CN113343562A - Fan power prediction method and system based on hybrid modeling strategy - Google Patents

Fan power prediction method and system based on hybrid modeling strategy Download PDF

Info

Publication number
CN113343562A
CN113343562A CN202110576264.1A CN202110576264A CN113343562A CN 113343562 A CN113343562 A CN 113343562A CN 202110576264 A CN202110576264 A CN 202110576264A CN 113343562 A CN113343562 A CN 113343562A
Authority
CN
China
Prior art keywords
power
fan
model
power prediction
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110576264.1A
Other languages
Chinese (zh)
Other versions
CN113343562B (en
Inventor
吴磊
于建成
董潇健
沈佳妮
贺益君
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.)
Shanghai Jiaotong University
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
Original Assignee
Shanghai Jiaotong University
State Grid Tianjin Electric Power Co Ltd
Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd
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 Shanghai Jiaotong University, State Grid Tianjin Electric Power Co Ltd, Electric Power Research Institute of State Grid Tianjin Electric Power Co Ltd filed Critical Shanghai Jiaotong University
Priority to CN202110576264.1A priority Critical patent/CN113343562B/en
Publication of CN113343562A publication Critical patent/CN113343562A/en
Application granted granted Critical
Publication of CN113343562B publication Critical patent/CN113343562B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/06—Wind turbines or wind farms
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A30/00—Adapting or protecting infrastructure or their operation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
    • Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Evolutionary Computation (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Computer Hardware Design (AREA)
  • General Business, Economics & Management (AREA)
  • Geometry (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Mathematical Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Algebra (AREA)
  • Computing Systems (AREA)
  • Fluid Mechanics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Wind Motors (AREA)

Abstract

The invention relates to a method and a system for predicting fan power based on a hybrid modeling strategy, which comprises the following steps: acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system, loading the weather forecast data and the data into a fan power prediction model, and acquiring a fan power prediction result; the fan power prediction model comprises a wind speed-power curve model, and the power conversion coefficient of the fan power prediction model is obtained through an artificial neural network model; the training process of the fan power prediction model comprises the following steps: acquiring training data; optimizing parameters in the artificial neural network model, and verifying the prediction performance of the fan power prediction model; and testing, correcting and updating the predicted performance of the generated power of the fan under different time scales by adopting different input parameters. Compared with the prior art, the method has the advantages of high model precision, low model complexity, strong generalization capability, capability of being used for multi-time scale power prediction and the like, and can effectively ensure the safe, stable and efficient operation of the wind power generation system.

