CN116757094B - Wind turbine wake field calculation method and device, electronic equipment and storage medium - Google Patents
Wind turbine wake field calculation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN116757094B CN116757094B CN202311008152.1A CN202311008152A CN116757094B CN 116757094 B CN116757094 B CN 116757094B CN 202311008152 A CN202311008152 A CN 202311008152A CN 116757094 B CN116757094 B CN 116757094B
- Authority
- CN
- China
- Prior art keywords
- wake
- machine learning
- model
- learning model
- wind turbine
- 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.)
- Active
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 154
- 238000010801 machine learning Methods 0.000 claims abstract description 174
- 239000012530 fluid Substances 0.000 claims abstract description 75
- 238000012549 training Methods 0.000 claims abstract description 73
- 238000000034 method Methods 0.000 claims abstract description 40
- 238000010977 unit operation Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims description 44
- 230000015654 memory Effects 0.000 claims description 28
- 238000010606 normalization Methods 0.000 claims description 13
- 238000005070 sampling Methods 0.000 claims description 9
- 238000010248 power generation Methods 0.000 abstract description 9
- 230000000694 effects Effects 0.000 description 9
- 238000000605 extraction Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000011217 control strategy Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Geometry (AREA)
- Wind Motors (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Public Health (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Water Supply & Treatment (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Medical Informatics (AREA)
Abstract
The invention relates to the technical field of wind power generation and discloses a method, a device, electronic equipment and a storage medium for calculating a wind turbine wake field, wherein the method comprises the steps of obtaining a wind power field model, inflow parameters and wind power field unit operation parameters, and generating a first calculation file and a second calculation file; respectively inputting a first calculation example file into an engineering wake model and a two-dimensional fluid dynamic wake model to generate a first data set, training a first machine learning model, and generating a first target machine learning model; respectively inputting the second calculation example file into a two-dimensional fluid dynamic wake model and a three-dimensional fluid dynamic wake model to generate a second data set, training a second machine learning model, and generating a second target machine learning model; calculating a wind turbine wake field in the wind power plant through the engineering wake model, the first target machine learning model and the second target machine learning model, and performing yaw control on the wind turbine; the invention can rapidly and accurately calculate the wind turbine wake field.
Description
Technical Field
The invention relates to the technical field of wind power generation, in particular to a method and a device for calculating a tail flow field of a wind turbine, electronic equipment and a storage medium.
Background
Wind energy is a clean renewable energy source, and the wind power generation can be effectively assisted in the sustainable development of economy. Wind farm development at present has covered various terrain environments such as offshore, inland plain and mountain land, and gradually developed to deep open sea. Tens or even hundreds of wind turbine generators are commonly installed in a wind power plant, a wake flow area can be generated at the downstream of the wind turbine generators due to absorption of incoming flow energy by a front-row wind turbine generator, the average wind speed in the wake flow area is reduced, the turbulence degree is increased, and the generated energy of the wind turbine generators in the downstream wake flow area is reduced. By adopting a field-level wake flow control technology in the wind power plant, the pitch, yaw and other cooperative control is carried out on each unit, and the strength and the range of wake flow can be adjusted, so that the influence of wake flow effect on the whole power generation capacity of the wind power plant is reduced. Because the incoming flow wind condition changes frequently, the rapid and accurate calculation of the wake field of the wind turbine is an important link of wake control technology.
Disclosure of Invention
In view of the above, the invention provides a method, a device, an electronic device and a storage medium for calculating a wind turbine wake field, so as to solve the technical problem of how to calculate the wind turbine wake field rapidly and accurately.
In a first aspect, the present invention provides a method for calculating a wake field of a wind turbine, including: acquiring a wind power plant model, inflow parameters and wind power plant unit operation parameters, and generating a first calculation file and a second calculation file based on the wind power plant model, the inflow parameters and the wind power plant unit operation parameters; respectively inputting a first example file into an engineering wake model and a two-dimensional fluid dynamic wake model to generate a first data set, training a first machine learning model based on the first data set, and generating a first target machine learning model; respectively inputting the second calculation example file into a two-dimensional fluid dynamic wake model and a three-dimensional fluid dynamic wake model to generate a second data set, training a second machine learning model based on the second data set, and generating a second target machine learning model; the method comprises the steps of obtaining real-time incoming flow conditions of a wind farm, calculating a wind turbine wake field in the wind farm through an engineering wake model, a first target machine learning model and a second target machine learning model based on the real-time incoming flow conditions of the wind farm, and generating a wind turbine wake field calculation result which is used for yaw control of a wind turbine.
According to the wind turbine wake field calculation method, the characteristics and advantages of an engineering wake model, a two-dimensional fluid dynamic wake model and a three-dimensional fluid dynamic wake model are combined, two sets of machine learning modules of a first target machine learning model and a second target machine learning model are built, the two sets of machine learning modules are associated with each other, physical models and machine learning fusion is formed, and rapid calculation of the wind turbine wake field is achieved; the wake flow field of the wind turbine can be rapidly calculated on the basis of improving the wake flow calculation precision, wake flow control is carried out by utilizing the wake flow field of the wind turbine, the generating capacity of the wind turbine is effectively improved, the influence of wake flow effect on the generating capacity of the wind turbine is reduced, and the generating level of the wind turbine is improved.
In an alternative embodiment, the method further comprises: acquiring wind power plant layout information, wind power plant topographic information and wind turbine data in a wind power plant, and constructing a wind power plant model based on the wind power plant layout information, the wind power plant topographic information and the wind turbine data; acquiring an inflow wind speed value interval and a wind turbine operation parameter value interval of the wind turbine, and randomly sampling the inflow wind speed value interval and the wind turbine operation parameter value interval of the wind turbine to generate inflow parameters and wind power plant unit operation parameters.
According to the wind turbine wake field calculation method, the wind power field model is constructed, inflow parameters and wind power field unit operation parameters are generated, data support is provided for training the machine learning model, and the training precision of the machine learning model is improved.
In an alternative embodiment, inputting the first computation file into the engineering wake model and the two-dimensional hydrodynamic wake model, respectively, generating a first dataset, and training the first machine learning model based on the first dataset, generating a first target machine learning model, comprising: inputting the first example file into an engineering wake model to generate an engineering wake cloud picture; inputting the first example file into a two-dimensional fluid dynamics wake model to generate a first two-dimensional fluid dynamics wake cloud picture; carrying out standardization processing on the engineering wake cloud picture and the first two-dimensional hydrodynamic wake cloud picture to generate a first data set; training a first machine learning model to be trained by using the first data set until a first preset condition is met, and generating a first target machine learning model; the first target machine learning model includes a first generator and a first arbiter.
