；
It should be noted that training sample data TrnInput sample X comprising modelnWith output sample Yn, input sample Xn With output sample YnIt is that wind power historical data continuous acquisition, input sample are represented by Wherein, m is that forecast model inputs number；Output sample is represented byWherein, l value can To be determined by forecast model output number；N is n-th of sample in sample set.In addition, test sample collection TenSelection mode with Training sample set TrnSelection mode it is identical.
S12：Test sample data are predicted using the extreme learning machine Optimized model pre-established, obtain wind-powered electricity generation work( Rate prediction result；
Wherein, Fig. 2 is refer to, Fig. 2 is the stream that a kind of extreme learning machine Optimized model provided in an embodiment of the present invention is established Journey schematic diagram, it is established process and is specially：
S21：Training sample data are added in extreme learning machine；
S22：The parameter of extreme learning machine is carried out at optimizing using the particle cluster algorithm for combining chaos crossover algorithm in length and breadth Reason, the extreme learning machine Optimized model after being trained；Parameter includes input weights and hidden layer biasing.
It should be noted that the node in hidden layer s of extreme learning machine can be set to 10, input number m can be set to 8 Individual, output number l can be set to 4, and hidden layer node activation primitive can use sigmoid functions, and extreme learning machine needs The variable number of optimization is num=m × s+s；
Wherein, sigmoid functions calculation formula is
Further, the particle cluster algorithm for combining chaos crossover algorithm in length and breadth is used in above-mentioned S22 to extreme learning machine Parameter carry out optimizing processing process, tool can using body as：
S220：PSO Population Sizes M to particle cluster algorithm, the first maximum iteration T in advancemax, initial weight w and add The fast factor is configured, and in advance to the CSO Population Sizes M of crossover algorithm in length and breadth0, the second maximum iteration CmaxWith it is vertical It is configured to crossover probability；
Specifically, the particle cluster algorithm of combination chaos in embodiment of the present invention crossover algorithm in length and breadth includes elementary particle group Algorithm (Particle Swarm Optimization, PSO) and in length and breadth crossover algorithm (CSO), using in crossover algorithm in length and breadth Crossed longitudinally operator avoids the local optimum problem of the full algorithm of elementary particle, wherein, the PSO Population Sizes M of particle cluster algorithm 60 can be set to, the first maximum iteration Tmax300, initial weight w can be taken to take 0.9, accelerated factor c1=2, c2= 2；The CSO Population Sizes M of crossover algorithm in length and breadth030 can be set to, the second maximum iteration CmaxIt can take 30, it is crossed longitudinally Probability can be set to 0.55.
S221：Input weights and hidden layer biasing to extreme learning machine initialize, and randomly generate PSO initialization Population；
It should be noted that each PSO particles in PSO initializes population can use XiRepresent, wherein,IW is input weights, shares m × s；Ib is hidden layer Biasing, share s；Common M particle in PSO populations.
S222：Each PSO in PSO initial populations is calculated by training sample data and fitness value calculation relational expression The fitness value of particle, and determine global optimum position P according to each fitness valuegbest；
Fitness value calculation relational expression isWherein, ptWithRespectively actual wind-powered electricity generation work( Rate exports and the output of target wind power, T are the quantity of training sample data；
S223：The position for initializing each PSO particles in population to PSO according to population operator is updated, and is counted The fitness value of each PSO particles after renewal is calculated, the fitness value renewal global optimum according to each PSO particles after renewal Position Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
Further, in above-mentioned S223, each particle in the first initialization population is carried out according to population operator The process of renewal, is specifically as follows：
Relational expression and location updating relational expression are updated by speed line position is entered to each particle in the first initialization population Put renewal；
Speed updates relational expressionWherein, r1、r2And r3Random number between respectively 0 and 1, w are inertia weight coefficient, c1、c2And c3It is accelerated factor, PgbestFor grain Sub- global optimum position,WithPosition and speed of i-th of particle in kth time iteration respectively in population,For The optimal location of i-th of particle in population,For speed of i-th of the particle in population in+1 iteration of kth；
In each iteration, the new position of each particle can be by location updating relational expressionObtain, Wherein,For position of i-th of the particle in population in+1 iteration of kth.
