CN117407773A - Digital twinning-based method, system and equipment for predicting icing state of fan blade - Google Patents
Digital twinning-based method, system and equipment for predicting icing state of fan blade Download PDFInfo
- Publication number
- CN117407773A CN117407773A CN202311713708.7A CN202311713708A CN117407773A CN 117407773 A CN117407773 A CN 117407773A CN 202311713708 A CN202311713708 A CN 202311713708A CN 117407773 A CN117407773 A CN 117407773A
- Authority
- CN
- China
- Prior art keywords
- fan blade
- data set
- icing
- model
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000012549 training Methods 0.000 claims abstract description 40
- 238000010801 machine learning Methods 0.000 claims abstract description 29
- 238000004088 simulation Methods 0.000 claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims description 12
- 238000007637 random forest analysis Methods 0.000 claims description 11
- 239000000463 material Substances 0.000 claims description 10
- 238000000547 structure data Methods 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 8
- 239000003570 air Substances 0.000 claims description 6
- 239000012080 ambient air Substances 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 description 7
- 238000006073 displacement reaction Methods 0.000 description 5
- 238000010248 power generation Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- 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
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/04—Constraint-based CAD
-
- 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
- G06F2119/08—Thermal analysis or thermal optimisation
-
- 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
-
- 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
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Geometry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Fluid Mechanics (AREA)
- Computing Systems (AREA)
- Algebra (AREA)
- Aerodynamic Tests, Hydrodynamic Tests, Wind Tunnels, And Water Tanks (AREA)
- Structures Of Non-Positive Displacement Pumps (AREA)
Abstract
The invention relates to the technical field of equipment detection, in particular to a digital twinning-based method, a digital twinning-based system and digital twinning-based equipment for predicting icing states of fan blades, which comprise the following steps: collecting historical running state data of the fan blade to obtain an icing state data set and a non-icing state data set; classifying and training the machine learning model by using the icing state data set and the non-icing state data set; constructing a digital twin model of the fan blade; establishing a meteorological data set according to weather conditions of future weather, and performing simulation operation on the meteorological data set through a digital twin model of the fan blade to obtain a new data sample of the running state of the fan blade; based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample. The judgment and prediction of the icing state of the fan blade are more accurate and comprehensive, and the prediction of the icing state is realized.
Description
Technical Field
The invention relates to the technical field of equipment detection, in particular to a digital twinning-based method, a digital twinning-based system and digital twinning-based equipment for predicting icing states of fan blades.
Background
Wind power generation is used as a novel renewable energy source, and has the advantages of environmental friendliness, no exhaustion and capability of realizing distributed power generation. However, in cold conditions, the fan blades are susceptible to icing, resulting in reduced fan performance, increased energy losses, and possibly severe equipment failure. Therefore, timely and accurate detection and determination of icing conditions of the wind turbine blades is critical to stable operation of the wind power generation system.
At present, the existing fan blade icing state detection method has some limitations:
the sensor monitoring method comprises the following steps: conventional methods typically employ temperature sensors, humidity sensors, vibration sensors, and the like to monitor the condition of the fan blades. However, since the icing condition of the surface of the fan blade is affected by various factors, such as meteorological conditions, blade materials, etc., the monitoring result of a single sensor may not be comprehensive and accurate enough.
The image recognition method comprises the following steps: and judging the icing state of the fan blade through an image recognition technology. But this approach requires the installation of a corresponding image capturing apparatus, and the accuracy and complexity of the image recognition algorithm is also a great challenge.
The physical model method comprises the following steps: some researches adopt a mathematical physical model to simulate the icing process of the fan blade, but the method needs a complex model and a large amount of data, and has poor real-time performance.
In summary, the existing method for monitoring the icing condition of the fan blade still has certain limitations in terms of accuracy, instantaneity and comprehensive optimization.
Disclosure of Invention
In order to solve the above problems, the first aspect of the present invention provides a method for predicting icing status of a fan blade based on digital twinning, which comprises the following steps:
collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through a sensor, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
establishing a meteorological data set according to weather conditions of future weather, and performing simulation operation on the meteorological data set through a digital twin model of the fan blade to obtain a new data sample of the running state of the fan blade;
based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
According to the method, by combining a digital twin technology and a machine learning technology, real-time prediction can be performed according to simulation data of the fan blade, the judgment accuracy of the icing state of the fan blade is improved, and the method has important significance for safe and stable operation of a wind power generation system.