According to the wind turbine wake field calculation method, the engineering wake flow cloud image and the first two-dimensional fluid dynamic wake flow cloud image are generated, and the engineering wake flow cloud image and the first two-dimensional fluid dynamic wake flow cloud image are subjected to standardized processing, so that on one hand, the cloud image is subjected to standardized processing, dimensions are unified, influence on the training precision of a machine learning model due to non-unification of the dimensions is avoided, and the training precision of the machine learning model is improved; on the other hand, the cloud image is beneficial to the feature extraction of the machine learning model, so that the self-learning capability of the machine learning model is conveniently exerted, and the precision and depth of the feature extraction of the cloud image are improved.
In an alternative embodiment, training the first machine learning model to be trained using the first data set until a first preset condition is met, generating a first target machine learning model, comprising: inputting the engineering wake flow cloud picture subjected to the standardization processing in the first data set into a first generator to generate a first cloud picture; and comparing the first cloud image with the first two-dimensional hydrodynamic wake cloud image after the normalization processing in the first data set through a first discriminator, and generating a first target machine learning model if the comparison result meets a first preset condition.
According to the wind turbine wake field calculation method, the first preset condition is set to serve as the termination condition of the first target machine learning model training, and the efficiency and the practical application requirements of the machine learning model training are considered.
In an alternative embodiment, inputting the second computation file into the two-dimensional hydrodynamic wake model and the three-dimensional hydrodynamic wake model, respectively, generating a second data set, and training the second machine learning model based on the second data set, generating a second target machine learning model, comprising: inputting the second calculation file into a two-dimensional fluid dynamics wake model to generate a second two-dimensional fluid dynamics wake cloud picture; inputting the second example file into a three-dimensional fluid dynamics wake model to generate a three-dimensional fluid dynamics wake cloud picture; performing standardization processing on the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image to generate a second data set; and training the second machine learning model to be trained by using the second data set until a second preset condition is met, and generating a second target machine learning model, wherein the second target machine learning model comprises a second generator and a second discriminator.
According to the wind turbine wake field calculation method, the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image are generated, and the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image are subjected to standardized processing, so that on one hand, the cloud image is subjected to standardized processing, dimensions are unified, influence on the training precision of a machine learning model due to non-uniformity of the dimensions is avoided, and the training precision of the machine learning model is improved; on the other hand, the cloud image is beneficial to the feature extraction of the machine learning model, so that the self-learning capability of the machine learning model is conveniently exerted, and the precision and depth of the feature extraction of the cloud image are improved.
In an alternative embodiment, training the second machine learning model to be trained using the second data set until a second preset condition is met, generating a second target machine learning model, including: inputting the normalized second two-dimensional hydrodynamic wake cloud image in the second data set into a second generator to generate a second cloud image; and comparing the second cloud image with the three-dimensional hydrodynamic wake cloud image subjected to the standardization processing in the second data set through a second discriminator, and generating a second target machine learning model if the comparison result meets a second preset condition.
According to the wind turbine wake field calculation method, the second preset condition is set to serve as the termination condition of the training of the second target machine learning model, and the training efficiency of the machine learning model and the actual application requirements are considered.
In an alternative embodiment, based on a real-time incoming wind condition of a wind farm, calculating a wind turbine wake field in the wind farm through an engineering wake model, a first target machine learning model and a second target machine learning model, and generating a wind turbine wake field calculation result, including: inputting real-time incoming flow conditions of a wind power plant into an engineering wake model to calculate a wind turbine wake field, and generating a first wake calculation result picture; inputting the first wake calculation result picture into a first target machine learning model to generate a second wake calculation result picture; inputting the second wake calculation result picture into a second target machine learning model to generate a third wake calculation result picture; and calculating a wind turbine wake field in the wind power plant based on the third wake flow calculation result picture to generate a wind turbine wake field calculation result.
According to the wind turbine wake field calculation method, the wind turbine wake field is calculated quickly through the engineering wake model, the first target machine learning model and the second target machine learning model are related to each other, the wind turbine wake field can be calculated quickly on the basis of improving wake calculation accuracy, the wind turbine wake field calculation method is used in the wake control field, the generated energy of a wind power plant is effectively improved, the influence of wake effects on the generated energy of the wind power plant is reduced, and the power generation level of the wind power plant is improved.
In a second aspect, the present invention provides a wind turbine wake field calculation device, comprising: the first generation module is used for acquiring a wind power plant model, inflow parameters and wind power plant unit operation parameters and generating a first calculation file and a second calculation file based on the wind power plant model, the inflow parameters and the wind power plant unit operation parameters; the first training module is used for inputting the first example file into the engineering wake model and the two-dimensional fluid dynamic wake model respectively, generating a first data set, training the first machine learning model based on the first data set and generating a first target machine learning model; the second training module is used for inputting a second calculation example file into the two-dimensional fluid dynamic wake model and the three-dimensional fluid dynamic wake model respectively, generating a second data set, training a second machine learning model based on the second data set and generating a second target machine learning model; the second generation module is used for acquiring real-time incoming flow conditions of the wind farm, calculating a wind turbine wake field in the wind farm through the engineering wake model, the first target machine learning model and the second target machine learning model based on the real-time incoming flow conditions of the wind farm, and generating a wind turbine wake field calculation result which is used for yaw control of the wind turbine.
In a third aspect, the present invention provides an electronic device, comprising: the wind turbine wake field calculation method according to the first aspect or any one of the embodiments corresponding to the first aspect is implemented by the processor and the memory, the memory and the processor are in communication connection with each other, and the memory stores computer instructions, and the processor executes the computer instructions.
In a fourth aspect, the present invention provides a computer readable storage medium having stored thereon computer instructions for causing a computer to perform the method of calculating a wind turbine wake field of the first aspect or any of its corresponding embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method of calculating a wake field of a wind turbine according to an embodiment of the invention;
FIG. 2 is a flow chart of another method of calculating a wake field of a wind turbine according to an embodiment of the invention;
FIG. 3 is a flow chart of a method of calculating a wake field of a wind turbine according to an embodiment of the invention;
FIG. 4 is a schematic diagram of a data set generation flow in accordance with an embodiment of the present invention;
FIG. 5 is a flowchart illustration of a first target machine learning module according to an embodiment of the invention;
FIG. 6 is a flowchart illustration of a second target machine learning module according to an embodiment of the present invention;
FIG. 7 is a block diagram of a wind turbine wake field calculation device according to an embodiment of the invention;
fig. 8 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The wind turbine wake field calculation method provided by the specification can be applied to electronic equipment for wind turbine wake field calculation in a wind power plant; the electronic device may include, but is not limited to, a notebook, desktop, mobile terminal, such as a cell phone, tablet, etc.; of course, the method for calculating the wake field of the wind turbine provided in the present specification may also be applied to an application program running in the above electronic device.
