CN117759490A - Wind turbine generator blade fault monitoring method and system - Google Patents
Wind turbine generator blade fault monitoring method and system Download PDFInfo
- Publication number
- CN117759490A CN117759490A CN202311701585.5A CN202311701585A CN117759490A CN 117759490 A CN117759490 A CN 117759490A CN 202311701585 A CN202311701585 A CN 202311701585A CN 117759490 A CN117759490 A CN 117759490A
- Authority
- CN
- China
- Prior art keywords
- noise
- wind turbine
- signal
- blade
- fault
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000012544 monitoring process Methods 0.000 title claims abstract description 54
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 33
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 238000003745 diagnosis Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 13
- 230000006870 function Effects 0.000 claims description 72
- 238000001914 filtration Methods 0.000 claims description 43
- 238000005259 measurement Methods 0.000 claims description 31
- 230000008569 process Effects 0.000 claims description 30
- 230000015654 memory Effects 0.000 claims description 26
- 239000011159 matrix material Substances 0.000 claims description 24
- 238000009432 framing Methods 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 13
- 230000004913 activation Effects 0.000 claims description 12
- 230000007787 long-term memory Effects 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 10
- 230000006403 short-term memory Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000009826 distribution Methods 0.000 claims description 8
- 230000036541 health Effects 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 5
- 239000000835 fiber Substances 0.000 claims description 4
- 238000012546 transfer Methods 0.000 claims description 4
- 230000000977 initiatory effect Effects 0.000 claims description 3
- 238000011161 development Methods 0.000 abstract description 3
- 230000008859 change Effects 0.000 abstract description 2
- 238000013528 artificial neural network Methods 0.000 description 7
- 210000004027 cell Anatomy 0.000 description 6
- 230000007774 longterm Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 5
- 239000013307 optical fiber Substances 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000008034 disappearance Effects 0.000 description 2
- 238000004880 explosion Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000011152 fibreglass Substances 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000005654 stationary process Effects 0.000 description 1
- -1 timber Substances 0.000 description 1
- 238000000489 vacuum metal deposition Methods 0.000 description 1
Classifications
-
- 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
Abstract
The invention discloses a wind turbine generator blade fault monitoring method and system, wherein the method comprises the following steps: collecting noise signals through a noise sensor arranged inside the wind turbine generator blade; performing data preprocessing on the noise signals; decomposing the preprocessed noise signal into a plurality of finite inherent mode function components through variation mode decomposition; and carrying out fault diagnosis through a plurality of inherent mode function components. After the blade is cracked, the damage condition of the blade can be monitored through the change of noise in the blade, the monitoring range of the wind turbine generator blade fault monitoring method is comprehensive, all-weather online monitoring can be realized, the generated data amount is less, the processing is convenient, and the algorithm development difficulty and the application cost are lower.
Description
Technical Field
The invention relates to the technical field of wind turbines, in particular to a wind turbine blade fault monitoring method and system.
Background
Because the wind turbine blade is made of composite materials such as timber, glass fiber reinforced plastic and the like, and is poured through a die and finally assembled through a die, the wind turbine blade can move at a high speed for a long time under severe environments along with the increase of the length and the mass of the wind turbine blade, and the wind turbine blade is cracked. In order to discover faults of the wind turbine generator blades in time and avoid the faults affecting the operation of the wind turbine generator, the wind turbine generator blades need to be monitored.
The existing wind turbine generator blade monitoring scheme mainly comprises optical monitoring, unmanned aerial vehicle monitoring and load monitoring. However, the optical monitoring scheme has the problem that the tip failure and the internal crack failure of the blade cannot be accurately identified; the unmanned aerial vehicle monitoring has the problems that the unmanned aerial vehicle monitoring is only suitable for the inspection function and long-term online monitoring cannot be realized; the load monitoring is carried out by installing more strain and acceleration sensors, calibrating a blade pneumatic model and static load, so that the problems of higher cost and more complicated data processing exist. Therefore, there is a need for a comprehensive and simpler solution for long-term on-line monitoring of wind turbine blades.
Disclosure of Invention
In order to overcome the problems in the related art, the invention provides a wind turbine generator blade fault monitoring method and system, which are used for solving the problems that in the prior art, the wind turbine generator blade monitoring scheme is limited in monitoring range and cannot realize long-term online monitoring, or is high in arrangement cost and complex in data processing.
According to a first aspect of the invention, a wind turbine generator blade fault monitoring method is provided, the method comprising:
collecting noise signals through a noise sensor arranged inside the wind turbine generator blade;
Performing data preprocessing on the noise signals;
decomposing the preprocessed noise signal into a plurality of finite inherent mode function components through variation mode decomposition;
and carrying out fault diagnosis through a plurality of inherent mode function components.
Preferably, the noise sensor is a fiber optic MEMS noise sensor.
Preferably, the noise sensor is installed in the following manner:
1 noise sensor is installed in the wind turbine blade at intervals of a specified distance from the blade root; and the number and the relative positions of the noise sensors arranged on each blade of the wind turbine are kept consistent.
Preferably, the data preprocessing of the noise signal comprises data framing of the noise signal; framing the noise signal, including:
dividing the equal angle of the motion track of the wind turbine blade on the wind turbine rotating plane into a plurality of angle ranges by taking the wind turbine as the circle center, and marking the noise signal acquired on each angle range with a corresponding angle label; and the angle labels marked on the noise signals on each blade of the wind turbine generator are different;
and carrying out data framing on the noise signal according to the angle label.
Preferably, the data preprocessing of the noise signal comprises data filtering the noise signal using kalman filtering; data filtering the noise signal using kalman filtering, comprising:
the state equation of the wind turbine generator system for collecting the noise signals is as follows:
x t ＝Ax t-1 +Bu t-1 +ω t-1
p(ω)～N(0,Q)
wherein x is t-1 For the state vector at time t-1, u t-1 For the input vector at time t-1, ω t-1 The method comprises the steps that A is a state transfer coefficient matrix of a noise signal, B is an optional control input matrix, Q is a covariance matrix of the process noise, and the process noise meets Gaussian distribution;
the measurement equation of the noise signal is:
z t ＝Hx t +v t
p(v)～N(0,R)
wherein z is t For the measurement vector at time t, v t For the corresponding measurement error, H is a measurement coefficient matrix, R is a covariance matrix of the measurement error, and the measurement error meets Gaussian distribution;
based on Kalman filtering, the time update equation of the noise signal is as follows:
wherein,p is the covariance matrix of the error, which is the predicted value of the state vector;
based on Kalman filtering, the measurement update equation of the noise signal is as follows:
where K is the kalman filter gain.