Specifically, above-mentioned inertia weight coefficient w calculating process, is specifically as follows：
According to chaos Inertia Weight calculation relational expression w=wmin+(wmax-wmin) × z (k) obtains Inertia Weight coefficient w；Its In, z (k)=μ z (k-1) (1-z (k-1)), μ=4, z (1)=0.8, wmaxFor 0.9, wminFor 0.4.It should be noted that μ, z (1)、wmaxAnd wminSpecific value can be determined according to actual conditions, the embodiment of the present invention does not do special limit to this It is fixed.
S224：Judge whether current iteration number k is more than preset times k', if it is, into S225, otherwise, return S223, to carry out next iteration；
S225：Judge the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep constant, if it is, entering Enter S226；Otherwise, into S228；
It should be noted that the k' in the embodiment of the present invention can be 10, it is of course also possible to be other concrete numerical values, Its specific value the application is not done, special restriction.
S226：To global optimum PSO particles XbestInitialized to obtain CSO populations, and using crossover algorithm pair in length and breadth Each CSO particles in CSO populations carry out local optimal searching, find optimal CSO particles；
Specifically, chaos sequence can be produced according to according to following Logistic mapping relations, crossover algorithm in length and breadth is constructed Population, and enter local optimal searching for g=1:Cmax。
Wherein, i-th of chaotic particle isCSO particles i is in crossover algorithm population in length and breadthM0For crossover algorithm Population Size in length and breadth, D is the dimension that particle includes, with Machine produces chaos primary C1, u values can be 2, and strength of turbulence s values can be 0.4.
Further, in above-mentioned S226, to global optimum PSO particles XbestInitialized to obtain CSO populations, and adopted Local optimal searching is carried out to each CSO particles in CSO populations with crossover algorithm in length and breadth, finds the process of optimal CSO particles, specifically Can be：
S2261：Two are carried out to each CSO particles in CSO populations neither to repeat to be grouped at random, and according to lateral cross meter Calculate relational expression to be updated the CSO particles in every group, and calculate the fitness value of each filial generation CSO particles after renewal, will The fitness value of each filial generation CSO particles compared with the fitness value of each self-corresponding parent CSO particles, preferentially preserve to In CSO populations；
Wherein, lateral cross calculation relational expression is：
MShc(i, d)=e1×F(i,d)+(1-e1)×F(j,d)+f1×(F(i,d)-F(j,d))
MShc(j, d)=e2×F(j,d)+(1-e2)×F(i,d)+f2×(F(j,d)-F(i,d))
i,j∈N(1,M0), d ∈ N (1, D)
Wherein, e1And e2For the random number between [0,1]；f1And f2Random number between [- 1,1]；M0It is big for CSO populations It is small；D is the dimension of CSO populations；The d that F (i, d) and F (j, d) is respectively parent CSO particle F (i) and F (j) is tieed up；MShc (i, d) and MShc(j, d) is respectively that F (i, d) and F (j, d) tie up filial generation CSO particles by d caused by lateral cross；
It should be noted that before each CSO particles progress two in CSO populations neither repeats random packet, need Every by each CSO particles obtained through lateral cross one-dimensional is normalized.
S2262：Two are carried out to all dimensions of each CSO particles in the CSO populations that obtain by lateral cross neither to weigh Multiple random packet, is updated by crossed longitudinally calculation relational expression to each CSO particles of CSO populations, and after calculating renewal Each current filial generation CSO particles fitness value, by the fitness value of each current filial generation CSO particles with it is respective corresponding respectively The fitness values of parent CSO particles be compared, preferentially preserve into CSO populations；
Wherein, crossed longitudinally calculation relational expression is：
MSvc(i,d1)=eF (i, d1)+(1-e)·F(i,d2)
i∈N(1,M 0), d1,d2∈ N (1, D), r ∈ [0,1]
Wherein, MSvc(i,d1) be parent CSO particle F (i) d1Peacekeeping d2Dimension passes through crossed longitudinally caused filial generation CSO particles, random numbers of the e between [0,1]；
S2262：Judge whether current iteration number reaches the second maximum iteration, if it is, according in CSO populations The fitness value of each current CSO particles determines optimal CSO particles；Otherwise, g=g+1, and S2261 is returned, it is next to carry out Secondary iteration.