The current wake calculation method is mainly proposed based on a physical model, and comprises an engineering wake model and a CFD (computational fluid dynamics) wake model; the engineering wake model based on the semi-empirical formula is a method mainly adopted in the wake calculation field at present, the calculation speed is high, but wake recovery factors in the model are empirical parameters, so that the model precision is lower than that of a CFD wake model, the model is only suitable for flat terrain and offshore wind fields, and the adaptability is poor under complex terrain environments; the computational accuracy of the CFD wake model is high, but the computational complexity is large and the wake distribution of the wind turbine cannot be calculated quickly, the CFD wake model can be divided into a CFD-2D (Two-dimensional fluid dynamics) wake model and a CFD-3D (Three-Dimensional Computational Fluid Dynamics, three-dimensional fluid dynamics) wake model, the CFD-2D wake model can calculate Two-dimensional flow field distribution, the computational complexity is less than that of the CFD-3D wake model, the CFD-3D wake model can calculate Three-dimensional flow field distribution, wind shear, complex topography and wake distribution under yaw can be calculated, and the computational complexity is maximum; therefore, how to combine the advantages of different wake models to perform wake fast computation is a technical problem to be solved.
Based on the technical problems, the invention provides a wind turbine wake field calculation method, which can be used for rapidly calculating a wind turbine wake field on the basis of improving wake calculation precision, is used for the wake control field, effectively improving the generated energy of a wind power plant, reducing the influence of wake effect on the generated energy of the wind power plant and improving the power generation level of the wind power plant.
According to an embodiment of the present invention, a wind turbine wake field calculation method embodiment is provided, it being noted that the steps shown in the flowchart of the figures may be performed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be performed in an order different from that shown or described herein.
In this embodiment, a method for calculating a wake field of a wind turbine is provided, which may be used in the above notebook computer, desktop computer, mobile terminal, such as a mobile phone, tablet computer, etc., fig. 1 is a flowchart of a method for calculating a wake field of a wind turbine according to an embodiment of the present invention, as shown in fig. 1, where the flowchart includes the following steps:
step S101, a wind power plant model, inflow parameters and wind power plant set operation parameters are obtained, and a first calculation file and a second calculation file are generated based on the wind power plant model, the inflow parameters and the wind power plant set operation parameters.
Specifically, wind power plant layout information, wind power plant topographic information and wind turbine data in a wind power plant are obtained, and a wind power plant model is constructed based on the wind power plant layout information, the wind power plant topographic information and the wind turbine data; acquiring an inflow wind speed value interval and a wind turbine running parameter value interval of a wind turbine, and randomly sampling the inflow wind speed value interval and the wind turbine running parameter value interval to generate an inflow parameter and a wind power plant unit running parameter; the wind farm model is constructed, inflow parameters and wind farm unit operation parameters are generated, data support is provided for training the machine learning model, and the training precision of the machine learning model is improved.
Further, setting a value interval of inflow parameters and unit operation parameters according to the actual operation rule of the wind field unit, wherein the value interval mainly comprises parameters such as inflow wind speed, pitch angle, yaw angle, thrust coefficient and the like, and then generating M, N inflow and unit operation parameter combinations in batches twice in the parameter value interval range by adopting a random sampling method such as Latin hypercube sampling method; on the basis, two kinds of calculation example files to be calculated are generated by combining a wind power plant model, wherein the first calculation example file comprises M calculation examples and the second calculation example file comprises N calculation examples.
Step S102, a first example file is respectively input into an engineering wake model and a two-dimensional fluid dynamic wake model to generate a first data set, and a first machine learning model is trained based on the first data set to generate a first target machine learning model.
Step S103, the second calculation example file is respectively input into the two-dimensional fluid dynamic wake model and the three-dimensional fluid dynamic wake model to generate a second data set, and the second machine learning model is trained based on the second data set to generate a second target machine learning model.
Step S104, obtaining real-time incoming flow conditions of the wind farm, calculating a wind turbine wake field in the wind farm through an engineering wake model, a first target machine learning model and a second target machine learning model based on the real-time incoming flow conditions of the wind farm, and generating a wind turbine wake field calculation result which is used for yaw control of the wind turbine.
According to the wind turbine wake field calculation method, the characteristics and advantages of an engineering wake model, a two-dimensional fluid dynamic wake model and a three-dimensional fluid dynamic wake model are combined, two sets of machine learning modules of a first target machine learning model and a second target machine learning model are built, the two sets of machine learning modules are associated with each other, physical models and machine learning fusion is formed, and rapid calculation of the wind turbine wake field is achieved; the wake flow field of the wind turbine can be rapidly calculated on the basis of improving the wake flow calculation precision, wake flow control is carried out by utilizing the wake flow field of the wind turbine, the generating capacity of the wind turbine is effectively improved, the influence of wake flow effect on the generating capacity of the wind turbine is reduced, and the generating level of the wind turbine is improved.
In this embodiment, a method for calculating a wake field of a wind turbine is provided, which may be used in the above notebook computer, desktop computer, mobile terminal, such as a mobile phone, tablet computer, etc., fig. 2 is a flowchart of a method for calculating a wake field of a wind turbine according to an embodiment of the present invention, as shown in fig. 2, where the flowchart includes the following steps:
step S201, a wind power plant model, inflow parameters and wind power plant set operation parameters are obtained, and a first calculation file and a second calculation file are generated based on the wind power plant model, the inflow parameters and the wind power plant set operation parameters; please refer to step S101 in the embodiment shown in fig. 1 in detail, which is not described herein.
Step S202, a first example file is respectively input into an engineering wake model and a two-dimensional fluid dynamic wake model to generate a first data set, and a first machine learning model is trained based on the first data set to generate a first target machine learning model.
Specifically, the step S202 includes:
step S2021, inputting the first example file into the engineering wake model to generate an engineering wake cloud image.
Step S2022, inputting the first example file into a two-dimensional hydrodynamic wake model, generating a first two-dimensional hydrodynamic wake cloud image.
Step S2023, performing a normalization process on the engineering wake cloud image and the first two-dimensional hydrodynamic wake cloud image to generate a first data set.
Specifically, the first calculation example file adopts two model methods of an engineering wake model and a two-dimensional fluid dynamic wake model for calculation, so that two results are generated; after all the calculation examples are calculated, the self-coding codes call the post-processing software to process the calculation results to obtain wake velocity cloud patterns of the hub of the wind turbine, wherein the wake velocity cloud patterns of the hub of the wind turbine are parallel to the ground, and the velocity cloud patterns generated by the two calculation results of the first calculation example file adopt the same color setting range; and finally, generating a first data set after the calculation result processing of the calculation example in the first calculation example file is completed.