Description

Fan power prediction method and system based on hybrid modeling strategy
Technical Field
The invention relates to the field of fan power prediction, in particular to a fan power prediction method and system based on a hybrid modeling strategy.
Background
Wind power generation is a novel power generation mode with great development potential and wide commercial application prospect. However, the wind power generation capability highly depends on meteorological conditions, and the wind power generation capability has extremely strong uncertainty and randomness, and the operation reliability of the wind power grid-connected system is seriously influenced. Therefore, an accurate and reliable wind turbine power generation prediction model needs to be constructed, and a reasonable scheduling and control scheme is formulated in advance on the basis, so that the safe, stable and efficient operation of the system is guaranteed.
Currently, proposed power prediction models for wind power generation systems can be classified into a physical model method, a wind speed-power curve method, a data-driven method, and the like. However, these methods generally have the following problems:
1) the physical model method needs to construct a complex fluid mechanics and thermodynamic equation set, has high computational complexity, is multipurpose for the design process of the fan, and is less applied to operation and scheduling scenes.
2) The wind speed-power curve method is widely applied to medium and long term prediction of the generated power of the fan by constructing a functional relation between the wind speed and the power, but the accuracy of short-term prediction is not high due to the fact that a model is simple.
3) The data driving method constructs the nonlinear relation between meteorological conditions and wind power based on historical data, and shows good performance in short-term prediction of the generated power of the fan, but the model is highly dependent on the meteorological conditions, and the data is relatively lack in medium-long term prediction, so that the practicability in medium-long term prediction is low.
Therefore, the method and the system for predicting the fan power, which can be applied to different time scales, are developed, and have important significance for realizing efficient and safe operation of a fan system.
Disclosure of Invention
The invention aims to provide a fan power prediction method and system based on a hybrid modeling strategy to overcome the defects of low fan power prediction precision and poor adaptability of multi-time scale scenes in the prior art.
The purpose of the invention can be realized by the following technical scheme:
a fan power prediction method based on a hybrid modeling strategy is characterized in that based on an existing common wind speed-power curve model, a power conversion coefficient item in the model is corrected by adopting an Artificial Neural Network (ANN), a power prediction model based on the hybrid modeling strategy is constructed, model parameters are optimized based on historical data, and the prediction performance of the model in different time scales is improved.
Specifically, the method comprises the steps of:
acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system, loading the weather forecast data and the data into a pre-established and trained fan power prediction model, and acquiring a fan power prediction result;
the fan power prediction model comprises a wind speed-power curve model, and a power conversion coefficient in the wind speed-power curve model is obtained through an artificial neural network model;
the training process of the fan power prediction model comprises the following steps:
a1: acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system as a historical database for model training and testing;
a2: optimizing parameters in the artificial neural network model based on the historical database in the step A1, and verifying the prediction performance of the fan power prediction model;
a3: and B, testing the fan power generation power prediction performance under different time scales by adopting different input parameters based on the fan power prediction model obtained in the step A2, and periodically correcting and updating parameters of the model according to a system real-time data updating historical database.
Further, the expression of the wind speed-power curve model is as follows:
Figure BDA0003084482470000021
Cp＝f(*)
where P is the output power, ρ is the air density, R is the wind turbine blade radius, CpThe wind power generation system comprises a power conversion coefficient, v is a wind speed, and f is an artificial neural network model for correlating the power conversion coefficient with meteorological conditions and fan operating conditions, wherein the meteorological conditions are obtained by weather forecast data, and the fan operating conditions are obtained by data of a wind power plant data acquisition and monitoring system.
Further, the calculation expression of the air density is as follows:
Figure BDA0003084482470000022
wherein Pa is air pressure, and the weather forecast data includes air pressure data, RgIs an ideal gas constant.
Further, the weather forecast data comprises wind speed, ambient temperature, humidity and air pressure, and the data of the wind power plant data acquisition and monitoring system comprises output power, blade pitch angle and fan rotation speed.
Further, the artificial neural network model is a radial basis function neural network model, and the expression of the selected radial basis function neural network model is as follows:
Figure BDA0003084482470000031
in the formula, CpAs power conversion factor, xkInput layer neuron of RBFNN, h is hidden layer node number, ciIs the position of the center point of the hidden layer node, sigma is the width of the kernel, betaiAs output layer weights, β0The output layer is biased.
Further, parameters in the artificial neural network model are optimized, specifically:
selecting the position of a center point of a hidden layer node based on a K-means clustering method;
optimizing the width of the kernel based on a maximum distance method;
optimizing the number of hidden layer nodes and a weight matrix based on a LASSO regression model;
and optimizing the weight matrix based on a BARON global algorithm.
Further, the calculation expression of the optimal kernel width based on the maximum distance method is as follows:
