CN103902813A - Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM - Google Patents

Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM Download PDF

Info

Publication number
CN103902813A
CN103902813A CN201410081740.2A CN201410081740A CN103902813A CN 103902813 A CN103902813 A CN 103902813A CN 201410081740 A CN201410081740 A CN 201410081740A CN 103902813 A CN103902813 A CN 103902813A
Authority
CN
China
Prior art keywords
model
steam
data
modeling
draught fan
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
CN201410081740.2A
Other languages
Chinese (zh)
Inventor
司风琪
邵壮
郭俊山
阎文生
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.)
CPI SHENTOU POWER GENERATION Co Ltd
Southeast University
Original Assignee
CPI SHENTOU POWER GENERATION Co Ltd
Southeast University
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 CPI SHENTOU POWER GENERATION Co Ltd, Southeast University filed Critical CPI SHENTOU POWER GENERATION Co Ltd
Priority to CN201410081740.2A priority Critical patent/CN103902813A/en
Publication of CN103902813A publication Critical patent/CN103902813A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Control Of Positive-Displacement Air Blowers (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a steam-driven draught fan full working condition online monitoring model modeling method based on a CPSO-LSSVM. The steam-driven draught fan full working condition online monitoring model modeling method based on the CPSO-LSSVM comprises the following steps that firstly, design data from a draught fan factory are analyzed and processed, and actual working condition points are selected as training data; a chaotic particle swarm optimization algorithm is used for optimizing modeling parameters and supporting modeling on a least square support vector machine to obtain a static model; a full working condition model of a steam-driven draught fan is obtained through training based on the existing static model by combining rotating speed variables; the established full working condition model is published in an online website mode by combing a webpage programming technique, so that the working points of the draught fan can be determined on line; finally, the actual operating data of the stem-driven draught fan are obtained by combining an SIS to carry out real-time online correction on the established model. Through the model obtained by using the method, the full working condition operating characteristics of the steam-driven draught fan can be accurately reflected, online correction can be achieved, the draught fan can operate correctly even after the characteristics of the draught fan are changed, and guidance is provided for actual operation.