In some implementations of the first aspect, the collecting state data of historical operation of the fan blade, wherein each piece of state data of historical operation includes the following characteristic parameters: blade surface temperature, blade vibration, ambient air temperature, ambient humidity, ambient wind speed, and icing conditions of the corresponding fan blade when the blade is in operation.
The method comprises the steps of collecting state data of fan blade historical operation, establishing a sample training set, preprocessing the sample training set, and preprocessing comprises the following steps:
data cleaning: processing state data missing values and abnormal values in a fan blade sample training set;
and (3) selecting data characteristics: selecting characteristic parameters affecting the icing of the fan blade from all the characteristics of the state data;
data feature scaling is performed: and scaling the selected data features with different scales to a uniform range.
In some implementations of the first aspect, the parameter characteristics of the fan blade include:
material characteristics including thermal conductivity, material density, and coefficient of thermal expansion of the fan blade;
geometric parameters including length, width, thickness and twist angle of the fan blade;
aerodynamic properties, including surface pressure distribution and wind drag of the surface of the fan blade affected by the wind.
The method comprises the steps of establishing a digital model of the fan blade according to the parameter characteristics of the fan blade, and specifically comprises an elastic model, a heat conduction model, a curved surface model and a hydrodynamic calculation model.
In some implementations of the first aspect, the method for establishing a weather dataset according to weather conditions of future weather, and performing simulated operation on the weather dataset by using a digital twin model of a fan blade includes:
acquiring weather conditions of the weather of the future week according to weather forecast, and calculating to obtain the average air temperature, average humidity and average wind speed of each day of the week of the future;
and establishing a meteorological data set by using the average air temperature, the average humidity and the average wind speed, and performing simulation operation on the meteorological data by using a digital twin model of the fan blade to obtain blade surface temperature and blade vibration data based on daily meteorological conditions.
Meanwhile, the simulation operation of the meteorological data set through the digital twin model of the fan blade further comprises the following steps: and combining the daily meteorological data set with the correspondingly acquired blade surface temperature and blade vibration data to establish a new data sample.
In some implementations of the first aspect, the classifying and predicting the icing state of the fan blade for the new data sample based on the trained machine learning model specifically includes:
the machine learning model adopts a random forest model, new sample data is input into the random forest model to judge results, and if the judging result is 1, the fan blade is in an icing state; if the judgment result is 0, the fan blade is in a non-icing state; and obtaining a final prediction result by voting or averaging the judgment results of all the classifiers in the random forest model.
A second aspect provides a digital twinning-based fan blade icing condition prediction system comprising:
and a collection module: collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
training module: classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
model construction module: establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through an actual sensor system, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
and (3) an analog module: establishing a meteorological data set according to weather conditions of future weather, performing simulation operation on the meteorological data set through a digital twin model of the fan blade, and acquiring a new data sample of the running state of the fan blade;
and a prediction module: based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
A third aspect provides a digital twin based fan blade icing condition prediction apparatus, comprising a processor and a memory, wherein the processor implements the digital twin based fan blade icing condition prediction method as described above when executing program data stored in the memory.
The beneficial effects are that:
according to the invention, the machine learning model is used for optimizing the identification and judgment of the icing state of the fan blade, and compared with the traditional method, the method can intelligently process the collected operation data, and improves the accuracy and instantaneity of judgment;
according to the invention, the correspondence between the fan blade and the twin model is realized through a digital twin technology, and the digital twin technology is combined with a machine learning technology, so that the judgment and the prediction of the icing state of the fan blade are more accurate and comprehensive, the prediction of the icing state is realized, the decision and the determination rationality of the model are optimized, and the capacity loss and the equipment fault risk are reduced;
according to the invention, the digital twin model of the fan blade is used for obtaining the real-time running state data of the blade according to the future weather condition simulation, so that the trend analysis and the state prediction of the icing state of the fan blade can be realized, the trend and the fault of the icing state of the blade can be found in time, corresponding preventive measures are taken, the reliability of the fan is improved, and the stability of wind power generation is ensured.