Preferably, the fault diagnosis by a plurality of natural mode function components includes:
Inputting a plurality of inherent mode function components into a designated long-period and short-period memory network to obtain a fault prediction result; the long-term and short-term memory network is obtained through training of a data set consisting of a plurality of fault signals and health signals in advance;
and if the fault prediction result indicates that the fault occurs, initiating a fault warning to prompt fault processing.
Preferably, the long-term memory network comprises an input layer, an implicit layer and an output layer; the hidden layer comprises a memory unit and a control unit, wherein the control unit comprises an input door, a forget door and an output door;
the long-term and short-term memory network sequentially performs the following calculation in the forward propagation process:
calculating and updating the value of the forgetting gate, wherein the formula is as follows:
f (t) ＝σ(W f h (t-1) +W i x (t) +b f )
wherein sigma is a sigmoid activation function, W f B is forgetting the gate weight f Is a forget gate threshold;
calculating and updating the value of the input gate, wherein the formula is as follows:
i (t) ＝σ(W i h (t-1) +W i x (t) +b i )
c ,(t) ＝tanh(W c h (t-1) +W c x (t) +b c )
wherein W is i 、W c 、b i And b c Respectively obtaining weights and thresholds corresponding to a sigmoid activation function and a tanh activation function of the input gate;
calculating and updating the value of the memory unit, wherein the formula is as follows:
wherein,is dot multiplied;
calculating and updating the value of the output gate, wherein the formula is as follows:
o (t) ＝σ(W o h (t-1) +W o x (t) +b o )
Wherein W is o To output the gate weight, b o To output the threshold value of h (t) An output vector for the hidden layer;
updating the prediction output at the current moment, wherein the formula is as follows:
wherein V and c are the weight and threshold of the connection of the hidden layer to the output layer, respectively.
Preferably, in the process of decomposing the noise signal after preprocessing into a limited number of inherent modal function components through a variational modal decomposition, the variational modal decomposition formula is as follows:
wherein f is inputNoise signal u k K is the number of the natural mode function components, omega k Delta (t) is a pulse function for the center frequency of the natural mode function component;
and introducing a Lagrange multiplication operator into the variation modal decomposition formula to solve, wherein the expression is as follows:
wherein lambda is Lagrange multiplier, alpha is penalty factor, f (t) is noise signal to be decomposed, u k (t) is the kth natural mode function component;
after the Lagrange multiplication operator is introduced, the solution method of the variation modal decomposition formula is as follows:
initializing the natural modal function componentCenter frequency of the intrinsic mode function component +. >And the Lagrangian multiplier +.>
Initializing the iteration number n=0, carrying out a plurality of iterative calculations with n=n+1, and carrying out the following formula on the inherent mode function component in each iterative calculationCenter frequency of the intrinsic mode function component +.>And the Lagrangian multiplier +.>Updating:
where τ is the noise margin;
stopping iterative calculation according to a preset precision value epsilon, and when the following conditions are met:
outputting a plurality of natural mode function components obtained by decomposition.
Preferably, the data preprocessing of the noise signal comprises data screening of the noise signal; data screening is carried out on the noise signals, wherein the data screening comprises at least one of the following steps:
judging the running state of the wind turbine generator according to the fan grid-connected signal, and filtering noise signals collected by the wind turbine generator in a non-grid-connected state;
judging whether the wind turbine generator is in a yaw state according to the yaw signal of the fan, and filtering noise signals collected by the wind turbine generator in the yaw state.
According to a second aspect of the invention, a wind turbine blade fault monitoring system is provided, the system comprises a plurality of noise sensors, a signal modem, a signal receiving station and a monitoring center, wherein the noise sensors, the signal modem, the signal receiving station and the monitoring center are arranged inside the wind turbine blade;
The noise sensor is used for collecting noise signals generated by the wind turbine generator blades;
the signal modem is used for receiving the noise signals acquired by the noise sensor and transmitting the noise signals to the signal receiving station through wireless connection;
the signal receiving station is used for transmitting the noise signal to the monitoring center through a ring network of the wind turbine generator;
the monitoring center is used for storing the noise signals and carrying out fault monitoring on the wind turbine blade through the wind turbine blade fault detection method according to any embodiment of the invention.
The invention discloses a wind turbine generator blade fault monitoring method and system, wherein noise sensors are arranged in blades to monitor noise signals generated at all parts in the blades to identify the fault condition of the blades, and the monitoring range is comprehensive; moreover, as the noise sensor is arranged in the blade and is not influenced by external environment, all-weather online monitoring can be realized; in addition, the invention only carries out fault diagnosis through the noise signals collected by the noise sensor, and compared with the existing load monitoring scheme, the invention has the advantages of less generated data quantity, single data type, more convenient processing and lower algorithm development difficulty and application cost.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.
Drawings
FIG. 1 is a flow chart illustrating a method for monitoring a wind turbine blade failure according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a long-short-term memory network according to an embodiment of the present invention.
FIG. 3 is a schematic structural diagram of a wind turbine blade failure monitoring system according to an embodiment of the present invention.
FIG. 4 is a flow chart illustrating operation of a wind turbine blade failure detection system according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any or all possible combinations of one or more of the associated listed items.
It should be understood that although the terms first, second, third, etc. may be used herein to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the invention. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.
The invention is described in detail below with reference to the drawings and the detailed description.