Figure BDA0003084482470000032
where σ is the kernel width, h is the number of hidden layer nodes, dmaxM is the maximum distance of the center point, and m is the number of types of input parameters.
Further, the LASSO regression model based optimization hidden layer node number and the weight matrix are optimized, an L1 regularization term with a penalty term is specifically adopted to correct the optimization target of the wind turbine power prediction model,
the expression of the optimization target of the fan power prediction model is as follows:
Figure BDA0003084482470000033
in the formula, RMSE is the root mean square error between the predicted power value and the actual measured value, lambda is the regularization parameter, h is the number of nodes in the hidden layer, betaiIs a weight matrix.
Further, in the process of verifying the prediction performance of the fan power prediction model, the root mean square error between the power prediction value and the actual measurement value is used as an evaluation index, and the calculation expression of the root mean square error between the power prediction value and the actual measurement value is as follows:
Figure BDA0003084482470000034
in the formula, RMSE is the root mean square error between the predicted power value and the actual measured value, K is the number of sampling points of the historical database, and Ppre,kIs a predicted value, P, of the fan generating power of the kth sampling pointmeas,kAnd measuring the generated power of the fan at the kth sampling point.
The invention also provides a wind turbine power prediction system based on the hybrid modeling strategy, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of the method.
Compared with the prior art, the fan power prediction method and system based on the hybrid modeling strategy, provided by the invention, couple two fan power prediction models, namely the artificial neural network model and the wind speed-power curve model, so that the prediction precision of the fan power generation power under different time scales is improved, and model parameters are updated by regularly acquiring real-time data of the NWP and SCADA systems, so that the long-term stability and reliability of the model prediction precision are ensured, and the efficient and safe operation of a fan grid-connected system is facilitated.
Drawings
Fig. 1 is a schematic flow chart of a method for predicting a wind turbine power based on a hybrid modeling strategy provided in an embodiment of the present invention;
FIG. 2 is a schematic diagram of a wind speed-power curve of a typical wind turbine provided in an embodiment of the present invention;
fig. 3 is a schematic diagram of a typical radial basis function neural network model provided in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The embodiment provides a wind turbine power prediction method based on a hybrid modeling strategy, which comprises the following steps:
acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system, loading the weather forecast data and the data into a pre-established and trained fan power prediction model, and acquiring a fan power prediction result;
the fan power prediction model comprises a wind speed-power curve model, and a power conversion coefficient in the wind speed-power curve model is obtained through an artificial neural network model;
the training process of the fan power prediction model comprises the following steps:
a1: acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system as a historical database for model training and testing;
a2: optimizing parameters in the artificial neural network model based on the historical database in the step A1, and verifying the prediction performance of the fan power prediction model;
a3: and B, testing the fan power generation power prediction performance under different time scales by adopting different input parameters based on the fan power prediction model obtained in the step A2, and periodically correcting and updating parameters of the model according to a system real-time data updating historical database.
Referring to fig. 1, in specific implementation, the method for predicting the fan power of the embodiment includes the following steps:
and S1, acquiring NWP and SCADA data of the location of the fan to be predicted within a certain time, and using the NWP and SCADA data as a historical database for model training and testing.
In the present embodiment, the collected NWP database includes data sets including wind speed (v), ambient temperature (T), humidity (RH), air pressure (Pa), and the like; the collected SCADA database contains data sets including output power (P), blade pitch angle (β), wind turbine rotational speed (n), and the like. The collected data range is all historical data of stable operation of the fan and power output.
And S2, constructing a fan power prediction model based on the artificial neural network model and the wind speed-power curve model.
Wind speed is an important influencing factor affecting the power of a fan. A typical wind speed-power curve for a wind turbine is shown in FIG. 2, where v isci,vcoAnd v andRrespectively setting the cut-in wind speed, the cut-out wind speed and the rated wind speed of the fan; p and PRThe actual output power and the rated power of the fan are respectively. It can be seen that when the real-time wind speed v is less than the cut-in wind speed, the fan does not work, and the output power is 0; when the actual wind speed v is greater than the cut-in wind speed and less than the rated wind speed, the output power of the fan is less than the rated power and is approximately cubic relation with the wind speed; when the actual wind speed v is greater than the rated wind speed and less than the cut-out wind speed, the fan runs at full power; and when the actual wind speed v is greater than the cut-out wind speed, the maximum wind speed which can be borne by the fan is exceeded, the fan stops working, and the output power is 0. How to construct [ v ] can be obtained as described aboveci,vR]The relation between the generated power and the wind speed in the interval is the key for realizing the accurate prediction of the generated power of the fan.
And (4) a fan power prediction model. The model structure is as follows:
Figure BDA0003084482470000051