Detailed Description
Examples: the invention provides a fan blade icing state prediction method based on digital twinning, which comprises the following steps:
step one: collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
preparing a sample training set for machine learning technology, wherein the sample training set comprises blade related data in an icing state of historical operation of a fan blade and blade related data in a normal non-icing state, and preprocessing the sample training set, and the preprocessing operation comprises the following steps:
data cleaning: the missing value and the abnormal value of the state data in the fan blade sample training set are processed, and the integrity and the reliability of the data are ensured; wherein the missing values indicate that certain data points in the data set lack the numerical value and information of specific characteristics, and in the state data of the fan blade, partial measured values may not be recorded and collected at some time points or under some conditions, so that the missing values exist in the sample training set; the processing method for the missing value specifically comprises the following steps: deleting the data sample with the missing value: if most of the characteristic data in the data sample is missing, deleting the data sample; or interpolation may be used to estimate missing values from information from known data points, such as using linear interpolation, mean, median, or other statistical methods to fill in missing values. The outliers are values in the data set that do not correspond to most data points, and may be due to measurement errors, equipment failure, or other anomalies; the abnormal value processing method specifically comprises the steps of deleting the abnormal value or correcting and replacing the abnormal value, and replacing the abnormal value with a reasonable value.
And (3) selecting data characteristics: selecting characteristic parameters affecting the icing of the fan blade from all the characteristics of the state data, wherein the characteristic parameters comprise the surface temperature and vibration of the blade, and the ambient air temperature, humidity and wind speed of the fan blade in the running state;
data feature scaling is performed: and scaling the selected data features with different scales to a uniform range, and adjusting the numerical ranges among the different data features to enable the data features to have similar scales. The method can ensure that different data features have relatively uniform influence on model training, and prevent certain data features from occupying dominant positions in model training to influence model training results.
And acquiring an icing state data set and a non-icing state data set by preprocessing the collected fan blade sample data set.
As a specific embodiment, the historical data defining the fan blade has N pieces, wherein each piece of historical data contains the following characteristic parameters: blade surface temperature (K), blade Vibration (Vibration), ambient air temperature (T), humidity (H), wind speed (V) and icing state of a corresponding fan blade when the blade operates, wherein an icing state Label (Label) adopts binary value to represent icing, 1 represents non-icing;
building each piece of history data into a matrixx n （n∈[1，N]) And establishing an icing condition data set D based on the icing condition labels 1 And a non-icing condition data set D 0 ，D 1 ={x 1 ,x 2 ,…,x n ,1}，D0={x 1 ,x 2 ,…,x n, 0}。
Step two: classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
the machine learning technique employed in this embodiment is a random forest learning algorithm by subtracting from the icing condition dataset D 1 Randomly select N ’ Group data samples, creating a new training subset D 1 ’ The method comprises the steps of carrying out a first treatment on the surface of the Randomly selecting M features from the feature parameters of the data sample, and establishing a new feature subset F; using the training subset D ’ Training a decision tree model by the feature subset F to obtain a base classifier; and repeating the above processes to obtain a plurality of base classifiers. Similarly, classification training is performed through the non-icing state data set (consistent with the above process, which is not described here in detail), and all the obtained base classifiers are used for constructing a random forest model.
Step three: establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through a sensor, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
s3.1, collecting parameter characteristics of the fan blade and establishing a digital model of the fan blade;
the parametric properties of the fan blade include material properties, geometric parameters and aerodynamic properties. Wherein, the material characteristics comprise the thermal conductivity, the material density and the thermal expansion coefficient of the fan blade; geometric parameters including length, width, thickness and twist angle of the fan blade; aerodynamic characteristics, including surface pressure distribution and wind resistance of the surface of the fan blade affected by wind, while collecting meteorological conditions of the environment in which the fan blade is located, including ambient temperature, humidity and wind speed, are required when collecting aerodynamic characteristics of the fan blade.