Referring to fig. 1, fig. 1 is a flowchart of a method for monitoring faults of a wind turbine blade according to an embodiment of the present invention, including the following steps:
s101, collecting noise signals through a noise sensor arranged inside a blade of a wind turbine generator;
step S102, data preprocessing is carried out on noise signals;
step S103, decomposing the preprocessed noise signal into a plurality of finite inherent mode function components through variation mode decomposition;
and step S104, performing fault diagnosis through a plurality of inherent mode function components.
In step S101, a noise sensor may be installed inside a wind turbine blade to collect noise information sent by the wind turbine blade during operation of the wind turbine in real time, so as to generate a noise signal, and the noise signal may be used to determine whether the wind turbine blade fails. Specifically, in order to avoid that the judgment of noise signals is influenced by the states or differences of the blades, before the wind turbine blade fault monitoring method is applied, the blade state inspection is required to be carried out on all the blades of the wind turbine, and the model consistency and the good state of each blade using the same noise identification method are ensured.
In particular, the noise sensor used in the present invention may be a fiber MEMS (Micro Electromechanical System ) noise sensor. The optical fiber MEMS noise sensor is a noise sensor manufactured by adopting an MEMS process to manufacture a sensing chip and packaging the sensing chip and a dual-optical fiber collimator. The MEMS chip has the characteristics of small volume, stable performance and good parameter consistency; and the optical fiber sensor is combined with the optical fiber sensing technology, so that the problem that the existing sensor is limited by broadband and high-precision is solved. Compared with the traditional sensor, the optical fiber MEMS noise sensor has the advantages of electrically insulating a sensing end, no power supply, electromagnetic interference resistance, small size, long transmission distance and the like, and is more suitable for collecting noise signals in the blades of the wind turbine generator. In particular, in some embodiments, the noise sensor used may also be other MEMS noise sensors, other fiber-optic noise sensors, or other types of noise sensors, as long as they can be mounted inside wind turbine blades and enable transmission of noise signals between wind turbines, as the invention is not limited in this regard.
Specifically, the installation mode of the noise sensor may be: 1 noise sensor is installed in the wind turbine blade from the blade root at intervals of a designated distance; and the number and the relative positions of the noise sensors arranged on each blade of the wind turbine are kept consistent.
Specifically, when installing the noise sensors, 1 noise sensor can be installed at intervals of a specified distance from the blade root inside the blade of the wind turbine generator, so that the number of the noise sensors to be installed is reduced as much as possible under the condition that a sufficient relative distance is ensured to be maintained between each noise sensor, so that noise signals acquired and obtained by all the noise sensors are more comprehensive. Specifically, the specified distance may be 10m, that is, 1 noise sensor is installed every 10m distance from the blade root; specifically, the distance between the noise sensors may be set to other values, which the present invention is not limited to.
Specifically, when the noise sensors are installed, the number and the relative positions of the noise sensors installed on each blade of the wind turbine generator can be kept consistent, so that the problem that the noise signal analysis and the fault condition judgment are affected by the installation difference of the noise sensors on different blades is avoided. Specifically, the mounting positions of all the noise sensors may be marked to determine and ensure whether the number and relative positions of the noise sensors mounted on each blade remain consistent based on the position marks.
Specifically, the noise sensor may be mounted in the blade in other mounting manners, so long as the mounted noise sensor can collect all noise information generated in the blade, which is not limited by the present invention.
Specifically, a signal modem can be arranged at the bottom of each blade of a wind turbine, a signal receiving station is arranged at the position under each wind turbine, such as the tower bottom or the engine room, and the like, the signal modem on each blade of the wind turbine is in communication connection with the signal receiving station under the wind turbine through wireless communication, and each signal receiving station is connected with a monitoring center through a ring network of the wind turbine; therefore, all the noise sensors can be connected into the signal modems of the corresponding blades, the collected noise signals are transmitted to the signal receiving station under the wind turbine generator set through wireless transmission, the signal receiving station transmits noise signal data to the monitoring center through the ring network of the wind turbine generator set, the monitoring center stores the noise signal data, and whether the blades have faults or not is judged according to the noise signals.
In step S102, before analyzing and judging the fault condition of the noise signal collected by the noise sensor, the noise signal may be subjected to data screening, data filtering, data framing and other processes, so as to reduce the error influence generated by the noise signal during the collection and transmission process, and improve the accuracy of the noise signal analysis result and the accuracy of the fault condition judgment.
Specifically, since the wind turbine generator generates additional noise in the non-running state or the yaw state, and noise signals collected by the additional noise cannot be used for analyzing the fault state, before analyzing the noise signals, available data, such as the unit data in the grid-connected state, can be reserved through data screening, and unless the available data, such as the unit data in the non-grid-connected state and the unit data in the yaw state, are excluded.
Specifically, the data preprocessing of the noise signal may include data screening of the noise signal; data filtering of the noise signal may include at least one of:
judging the running state of the wind turbine according to the fan grid-connected signal, and filtering noise signals collected by the wind turbine in a non-grid-connected state;
judging whether the wind turbine generator is in a yaw state according to the yaw signal of the fan, and filtering noise signals collected by the wind turbine generator in the yaw state.
In particular, when data filtering is performed on the noise signal, other unavailable data not mentioned in the specification of the present invention can be also eliminated, and the present invention is not limited to the type of data to be filtered.
Specifically, because the wind turbine generator system can generate aerodynamic noise in the operation process and mix with other background environmental noise, in order to improve the quality of noise signals, the interference noise characteristics in the noise signals can be removed through data filtering, and the signal-to-noise ratio of the noise signals is improved.