wherein ρ is an air density; r is the wind turbine blade radius; cpIs a power conversion coefficient; f () is an artificial neural network model used to correlate power conversion coefficients with meteorological conditions, wind turbine operating conditions.
Preferably, in the step S2, the artificial Neural Network model is a Radial Basis Function Neural Network (RBFNN) model. The power conversion coefficient correlation formula based on the RBFNN is as follows:
Figure BDA0003084482470000061
in the formula, xkIs an input layer neuron of RBFNN; h. c. CiAnd σ is the number of hidden layer nodes, the position of the hidden layer node center point and the kernel width, respectively, βiAnd beta0Output layer weights and offsets, respectively.
The wind speed-power curve model shown in the formula (1) is characterized in that the air density rho is used for representing the wind energy of the airflow vertically passing through the fan blade surface of a unit area, and has a significant correlation with the altitude of the fan. The calculation formula of the air density ρ in the present embodiment is as follows:
Figure BDA0003084482470000062
in the formula, RgIs an ideal gas constant.
Because partial resistance exists in the wind wheel, the wind energy is difficult to be completely converted into electric energy, and the power conversion coefficient C ispFor indicating the conversion loss of wind energy. The fan power conversion factor is related to operating parameters of the fan, such as Tip Speed Ratio (TSR), blade pitch angle, and the like. In addition to local meteorological conditions, geographical factors, wind of a wind farmThe actual wind speed and the final output power of the fan are affected by the machine arrangement mode, the wake effect, the turbulence effect and the like. In the present model, the influence of different operating parameters and weather condition parameters is described as the influence on the power conversion coefficient, and a hybrid model as shown in formula (1) is constructed, while other influence factors not considered are corrected by the bias term in formula (2). In this embodiment, the operation parameters and weather condition parameters considered include: ambient temperature, humidity, air pressure, blade pitch angle, fan rotational speed. The fan rotating speed is related to the blade tip speed ratio, and the relationship is as follows:
Figure BDA0003084482470000063
in this embodiment, the relationship between the power conversion coefficient and the meteorological and operational parameters is associated by a radial basis function neural network model, and the association formula is shown in formula (2). A typical radial basis function neural network model is schematically illustrated in fig. 3.
S3, optimizing parameters in the artificial neural network model based on the historical database of the step S1, and verifying the power prediction performance of the model.
In order to verify the prediction performance of the proposed wind turbine power prediction method based on the hybrid modeling strategy, the present embodiment verifies the model performance by using a ten-fold cross-validation method. Namely, the historical data obtained in step S1 is randomly divided into 10 groups, wherein 9 groups are used for the training set of the model, and the remaining 1 group is used as the test set of the model, and the process is repeated 10 times. In this embodiment, a Root Mean Square Error (RMSE) between a power predicted value obtained by 10 tests and an actual measured value is used as an evaluation index of model performance, and a formula thereof is as follows:
Figure BDA0003084482470000071
where K is the number of sampling points in the historical database, Ppre,kAnd Pmeas,kAre respectively the kth sampleAnd (4) a predicted value and a measured value of the point fan generated power.
Model parameters in the RBFNN model in the formula (2) are numerous and have a significant influence on the model prediction accuracy. The embodiment adopts different methods to align the position c of the central pointiKernel width σ, number of hidden layer nodes h, and weight matrix βiAnd the like are preferable. Firstly, selecting the position of a center point of a hidden layer node based on a K-means clustering method; secondly, based on the maximum distance method, the kernel width is optimized, and the formula is as follows:
Figure BDA0003084482470000072
in the formula (d)maxThe maximum distance from the center point, m is the number of types of input parameters, and m is 5 in this embodiment.
Then, optimizing the number of hidden layer nodes and the weight matrix based on the LASSO regression model, and specifically adopting an L1 regularization term with a penalty term to correct the optimization target of the model, which is specifically as follows:
Figure BDA0003084482470000073
in the formula, λ is a regularization parameter.
In this embodiment, based on the test set selected by the cross validation method, the weight matrix is optimized based on the BARON global algorithm with the optimization objective shown in the above formula (7).
And S4, testing the predicted performance of the generated power of the fan under different time scales by adopting different input parameters based on the model obtained in the step S3, and periodically correcting and updating the model according to a real-time data updating database of the NWP and SCADA systems.
In this embodiment, the selection of the input parameters under different time scales is specifically as follows:
1) aiming at long-term power generation power prediction, meteorological condition parameters adopt the average value of recent meteorological conditions, and fan operating condition parameters adopt the average design value;
2) aiming at medium-term power generation power prediction, adopting a predicted value of an NWP (non-Newton P) system for meteorological condition parameters, and adopting an average design value for fan operation condition parameters;
3) and aiming at the short-term power generation power prediction, the meteorological condition and fan operation condition parameters adopt the predicted values of a given time scale.
The embodiment also provides a wind turbine power prediction system based on the hybrid modeling strategy, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of the wind turbine power prediction method based on the hybrid modeling strategy.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.