S3.1.1 based on the material characteristics of the fan blade, establishing an elastic model and a heat conduction model of the fan blade;
the elastic model is used for describing deformation behavior of the blade under the action of stress, and is established by using a finite element method (Finite Element Method, abbreviated as FEM). FEM divides the fan blade into a plurality of finite element units; setting the elastic properties of the material of each finite element unit, including the elastic modulus and poisson's ratio; and define the load and boundary constraints of each finite element, the load refers to the external force or external moment acting on the stressed object, and in the elastic model of the fan blade, the load is derived from the force or moment of various external factors, such as wind load, mechanical vibration load, etc., which can cause the deformation of the blade. And the loads may be static (time-invariant) or dynamic (time-variant), the magnitude, direction and point of action of these loads need to be precisely defined when the elastic model is built in order to simulate the response of the blade under different conditions. The boundary constraint is a condition for limiting displacement or displacement derivative (such as velocity or acceleration) of certain points on the stressed object. In an elastic model of a fan blade, the support points or connection points of the blade are simulated by boundary constraints, and displacements at these points can be limited to simulate the stress conditions of the blade at these points. These constraints may be fixed points (displacements are not allowed) or points that allow limited displacements, depending on the complexity of the model and the physical system reality; the deformation behavior of the fan blade under the action of stress is calculated through an elastic equation set, wherein the elastic equation set is as follows:
，
wherein,σis the stress tensor of the stress, which is the stress,εis the strain tensor, which is the strain tensor,Eis the modulus of elasticity.
The heat conduction and conduction model is used for describing the distribution and transmission of the temperature field inside the blade, and is established according to a heat conduction equation, and the specific formula is as follows:
，
wherein,indicating the rate of change of temperature over time;αis the thermal conductivity, which indicates the rate of heat transfer in a substance; />Representing the gradient and divergence of temperature in space.
S3.1.2 based on the geometric parameters of the fan blade, establishing a NURBS (Non-Uniform Rational B-Spline) curved surface model;
the NURBS surface model is used to describe the geometry of the blade, in particular:
a. determining a control point: selecting a proper number of control points to control the shape of the curved surface;
b. determining the weight of the control point: distributing the weight of each control point for controlling the local influence of the points on the curved surface;
c. defining a node vector: according to the overall shape of the blade curved surface, defining a node vector to determine a NURBS curved surface definition formula is as follows:
，
wherein S (u, v) represents a curved surface control point, N i,p (u) and N j,p (v) Is a B spline basis function, P i,j Is the control point omega i,j Is a weight.
S3.1.3 establishing a hydrodynamic calculation model to determine the aerodynamic characteristics of the fan blade according to the meteorological conditions of the environment to which the fan blade belongs;
the hydrodynamic calculation model is used for describing the surface pressure distribution and wind resistance of the fan blade affected by wind power; the hydrodynamic computational model is represented by the Navier-Stokes equation and Reynolds average Navier-Stokes equation (RANS), in particular:
a. dividing grid units on the surface and surrounding space of the fan blade;
b. setting boundary conditions for the specified blade surface and its surrounding space;
c. solving a Navier-Stokes equation by adopting a numerical method to obtain the speed distribution of the wind field:
the Navier-Stokes equation (simplified, two-dimensional problem) is:
，
where u is the velocity field, p is the pressure field, ρ is the liquid density, v is the kinematic viscosity,is a gradient operator, < >>Is a laplace operator.
d. According to the speed distribution of the wind field, solving the RANS equation to obtain the pressure distribution and wind resistance of the surface of the fan blade, wherein the specific equation formula is as follows:
，
wherein,、/>representing different reynolds average velocity components +.>Is the average pressure of the reynolds number,vis a dynamic viscosity system, ">Is the turbulent Reynolds stress tensor, τ ij Is the turbulent stress tensor;x i 、x j representing different spatial coordinates.
S3.2, acquiring running state data of the fan blade through a sensor system connected with the actual fan blade, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set;
relevant sensors are installed on an actual fan blade, the relevant sensors comprise a temperature sensor, a humidity sensor, a wind speed sensor and a vibration sensor, relevant data of the surface and the structure of the fan blade in the running state are collected in real time, the collected data are preprocessed, data noise is removed, abnormal values of the data are filtered and corrected, the accuracy and the reliability of the data are ensured, and a processed structure data set, a processed heat conductivity data set and a processed pneumatic data set are obtained.
S3.3, carrying out association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
and correlating the preprocessed structure data set, the heat conductivity data set and the pneumatic data set with a pre-constructed digital model of the fan blade, so that high-precision mapping of the actual fan blade and the digital model is realized, and a digital twin model of the fan blade is constructed by continuously updating the digital model.