Step S2024, training the first machine learning model to be trained by using the first data set until a first preset condition is satisfied, generating a first target machine learning model; the first target machine learning model includes a first generator and a first arbiter.
In some alternative embodiments, step S2024 described above comprises:
and a step a1, inputting the engineering wake cloud pictures subjected to the standardization processing in the first data set into a first generator, and generating a first cloud picture.
And a step a2 of comparing the first cloud image with the first two-dimensional hydrodynamic wake cloud image after the standardization processing in the first data set through a first discriminator, and generating a first target machine learning model if the comparison result meets a first preset condition.
Specifically, the first preset condition may be that a similarity between the first cloud image and the first two-dimensional hydrodynamic wake cloud image is greater than a similarity preset threshold, or a loss value of the first cloud image relative to the first two-dimensional hydrodynamic wake cloud image is less than a loss value preset threshold; the first preset condition is set to serve as a termination condition of training of the first target machine learning model, so that the training efficiency of the machine learning model and the actual application requirement are considered.
According to the wind turbine wake field calculation method, the engineering wake flow cloud image and the first two-dimensional fluid dynamic wake flow cloud image are generated, and the engineering wake flow cloud image and the first two-dimensional fluid dynamic wake flow cloud image are subjected to standardized processing, so that on one hand, the cloud image is subjected to standardized processing, dimensions are unified, influence on the training precision of a machine learning model due to non-unification of the dimensions is avoided, and the training precision of the machine learning model is improved; on the other hand, the cloud image is beneficial to the feature extraction of the machine learning model, so that the self-learning capability of the machine learning model is conveniently exerted, and the precision and depth of the feature extraction of the cloud image are improved.
Step S203, the second calculation example file is respectively input into the two-dimensional fluid dynamic wake model and the three-dimensional fluid dynamic wake model to generate a second data set, and the second machine learning model is trained based on the second data set to generate a second target machine learning model.
Specifically, the step S203 includes:
step S2031, inputting the second example file into the two-dimensional hydrodynamic wake model, and generating a second two-dimensional hydrodynamic wake cloud image.
Step S2032, inputting the second example file into the three-dimensional hydrodynamic wake model to generate a three-dimensional hydrodynamic wake cloud image.
Step S2033, performing standardization processing on the second two-dimensional hydrodynamic wake cloud image and the three-dimensional hydrodynamic wake cloud image to generate a second data set.
Specifically, the second calculation example file adopts two model methods of a two-dimensional fluid dynamic wake model and a three-dimensional fluid dynamic wake model to calculate, so that two results are generated; after all the calculation examples are calculated, the self-coding codes call the post-processing software to process the calculation results to obtain wake velocity cloud patterns of the hub of the wind turbine, wherein the height position of the hub of the wind turbine is parallel to the ground, and the velocity cloud patterns generated by the two calculation results of the second calculation example file adopt the same color setting range; and finally, generating a second data set after the calculation result processing of the calculation example in the first calculation example file is completed.
Step S2034, training the second machine learning model to be trained using the second data set until a second preset condition is satisfied, generating a second target machine learning model, where the second target machine learning model includes a second generator and a second arbiter.
In some optional embodiments, step S2034 includes:
and b1, inputting a second two-dimensional hydrodynamic wake cloud image after the normalization processing in the second data set into a second generator to generate a second cloud image.
Step b2, comparing the second cloud image with the three-dimensional hydrodynamic wake cloud image subjected to the standardization processing in the second data set through a second discriminator, and generating a second target machine learning model if a comparison result meets a second preset condition; and setting a second preset condition as a termination condition of training of the second target machine learning model, and considering the training efficiency of the machine learning model and the actual application requirement.
Specifically, the second preset condition may be that the similarity between the second cloud image and the three-dimensional hydrodynamic wake cloud image is greater than a similarity preset threshold, or that the loss value of the second cloud image relative to the three-dimensional hydrodynamic wake cloud image is less than a loss value preset threshold.
According to the wind turbine wake field calculation method, the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image are generated, and the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image are subjected to standardized processing, so that on one hand, the cloud image is subjected to standardized processing, dimensions are unified, influence on the training precision of a machine learning model due to non-uniformity of the dimensions is avoided, and the training precision of the machine learning model is improved; on the other hand, the cloud image is beneficial to the feature extraction of the machine learning model, so that the self-learning capability of the machine learning model is conveniently exerted, and the precision and depth of the feature extraction of the cloud image are improved.
Step S204, obtaining real-time incoming flow conditions of the wind farm, calculating a wind turbine wake field in the wind farm through an engineering wake model, a first target machine learning model and a second target machine learning model based on the real-time incoming flow conditions of the wind farm, and generating a wind turbine wake field calculation result which is used for yaw control of the wind turbine.
Specifically, the step S204 includes:
step S2041, inputting real-time incoming flow conditions of the wind farm into an engineering wake model to calculate a wind turbine wake field, and generating a first wake calculation result picture.
Step S2042, inputting the first wake calculation result picture into the first target machine learning model, and generating a second wake calculation result picture.
Step S2043, the second wake calculation result picture is input into the second target machine learning model, and a third wake calculation result picture is generated.
Step S2044, based on the third wake flow calculation result picture, calculating a wind turbine wake flow field in the wind power plant to generate a wind turbine wake flow field calculation result.
Specifically, the wind turbine wake field calculation result is used for carrying out yaw control on the wind turbine, and adjusting a yaw angle and a control strategy, so that the influence of wake effects of the front-row wind turbine on the downstream wind turbine is reduced, and the power generation level of the whole wind power plant is improved.
According to the wind turbine wake field calculation method, the wind turbine wake field is calculated quickly through the engineering wake model, the first target machine learning model and the second target machine learning model are related to each other, the wind turbine wake field can be calculated quickly on the basis of improving wake calculation accuracy, the wind turbine wake field calculation method is used in the wake control field, the generated energy of a wind power plant is effectively improved, the influence of wake effects on the generated energy of the wind power plant is reduced, and the power generation level of the wind power plant is improved.
A method of calculating a wind turbine wake field is described below with reference to a specific example.
Example 1:
the wind turbine wake field calculation method comprises the following calculation steps:
step 1, modeling of wind farm
For a given wind farm, the wind farm is modeled first according to the basic information of the wind farm (wind wheel diameter, hub height, rated power and performance curves, etc.), wind farm layout (number and spacing of wind farms) and terrain information (ground roughness).