In particular, since Kalman filtering is more suitable for non-stationary random processes, data preprocessing of the noise signal may include data filtering the noise signal, e.g., using Kalman filtering; data filtering of noise signals using kalman filtering may include:
the state equation for the wind turbine generator to collect the noise signal is as follows:
x t ＝Ax t-1 +Bu t-1 +ω t-1
p(ω)～N(0,Q)
wherein x is t-1 For the state vector at time t-1, u t-1 For the input vector at time t-1, ω t-1 The method is characterized in that the method comprises the steps that A is a state transfer coefficient matrix of a noise signal, B is an optional control input matrix, Q is a covariance matrix of the process noise, and the process noise meets Gaussian distribution;
the measurement equation for the noise signal is:
z t ＝Hx t +v t
p(v)～N(0,R)
wherein z is t For the measurement vector at time t, v t For the corresponding measurement error, H is a measurement coefficient matrix, R is a covariance matrix of the measurement error, and the measurement error meets Gaussian distribution;
In the invention, the Kalman filtering process is mainly divided into two parts of time updating and measurement updating, wherein the time updating equation of the noise signal based on the Kalman filtering is as follows:
wherein,p is the covariance moment of the error, which is the predicted value of the state vectorAn array; the formula is used for updating the predicted value of the state vector at the moment t by using the estimated value of the covariance matrix of the state vector and the error at the moment t-1, namely, correcting the state predicted value by using the measured data;
based on Kalman filtering, the measurement update equation of the noise signal is:
where K is the kalman filter gain. Wherein the difference between the measurement vector and the measurement vector predicted value predicted based on the predicted value of the state vector of the previous stepCalled innovation, the product of the innovation and the Kalman gain K is used for correcting the error of the predicted value of the state vector, and the smaller the Kalman gain K is, the more the predicted value of the state vector is trusted by the result of the filter estimation, otherwise, the result of the filter estimation is more dependent on the measured value. It can be seen that the Kalman filtering uses the state vector predictor +.>Estimating the next system state +.>The actual measured value output is then used to correct the estimate.
In particular, when data filtering is performed on the noise signal, other filters may be used to perform data filtering, which is not limited by the present invention.
Specifically, because the rotation of the fan blade is a non-stationary process, the noise signal collected by the noise sensor is also a non-stationary signal, but the frequency of the noise signal is considered to be relatively fixed in a short time range, and the noise signal has a short time stationary characteristic, so that before the noise signal is decomposed, the noise signal can be subjected to data framing according to the operation rule of the wind turbine generator, and fault condition analysis is performed according to the noise signal of each frame.
Specifically, due to the periodical rotation of the fan impeller, uneven wind load caused by gravity load factors and wind shear brings non-stationary impact to the fan blades; and because the rotation speed of the wind wheel is time-varying, the noise signals collected by the noise sensor still have volatility in a short time. The motion track of the blades on the wind wheel rotating plane of the fan is a circle taking the wind wheel as the center of a circle, so that a starting point can be selected on the wind wheel rotating plane, for example, the top end of the wind wheel rotating plane is marked with 0 degrees, the wind wheel is taken as the center of a circle, the wind wheel is divided into a plurality of angle ranges clockwise, and each angle range is marked as a label of a noise signal to carry out data framing. In addition, due to process differences of three blades of the fan, etc., the three blades also need to be marked when data framing is performed.
Specifically, data preprocessing the noise signal may include data framing the noise signal; framing the noise signal may include: dividing the motion track of the wind turbine blade on the wind turbine rotation plane into a plurality of angle ranges by taking the wind turbine as the circle center, and marking the noise signal collected on each angle range with a corresponding angle label; the angle labels marked on the noise signals on each blade of the wind turbine generator are different; and then carrying out data framing on the noise signals according to the angle labels.
Specifically, each angle range may include only one angle, or may include a plurality of angles in succession, where the number of angles specifically included may be set according to the granularity of dividing the data as required, which is not limited in the present invention.
Specifically, in order to ensure that the sound characteristic parameters of the noise signal are not lost and smooth, an overlapping frame taking mode can be adopted when data framing is performed.
Specifically, when the noise signal is subjected to data framing, the data can be divided according to other rules, which is not limited by the invention.
In step S103, the preprocessed noise signal may be decomposed into a limited number of natural mode function (Intrinsic Mode Function, IMF) components by means of a variational mode decomposition (Variational Mode Decomposition, VMD) to determine the fault condition of the wind turbine according to the characteristics of each IMF component.
VMD is a method of adaptive, completely non-recursive modal variation and signal processing. The VMD has the advantage of determining the number of modal decomposition, the adaptivity of the VMD is represented by determining the number of modal decomposition of a given sequence according to actual conditions, the optimal center frequency and the limited bandwidth of each modal can be adaptively matched in the subsequent searching and solving process, the effective separation of inherent modal components (IMFs), the frequency domain division of signals and further the effective decomposition components of the given signals can be realized, and finally the optimal solution of the variation problem is obtained. The VMD overcomes the problems of end effect and modal component aliasing of an EMD (Empirical Mode Decomposition) method, has a firmer mathematical theory basis, can reduce the non-stationarity of a time sequence with high complexity and strong nonlinearity, and is suitable for a non-stationarity sequence by decomposing to obtain a relatively stable subsequence containing a plurality of different frequency scales.
Specifically, in the process of decomposing the noise signal after preprocessing into a limited number of IFM components through VMD, the formula of the decomposition mode of the variation may be:
where f is the input noise signal, u k K is the number of the natural mode function components, omega k The center frequency of the intrinsic mode function component is shown, and delta (t) is a pulse function;
the Lagrange multiplication operator (Lagrange) is introduced into a variation modal decomposition formula to solve, and the expression is as follows:
wherein lambda is Lagrange multiplier, alpha is penalty factor, f (t) is noise signal to be decomposed, u k (t) is the kth natural mode function component;
after the Lagrange multiplication operator is introduced, the resolving method of the variation modal decomposition formula comprises the following steps:
initializing natural modal function componentsCenter frequency of the intrinsic mode function component +.>And the Lagrangian multiplier +.>
Initializing the iteration number n=0, carrying out a plurality of iterative calculations with n=n+1, and carrying out the following formula on the inherent mode function component in each iterative calculationCenter frequency of intrinsic mode function component +.>Lagrangian multiplierUpdating:
where τ is the noise margin; in each update, stopping continuing to decompose when the number of IMF components obtained by decomposing the noise signals reaches K;
stopping iterative calculation according to a preset precision value epsilon, and when the following conditions are met:
outputting a plurality of natural mode function components obtained by decomposition. Specifically, the precision value epsilon can be set according to actual requirements, and the invention is not limited to this.