Step four: establishing a meteorological data set according to weather conditions of future weather, and performing simulation operation on the meteorological data set through a digital twin model of the fan blade to obtain a new data sample of the running state of the fan blade;
acquiring weather conditions of the weather of the future week according to weather forecast, and calculating to obtain the average air temperature, average humidity and average wind speed of each day of the week of the future;
establishing a meteorological data set by using the average air temperature, the average humidity and the average wind speed, and performing simulation operation on the meteorological data by using a digital twin model of the fan blade to obtain blade surface temperature and blade vibration data based on daily meteorological conditions;
and combining the daily meteorological data set with the correspondingly acquired blade surface temperature and blade vibration data to establish a new data sample.
Step five: based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
Inputting new sample data into the random forest model to judge the result, and if the judging result is 1, indicating that the fan blade is in an icing state; if the judgment result is 0, the fan blade is in a non-icing state; and obtaining a final prediction result by voting or averaging the judgment results of all the classifiers in the random forest model.
And when the final result of the fan blade predicted by the random forest model is in an icing state, starting a heating device of the actual fan blade or adjusting the operation modulus of the fan to physically remove ice.
In addition, the invention also provides a fan blade icing state prediction system based on digital twinning, which comprises the following steps:
and a collection module: collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
training module: classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
model construction module: establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through an actual sensor system, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
and (3) an analog module: establishing a meteorological data set according to weather conditions of future weather, performing simulation operation on the meteorological data set through a digital twin model of the fan blade, and acquiring a new data sample of the running state of the fan blade;
and a prediction module: based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
Finally, a device for predicting icing conditions of fan blades based on digital twinning is provided, which comprises a processor and a memory, wherein the processor realizes the method for predicting icing conditions of fan blades based on digital twinning when executing program data stored in the memory.
Claims (10)
1. The method for predicting the icing state of the fan blade based on digital twinning is characterized by comprising the following steps:
collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through a sensor, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
establishing a meteorological data set according to weather conditions of future weather, and performing simulation operation on the meteorological data set through a digital twin model of the fan blade to obtain a new data sample of the running state of the fan blade;
based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
2. The method for predicting icing conditions of a digitally twinned fan blade according to claim 1, wherein said collecting historical operational status data of the fan blade, wherein each historical operational status data comprises the following characteristics: blade surface temperature, blade vibration, ambient air temperature, ambient humidity, ambient wind speed, and icing conditions of the corresponding fan blade when the blade is in operation.
3. The method for predicting icing conditions of a fan blade based on digital twinning of claim 1, wherein collecting the historical operational status data of the fan blade creates a sample training set, and preprocessing the sample training set comprises:
data cleaning: processing state data missing values and abnormal values in a fan blade sample training set;
and (3) selecting data characteristics: selecting characteristic parameters affecting the icing of the fan blade from all the characteristics of the state data;
data feature scaling is performed: and scaling the selected data features with different scales to a uniform range.
4. The digital twinning-based fan blade icing condition prediction method of claim 1, wherein the parameter characteristics of the fan blade include:
material characteristics including thermal conductivity, material density, and coefficient of thermal expansion of the fan blade;
geometric parameters including length, width, thickness and twist angle of the fan blade;
aerodynamic properties, including surface pressure distribution and wind drag of the surface of the fan blade affected by the wind.
5. The method for predicting icing condition of a fan blade based on digital twinning of claim 1, wherein the establishing a digital model of the fan blade based on the parameter characteristics of the fan blade specifically comprises establishing an elastic model, a heat conduction model, a curved surface model and a hydrodynamic calculation model.
6. The method for predicting icing state of a fan blade based on digital twin according to claim 1, wherein the method for establishing a meteorological data set according to weather conditions of future weather and performing simulated operation on the meteorological data set by using a fan blade digital twin model comprises the following specific steps:
acquiring weather conditions of the weather of the future week according to weather forecast, and calculating to obtain the average air temperature, average humidity and average wind speed of each day of the week of the future;
and establishing a meteorological data set by using the average air temperature, the average humidity and the average wind speed, and performing simulation operation on the meteorological data by using a digital twin model of the fan blade to obtain blade surface temperature and blade vibration data based on daily meteorological conditions.