S227：Judge whether the fitness of optimal CSO particles is better than the fitness of global optimum's PSO particles, if it is, By global optimum PSO particles XbestIt is updated to optimal CSO particles；Otherwise, replaced in PSO populations and adapted to using optimal CSO particles Worst PSO particles are spent, and return to S223, to carry out next iteration；
S228：Judge whether current iteration number reaches the first maximum iteration, if it is, by global optimum's particle XbestOutput, and according to global optimum particle XbestOptimized parameter is obtained, otherwise, S223 is returned to, to carry out next iteration.
The embodiments of the invention provide a kind of Forecasting Methodology of short-term wind-electricity power, including：Obtain wind power history number According to, and wind power historical data is pre-processed to obtain training sample data and test sample data；Using pre-establishing Extreme learning machine Optimized model test sample data are predicted, obtain wind power prediction result；Wherein, the limit learns The process of establishing of machine Optimized model is that training sample data are added in extreme learning machine；Intersect calculation in length and breadth using with reference to chaos The particle cluster algorithm of method carries out optimizing processing, the extreme learning machine Optimized model after being trained to the parameter of extreme learning machine； Parameter includes input weights and hidden layer biasing.
The embodiment of the present invention is optimized by the particle cluster algorithm of chaos crossover algorithm in length and breadth to extreme learning machine model, The local search ability and global convergence precision of model are improved, and using the extreme learning machine Optimized model after optimization to through pre- Test sample data after processing are predicted, and obtain short-term wind-electricity power prediction result, make prediction result more accurate.
Accordingly the embodiment of the invention also discloses a kind of prediction meanss of short-term wind-electricity power, Fig. 3 is specifically refer to, is schemed 3 be a kind of structural representation of the prediction meanss of short-term wind-electricity power provided in an embodiment of the present invention.In the base of above-described embodiment On plinth：
The device includes：
Pretreatment module 11, pre-processed for obtaining wind power historical data, and to wind power historical data Obtain training sample data and test sample data；
Prediction module 12, it is pre- for being carried out using the extreme learning machine Optimized model pre-established to test sample data Survey, obtain wind power prediction result；
Wherein, extreme learning machine Optimized model includes add module 21 and training module 22, refer to Fig. 4, specifically：
Add module 21, for training sample data to be added in extreme learning machine；
Training module 22, for using combining the particle cluster algorithm of chaos crossover algorithm in length and breadth to the parameter of extreme learning machine Carry out optimizing processing, the extreme learning machine Optimized model after being trained；Parameter includes input weights and hidden layer biasing.
Optionally, training module 22 includes：
Default unit, in advance to the PSO Population Sizes of particle cluster algorithm, the first maximum iteration, initial weight It is configured with accelerated factor, and in advance to the CSO Population Sizes of crossover algorithm in length and breadth, the second maximum iteration and longitudinal direction Crossover probability is configured；
Initialization unit, initialize for the input weights to extreme learning machine and hidden layer biasing, and produce at random Raw PSO initialization population；
Fitness value calculation unit, at the beginning of calculating PSO by training sample data and fitness value calculation relational expression The fitness value of each PSO particles in beginning population, and determine global optimum position P according to each fitness valuegbest；
Fitness value calculation relational expression isWherein, ptWithRespectively actual wind power Output and the output of target wind power, T are the quantity of training sample data；
Updating block, the position for initializing each PSO particles in population to PSO according to population operator are carried out more Newly, and the fitness value of each PSO particles after renewal is calculated, the fitness value according to each PSO particles after renewal updates Global optimum position Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
First judging unit, for judging whether current iteration number k is more than preset times k', if it is, triggering the Two judging units, otherwise, updating block is triggered, to carry out next iteration；
Second judging unit, for judging the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep constant, If it is, into crossover algorithm optimizing unit in length and breadth；Otherwise, the 4th judging unit is triggered；
Crossover algorithm optimizing unit in length and breadth, for global optimum PSO particles XbestInitialized to obtain CSO populations, And local optimal searching is carried out to each CSO particles in CSO populations using crossover algorithm in length and breadth, find optimal CSO particles；
3rd judging unit, for judging whether the fitness of optimal CSO particles is better than the adaptation of global optimum's PSO particles Degree, if it is, by global optimum PSO particles XbestIt is updated to optimal CSO particles；Otherwise, replaced using optimal CSO particles The worst PSO particles of fitness in PSO populations, and updating block is triggered, to carry out next iteration；
4th judging unit, for judging whether current iteration number reaches the first maximum iteration, if it is, will Global optimum particle XbestOutput, and according to global optimum particle XbestOptimized parameter is obtained, otherwise, updating block is triggered, to enter Row next iteration.