Description

The full operating mode monitoring model online of steam-driven induced draft fan modeling method based on CPSO-LSSVM
Technical field
The full operating mode monitoring model online of a kind of steam-driven induced draft fan based on CPSO-LSSVM of the present invention modeling method, relates to supporting vector machine model, belongs to machine learning modeling field.
Background technology
Power Plant Feedwater pump adopts after Steam Turbine Driven, and induced draft fan has become the subsidiary engine of power consumption maximum.Under electric motor driven pattern, the separate unit id-motor maximum electric power of 1036MW unit reaches 7400kW, accounts for 1.48% of unit generated energy.And under motor drive mode, induced draft fan adopts stator blade to regulate, and motor power (output) is constant, in the time that unit load changes, the extra station service loss that motor causes is very large, and energy dissipation is serious.Adopting steam turbine to replace motor to drive induced draft fan is thoroughly to overcome the above problems splendid scheme.Adopt steam-driven induced draft fan scheme, the impact to station service electrical system electric motor starting can be reduced on the one hand time; In addition, thus adopting variable speed to regulate also can ensure higher efficiency when variable parameter operation.
Due to the regulative mode of steam-driven induced draft fan employing stator blade one rotating speed combination, its overall performance is than determining the many of rotating speed steam-driven induced draft fan complexity.Must be able to monitor the working point of steam-driven induced draft fan in order to make steam-driven induced draft fan efficient operation.But the steam-driven induced draft fan monitoring model online that can not use at present.Steam-driven induced draft fan family curve relates to the nonlinear model of specific pressure Y (p/ ρ), volume flow Q and 3 parameter complexity of blade angle, and traditional modeling method is difficult to accomplish determine exactly steam-driven induced draft fan working point; And for steam-driven induced draft fan, regulate complexity further to increase owing to having increased variable speed.
Summary of the invention
Goal of the invention: for the deficiencies in the prior art, the invention provides the full operating mode monitoring model online of a kind of steam-driven induced draft fan based on CPSO-LSSVM modeling method, realize the full operating mode on-line monitoring of steam-driven induced draft fan, ensured steam-driven induced draft fan efficient operation.
Technical scheme: in order to realize foregoing invention object, the technical solution used in the present invention is:
The full operating mode monitoring model online of a steam-driven induced draft fan modeling method based on CPSO-LSSVM, comprises the steps:
(1) data processing; The design data that blower fan producer is provided arranges calculating, choose and enough can determine the characteristic data of steam-driven induced draft fan, arrangement draws following parameters, specific pressure Y, volume flow Q, blade angle β and rotation speed of fan n, wherein specific pressure Y is the ratio of fluid total head p and fluid density ρ, i.e. Y=p/ ρ; In all design datas, choose in right amount as training data for modeling, all design datas are all used for test model accuracy;
(2) least square method supporting vector machine training modeling; Blade angle β and specific pressure Y are as the input of training pattern, and volume flow Q, as output, utilizes Chaos particle swarm optimization algorithm, set up about modeling method of least squares support parameter γ and σ 2optimizing population, utilize the individual γ of the blade angle β of training data and specific pressure Y and population and σ 2parameter is carried out Experiment Modeling, taking the error minimum of model delivery volume flow Q and actual delivery volume flow Qm as target, iterates and optimizes population structure, obtains optimum supporting vector machine model; Utilize test data detection model precision, obtain the standard compliant Q=f of precision (Y, β) static model;
(3) training pattern again; On the basis of Q=f (Y, β) static model, increase variable rotation speed of fan n, obtain the full condition model of Q=f (Y, β, n);
(4) issue model; The full condition model Q=f (Y, β, n) that utilizes training to obtain, in conjunction with ASP.NET webpage programming technique, issues institute's established model with the form of visual website, determine online steam-driven induced draft fan working point;
(5) correction model; Step (four) gained steam-driven induced draft fan operating point data and SIS system (Supervisory Information System inplant level) are gathered to actual operating data compares, if the error of existence, replaces design data to re-establish model according to above-mentioned steps one to step 4 by actual operating data.
Support vector machine (SVM) has obvious advantage solving in small sample, non-linear, higher-dimension problem, combine VC dimension theory and the structure risk minimum theoretical of statistical learning, utilize kernel function, sample vector is mapped to high-dimensional feature space, make former spatial data at higher dimensional space linear separability, and construct optimal classification face.The algorithm complex of support vector machine and sample dimension are irrelevant, and model is decided by minority support vector more, therefore strong robustness, and generalization ability is strong.Least square method supporting vector machine (LSSVM) is by choosing different loss functions, the quadratic programming problem of standard support vector machine is changed into the problem that solves linear equation, accelerate speed of convergence, be more applicable for engineering application, can realize better Real-time modeling set and model correction.Utilize Chaos particle swarm optimization algorithm (CPSO) to carry out modeling parameters optimizing, make model have stronger controllability, facilitate the later stage according to different objects, determine different modeling strategy, than other optimizing algorithms, speed of searching optimization and optimizing precision are greatly accelerated.
Use steam-driven induced draft fan design data training least square method supporting vector machine, after empirical tests, can have more intuitive understanding to the acceptable operating point of steam-driven induced draft fan, thereby the rotating speed to steam-driven induced draft fan, stator setting angles adjustment propose guiding opinion; In addition, in the time of steam-driven induced draft fan operation off-design operating mode, can from SIS system (Supervisory Information System in plant level), again choose data model is revised.