Step 2, generating an example to be calculated
Setting a value interval of inflow and unit operation parameters according to the actual operation rule of the wind field unit, mainly comprising parameters such as inflow wind speed, pitch angle, yaw angle, thrust coefficient and the like, and then generating M, N inflow and unit operation parameter combinations in batches twice in the parameter value interval range by using a Latin hypercube sampling method. On the basis, modeling is combined with a wind farm, two example files to be calculated are generated, wherein the first example file comprises M examples, and the second example file comprises N examples, as shown in fig. 3; the recommended values of the parameters are shown in the following table 1:
table 1:
| sampling parameters | Suggested range | Suggested intervals |
| Inflow wind speed | 3-25m/s | 0.5m/s |
| Yaw angle | -20-20deg | 5deg |
| Pitch angle | -5-15deg | 2deg |
Note that: deg (degree)
Step 3, classifying and calculating the example file
Classifying and calculating the two kinds of calculation files; the first example of the example file adopts two methods of an engineering wake model and a CFD-2D wake model for calculation, and the second example of the example file adopts a CFD-2D wake model and a CFD-3D wake model for calculation.
Taking a near sea wind power plant as an example, the layout of each unit in the wind power plant is shown in fig. 4, and the wind power plant consists of 15 4MW (megawatt) units in total, wherein the 3 rows and the 5 columns are arranged.
Carrying out wake field calculation on the wind power plant by using a work Cheng Weiliu model, a CFD-2D wake model and a CFD-3D wake model, wherein the wind speed is 9m/s, and the wind directions are respectively the south wind; and outputting three sets of wake cloud charts, wherein the calculation precision of the three sets of wake cloud charts output by the engineering wake model, the CFD-2D wake model and the CFD-3D wake model meets the requirements of the CFD-3D wake model > the CFD-2D wake model > the engineering wake model.
Step 4, data processing and data set establishment
Because each calculation example adopts two model methods for calculation, two results are generated; after all the calculation examples are calculated, the self-coding codes call the post-processing software to process the calculation results to obtain wake velocity cloud patterns of the hub of the wind turbine, wherein the height position of the hub of the wind turbine is parallel to the ground, and the velocity cloud patterns generated by the two calculation results of each calculation example adopt the same color setting range; finally, generating a first data set after the calculation result processing of the first calculation example in the calculation example file is finished, and generating a second data set after the calculation result processing of the second calculation example in the calculation example file is finished; each data set is internally divided into a training set, a verification set and a test set.
Step 5, establishing a machine learning module
For the first data set and the second data set, two GAN (Generative Adversarial Networks) machine learning modules, namely GAN-I (first target machine learning model) and GAN-II (second target machine learning model), are established, and schematic diagrams of the two machine learning modules are shown in fig. 5 and fig. 6.
The GAN-I carries out learning training on the first data set, the first data set is divided into a generator and a discriminator, the generator is input into an engineering wake model cloud image, the cloud image is generated through output of an encoder, a vector and a decoder submodule, then the cloud image generated and the cloud image of the CFD-2D wake model are input into the discriminator, the approaching degree of the cloud image generated and the cloud image of the CFD-2D wake model is judged, and the cloud image generated is enabled to be continuously close to the wake cloud image of the CFD-2D wake model through training; because the CFD-2D wake model can accurately calculate the two-dimensional flow field distribution, the wake calculation accuracy of the engineering wake model can be improved through the generated cloud image output by the GAN-I machine learning module.
The GAN-II carries out learning training on the second data set, the second data set is divided into a generator and a discriminator, the generator is input into a CFD-2D wake model cloud image, the cloud image is output through a encoder, vector and decoder sub-module, then the generated cloud image and the CFD-3D wake model cloud image are input into the discriminator, the approaching degree of the cloud image and the CFD-3D wake model cloud image is judged, and the cloud image is continuously approaching to the wake cloud image of the CFD-3D wake model through training, so that the CFD-3D wake model can consider the influence of wind shearing, complex topography, fan yaw and the like on wake results, and the accuracy of the CFD-2D wake model can be further improved through the cloud image output by the GAN-II machine learning module.
Step 6, fast and accurate calculation of wind turbine wake field
After the machine learning module is trained, the wind turbine wake field can be rapidly calculated.
(1) According to the real-time incoming flow condition, combining a wind power plant model, firstly calculating a wind turbine wake field by using an engineering wake model to obtain a wake calculation result picture;
(2) Inputting the flow field result picture in the step (1) into a GAN-I machine learning module to generate a wake flow calculation result picture with higher accuracy;
(3) Inputting the flow field result picture in the step (2) into a GAN-II machine learning module to generate a flow field result picture which can finally consider the three-dimensional effect.
Step 7, implementing wake control
When the wind farm operates, yaw control is carried out on the front-row wind turbines, wake flow distribution of the front-row wind turbines is rapidly calculated according to real-time incoming flow conditions, yaw angles and control strategies are adjusted, the influence of wake flow effects of the front-row wind turbines on downstream wind turbines is further reduced, and the power generation level of the whole wind farm is improved.
The embodiment also provides a wind turbine wake field calculation device, which is used for realizing the embodiment and the preferred implementation manner, and is not described in detail; as used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function; while the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
The present embodiment provides a wind turbine wake field calculation device, as shown in fig. 7, including:
the first generating module 701 is configured to obtain a wind farm model, an inflow parameter, and a wind farm unit operation parameter, and generate a first calculation file and a second calculation file based on the wind farm model, the inflow parameter, and the wind farm unit operation parameter;
the first training module 702 is configured to input the first example file into the engineering wake model and the two-dimensional hydrodynamic wake model, generate a first data set, train the first machine learning model based on the first data set, and generate a first target machine learning model;
a second training module 703, configured to input a second example file into the two-dimensional hydrodynamic wake model and the three-dimensional hydrodynamic wake model, generate a second data set, and train the second machine learning model based on the second data set, to generate a second target machine learning model;
the second generating module 704 is configured to obtain a real-time wind incoming flow condition of the wind farm, calculate a wind turbine wake field in the wind farm through the engineering wake model, the first target machine learning model and the second target machine learning model based on the real-time wind incoming flow condition of the wind farm, and generate a wind turbine wake field calculation result, where the wind turbine wake field calculation result is used to perform yaw control on the wind turbine.
In some alternative embodiments, further comprising:
the building module is used for acquiring wind power plant layout information, wind power plant topographic information and wind turbine data in the wind power plant and building a wind power plant model based on the wind power plant layout information, the wind power plant topographic information and the wind turbine data;
the third generation module is used for acquiring an inflow wind speed value interval of the wind turbine and an operation parameter value interval of the wind turbine, randomly sampling the inflow wind speed value interval of the wind turbine and the operation parameter value interval of the wind turbine, and generating inflow parameters and operation parameters of a wind power plant unit.