It should be noted that the embodiment of the present invention by the particle cluster algorithm of chaos crossover algorithm in length and breadth to extreme learning machine Model optimizes, and improves the local search ability and global convergence precision of model, and using the extreme learning machine after optimization Optimized model is predicted to test sample data after pretreatment, is obtained short-term wind-electricity power prediction result, is tied prediction Fruit is more accurate.
In addition, the specific introduction for the Forecasting Methodology of short-term wind-electricity power involved in the embodiment of the present invention please be joined According to above method embodiment, the application will not be repeated here.
The embodiment of the present invention additionally provides a kind of forecasting system of short-term wind-electricity power, including short-term wind-electricity work(described above The prediction meanss of rate.
It should be noted that the embodiment of the present invention by the particle cluster algorithm of chaos crossover algorithm in length and breadth to extreme learning machine Model optimizes, and improves the local search ability and global convergence precision of model, and using the extreme learning machine after optimization Optimized model is predicted to test sample data after pretreatment, is obtained short-term wind-electricity power prediction result, is tied prediction Fruit is more accurate.In addition, the specific introduction of the Forecasting Methodology for short-term wind-electricity power involved in the embodiment of the present invention Above method embodiment is refer to, the application will not be repeated here.
Each embodiment is described by the way of progressive in this specification, what each embodiment stressed be and other The difference of embodiment, between each embodiment identical similar portion mutually referring to.For device disclosed in embodiment For, because it is corresponded to the method disclosed in Example, so description is fairly simple, related part is said referring to method part It is bright.
It should also be noted that, in this manual, such as first and second or the like relational terms be used merely to by One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning Covering including for nonexcludability, so that process, method, article or equipment including a series of elements not only include that A little key elements, but also the other element including being not expressly set out, or also include for this process, method, article or The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", is not arranged Except other identical element in the process including the key element, method, article or equipment being also present.
Professional further appreciates that, with reference to the unit of each example of the embodiments described herein description And algorithm steps, can be realized with electronic hardware, computer software or the combination of the two, in order to clearly demonstrate hardware and The interchangeability of software, the composition and step of each example are generally described according to function in the above description.These Function is performed with hardware or software mode actually, application-specific and design constraint depending on technical scheme.Specialty Technical staff can realize described function using distinct methods to each specific application, but this realization should not Think beyond the scope of this invention.
Directly it can be held with reference to the step of method or algorithm that the embodiments described herein describes with hardware, processor Capable software module, or the two combination are implemented.Software module can be placed in random access memory (RAM), internal memory, read-only deposit Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology In the storage medium of any other forms well known in field.
The foregoing description of the disclosed embodiments, professional and technical personnel in the field are enable to realize or using the present invention. A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, it is of the invention The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one The most wide scope caused.

Claims (8)

  1. A kind of 1. Forecasting Methodology of short-term wind-electricity power, it is characterised in that including：
    S11：Wind power historical data is obtained, and the wind power historical data is pre-processed to obtain number of training According to and test sample data；
    S12：The test sample data are predicted using the extreme learning machine Optimized model pre-established, obtain wind-powered electricity generation work( Rate prediction result；
    Wherein, the process of establishing of the extreme learning machine Optimized model is：
    S21：The training sample data are added in extreme learning machine；
    S22：The parameter of the extreme learning machine is carried out at optimizing using the particle cluster algorithm with reference to chaos crossover algorithm in length and breadth Reason, the extreme learning machine Optimized model after being trained；The parameter includes input weights and hidden layer biasing.