Beneficial effect: 1, the present invention utilizes least square method supporting vector machine to set up steam-driven induced draft fan model and can avoid the poor problem of Polynomial modeling accuracy in the past, realizes the accurate simulation of steam-driven induced draft fan working point.
2, the present invention has realized the full state modeling of steam-driven induced draft fan, provides good directive function for carrying out rotational speed regulation in steam-driven induced draft fan actual moving process.
3, owing to adopting machine learning modeling method, the foundation of model of the present invention and correction only depend on training data, in the time that fan operation characteristic changes, can realize real-time correction.
Brief description of the drawings
Fig. 1 is that volume flow Q-specific pressure Y is in different stator blade aperture β (75 ° to 30 °) performance curve group.
Fig. 2 is least square method supporting vector machine training modeling realization flow figure of the present invention.
Embodiment
Below in conjunction with embodiment, technical scheme of the present invention is elaborated; be noted that the following stated is only the preferred embodiment of the present invention; but protection scope of the present invention is not limited to embodiment; for those skilled in the art; under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Embodiment
Taking certain power station 660MW unit steam-driven induced draft fan monitoring model online as example, obtain the steam-driven induced draft fan operation characteristic data that blower fan producer provides, the data under the data while comprising each aperture of 995r/min and other rotating speeds of part.Modeling mainly comprise data processing, least square method supporting vector machine training modeling, again training pattern, issue the nucleus modules such as model and correction model.
(1) data processing; The 995r/min performance data that design producer is provided is screened to analyze and is chosen 538 data, the following parameters of specified data, specific pressure Y (p/ ρ, wherein p is fluid total head, ρ is fluid density), volume flow Q, blade angle β and rotation speed of fan n, choose that wherein 152 data are as training data, all 538 data are all as test data.
(2) least square method supporting vector machine training modeling, idiographic flow as illustrated in fig. 1 and 2,
A) initialization training parameter, blade angle β and specific pressure Y are as the input of training pattern, and volume flow Q is as the output of training pattern.
B) determine LSSVM correlation parameter, wherein selecting kernel function type is RBF kernel:
K ( x i , x j ) = e - | | x i - x j | | 2 σ 2 - - - ( 1 )
Wherein x i, x jbe two samples; σ 2for the variance of Gaussian kernel.
For modeling parameters γ and σ 2, utilize Chaos particle swarm optimization algorithm optimizing to determine: in a big way, set up primary group by the chaos factor, total number of particles is made as 50, and particle is bivector, stores modeling parameters γ and the σ of current particle 2, by model delivery volume flow Q mwith actual delivery volume flow Q error as fitness function
G = Σ k = 1 N ( Q - Q m ) 2 - - - ( 2 )
Wherein, Q mfor the model output taking 538 test datas as input, the actual output that Q is these test datas, N is data bulk 538, and G is fitness.
Taking this fitness function minimum as optimization aim, iterative formula is:
v i , j k + 2 = r k + 2 × ω k + 1 × v i , j k + c 1 × h 1 × ( pbest i , j k - x i , j k ) + c 2 × h 2 × ( gbest j k - x i , j k ) - - - ( 3 )
x i , j k + 2 = x i , j k + v i , j k + 2 - - - ( 4 )
Wherein
r k+2=4×r k×(1-r k)
ω k + 1 = ω max - ( ω max - ω min ) ITERA _ N × ( k + 1 )
Figure BDA0000473745000000045
for k is for particle position,
Figure BDA0000473745000000046
for k is for particle rapidity, r kfor k is for the chaos factor, ω kfor k is for particle, ω max, ω minbe respectively in the bound, this example of weight and get respectively 0.9,0.1, ITERA_N is total iterations, gets 1000, c in this example 1with c 2for the study factor, be all taken as in this example 2, h 1with h 2for the random number in (0,1) interval,
Figure BDA0000473745000000047
for arriving k on behalf of stopping, the optimum modeling parameter that each particle has searched, for arriving k on behalf of stopping, the optimum modeling parameter that all particles have searched.The output γ of global optimum and σ 2.
C) use training data training pattern, draw static model Q=f (Y, β); Under drafting rotating speed 995r/min, as shown in Figure 1, modeling parameters is [γ, σ to steam-driven induced draft fan performance curved surface 2]=[1213.27970.78760461].
D) testing model accuracy.For accuracy and the generalization ability of verification model, we have obtained 152 training data errors, choose again whole 538 test datas, obtain test data error.Error calculation formula is:
err = 1 N Σ i = 1 N ( Q - Q m ) 2 - - - ( 5 )
Wherein Q is experiment condition volume flow measured value, Q mfor the volume flow value that model calculates accordingly.
Be listed as follows:
? Training data error Test data error
CPSO-LSSVM 10.85342 10.50394
From upper table, we can find out, no matter are training data standard deviation or test data standard deviation, all smaller, can say that model still more accurately.In addition, test data error is even also smaller than training data error, illustrates that model generalization ability is also very strong.
(3) set up the full condition model of steam-driven induced draft fan; On the basis of step 2 gained static model Q=f (Y, β), provide the data of rotation speed of fan n and the data that part calculates according to blower fan scaling law, the training full condition model Q=f of steam-driven induced draft fan (Y, β, n) in conjunction with producer.
(4) issue model; The established model Q=f of step 3 institute (Y, β, n) is issued with form of websites in conjunction with programming technique, realize the on-line monitoring to steam-driven induced draft fan.
(5) correction model; The induced draft fan working point information obtaining in step 4 website is compared with the actual operating data in SIS system, if precision undesirable adopt actual operating data replace design data again Training Support Vector Machines revise.
So far,, from Data Management Analysis to CPSO-LSSVM modeling, finally built to the monitoring model of determining steam-driven induced draft fan working point.