In some alternative embodiments, first training module 702 includes:
the first input sub-module is used for inputting the first example file into the engineering wake model to generate an engineering wake cloud picture;
the second input submodule is used for inputting the first example file into the two-dimensional fluid dynamics wake model to generate a first two-dimensional fluid dynamics wake cloud picture;
the first normalization processing submodule is used for performing normalization processing on the engineering wake cloud image and the first two-dimensional hydrodynamic wake cloud image to generate a first data set;
the first training sub-module is used for training the first machine learning model to be trained by utilizing the first data set until a first preset condition is met, so as to generate a first target machine learning model; the first target machine learning model includes a first generator and a first arbiter.
In some alternative embodiments, the first training sub-module comprises:
the first input unit is used for inputting the engineering wake flow cloud pictures subjected to the standardized processing in the first data set into the first generator to generate a first cloud picture;
and the first comparison unit is used for comparing the first cloud image with the first two-dimensional hydrodynamic wake cloud image subjected to the standardized processing in the first data set through the first discriminator, and generating a first target machine learning model if the comparison result meets a first preset condition.
In some alternative embodiments, the second training module 703 includes:
the third input submodule is used for inputting the second example file into the two-dimensional fluid dynamics wake model to generate a second two-dimensional fluid dynamics wake cloud picture;
the fourth input submodule is used for inputting the second example file into the three-dimensional fluid dynamics wake model to generate a three-dimensional fluid dynamics wake cloud picture;
the second normalization processing submodule is used for performing normalization processing on the second two-dimensional fluid dynamic wake cloud picture and the three-dimensional fluid dynamic wake cloud picture to generate a second data set;
the second training sub-module is used for training a second machine learning model to be trained by using a second data set until a second preset condition is met, and a second target machine learning model is generated, wherein the second target machine learning model comprises a second generator and a second discriminator.
In some alternative embodiments, the second training sub-module comprises:
the second input unit is used for inputting the second two-dimensional hydrodynamic wake cloud image subjected to the normalization processing in the second data set into the second generator to generate a second cloud image;
and the second comparison unit is used for comparing the second cloud image with the three-dimensional hydrodynamic wake cloud image subjected to the standardized processing in the second data set through the second discriminator, and generating a second target machine learning model if the comparison result meets a second preset condition.
In some alternative embodiments, the second generating module 704 includes:
the fifth input submodule is used for inputting real-time incoming wind conditions of the wind farm into the engineering wake model to calculate the wind turbine wake field and generating a first wake calculation result picture;
the sixth input submodule is used for inputting the first wake flow calculation result picture into the first target machine learning model and generating a second wake flow calculation result picture;
a seventh input sub-module, configured to input a second wake computation result picture into the second target machine learning model, and generate a third wake computation result picture;
the generation submodule is used for calculating a wind turbine wake field in the wind power plant based on the third wake flow calculation result picture to generate a wind turbine wake field calculation result.
Further functional descriptions of the above respective modules and units are the same as those of the above corresponding embodiments, and are not repeated here.
The wind turbine wake field calculation means in this embodiment is presented in the form of functional units, here referred to as ASIC (Application Specific Integrated Circuit ) circuits, processors and memories executing one or more software or fixed programs, and/or other devices that can provide the above described functions.
The embodiment of the invention also provides electronic equipment, which is provided with the wind turbine wake field calculation device shown in the figure 7.
Referring to fig. 8, fig. 8 is a schematic structural diagram of an electronic device according to an alternative embodiment of the present invention, as shown in fig. 8, the electronic device includes: one or more processors 10, memory 20, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The various components are communicatively coupled to each other using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the electronic device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In some alternative embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple electronic devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). One processor 10 is illustrated in fig. 8.
The processor 10 may be a central processor, a network processor, or a combination thereof. The processor 10 may further include a hardware chip, among others. The hardware chip may be an application specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general-purpose array logic, or any combination thereof.
Wherein the memory 20 stores instructions executable by the at least one processor 10 to cause the at least one processor 10 to perform the methods shown in implementing the above embodiments.
The memory 20 may include a storage program area that may store an operating system, at least one application program required for functions, and a storage data area; the storage data area may store data created according to the use of the electronic device, etc. In addition, the memory 20 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, memory 20 may optionally include memory located remotely from processor 10, which may be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
Memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk, or solid state disk; the memory 20 may also comprise a combination of the above types of memories.
The electronic device further comprises input means 30 and output means 40. The processor 10, memory 20, input device 30, and output device 40 may be connected by a bus or other means, for example in fig. 8.
The input device 30 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the electronic device, such as a touch screen, keypad, mouse, trackpad, touchpad, pointer stick, one or more mouse buttons, trackball, joystick, and the like. The output means 40 may include a display device, auxiliary lighting means (e.g., LEDs), tactile feedback means (e.g., vibration motors), and the like. Such display devices include, but are not limited to, liquid crystal displays, light emitting diodes, displays and plasma displays. In some alternative implementations, the display device may be a touch screen.
The embodiments of the present invention also provide a computer readable storage medium, and the method according to the embodiments of the present invention described above may be implemented in hardware, firmware, or as a computer code which may be recorded on a storage medium, or as original stored in a remote storage medium or a non-transitory machine readable storage medium downloaded through a network and to be stored in a local storage medium, so that the method described herein may be stored on such software process on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware. The storage medium can be a magnetic disk, an optical disk, a read-only memory, a random access memory, a flash memory, a hard disk, a solid state disk or the like; further, the storage medium may also comprise a combination of memories of the kind described above. It will be appreciated that a computer, processor, microprocessor controller or programmable hardware includes a storage element that can store or receive software or computer code that, when accessed and executed by the computer, processor or hardware, implements the methods illustrated by the above embodiments.
Although embodiments of the present invention have been described in connection with the accompanying drawings, various modifications and variations may be made by those skilled in the art without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope of the invention as defined by the appended claims.