Claims (10)

1. A wind turbine power prediction method based on a hybrid modeling strategy is characterized by comprising the following steps:
acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system, loading the weather forecast data and the data into a pre-established and trained fan power prediction model, and acquiring a fan power prediction result;
the fan power prediction model comprises a wind speed-power curve model, and a power conversion coefficient in the wind speed-power curve model is obtained through an artificial neural network model;
the training process of the fan power prediction model comprises the following steps:
a1: acquiring weather forecast data of the location of a fan to be tested and data of a wind power plant data acquisition and monitoring system as a historical database for model training and testing;
a2: optimizing parameters in the artificial neural network model based on the historical database in the step A1, and verifying the prediction performance of the fan power prediction model;
a3: and B, testing the fan power generation power prediction performance under different time scales by adopting different input parameters based on the fan power prediction model obtained in the step A2, and periodically correcting and updating parameters of the model according to a system real-time data updating historical database.
2. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 1, wherein the expression of the wind speed-power curve model is as follows:
Figure FDA0003084482460000011
Cp＝f(*)
where P is the output power, ρ is the air density, R is the wind turbine blade radius, CpThe wind power generation system comprises a power conversion coefficient, v is a wind speed, and f is an artificial neural network model for correlating the power conversion coefficient with meteorological conditions and fan operating conditions, wherein the meteorological conditions are obtained by weather forecast data, and the fan operating conditions are obtained by data of a wind power plant data acquisition and monitoring system.
3. The method of claim 2, wherein the air density is calculated by the following expression:
Figure FDA0003084482460000012
wherein Pa is air pressure, and the weather forecast data includes air pressure data, RgIs an ideal gas constant.
4. The method of claim 1, wherein the weather forecast data includes wind speed, ambient temperature, humidity, and barometric pressure, and the wind farm data collection and monitoring system data includes output power, blade pitch angle, and wind turbine rotational speed.
5. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 1, wherein the artificial neural network model is a radial basis function neural network model, and the expression of the radial basis function neural network model is as follows:
Figure FDA0003084482460000021
in the formula, CpAs power conversion factor, xkInput layer neuron of RBFNN, h is hidden layer node number, ciIs the position of the center point of the hidden layer node, sigma is the width of the kernel, betaiAs output layer weights, β0The output layer is biased.
6. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 5, wherein parameters in the artificial neural network model are optimized, specifically:
selecting the position of a center point of a hidden layer node based on a K-means clustering method;
optimizing the width of the kernel based on a maximum distance method;
optimizing the number of hidden layer nodes and a weight matrix based on a LASSO regression model;
and optimizing the weight matrix based on a BARON global algorithm.
7. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 6, wherein the calculation expression of the optimal kernel width based on the maximum distance method is as follows:
Figure FDA0003084482460000022
where σ is the kernel width, h is the number of hidden layer nodes, dmaxM is the maximum distance of the center point, and m is the number of types of input parameters.
8. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 6, wherein the LASSO regression model is used to optimize the number of hidden layer nodes and the weight matrix, and an L1 regularization term with a penalty term is used to correct the optimization objective of the wind turbine power prediction model,
the expression of the optimization target of the fan power prediction model is as follows:
Figure FDA0003084482460000023
in the formula, RMSE is the root mean square error between the predicted power value and the actual measured value, lambda is the regularization parameter, h is the number of nodes in the hidden layer, betaiIs a weight matrix.
9. The wind turbine power prediction method based on the hybrid modeling strategy according to claim 1, wherein in the process of verifying the prediction performance of the wind turbine power prediction model, a root mean square error between a power prediction value and an actual measurement value is used as an evaluation index, and a calculation expression of the root mean square error between the power prediction value and the actual measurement value is as follows:
Figure FDA0003084482460000031
in the formula, RMSE is the root mean square error between the predicted power value and the actual measured value, K is the number of sampling points of the historical database, and Ppre,kIs a predicted value, P, of the fan generating power of the kth sampling pointmeas,kAnd measuring the generated power of the fan at the kth sampling point.
10. A wind turbine power prediction system based on a hybrid modeling strategy, comprising a memory storing a computer program and a processor invoking the computer program to perform the steps of the method according to any one of claims 1 to 9.
CN202110576264.1A 2021-05-26 2021-05-26 Fan power prediction method and system based on hybrid modeling strategy Active CN113343562B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110576264.1A CN113343562B (en) 2021-05-26 2021-05-26 Fan power prediction method and system based on hybrid modeling strategy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110576264.1A CN113343562B (en) 2021-05-26 2021-05-26 Fan power prediction method and system based on hybrid modeling strategy

Publications (2)

Publication Number Publication Date
CN113343562A true CN113343562A (en) 2021-09-03
CN113343562B CN113343562B (en) 2023-04-07

Family

ID=77471508

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110576264.1A Active CN113343562B (en) 2021-05-26 2021-05-26 Fan power prediction method and system based on hybrid modeling strategy

Country Status (1)

Country Link
CN (1) CN113343562B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114139791A (en) * 2021-11-24 2022-03-04 北京华能新锐控制技术有限公司 Wind generating set power prediction method, system, terminal and storage medium
CN114718808A (en) * 2022-03-31 2022-07-08 广西电网有限责任公司电力科学研究院 Fan control method and system