7. The digital twinning-based fan blade icing condition prediction method of claim 6, further comprising: and combining the daily meteorological data set with the correspondingly acquired blade surface temperature and blade vibration data to establish a new data sample.
8. The method for predicting icing conditions of a fan blade based on digital twinning according to claim 1, wherein the classifying and predicting the icing conditions of the fan blade for the new data sample based on the trained machine learning model specifically comprises:
the machine learning model adopts a random forest model, new sample data is input into the random forest model to judge results, and if the judging result is 1, the fan blade is in an icing state; if the judgment result is 0, the fan blade is in a non-icing state; and obtaining a final prediction result by voting or averaging the judgment results of all the classifiers in the random forest model.
9. A digital twinning-based fan blade icing condition prediction system, comprising:
and a collection module: collecting historical running state data of the fan blades, establishing a sample training set, and preprocessing the sample training set to obtain an icing state data set and a non-icing state data set;
training module: classifying and training the machine learning model by using the icing state data set and the non-icing state data set to obtain a trained machine learning model;
model construction module: establishing a digital model of the fan blade according to the parameter characteristics of the fan blade; acquiring the running state data of the fan blade through an actual sensor system, and acquiring a preprocessed structure data set, a preprocessed thermal conductivity data set and a preprocessed pneumatic data set; performing association mapping on the structural data set, the thermal conductivity data set and the pneumatic data set and the fan blade digital model to construct a fan blade digital twin model;
and (3) an analog module: establishing a meteorological data set according to weather conditions of future weather, performing simulation operation on the meteorological data set through a digital twin model of the fan blade, and acquiring a new data sample of the running state of the fan blade;
and a prediction module: based on the trained machine learning model, the classification prediction of the icing state of the fan blade is carried out on the new data sample.
10. A digital twin based fan blade icing condition prediction apparatus comprising a processor and a memory, wherein the processor implements the digital twin based fan blade icing condition prediction method of any of claims 1-8 when executing program data stored in the memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311713708.7A CN117407773B (en) | 2023-12-14 | 2023-12-14 | Digital twinning-based method, system and equipment for predicting icing state of fan blade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311713708.7A CN117407773B (en) | 2023-12-14 | 2023-12-14 | Digital twinning-based method, system and equipment for predicting icing state of fan blade |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117407773A true CN117407773A (en) | 2024-01-16 |
| CN117407773B CN117407773B (en) | 2024-06-18 |
Family
ID=89500286
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311713708.7A Active CN117407773B (en) | 2023-12-14 | 2023-12-14 | Digital twinning-based method, system and equipment for predicting icing state of fan blade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117407773B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109209790A (en) * | 2018-10-09 | 2019-01-15 | 浙江运达风电股份有限公司 | A kind of wind electricity blade icing estimation method based on data modeling |
| US20210332796A1 (en) * | 2017-12-14 | 2021-10-28 | Beijing Goldwind Science & Creation Windpower Equipment Co., Ltd. | Method and apparatus for icing prediction, method and apparatus for generatng icing prediction model |
| CN114382662A (en) * | 2022-01-21 | 2022-04-22 | 华电安诺(北京)信息科技有限公司 | Fan state early warning method based on digital twinning |
| CN114757048A (en) * | 2022-04-28 | 2022-07-15 | 北京千尧新能源科技开发有限公司 | Health state assessment method, device, equipment and medium for fan foundation |
| KR20230032804A (en) * | 2021-08-31 | 2023-03-07 | 주식회사 아이지아이 코리아 | Air quality management method using digital twin model |
| CN115859837A (en) * | 2023-02-23 | 2023-03-28 | 山东大学 | Fan blade dynamic impact detection method and system based on digital twin modeling |
| CN116167531A (en) * | 2023-04-26 | 2023-05-26 | 天津福天科技有限公司 | Photovoltaic power generation prediction method based on digital twin |