Claims (10)
1. A method for calculating a wake field of a wind turbine, the method comprising:
acquiring a wind power plant model, inflow parameters and wind power plant unit operation parameters, and generating a first calculation file and a second calculation file based on the wind power plant model, the inflow parameters and the wind power plant unit operation parameters;
respectively inputting the first calculation example file into an engineering wake model and a two-dimensional fluid dynamic wake model to generate a first data set, and training a first machine learning model based on the first data set until a first preset condition is met to generate a first target machine learning model, wherein the first target machine learning model comprises a first generator and a first discriminator;
respectively inputting the second calculation example file into the two-dimensional fluid dynamic wake model and the three-dimensional fluid dynamic wake model to generate a second data set, and training a second machine learning model based on the second data set until a second preset condition is met to generate a second target machine learning model, wherein the second target machine learning model comprises a second generator and a second discriminator;
acquiring a real-time incoming flow condition of a wind power plant, and calculating a wind turbine wake field in the wind power plant through the engineering wake model, the first target machine learning model and the second target machine learning model based on the real-time incoming flow condition of the wind power plant to generate a wind turbine wake field calculation result, wherein the wind turbine wake field calculation result is used for yaw control of a wind turbine;
Training the first machine learning model based on the first data set until a first preset condition is met, and generating a first target machine learning model, including:
inputting the engineering wake flow cloud pictures subjected to the standardization processing in the first data set into the first generator to generate a first cloud picture;
comparing the first cloud image with the first two-dimensional fluid dynamic wake cloud image after the standardization processing in the first data set through the first discriminator, and generating the first target machine learning model if a comparison result meets the first preset condition;
training a second machine learning model based on the second data set until a second preset condition is met, and generating a second target machine learning model, including:
inputting the second two-dimensional hydrodynamic wake cloud image after the standardization processing in the second data set into the second generator to generate a second cloud image;
comparing the second cloud image with the three-dimensional hydrodynamic wake cloud image subjected to the standardization processing in the second data set through the second discriminator, and generating a second target machine learning model if a comparison result meets the second preset condition;
The calculating a wind turbine wake field in the wind power plant based on the wind power plant real-time incoming flow condition through the engineering wake model, the first target machine learning model and the second target machine learning model, and generating a wind turbine wake field calculation result comprises the following steps:
inputting real-time incoming wind conditions of the wind power plant into the engineering wake model to calculate a wind turbine wake field, and generating a first wake calculation result picture;
inputting the first wake calculation result picture into the first target machine learning model to generate a second wake calculation result picture;
inputting the second wake calculation result picture into the second target machine learning model to generate a third wake calculation result picture;
and calculating a wind turbine wake field in the wind power plant based on the third wake flow calculation result picture to generate a wind turbine wake field calculation result.
2. The method of claim 1, wherein prior to obtaining the wind farm model, the inflow parameters, and the wind farm operational parameters, generating the first and second example files based on the wind farm model, the inflow parameters, and the wind farm operational parameters, further comprises:
Acquiring wind power plant layout information, wind power plant topographic information and wind turbine data in a wind power plant, and constructing a wind power plant model based on the wind power plant layout information, the wind power plant topographic information and the wind turbine data;
acquiring an inflow wind speed value interval and a wind turbine operation parameter value interval of a wind turbine, and randomly sampling the inflow wind speed value interval and the wind turbine operation parameter value interval to generate the inflow parameters and the wind power plant unit operation parameters.
3. The method of claim 2, wherein the inputting the first example file into the engineering wake model and the two-dimensional hydrodynamic wake model, respectively, generating a first dataset, and training a first machine learning model based on the first dataset, generating a first target machine learning model, comprises:
inputting the first example file into an engineering wake model to generate an engineering wake cloud picture;
inputting the first example file into a two-dimensional fluid dynamic wake model to generate a first two-dimensional fluid dynamic wake cloud picture;
carrying out standardized processing on the engineering wake cloud image and the first two-dimensional hydrodynamic wake cloud image to generate a first data set;
And training the first machine learning model by using the first data set until a first preset condition is met, and generating a first target machine learning model.
4. The method of claim 3, wherein the inputting the second example file into the two-dimensional and three-dimensional hydrodynamic wake models, respectively, generating a second data set, and training a second machine learning model based on the second data set, generating a second target machine learning model, comprises:
inputting the second example file into the two-dimensional fluid dynamics wake model to generate a second two-dimensional fluid dynamics wake cloud image;
inputting the second example file into the three-dimensional fluid dynamics wake model to generate a three-dimensional fluid dynamics wake cloud picture;
performing standardization processing on the second two-dimensional fluid dynamic wake cloud image and the three-dimensional fluid dynamic wake cloud image to generate a second data set;
and training the second machine learning model by using the second data set until a second preset condition is met, and generating a second target machine learning model.
5. A wind turbine wake field computing device, the device comprising:
The first generation module is used for acquiring a wind power plant model, inflow parameters and wind power plant unit operation parameters and generating a first calculation file and a second calculation file based on the wind power plant model, the inflow parameters and the wind power plant unit operation parameters;
the first training module is used for inputting the first calculation file into an engineering wake model and a two-dimensional fluid dynamic wake model respectively, generating a first data set, training a first machine learning model based on the first data set until a first preset condition is met, and generating a first target machine learning model, wherein the first target machine learning model comprises a first generator and a first discriminator;
the second training module is used for inputting the second calculation case file into the two-dimensional fluid dynamic wake model and the three-dimensional fluid dynamic wake model respectively, generating a second data set, training a second machine learning model based on the second data set until a second preset condition is met, and generating a second target machine learning model, wherein the second target machine learning model comprises a second generator and a second discriminator;
the second generation module is used for acquiring real-time incoming flow conditions of the wind farm, calculating a wind turbine wake field in the wind farm through the engineering wake model, the first target machine learning model and the second target machine learning model based on the real-time incoming flow conditions of the wind farm, and generating a wind turbine wake field calculation result, wherein the wind turbine wake field calculation result is used for yaw control of a wind turbine;
The first training module includes a first training sub-module including:
the first input unit is used for inputting the engineering wake cloud pictures subjected to the standardized processing in the first data set into the first generator to generate a first cloud picture;
a first comparing unit, configured to compare, by using the first discriminator, the first cloud image with the first two-dimensional hydrodynamic wake cloud image after the normalization processing in the first dataset, and generate the first target machine learning model if a comparison result meets the first preset condition;
the second training module includes a second training sub-module including:
a second input unit, configured to input a second two-dimensional hydrodynamic wake cloud image after the normalization processing in the second data set into the second generator, and generate a second cloud image;
the second comparing unit is used for comparing the second cloud image with the three-dimensional hydrodynamic wake cloud image after the standardization processing in the second data set through the second discriminator, and generating a second target machine learning model if the comparison result meets the second preset condition;
the second generation module includes:
The fifth input submodule is used for inputting the real-time incoming flow condition of the wind power plant into the engineering wake model to calculate the wind turbine wake field and generating a first wake calculation result picture;
a sixth input sub-module, configured to input the first wake computation result picture into the first target machine learning model, and generate a second wake computation result picture;
a seventh input sub-module, configured to input the second wake calculation result picture into the second target machine learning model, and generate a third wake calculation result picture;
the generation submodule is used for calculating a wind turbine wake field in the wind power plant based on the third wake flow calculation result picture to generate a wind turbine wake field calculation result.
6. The apparatus as recited in claim 5, further comprising:
the building module is used for acquiring wind power plant layout information, wind power plant topographic information and wind turbine data in a wind power plant and building the wind power plant model based on the wind power plant layout information, the wind power plant topographic information and the wind turbine data;
the third generation module is used for acquiring an inflow wind speed value interval of the wind turbine and an operation parameter value interval of the wind turbine, randomly sampling the inflow wind speed value interval of the wind turbine and the operation parameter value interval of the wind turbine, and generating the inflow parameters and the operation parameters of the wind power plant unit.