  2. 2. the Forecasting Methodology of short-term wind-electricity power according to claim 1, it is characterised in that described use is indulged with reference to chaos The process that the particle cluster algorithm of horizontal crossover algorithm carries out optimizing processing to the parameter of the extreme learning machine is specially：
    S220：The PSO Population Sizes of particle cluster algorithm, the first maximum iteration, initial weight and accelerated factor are entered in advance Row is set, and the CSO Population Sizes of crossover algorithm in length and breadth, the second maximum iteration and crossed longitudinally probability are carried out in advance Set；
    S221：Input weights and hidden layer biasing to extreme learning machine initialize, and randomly generate PSO initialization kind Group；
    S222：Calculated by the training sample data and fitness value calculation relational expression each in the PSO initial populations The fitness value of individual PSO particles, and determine global optimum position P according to each fitness valuegbest；
    The fitness value calculation relational expression isWherein, ptWithRespectively actual wind power Output and the output of target wind power, T are the quantity of the training sample data；
    S223：The position for initializing each PSO particles in population to the PSO according to population operator is updated, and is counted The fitness value of each PSO particles after renewal is calculated, the fitness value according to each PSO particles after renewal updates the overall situation Optimal location Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
    S224：Judge whether current iteration number k is more than preset times k', if it is, into S225, otherwise, return to S223, To carry out next iteration；
    S225：Judge the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep constant, if it is, into S226；Otherwise, into S228；
    S226：To the global optimum PSO particles XbestInitialized to obtain CSO populations, and using crossover algorithm in length and breadth to institute The each CSO particles stated in CSO populations carry out local optimal searching, find optimal CSO particles；
    S227：Judge whether the fitness of the optimal CSO particles is better than the fitness of the global optimum PSO particles, if It is, then by the global optimum PSO particles XbestIt is updated to the optimal CSO particles；Otherwise, using the optimal CSO particles The PSO particles that fitness is worst in PSO populations are replaced, and return to S223, to carry out next iteration；
    S228：Judge whether current iteration number reaches the first maximum iteration, if it is, by global optimum particle Xbest Output, and according to global optimum's particle XbestOptimized parameter is obtained, otherwise, S223 is returned to, to carry out next iteration.
  3. 3. the Forecasting Methodology of short-term wind-electricity power according to claim 2, it is characterised in that described to the global optimum PSO particles XbestInitialized to obtain CSO populations, and using crossover algorithm in length and breadth to each CSO grains in the CSO populations Son carries out local optimal searching, and the process for finding optimal CSO particles is：
    S2261：Two are carried out to each CSO particles in CSO populations neither to repeat to be grouped at random, and are calculated and closed according to lateral cross It is that formula is updated to the CSO particles in every group, and calculates the fitness value of each filial generation CSO particles after renewal, will be each The fitness value of filial generation CSO particles is preferentially preserved to CSO compared with the fitness value of each self-corresponding parent CSO particles In population；
    Wherein, the lateral cross calculation relational expression is：
    MShc(i, d)=e1×F(i,d)+(1-e1)×F(j,d)+f1×(F(i,d)-F(j,d))
    MShc(j, d)=e2×F(j,d)+(1-e2)×F(i,d)+f2×(F(j,d)-F(i,d))
    i,j∈N(1,M0), d ∈ N (1, D)
    Wherein, e1And e2For the random number between [0,1]；f1And f2Random number between [- 1,1]；M0For CSO Population Sizes；D For the dimension of CSO populations；The d that F (i, d) and F (j, d) is respectively parent CSO particle F (i) and F (j) is tieed up；MShc(i,d) And MShc(j, d) is respectively that F (i, d) and F (j, d) tie up filial generation CSO particles by d caused by lateral cross；
    S2262：All dimensions of each CSO particles in the CSO populations that are obtained by lateral cross are carried out two neither repeat with Machine is grouped, and each CSO particles of CSO populations are updated by crossed longitudinally calculation relational expression, and is calculated each after renewal The fitness value of individual current filial generation CSO particles, by the fitness value of each current filial generation CSO particles father corresponding with each difference It is compared, is preferentially preserved into CSO populations for the fitness value of CSO particles；
    Wherein, the crossed longitudinally calculation relational expression is：
    MSvc(i,d1)=eF (i, d1)+(1-e)·F(i,d2)
    i∈N(1,M0), d1,d2∈ N (1, D), r ∈ [0,1]
    Wherein, MSvc(i,d1) be parent CSO particle F (i) d1Peacekeeping d2Dimension passes through crossed longitudinally caused filial generation CSO grains Son, random numbers of the e between [0,1]；
    S2262：Judge whether current iteration number reaches the second maximum iteration, if it is, according to each in CSO populations The fitness value of current CSO particles determines optimal CSO particles；Otherwise, S2261 is returned to, to carry out next iteration.