Claims (2)

1. the full operating mode monitoring model online of the steam-driven induced draft fan based on a CPSO-LSSVM modeling method, is characterized in that, comprises the steps:
(1) data processing; The design data that blower fan producer is provided arranges calculating, choose and enough can determine the characteristic data of steam-driven induced draft fan, arrangement draws following parameters, specific pressure Y, volume flow Q, blade angle β and rotation speed of fan n, wherein specific pressure Y is the ratio of fluid total head p and fluid density ρ, i.e. Y=p/ ρ; In all design datas, choose in right amount as training data for modeling, all design datas are all used for test model accuracy;
(2) least square method supporting vector machine training modeling; Blade angle β and specific pressure Y are as the input of training pattern, and volume flow Q, as the output of training pattern, utilizes Chaos particle swarm optimization algorithm, set up about modeling method of least squares support parameter γ and σ 2optimizing population, utilize the individual γ of the blade angle β of training data and specific pressure Y and population and σ 2parameter is carried out Experiment Modeling, taking the error minimum of model delivery volume flow Q and actual delivery volume flow Qm as target, iterate and optimize population structure, obtain optimum supporting vector machine model, utilize test data detection model precision, obtain the standard compliant Q=f of precision (Y, β) static model;
(3) training pattern again; On the basis of Q=f (Y, β) static model, increase variable rotation speed of fan n, obtain the full condition model of Q=f (Y, β, n);
(4) issue model; The full condition model Q=f (Y, β, n) that utilizes training to obtain, in conjunction with ASP.NET webpage programming technique, issues institute's established model with the form of visual website, determine online steam-driven induced draft fan working point;
(5) correction model; Step (four) gained steam-driven induced draft fan operating point data are compared with SIS system acquisition actual operating data, if there is larger error, replace design data by actual operating data, re-establish model according to above-mentioned steps one to step 4.
2. the full operating mode monitoring model online of the steam-driven induced draft fan based on CPSO-LSSVM according to claim 1 modeling method, is characterized in that, described step (2) comprises the steps:
A) initialization training parameter, blade angle β and specific pressure Y are as the input of training pattern, and volume flow Q is as the output of training pattern;
B) determine LSSVM correlation parameter, wherein selecting kernel function type is RBF kemel:
K ( x i , x j ) = e - | | x i - x j | | 2 σ 2 - - - ( 1 ) Wherein x i, x ibe two samples; σ 2for the variance of Gaussian kernel.
For modeling parameters γ and σ 2, utilize Chaos particle swarm optimization algorithm optimizing to determine: in a big way, set up primary group by the chaos factor, particle is bivector, stores modeling parameters γ and the σ of current particle 2, by model delivery volume flow Q mwith the error of actual delivery volume flow Q as fitness function
G = Σ k = 1 N ( Q - Q m ) 2 - - - ( 2 )
Wherein, Q mfor the model output taking test data as input, the actual output that Q is these test datas, N is data bulk, and G is fitness;
Taking this fitness function minimum as optimization aim, iterative formula is:
v i , j k + 2 = r k + 2 × ω k + 1 × v i , j k + c 1 × h 1 × ( pbest i , j k - x i , j k ) + c 2 × h 2 × ( gbest j k - x i , j k ) - - - ( 3 )
x i , j k + 2 = x i , j k + v i , j k + 2 - - - ( 4 ) Wherein
r k+2=4×r k×(1-r k)
ω k + 1 = ω max - ( ω max - ω min ) ITERA _ N × ( k + 1 )
Figure FDA0000473744990000025
for k is for particle position,
Figure FDA0000473744990000026
for k is for particle rapidity, r kfor k is for the chaos factor, ω kfor k is for particle, ω max, ω minbe respectively weight factor bound, get respectively 0.9,0.1 in this example, ITERA_N is total iterations, c 1with c 2for the study factor, h 1with h 2for the random number in (0,1) interval, for arriving k on behalf of stopping, the optimum modeling parameter that each particle has searched,
Figure FDA0000473744990000028
for arriving k on behalf of stopping, the optimum modeling parameter that all particles have searched, the output γ of global optimum and σ 2;
C) use training data training pattern, draw static model Q=f (Y, β);
D) testing model accuracy: for accuracy and the generalization ability of verification model, obtain test data error, error calculation formula is:
err = 1 N Σ i = 1 N ( Q - Q m ) 2 - - - ( 5 )
Wherein Q is experiment condition volume flow measured value, Q mfor the volume flow value that model calculates accordingly.
CN201410081740.2A 2014-03-06 2014-03-06 Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM Pending CN103902813A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410081740.2A CN103902813A (en) 2014-03-06 2014-03-06 Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410081740.2A CN103902813A (en) 2014-03-06 2014-03-06 Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM

Publications (1)

Publication Number Publication Date
CN103902813A true CN103902813A (en) 2014-07-02

Family

ID=50994130

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410081740.2A Pending CN103902813A (en) 2014-03-06 2014-03-06 Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM

Country Status (1)

Country Link
CN (1) CN103902813A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104899665A (en) * 2015-06-19 2015-09-09 国网四川省电力公司经济技术研究院 Wind power short-term prediction method
CN106777642A (en) * 2016-12-08 2017-05-31 南京航空航天大学 A kind of Forecasting Methodology of film cooling structure discharge coefficient
CN108647478A (en) * 2018-06-29 2018-10-12 国网山东省电力公司电力科学研究院 Cogeneration units SCR inlet smoke temperature on-line calculation method based on BP neural network
CN109765331A (en) * 2018-10-08 2019-05-17 陕西科技大学 A kind of oxygen content of smoke gas hard measurement system based on least square method supporting vector machine
CN110163269A (en) * 2019-05-09 2019-08-23 北京迈格威科技有限公司 Model generating method, device and computer equipment based on deep learning

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101719230A (en) * 2009-12-18 2010-06-02 东南大学 Method for optimizing main steam pressure of large steam turbine set

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101719230A (en) * 2009-12-18 2010-06-02 东南大学 Method for optimizing main steam pressure of large steam turbine set

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
QI WU: "A hybrid-forecasting model based on Gaussian support vector machine and chaotic particle swarm optimization", 《EXPERT SYSTEMS WITH APPLICATIONS》 *
YONG FENG ET AL.: "COMPARING WITH CHAOTIC INERTIA WEIGHTS IN PARTICLE SWARM OPTIMIZATION", 《PROCEEDINGS OF THE SIXTH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS》 *
朱予东 等: "基于NWAPSO参数全局寻优的LS-SVM电站风机性能预测", 《应用能源技术》 *
王松岭 等: "基于改进支持向量机的风机性能在线监测平台研究", 《流体机械》 *
马仪炜: "基于神经网络的轴流风机建模方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
高异 等: "基于粒子群优化LS-SVM的裂纹图像检测", 《PROCEEDINGS OF THE 32ND CHINESE CONTROL CONFERENCE》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104899665A (en) * 2015-06-19 2015-09-09 国网四川省电力公司经济技术研究院 Wind power short-term prediction method
CN106777642A (en) * 2016-12-08 2017-05-31 南京航空航天大学 A kind of Forecasting Methodology of film cooling structure discharge coefficient
CN106777642B (en) * 2016-12-08 2018-05-08 南京航空航天大学 A kind of Forecasting Methodology of film cooling structure discharge coefficient
CN108647478A (en) * 2018-06-29 2018-10-12 国网山东省电力公司电力科学研究院 Cogeneration units SCR inlet smoke temperature on-line calculation method based on BP neural network
CN108647478B (en) * 2018-06-29 2022-09-20 国网山东省电力公司电力科学研究院 BP neural network-based SCR inlet smoke temperature online calculation method for cogeneration unit
CN109765331A (en) * 2018-10-08 2019-05-17 陕西科技大学 A kind of oxygen content of smoke gas hard measurement system based on least square method supporting vector machine
CN110163269A (en) * 2019-05-09 2019-08-23 北京迈格威科技有限公司 Model generating method, device and computer equipment based on deep learning

Similar Documents

Publication Publication Date Title
CN103902813A (en) Steam-driven draught fan full working condition online monitoring model modeling method based on CPSO-LSSVM
CN103440528B (en) Thermal power unit operation optimization method and device based on power consumption analysis
CN103887815A (en) Wind power plant parameter identification and dynamic equivalence method based on operation data
Liu et al. Wind farm macro-siting optimization with insightful bi-criteria identification and relocation mechanism in genetic algorithm
CN104036073B (en) Double-fed wind power plant dynamic equivalence modeling method suitable for active power characteristic analysis
CN109670213B (en) Multi-machine equivalent modeling method and device for wind power plant
CN103219725A (en) Wind power plant equivalent modeling method based on real-time operation data
CN109740962A (en) Voltage stabilization probability evaluation method of failure based on scene subregion and cumulant
CN105184683A (en) Probability clustering method based on wind electric field operation data
CN106684905A (en) Wind power plant dynamic equivalence method with wind power forecast uncertainty considered
Deng et al. Whole flow field performance prediction by impeller parameters of centrifugal pumps using support vector regression
CN100535911C (en) Ventilator energy efficiency optimized design method based on systematic mode identification technology
CN108710764A (en) A kind of double entry pump multi-objective optimization design of power method based on mixing approximate model
CN105279384A (en) Wind turbine cabin wind speed-based method and device for calculating wind speed of incoming flow
CN109921426A (en) Wind-electricity integration system probability load flow calculation method based on CV-KDE
CN116771610A (en) Method for adjusting fault evaluation value of variable pitch system of wind turbine
CN103903087A (en) Steam-driven induced draft fan all-working-condition online monitoring method based on BP neural network
CN108446452A (en) A kind of mixed-flow pump impeller Robust Optimal Design
CN116292057A (en) Control method, device, equipment and medium for variable-speed pumped storage unit
CN114511158B (en) Wind turbine power prediction method based on wake deflection effect and 2DJensen model
CN116595874A (en) Impeller mechanical performance prediction model parameter optimization method and device and storage medium
Shen et al. Coordinated optimal control of active power of wind farms considering wake effect
CN116316629A (en) Data-driven optimal power flow calculation method considering topology feature learning
CN109388884A (en) A kind of generalized regression extreme value response phase method calculating coupling leaf dish fatigue life
CN109766611A (en) A kind of simplified prediction technique of the wind power considering terrain slope

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into 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: 20140702