| CN116451594A (en) * | 2023-06-15 | 2023-07-18 | 北京东润环能科技股份有限公司 | Training method and device of icing prediction model, prediction method and device and electronic equipment |
-
2023
- 2023-12-14 CN CN202311713708.7A patent/CN117407773B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210332796A1 (en) * | 2017-12-14 | 2021-10-28 | Beijing Goldwind Science & Creation Windpower Equipment Co., Ltd. | Method and apparatus for icing prediction, method and apparatus for generatng icing prediction model |
| CN109209790A (en) * | 2018-10-09 | 2019-01-15 | 浙江运达风电股份有限公司 | A kind of wind electricity blade icing estimation method based on data modeling |
| KR20230032804A (en) * | 2021-08-31 | 2023-03-07 | 주식회사 아이지아이 코리아 | Air quality management method using digital twin model |
| CN114382662A (en) * | 2022-01-21 | 2022-04-22 | 华电安诺(北京)信息科技有限公司 | Fan state early warning method based on digital twinning |
| CN114757048A (en) * | 2022-04-28 | 2022-07-15 | 北京千尧新能源科技开发有限公司 | Health state assessment method, device, equipment and medium for fan foundation |
| CN115859837A (en) * | 2023-02-23 | 2023-03-28 | 山东大学 | Fan blade dynamic impact detection method and system based on digital twin modeling |
| CN116167531A (en) * | 2023-04-26 | 2023-05-26 | 天津福天科技有限公司 | Photovoltaic power generation prediction method based on digital twin |
| CN116451594A (en) * | 2023-06-15 | 2023-07-18 | 北京东润环能科技股份有限公司 | Training method and device of icing prediction model, prediction method and device and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| RUPFLE, J. ET AL.: "Long-Term MultipleSensorMonitoring of a Hybrid Tower Wind Turbine - Lessons Learned", EXPERIMENTAL VIBRATION ANALYSIS FOR CIVIL ENGINEERING STRUCTURES: EVACES 2023 . LECTURE NOTES IN CIVIL ENGINEERING (432), 2 August 2023 (2023-08-02), pages 539 * |
| 李大中;刘家瑞;张华英;: "基于深度全连接神经网络的风机叶片结冰预测方法", 电力科学与工程, no. 04, 28 April 2019 (2019-04-28), pages 43 - 48 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117407773B (en) | 2024-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022151890A1 (en) | Method for establishing power generating capacity evaluation and micro-siting model for wind farms | |
| WO2019114160A1 (en) | Method and device for predicting ice formation, and model generation method and device | |
| CN109751206B (en) | Fan blade icing fault prediction method and device and storage medium | |
| WO2018171165A1 (en) | Fault prediction method and device for fan | |
| Pimenta et al. | Development of a digital twin of an onshore wind turbine using monitoring data | |
| JP4875661B2 (en) | Aircraft soundness diagnosis apparatus and method, and program | |
| CN109522627B (en) | Fan blade icing prediction method based on SCADA (Supervisory control and data acquisition) data | |
| CN110006552B (en) | Method for detecting abnormal temperature of equipment | |
| CN112818604A (en) | Wind turbine generator risk degree assessment method based on wind power prediction | |
| Shankar Verma et al. | A probabilistic long‐term framework for site‐specific erosion analysis of wind turbine blades: A case study of 31 Dutch sites | |
| CN113847216B (en) | Fan blade state prediction method, device, equipment and storage medium | |
| CN106875037A (en) | Wind-force Forecasting Methodology and device | |
| CN115438726A (en) | Device life and fault type prediction method and system based on digital twin technology | |
| CN115575104A (en) | Fan blade damage rapid detection method based on inverse finite element reconstruction image recognition | |
| Li et al. | Icing condition prediction of wind turbine blade by using artificial neural network based on modal frequency | |
| TWI546762B (en) | Wind power generation equipment of the stress estimation device and wind power equipment, the stress estimation method, wind power generation system | |
| Azzam et al. | Development of a wind turbine gearbox virtual load sensor using multibody simulation and artificial neural networks | |
| CN111340307A (en) | Method for predicting wind power generation power of fan and related device | |
| CN117407773B (en) | Digital twinning-based method, system and equipment for predicting icing state of fan blade | |
| CN114742363A (en) | Energy efficiency state evaluation method, system and medium for wind turbine generator | |
| CN113722860A (en) | Transient thermodynamic state online evaluation method, device and medium based on reduced order model | |
| CN117056402B (en) | Motor diagnosis method and device based on multi-source signals and storage medium | |
| GB2555010B (en) | Determining loads on a wind turbine | |
| TWI780434B (en) | Abnormal diagnosis device and method | |
| CN110414734B (en) | Method for forecasting and evaluating wind resource utilization rate |
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 |