  4. 4. the Forecasting Methodology of short-term wind-electricity power according to claim 2, it is characterised in that described according to population operator The process that each particle in population is updated is initialized to described first is：
    Relational expression and location updating relational expression are updated by speed line position is entered to each particle in the described first initialization population Put renewal；
    The speed updates relational expressionWherein, r1、r2And r3Random number between respectively 0 and 1, w are inertia weight coefficient, c1、c2And c3It is accelerated factor, PgbestFor institute Particle global optimum position is stated,WithPosition and speed of i-th of particle in kth time iteration respectively in population, For the optimal location of i-th of particle in population,For speed of i-th of the particle in population in+1 iteration of kth；
    The location updating relational expression isWherein,Changed for+1 time in kth for i-th of particle in population Position in generation.
  5. 5. the Forecasting Methodology of short-term wind-electricity power according to claim 4, it is characterised in that the inertia weight coefficient w Calculating process be：
    According to chaos Inertia Weight calculation relational expression w=wmin+(wmax-wmin) × z (k) obtains the Inertia Weight coefficient w；Its In, z (k)=μ z (k-1) (1-z (k-1)), μ=4, z (1)=0.8, wmaxFor 0.9, wminFor 0.4.
  6. A kind of 6. prediction meanss of short-term wind-electricity power, it is characterised in that including：
    Pretreatment module, for obtaining wind power historical data, and to the wind power historical data pre-process To training sample data and test sample data；
    Prediction module, for being predicted using the extreme learning machine Optimized model pre-established to the test sample data, Obtain wind power prediction result；
    Wherein, the extreme learning machine Optimized model includes：
    Add module, for the training sample data to be added in extreme learning machine；
    Training module, for being entered using the particle cluster algorithm with reference to chaos crossover algorithm in length and breadth to the parameter of the extreme learning machine Row optimizing is handled, the extreme learning machine Optimized model after being trained；The parameter includes input weights and hidden layer is inclined Put.
  7. 7. the prediction meanss of short-term wind-electricity power according to claim 6, it is characterised in that the training module includes：
    Default unit, for the PSO Population Sizes of particle cluster algorithm, the first maximum iteration, initial weight and adding in advance The fast factor is configured, and in advance to the CSO Population Sizes of crossover algorithm in length and breadth, the second maximum iteration and crossed longitudinally Probability is configured；
    Initialization unit, initialized for the input weights to extreme learning machine and hidden layer biasing, and randomly generate PSO Initialize population；
    Fitness value calculation unit, it is described for being calculated by the training sample data and fitness value calculation relational expression The fitness value of each PSO particles in PSO initial populations, and determine global optimum position according to each fitness value Pgbest；
    The fitness value calculation relational expression isWherein, ptWithRespectively actual wind power Output and the output of target wind power, T are the quantity of the training sample data；
    Updating block, the position for initializing each PSO particles in population to the PSO according to population operator are carried out more Newly, and the fitness value of each PSO particles after renewal is calculated, the fitness value according to each PSO particles after renewal updates The global optimum position Pgbest, global optimum PSO particles XbestAnd the optimal value of each particle
    First judging unit, for judging whether current iteration number k is more than preset times k', if it is, triggering second is sentenced Disconnected unit, otherwise, triggers the updating block, to carry out next iteration；
    Second judging unit, for judging the global optimum position P in POS populationsgbestWhether in continuous k' generations, keep constant, If it is, crossover algorithm optimizing unit in length and breadth described in entering；Otherwise, the 4th judging unit is triggered；
    Crossover algorithm optimizing unit in length and breadth, for the global optimum PSO particles XbestInitialized to obtain CSO populations, and Local optimal searching is carried out to each CSO particles in the CSO populations using crossover algorithm in length and breadth, finds optimal CSO particles；
    3rd judging unit, for judging the fitness of the optimal CSO particles whether better than the global optimum PSO particles Fitness, if it is, by the global optimum PSO particles XbestIt is updated to the optimal CSO particles；Otherwise, using described Optimal CSO particles replace the PSO particles that fitness is worst in PSO populations, and trigger the updating block, to be changed next time Generation；
    4th judging unit, for judging whether current iteration number reaches the first maximum iteration, if it is, will Global optimum particle XbestOutput, and according to global optimum's particle XbestOptimized parameter is obtained, otherwise, triggers the renewal Unit, to carry out next iteration.
  8. 8. a kind of forecasting system of short-term wind-electricity power, it is characterised in that including short-term wind-electricity as claimed in claims 6 or 7 The prediction meanss of power.
CN201710742259.7A 2017-08-25 2017-08-25 A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system Pending CN107516150A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710742259.7A CN107516150A (en) 2017-08-25 2017-08-25 A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710742259.7A CN107516150A (en) 2017-08-25 2017-08-25 A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system

Publications (1)

Publication Number Publication Date
CN107516150A true CN107516150A (en) 2017-12-26

Family

ID=60724137

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710742259.7A Pending CN107516150A (en) 2017-08-25 2017-08-25 A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system

Country Status (1)

Country Link
CN (1) CN107516150A (en)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108388115A (en) * 2018-02-08 2018-08-10 南京邮电大学 NCS method for compensating network delay based on generalized predictive control
CN108985514A (en) * 2018-07-17 2018-12-11 广东工业大学 Load forecasting method, device and equipment based on EEMD and LSTM
CN109063916A (en) * 2018-08-13 2018-12-21 广东工业大学 A kind of wind power forecasting method, device, equipment and storage medium
CN109146183A (en) * 2018-08-24 2019-01-04 广东工业大学 Short-term impact load forecasting model method for building up based on signal decomposition and intelligent optimization algorithm
CN109522627A (en) * 2018-11-01 2019-03-26 西安电子科技大学 Fan blade icing prediction technique based on SCADA data
CN110048420A (en) * 2019-05-15 2019-07-23 广东电网有限责任公司 A kind of method, apparatus and medium of the scheduling of power distribution network random optimization
CN110210657A (en) * 2019-05-22 2019-09-06 国网冀北电力有限公司 A kind of power of fan prediction technique, system and computer storage medium based on single machine model
CN110363363A (en) * 2019-07-31 2019-10-22 广东电网有限责任公司 A kind of dual-stage scheduling model optimization method for integrated energy system
CN110969262A (en) * 2019-12-03 2020-04-07 广东电网有限责任公司 Transformer fault diagnosis method
CN111075647A (en) * 2019-12-04 2020-04-28 浙江大学 ELM-based maximum wind energy capture method for variable-speed wind turbine generator
CN113449920A (en) * 2021-06-30 2021-09-28 上海电机学院 Wind power prediction method, system and computer readable medium
CN113468817A (en) * 2021-07-13 2021-10-01 淮阴工学院 Ultra-short-term wind power prediction method based on IGOA (optimized El-electric field model)
CN113537566A (en) * 2021-06-16 2021-10-22 广东工业大学 Ultra-short-term wind power prediction method based on DCCSO optimization deep learning model
CN114169251A (en) * 2021-12-27 2022-03-11 广东工业大学 Ultra-short-term wind power prediction method
CN116388865A (en) * 2023-06-05 2023-07-04 深圳市飞思卓科技有限公司 PON optical module-based automatic screening method for abnormal optical power

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108388115A (en) * 2018-02-08 2018-08-10 南京邮电大学 NCS method for compensating network delay based on generalized predictive control
CN108985514A (en) * 2018-07-17 2018-12-11 广东工业大学 Load forecasting method, device and equipment based on EEMD and LSTM
CN109063916A (en) * 2018-08-13 2018-12-21 广东工业大学 A kind of wind power forecasting method, device, equipment and storage medium
CN109146183A (en) * 2018-08-24 2019-01-04 广东工业大学 Short-term impact load forecasting model method for building up based on signal decomposition and intelligent optimization algorithm
CN109522627A (en) * 2018-11-01 2019-03-26 西安电子科技大学 Fan blade icing prediction technique based on SCADA data
CN109522627B (en) * 2018-11-01 2022-12-02 西安电子科技大学 Fan blade icing prediction method based on SCADA (Supervisory control and data acquisition) data
CN110048420B (en) * 2019-05-15 2021-08-13 广东电网有限责任公司 Method, device and medium for random optimal scheduling of power distribution network
CN110048420A (en) * 2019-05-15 2019-07-23 广东电网有限责任公司 A kind of method, apparatus and medium of the scheduling of power distribution network random optimization
CN110210657A (en) * 2019-05-22 2019-09-06 国网冀北电力有限公司 A kind of power of fan prediction technique, system and computer storage medium based on single machine model