Specifically, in step S103, the noise signal in the healthy state, and the noise signal acquired in the known fault state may be decomposed into a limited number of IMF components such as IMF1, IMF2, and the like by the VMD. Specifically, for different fault types, the difference between the fault signal and the normal signal with 1-2 or more IMF components can be maximized by setting different numbers of IMF components K and penalty factors α, so as to distinguish different fault signals from normal signals.
In step S104, it may be determined whether there is a fault and a corresponding fault type by the characteristics of the several IMF components decomposed by the VMD. Specifically, fault diagnosis can be performed through a neural network model method, namely, a training set is formed by acquiring a health signal in a health state and a plurality of fault signals acquired under known faults, the training set is input into a preset neural network for training, so that a neural network capable of being used for judging fault monitoring of wind turbine generator blades is obtained, and the neural network is utilized to analyze a plurality of IMF components acquired by VMD decomposition of noise signals acquired in real time, so that whether faults occur or not and the types of the faults are predicted.
Specifically, the neural network used in the present invention may be a Long Short-Term Memory network (LSTM), which is a special cyclic neural network (Recurrent Neural Network, RNN) structure used for processing tasks such as speech recognition, natural language processing, video analysis, etc. The LSTM network incorporates a memory cell (memory cell) for storing and updating information in the sequence, and three gates (gates) to control the flow of information in the memory cell: input gate (input gate), forget gate (for gate), and output gate (output gate). The input gate controls the inflow of new input, forgets the forgets of gate control history information, and outputs information output in the gate control memory unit. The input gate is used for selectively recording new information into the cell state, the forgetting gate is used for selectively forgetting the information in the cell, and the output gate is used for carrying the stored information to the next neuron. The switching states of the three gates are controlled by a sigmoid function so that the information flow can be adaptively controlled. By introducing the memory unit and the gate structure, the LSTM network can better process long sequence data and alleviate the problems of gradient disappearance and gradient explosion to a certain extent.
Specifically, the process of fault diagnosis through several natural mode function components may include:
inputting a plurality of inherent mode function components into a designated long-short-period memory network to obtain a fault prediction result; the long-term and short-term memory network is obtained through training a data set consisting of a plurality of fault signals and health signals in advance;
if the fault prediction result indicates that a fault occurs, a fault warning is initiated to prompt fault processing.
Specifically, as shown in fig. 2, fig. 2 is a schematic structural diagram of a long-short-term memory network according to an embodiment of the present invention, where the long-short-term memory network may include an input layer, an implicit layer, and an output layer; wherein the hidden layer may include a memory unit and a control unit, and the control unit may include an input gate, a forget gate, and an output gate.
The input gate is used for controlling the inflow of new input and updating the information in the memory unit; the input gate allows the LSTM network to control the inflow of new inputs and to selectively update information in the memory cells to better capture long-term dependencies in the sequence data. The forgetting door is used for controlling forgetting of the history information and updating the information in the memory unit; the forget gate allows the LSTM network to selectively forget the history information and update the information in the memory unit; the mechanism can better capture the long-term dependency in the sequence data and alleviate the problems of gradient disappearance and gradient explosion. The output gate is used for controlling information output in the memory unit; the output gate allows the LSTM network to control information output in the memory unit, thereby better capturing long-term dependencies in the sequence data and improving the prediction accuracy of the model.
The long-term and short-term memory network sequentially performs the following calculation in the forward propagation process:
the value of the updated forgetting gate is calculated, and the formula is as follows:
f (t) ＝σ(W f h (t-1) +W i x (t) +b f )
wherein sigma is a sigmoid activation function, W f B is forgetting the gate weight f Is a forget gate threshold;
the value of the updated input gate is calculated by the following formula:
i (t) ＝σ(W i h (t-1) +W i x (t) +b i )
c ’(t) ＝tanh(W c h (t-1) +W c x (t) +b c )
wherein W is i 、W c 、b i And b c The weights and the thresholds are respectively corresponding to a sigmoid activation function and a tanh activation function of the input gate;
calculating and updating the value of the memory unit, wherein the formula is as follows:
wherein,is dot multiplied;
the value of the update output gate is calculated, and the formula is:
o (t) ＝σ(W o h (t-1) +W o x (t) +b o )
wherein W is o To output the gate weight, b o To output the threshold value of h (t) An output vector that is an implicit layer;
updating the prediction output at the current moment, wherein the formula is as follows:
wherein V and c are the weights and thresholds, respectively, of the implicit layer to output layer connection.
After each forward propagation process, the LSTM network performs reverse calculation by using the error between the predicted value and the actual value, so as to update each weight and threshold value until the maximum iteration number is satisfied.
Before fault monitoring, the pre-collected fault signal and health signal data can be imported into the LSTM network model for iterative training, and then the trained LSTM network can be used for fault diagnosis. In step S104, the fault condition of the blade corresponding to the noise signal, including whether the fault occurs and the type of the fault occurring, may be obtained only by inputting a plurality of IMF components obtained by decomposing the noise signal by the VMD into a preset LSTM network model.
Specifically, in other embodiments, other neural network models other than LSTM networks may be used to perform fault diagnosis, and other algorithms other than neural networks may be used to analyze IMF components to determine fault conditions, so long as they may be used to perform noise identification and may be classified according to IMF components decomposed by VMDs, which is not a limitation of the present invention.
Specifically, after fault diagnosis is performed on a plurality of IMF components to obtain a fault prediction result, corresponding fault processing may be performed according to the fault prediction result. For example, if the failure prediction result indicates that a failure occurs, a failure warning may be initiated to prompt corresponding failure processing to repair the blade failure. In particular, the manner of initiating the fault alert may include any one or more of whistling the alert, sending a warning message, outputting a fault report, or other alert manners, as the invention is not limited in this regard.
According to the wind turbine generator blade fault monitoring method disclosed by the invention, the noise sensor is arranged in the blade, and after the blade is cracked, the damage condition of the blade can be monitored through the change of the noise in the blade. For example, when the blade is cracked, the natural frequency of the blade changes, and the noise generated by the blade also changes; when the blade cracks seriously and is accompanied by air leakage, whistle appears in the blade; therefore, by monitoring noise differences generated during no faults and after the faults and comparing the noise differences of three blades of the wind turbine generator, fault conditions such as cracking of the blades can be identified through methods such as a noise identification model. Compared with the existing optical monitoring scheme, the wind turbine generator blade fault monitoring method provided by the invention can identify various fault conditions according to noise changes, has a comprehensive monitoring range, generates less data, is convenient to process, and has lower technical difficulty and application cost; compared with an unmanned aerial vehicle monitoring scheme, the wind turbine generator blade fault monitoring method provided by the invention has the advantages that the noise sensor is arranged in the blade, so that the wind turbine generator blade fault monitoring method is not influenced by external environment, and all-weather online monitoring can be realized; compared with the load monitoring scheme, the wind turbine generator blade fault monitoring method provided by the invention has the advantages that only a certain number of noise sensors are arranged, the number of the sensors is small, the construction period is short, the cost is low, meanwhile, the data volume generated in the later period is also small, the processing is convenient, and the algorithm development difficulty is also lower.
Corresponding to the embodiment of the wind turbine blade fault monitoring method, the invention further provides a wind turbine blade fault monitoring system.
FIG. 3 is a schematic structural view of a wind turbine blade fault monitoring system according to an embodiment of the present invention, the system including a plurality of noise sensors, a signal modem, a signal receiving station, and a monitoring center mounted inside a wind turbine blade;
the noise sensor is used for collecting noise signals generated by the blades of the wind turbine generator;
the signal modem is used for receiving the noise signals acquired by the noise sensor and transmitting the noise signals to the signal receiving station through wireless connection;
the signal receiving station is used for transmitting noise signals to the monitoring center through the ring network of the wind turbine generator;
the monitoring center is used for storing noise signals and carrying out fault monitoring on the wind turbine blade through the wind turbine blade fault detection method according to any embodiment of the invention.
Specifically, the process of the monitoring center for fault monitoring of the wind turbine blade may include:
collecting noise signals through a noise sensor arranged inside a blade of the wind turbine generator;
Carrying out data preprocessing on the noise signals;
decomposing the preprocessed noise signal into a plurality of finite inherent mode function components through variation mode decomposition;
and carrying out fault diagnosis through a plurality of inherent mode function components.
In particular, the noise sensor may be a fiber optic MEMS noise sensor.
Specifically, the installation mode of the noise sensor may be: 1 noise sensor is installed in the wind turbine blade from the blade root at intervals of a designated distance; and the number and the relative positions of the noise sensors arranged on each blade of the wind turbine are kept consistent.
Specifically, data preprocessing the noise signal may include data framing the noise signal; framing the noise signal may include:
dividing the motion track of the wind turbine blade on the wind turbine rotation plane into a plurality of angle ranges by taking the wind turbine as the circle center, and marking the noise signal collected on each angle range with a corresponding angle label; the angle labels marked on the noise signals on each blade of the wind turbine generator are different;
and carrying out data framing on the noise signals according to the angle labels.
In particular, data preprocessing the noise signal may include data filtering the noise signal, for example using kalman filtering; data filtering of noise signals using kalman filtering may include:
The state equation for the wind turbine generator to collect the noise signal is as follows:
x t ＝Ax t-1 +Bu t-1 +ω t-1
p(ω)～N(0,Q)
wherein x is t-1 For the state vector at time t-1, u t-1 For the input vector at time t-1, ω t-1 The method is characterized in that the method comprises the steps that A is a state transfer coefficient matrix of a noise signal, B is an optional control input matrix, Q is a covariance matrix of the process noise, and the process noise meets Gaussian distribution;
the measurement equation for the noise signal is:
z t ＝Hx t +v t
p(v)～N(0,R)
wherein z is t For the measurement vector at time t, v t For the corresponding measurement error, H is the measurement coefficient matrix, R is the covariance matrix of the measurement error, and the measurement error is fullA foot gaussian distribution;
based on Kalman filtering, the time update equation of the noise signal is:
wherein,p is the covariance matrix of the error, which is the predicted value of the state vector;
based on Kalman filtering, the measurement update equation of the noise signal is:
where K is the kalman filter gain.
Specifically, the process of fault diagnosis through several natural mode function components may include:
inputting a plurality of inherent mode function components into a designated long-short-period memory network to obtain a fault prediction result; the long-term and short-term memory network is obtained through training a data set consisting of a plurality of fault signals and health signals in advance;
If the fault prediction result indicates that a fault occurs, a fault warning is initiated to prompt fault processing.
In particular, the long-term memory network may include an input layer, an hidden layer, and an output layer; wherein the hidden layer can comprise a memory unit and a control unit, and the control unit can comprise an input door, a forget door and an output door;
the long-term and short-term memory network sequentially performs the following calculation in the forward propagation process:
the value of the updated forgetting gate is calculated, and the formula is as follows:
f (t) ＝σ(W f h (t-1) +W i x (t) +b f )
wherein sigma is a sigmoid activation function, W f B is forgetting the gate weight f Is a forget gate threshold;
the value of the updated input gate is calculated by the following formula:
i (t) ＝σ(W i h (t-1) +W i x (t) +b i )
c ’(t) ＝tanh(W c h (t-1) +W c x (t) +b c )
wherein W is i 、W c 、b i And b c The weights and the thresholds are respectively corresponding to a sigmoid activation function and a tanh activation function of the input gate;
calculating and updating the value of the memory unit, wherein the formula is as follows:
wherein,is dot multiplied;
the value of the update output gate is calculated, and the formula is:
o (t) ＝σ(W o h (t-1) +W o x (t) +b o )
wherein W is o To output the gate weight, b o To output the threshold value of h (t) An output vector that is an implicit layer;
updating the prediction output at the current moment, wherein the formula is as follows:
wherein V and c are the weights and thresholds, respectively, of the implicit layer to output layer connection.
Specifically, in the process of decomposing the preprocessed noise signal into a limited number of natural mode function components through the variational mode decomposition, the variational mode decomposition formula can be as follows:
Where f is the input noise signal, u k K is the number of the natural mode function components, omega k The center frequency of the intrinsic mode function component is shown, and delta (t) is a pulse function;
the Lagrange multiplication operator is introduced into a variation modal decomposition formula to solve, and the expression is as follows:
wherein lambda is Lagrange multiplier, alpha is penalty factor, f (t) is noise signal to be decomposed, u k (t) is the kth natural mode function component;
after the Lagrange multiplication operator is introduced, the resolving method of the variation modal decomposition formula comprises the following steps:
initializing natural modal function componentsCenter frequency of the intrinsic mode function component +.>And the Lagrangian multiplier +.>
Initializing the iteration number n=0, carrying out a plurality of iterative calculations with n=n+1, and carrying out the following formula on the inherent mode function component in each iterative calculationCenter frequency of intrinsic mode function component +.>Lagrangian multiplierUpdating:
where τ is the noise margin;
stopping iterative calculation according to a preset precision value epsilon, and when the following conditions are met:
/>
outputting a plurality of natural mode function components obtained by decomposition.
Specifically, the data preprocessing of the noise signal may include data screening of the noise signal; data filtering of the noise signal may include at least one of:
Judging the running state of the wind turbine according to the fan grid-connected signal, and filtering noise signals collected by the wind turbine in a non-grid-connected state;
judging whether the wind turbine generator is in a yaw state according to the yaw signal of the fan, and filtering noise signals collected by the wind turbine generator in the yaw state.
Specifically, the implementation process of the functions and roles of each device in the above system is specifically shown in the implementation process of the corresponding steps in the above method, which is not described herein again.
The method and system of the invention are further described below with reference to an operational embodiment of a wind turbine blade failure monitoring system. FIG. 4 is a flow chart illustrating operation of a wind turbine blade failure detection system according to one embodiment of the present invention. In the system, firstly, a noise sensor arranged in a blade collects noise signals, and the noise signals are transmitted to a monitoring center for analysis through a signal modem and a signal receiving station in sequence; then, removing interference noise characteristics in the noise signal through Kalman filtering; then, decomposing the processed noise signal into a plurality of IMF components through VMD; then, inputting a plurality of IMF components obtained through decomposition into an LSTM network for training, or inputting a trained LSTM network for recognition; then, performing model deployment through a trained LSTM network, or performing fault alarm according to a fault prediction result obtained by LSTM network identification; and under the condition of faults, notifying related personnel to perform corresponding fault treatment; and if no fault occurs, continuing VMD decomposition and fault diagnosis on the noise signal data of the next frame.
The above examples of the present invention are merely illustrative of the present invention and are not intended to limit the embodiments of the present invention. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the invention are desired to be protected by the following claims.
Claims (10)
1. A method for monitoring blade faults of a wind turbine, the method comprising:
collecting noise signals through a noise sensor arranged inside the wind turbine generator blade;
performing data preprocessing on the noise signals;
decomposing the preprocessed noise signal into a plurality of finite inherent mode function components through variation mode decomposition;
and carrying out fault diagnosis through a plurality of inherent mode function components.
2. The method of claim 1, wherein the noise sensor is a fiber optic MEMS noise sensor.
3. The method according to claim 1, wherein the noise sensor is installed in the following manner:
1 noise sensor is installed in the wind turbine blade at intervals of a specified distance from the blade root; and the number and the relative positions of the noise sensors arranged on each blade of the wind turbine are kept consistent.
4. The method of claim 1, wherein data preprocessing the noise signal comprises data framing the noise signal; framing the noise signal, including:
dividing the equal angle of the motion track of the wind turbine blade on the wind turbine rotating plane into a plurality of angle ranges by taking the wind turbine as the circle center, and marking the noise signal acquired on each angle range with a corresponding angle label; and the angle labels marked on the noise signals on each blade of the wind turbine generator are different;
and carrying out data framing on the noise signal according to the angle label.
5. The method of claim 1, wherein data preprocessing the noise signal comprises data filtering the noise signal using kalman filtering; data filtering the noise signal using kalman filtering, comprising:
the state equation of the wind turbine generator system for collecting the noise signals is as follows:
x t ＝Ax t-1 +Bu t-1 +ω t-1
p(ω)～N(0，Q)
wherein x is t-1 For the state vector at time t-1, u t-1 For the input vector at time t-1, ω t-1 The method comprises the steps that A is a state transfer coefficient matrix of a noise signal, B is an optional control input matrix, Q is a covariance matrix of the process noise, and the process noise meets Gaussian distribution;
The measurement equation of the noise signal is:
z t ＝Hx t +v t
p(v)～N(0，R)
wherein z is t For the measurement vector at time t, v t For the corresponding measurement error, H is a measurement coefficient matrix, R is a covariance matrix of the measurement error, and the measurement error meets Gaussian distribution;
based on Kalman filtering, the time update equation of the noise signal is as follows:
wherein,p is the covariance matrix of the error, which is the predicted value of the state vector;
based on Kalman filtering, the measurement update equation of the noise signal is as follows:
where K is the kalman filter gain.
6. The method of claim 1, wherein said diagnosing a fault with a plurality of said natural mode function components comprises:
inputting a plurality of inherent mode function components into a designated long-period and short-period memory network to obtain a fault prediction result; the long-term and short-term memory network is obtained through training of a data set consisting of a plurality of fault signals and health signals in advance;
and if the fault prediction result indicates that the fault occurs, initiating a fault warning to prompt fault processing.
7. The method of claim 6, wherein the long-term memory network comprises an input layer, an implied layer, and an output layer; the hidden layer comprises a memory unit and a control unit, wherein the control unit comprises an input door, a forget door and an output door;
The long-term and short-term memory network sequentially performs the following calculation in the forward propagation process:
calculating and updating the value of the forgetting gate, wherein the formula is as follows:
f (t) ＝σ(W f h (t-1) +W i x (t) +b f )
wherein sigma is a sigmoid activation function, W f B is forgetting the gate weight f Is a forget gate threshold;
calculating and updating the value of the input gate, wherein the formula is as follows:
i (t) ＝σ(W i h (t-1) +W i x (t) +b i )
c ’(t) ＝tanh(W c h (t-1) +W c x (t) +b c )
wherein W is i 、W c 、b i And b c Respectively obtaining weights and thresholds corresponding to a sigmoid activation function and a tanh activation function of the input gate;
calculating and updating the value of the memory unit, wherein the formula is as follows:
C (t) ＝C (t-1) ⊙f (t) +i(t)⊙c′ (t)
wherein, as follows;
calculating and updating the value of the output gate, wherein the formula is as follows:
o (t) ＝σ(W o h (t-1) +W o x (t) +b o )
h (t) ＝o (t) ⊙tanh(C (t) )
wherein W is o To output the gate weight, b o To output the threshold value of h (t) An output vector for the hidden layer;
updating the prediction output at the current moment, wherein the formula is as follows:
wherein V and c are the weight and threshold of the connection of the hidden layer to the output layer, respectively.
8. The method according to claim 1, wherein in decomposing the noise signal after preprocessing into a finite number of intrinsic mode function components by a variational mode decomposition, the variational mode decomposition is formulated as:
where f is the input noise signal, u k K is the number of the natural mode function components, omega k Delta (t) is a pulse function for the center frequency of the natural mode function component;
and introducing a Lagrange multiplication operator into the variation modal decomposition formula to solve, wherein the expression is as follows:
wherein lambda is Lagrange multiplier, alpha is penalty factor, f (t) is noise signal to be decomposed, u k (t) is the kth natural mode function component;
after the Lagrange multiplication operator is introduced, the solution method of the variation modal decomposition formula is as follows:
initializing the natural modal function componentCenter frequency of the intrinsic mode function component +.>And the Lagrangian multiplier +.>
Initializing the iteration number n=0, carrying out a plurality of iterative calculations with n=n+1, and carrying out the following formula on the inherent mode function component in each iterative calculationCenter frequency of the intrinsic mode function component +.>And the Lagrangian multiplier +.>Updating:
where τ is the noise margin;
stopping iterative calculation according to a preset precision value epsilon, and when the following conditions are met:
outputting a plurality of natural mode function components obtained by decomposition.
9. The method of claim 1, wherein data preprocessing the noise signal comprises data screening the noise signal; data screening is carried out on the noise signals, wherein the data screening comprises at least one of the following steps:
Judging the running state of the wind turbine generator according to the fan grid-connected signal, and filtering noise signals collected by the wind turbine generator in a non-grid-connected state;
judging whether the wind turbine generator is in a yaw state according to the yaw signal of the fan, and filtering noise signals collected by the wind turbine generator in the yaw state.
10. The wind turbine blade fault monitoring system is characterized by comprising a plurality of noise sensors, a signal modem, a signal receiving station and a monitoring center, wherein the noise sensors, the signal modem, the signal receiving station and the monitoring center are arranged inside the wind turbine blade;
the noise sensor is used for collecting noise signals generated by the wind turbine generator blades;
the signal modem is used for receiving the noise signals acquired by the noise sensor, and
transmitting the noise signal to the signal receiving station via a wireless connection;
the signal receiving station is used for transmitting the noise signal to the monitoring center through a ring network of the wind turbine generator;
the monitoring center is used for storing the noise signals and carrying out fault monitoring on the wind turbine blades through the wind turbine blade fault detection method according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311701585.5A CN117759490A (en) | 2023-12-11 | 2023-12-11 | Wind turbine generator blade fault monitoring method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311701585.5A CN117759490A (en) | 2023-12-11 | 2023-12-11 | Wind turbine generator blade fault monitoring method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117759490A true CN117759490A (en) | 2024-03-26 |
Family
ID=90313625
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311701585.5A Pending CN117759490A (en) | 2023-12-11 | 2023-12-11 | Wind turbine generator blade fault monitoring method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117759490A (en) |
-
2023
- 2023-12-11 CN CN202311701585.5A patent/CN117759490A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6313929B2 (en) | Method and system for monitoring structures | |
| CN108984893B (en) | Gradient lifting method-based trend prediction method | |
| CN104773304B (en) | Load estimation system for aerodynamic structures | |
| US10458863B2 (en) | Hybrid virtual load monitoring system and method | |
| US10739736B2 (en) | Apparatus and method for event detection and duration determination | |
| CN110704801B (en) | Bridge cluster structure operation safety intelligent monitoring and rapid detection complete method | |
| EP3407271A1 (en) | Fault prediction method and device for fan | |
| CN108427400B (en) | Aircraft airspeed head fault diagnosis method based on neural network analytic redundancy | |
| CN110006552B (en) | Method for detecting abnormal temperature of equipment | |
| Das et al. | Structural health monitoring techniques implemented on IASC–ASCE benchmark problem: a review | |
| CN110678816A (en) | Method and control device for controlling a technical system | |
| CA3035871A1 (en) | Method and device for monitoring a status of at least one wind turbine and computer program product | |
| CN110570013B (en) | Single-station online wave period data prediction diagnosis method | |
| CN116380445B (en) | Equipment state diagnosis method and related device based on vibration waveform | |
| CN114611372A (en) | Industrial equipment health prediction method based on Internet of things edge calculation | |
| Smarsly et al. | A decentralized approach towards autonomous fault detection in wireless structural health monitoring systems | |
| EP3721197B1 (en) | Estimating fatigue damage in a structure | |
| Valeti et al. | Remaining useful life estimation of wind turbine blades under variable wind speed conditions using particle filters | |
| CN117759490A (en) | Wind turbine generator blade fault monitoring method and system | |
| WO2023101757A1 (en) | Multicopter online rotor fault diagnosis system | |
| Asadollahi | Bayesian-based finite element model updating, damage detection, and uncertainty quantification for cable-stayed bridges | |
| CN117708553A (en) | Bridge monitoring data inspection method and system based on PSO-BP neural network | |
| CN117216677A (en) | Real-time identification method of engineering structure anomaly monitoring data based on Bayesian filtering | |
| CN116579383A (en) | Wind turbine blade icing prediction method based on deep neural network | |
| Weiss et al. | Uncertainty Quantification for Full-Flight Data Based Engine Fault Detection with Neural Networks. Machines 2022, 10, 846 |
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 |