7. The apparatus of claim 6, wherein the first training module comprises:
the first input sub-module is used for inputting the first example file into an engineering wake model to generate an engineering wake cloud picture;
the second input submodule is used for inputting the first example file into a two-dimensional fluid dynamics wake model to generate a first two-dimensional fluid dynamics wake cloud picture;
a first normalization processing sub-module, configured to normalize the engineering wake cloud image and the first two-dimensional hydrodynamic wake cloud image to generate a first data set;
the first training sub-module is used for training the first machine learning model by utilizing the first data set until a first preset condition is met, so as to generate a first target machine learning model; the first target machine learning model includes a first generator and a first arbiter.
8. The apparatus of claim 7, wherein the second training module comprises:
the third input submodule is used for inputting the second example file into the two-dimensional fluid dynamics wake model to generate a second two-dimensional fluid dynamics wake cloud picture;
a fourth input sub-module for inputting the second example file into the three-dimensional hydrodynamic wake model to generate a three-dimensional hydrodynamic wake cloud image;
The second normalization processing submodule is used for performing normalization processing on the second two-dimensional fluid dynamic wake cloud picture and the three-dimensional fluid dynamic wake cloud picture to generate a second data set;
and the second training sub-module is used for training a second machine learning model by using the second data set until a second preset condition is met, so as to generate a second target machine learning model, wherein the second target machine learning model comprises a second generator and a second discriminator.
9. An electronic device, comprising:
a memory and a processor, the memory and the processor being communicatively connected to each other, the memory having stored therein computer instructions, the processor executing the computer instructions to perform the method of calculating a wind turbine wake field of any one of claims 1 to 4.
10. A computer-readable storage medium having stored thereon computer instructions for causing a computer to perform the wind turbine wake field calculation method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311008152.1A CN116757094B (en) | 2023-08-11 | 2023-08-11 | Wind turbine wake field calculation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311008152.1A CN116757094B (en) | 2023-08-11 | 2023-08-11 | Wind turbine wake field calculation method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116757094A CN116757094A (en) | 2023-09-15 |
| CN116757094B true CN116757094B (en) | 2023-11-03 |
Family
ID=87948088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311008152.1A Active CN116757094B (en) | 2023-08-11 | 2023-08-11 | Wind turbine wake field calculation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116757094B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111046533A (en) * | 2019-11-22 | 2020-04-21 | 中国华能集团清洁能源技术研究院有限公司 | Wind turbine generator single tail stream distribution simulation method based on CFD pre-calculation |
| CN115587470A (en) * | 2022-09-07 | 2023-01-10 | 广东电网有限责任公司 | Training method, device, equipment and storage medium for single-fan wake agent model |
| CN115657467A (en) * | 2022-09-07 | 2023-01-31 | 广东电网有限责任公司 | Wind power plant wake flow control optimization method, device, equipment and storage medium |
| WO2023014984A1 (en) * | 2021-08-06 | 2023-02-09 | Texas Tech University System | Methods and systems for thrust-based turbine wake control |
| CN116522807A (en) * | 2022-11-02 | 2023-08-01 | 华北电力大学（保定） | Three-dimensional non-uniform yaw full wake model of wind turbine |
-
2023
- 2023-08-11 CN CN202311008152.1A patent/CN116757094B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111046533A (en) * | 2019-11-22 | 2020-04-21 | 中国华能集团清洁能源技术研究院有限公司 | Wind turbine generator single tail stream distribution simulation method based on CFD pre-calculation |
| WO2023014984A1 (en) * | 2021-08-06 | 2023-02-09 | Texas Tech University System | Methods and systems for thrust-based turbine wake control |
| CN115587470A (en) * | 2022-09-07 | 2023-01-10 | 广东电网有限责任公司 | Training method, device, equipment and storage medium for single-fan wake agent model |
| CN115657467A (en) * | 2022-09-07 | 2023-01-31 | 广东电网有限责任公司 | Wind power plant wake flow control optimization method, device, equipment and storage medium |
| CN116522807A (en) * | 2022-11-02 | 2023-08-01 | 华北电力大学（保定） | Three-dimensional non-uniform yaw full wake model of wind turbine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116757094A (en) | 2023-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Short-term wind power prediction based on multidimensional data cleaning and feature reconfiguration | |
| CN112836064A (en) | Knowledge graph complementing method and device, storage medium and electronic equipment | |
| CN109992889B (en) | Wind power plant model building method and system and wake value calculating method and system | |
| Wang et al. | A computationally-efficient layout optimization method for real wind farms considering altitude variations | |
| CN106778910A (en) | Deep learning system and method based on local training | |
| Stein et al. | Operator-and template-based modeling of solid geometry for Isogeometric Analysis with application to Vertical Axis Wind Turbine simulation | |
| CN102496033B (en) | Image SIFT feature matching method based on MR computation framework | |
| CN116757094B (en) | Wind turbine wake field calculation method and device, electronic equipment and storage medium | |
| CN108880872B (en) | Method and device for decomposing topological structure of Internet test bed | |
| CN102541811A (en) | On-demand computing-based data analysis device and method for analysis factors | |
| CN109934928B (en) | Three-dimensional model simplification method based on skeletonization | |
| Sichani et al. | First passage probability estimation of wind turbines by Markov Chain Monte Carlo | |
| CN116306849A (en) | Training of reverse neural network model and determining method and device of optical processor | |
| WO2023099414A1 (en) | A method for controlling a wind farm | |
| CN113449968B (en) | New energy power grid frequency risk assessment method and device based on deep learning | |
| CN115408958A (en) | Wind turbine simulation method and system based on grid-independent actuating line model | |
| CN113688554B (en) | Flow field calculation method combining discontinuous finite element and immersion boundary method | |
| CN109344875A (en) | Based on clustering day wind power output timing generation method, device and medium | |
| CN114398711A (en) | Concrete gravity dam CAE integrated analysis method and device and electronic equipment | |
| Deskos | Numerical simulations of wind turbine wakes | |
| Hassanzadeh | Best location of the small wind turbine around a single high-rise building | |
| CN116720027B (en) | Wind power plant generating capacity calculation method and device considering wind turbine blocking effect | |
| Yan et al. | Hybrid physical and data driven modeling for dynamic operation characteristic simulation of wind turbine | |
| CN111523657A (en) | Neural network accelerator creating method and device, electronic device and storage medium | |
| CN110990963A (en) | Fan interval optimization method and device and computer readable storage medium |
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 |