CN110363363A (en) * 2019-07-31 2019-10-22 广东电网有限责任公司 A kind of dual-stage scheduling model optimization method for integrated energy system
CN110969262A (en) * 2019-12-03 2020-04-07 广东电网有限责任公司 Transformer fault diagnosis method
CN111075647B (en) * 2019-12-04 2021-06-22 浙江大学 ELM-based maximum wind energy capture method for variable-speed wind turbine generator
CN111075647A (en) * 2019-12-04 2020-04-28 浙江大学 ELM-based maximum wind energy capture method for variable-speed wind turbine generator
CN113537566A (en) * 2021-06-16 2021-10-22 广东工业大学 Ultra-short-term wind power prediction method based on DCCSO optimization deep learning model
CN113537566B (en) * 2021-06-16 2022-05-06 广东工业大学 Ultra-short-term wind power prediction method based on DCCSO optimization deep learning model
CN113449920A (en) * 2021-06-30 2021-09-28 上海电机学院 Wind power prediction method, system and computer readable medium
CN113468817A (en) * 2021-07-13 2021-10-01 淮阴工学院 Ultra-short-term wind power prediction method based on IGOA (optimized El-electric field model)
CN113468817B (en) * 2021-07-13 2023-08-22 淮阴工学院 Ultra-short-term wind power prediction method based on IGOA (insulated gate bipolar transistor) optimized ELM (ELM)
CN114169251A (en) * 2021-12-27 2022-03-11 广东工业大学 Ultra-short-term wind power prediction method
CN116388865A (en) * 2023-06-05 2023-07-04 深圳市飞思卓科技有限公司 PON optical module-based automatic screening method for abnormal optical power
CN116388865B (en) * 2023-06-05 2023-09-15 深圳市飞思卓科技有限公司 PON optical module-based automatic screening method for abnormal optical power

Similar Documents

Publication Publication Date Title
CN107516150A (en) A kind of Forecasting Methodology of short-term wind-electricity power, apparatus and system
CN107392397A (en) A kind of short-term wind speed forecasting method, apparatus and system
Cong et al. Traffic flow forecasting by a least squares support vector machine with a fruit fly optimization algorithm
CN105354646B (en) Power load forecasting method for hybrid particle swarm optimization and extreme learning machine
CN105449713B (en) Consider the intelligent Sofe Switch planing method of active power distribution network of distributed power source characteristic
CN102567391B (en) Method and device for building classification forecasting mixed model
CN102819772B (en) Power matching network builds material requirements Forecasting Methodology and device
CN105868853B (en) Method for predicting short-term wind power combination probability
CN108985521A (en) Output power predicting method, device, equipment and the medium of photovoltaic generating system
CN104217258B (en) A kind of electric load sigma-t Forecasting Methodology
CN102831489B (en) Power matching network builds material requirements Forecasting Methodology and device
CN107274038A (en) A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion
CN104951834A (en) LSSVM (least squares support vector machine) wind speed forecasting method based on integration of GA (genetic algorithm) and PSO (particle swarm optimization)
CN109215344B (en) Method and system for urban road short-time traffic flow prediction
CN107506861A (en) A kind of short-term wind speed forecasting method, apparatus and system
CN106680669A (en) Power distribution network fault section positioning method based on binary mixing algorithm
CN106779198A (en) A kind of congestion in road situation analysis method
CN107317338A (en) The optimal load flow computational methods and device of a kind of power system
CN106355294A (en) Electric vehicle charging station site selection and volume fixing method for large-scale complex power distribution network
CN109272139A (en) It is a kind of based on Nonlinear Set at the short-term wind speed forecasting method of deep learning
CN105574586B (en) General-purpose aircraft boat material needing forecasting method based on MPSO-BP networks
CN107292462A (en) A kind of short-term load forecasting method, apparatus and system
CN109754168A (en) Charging station site selecting method and device
CN106327357A (en) Load identification method based on improved probabilistic neural network
CN106529732A (en) Carbon emission efficiency prediction method based on neural network and random frontier analysis

Legal Events

Date Code Title Description
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: 20171226