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102880810A (en) * 2012-10-25 2013-01-16 山东电力集团公司电力科学研究院 Wind power prediction method based on time sequence and neural network method
CN103323772A (en) * 2012-03-21 2013-09-25 北京光耀能源技术股份有限公司 Wind driven generator operation state analyzing method based on neural network model
US20140195159A1 (en) * 2013-01-09 2014-07-10 Iteris, Inc. Application of artificial intelligence techniques and statistical ensembling to forecast power output of a wind energy facility
CN107133695A (en) * 2017-04-28 2017-09-05 北京天诚同创电气有限公司 A kind of wind power forecasting method and system
CN109242143A (en) * 2018-07-31 2019-01-18 中国电力科学研究院有限公司 A kind of neural network wind power forecasting method and system
CN109978280A (en) * 2019-04-19 2019-07-05 上海交通大学 A kind of generalization photovoltaic cell operating temperature prediction technique and device
CN110070223A (en) * 2019-04-19 2019-07-30 中能电力科技开发有限公司 A kind of short term power prediction technique applied to newly-built wind power plant

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103323772A (en) * 2012-03-21 2013-09-25 北京光耀能源技术股份有限公司 Wind driven generator operation state analyzing method based on neural network model
CN102880810A (en) * 2012-10-25 2013-01-16 山东电力集团公司电力科学研究院 Wind power prediction method based on time sequence and neural network method
US20140195159A1 (en) * 2013-01-09 2014-07-10 Iteris, Inc. Application of artificial intelligence techniques and statistical ensembling to forecast power output of a wind energy facility
CN107133695A (en) * 2017-04-28 2017-09-05 北京天诚同创电气有限公司 A kind of wind power forecasting method and system
CN109242143A (en) * 2018-07-31 2019-01-18 中国电力科学研究院有限公司 A kind of neural network wind power forecasting method and system
CN109978280A (en) * 2019-04-19 2019-07-05 上海交通大学 A kind of generalization photovoltaic cell operating temperature prediction technique and device
CN110070223A (en) * 2019-04-19 2019-07-30 中能电力科技开发有限公司 A kind of short term power prediction technique applied to newly-built wind power plant

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114139791A (en) * 2021-11-24 2022-03-04 北京华能新锐控制技术有限公司 Wind generating set power prediction method, system, terminal and storage medium
CN114718808A (en) * 2022-03-31 2022-07-08 广西电网有限责任公司电力科学研究院 Fan control method and system

Also Published As

Publication number Publication date
CN113343562B (en) 2023-04-07

Similar Documents

Publication Publication Date Title
CN104021424B (en) Method and apparatus for predicting the output power of the blower in wind field
CN113343562B (en) Fan power prediction method and system based on hybrid modeling strategy
Mabel et al. Estimation of energy yield from wind farms using artificial neural networks
CN107944175B (en) Method for acquiring real power curve of fan by considering wind turbulence intensity
CN109784563B (en) Ultra-short-term power prediction method based on virtual anemometer tower technology
Xu et al. Correlation based neuro-fuzzy Wiener type wind power forecasting model by using special separate signals
CN111340307B (en) Method for predicting wind power generation power of fan and related device
CN115358606B (en) Method and system for evaluating energy efficiency of in-service wind power plant under gentle terrain
CN113449847A (en) Offshore wind power rolling prediction method considering second-level time series wind speed change
Cao et al. Development of short-term wind power forecasting methods
CN116307282A (en) Wind power prediction method and system based on mechanism model and data driving
CN111090932A (en) On-site wind power plant field calibration method suitable for medium and complex terrain
CN115640737A (en) Wind power prediction method and system for abnormal weather state
CN115898787A (en) Method and device for dynamically identifying static yaw error of wind turbine generator
CN111738556B (en) Method for evaluating power generation capacity of fan based on machine head microclimate
Cascianelli et al. Experimental prediction intervals for monitoring wind turbines: An ensemble approach
CN110942218B (en) Method, device and system for calculating online electric quantity of wind power plant
CN113408076A (en) Small sample mechanical residual life prediction method based on support vector machine model
CN114298512A (en) Wind turbine generator set power characteristic analysis method
Duan et al. Study on performance evaluation and prediction of mixed-flow hydraulic turbine units under variable operating conditions
CN112070320A (en) Ultra-short-term wind power prediction method and system based on dynamic harmonic regression
Dai et al. Study on Obtaining Real Power Curve of Wind Turbines Using SCADA Data
CN113309661B (en) Method for improving generating capacity of wind generating set
CN116757101B (en) Cabin wind speed correction method and system based on mechanism model and neural network
CN117892099B (en) Wind turbine generator service efficiency evaluation method and system based on probabilistic neural network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant