[go: up one dir, main page]

CN120969089A - Blade fault diagnosis method and device based on vibration of wind generating set - Google Patents

Blade fault diagnosis method and device based on vibration of wind generating set

Info

Publication number
CN120969089A
CN120969089A CN202511494495.2A CN202511494495A CN120969089A CN 120969089 A CN120969089 A CN 120969089A CN 202511494495 A CN202511494495 A CN 202511494495A CN 120969089 A CN120969089 A CN 120969089A
Authority
CN
China
Prior art keywords
blade
modal
frequency
trajectory
vibration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202511494495.2A
Other languages
Chinese (zh)
Other versions
CN120969089B (en
Inventor
杨安韬
马振江
陈曦
黄宇豪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CCCC Shanghai Third Harbor Engineering Science and Technology Research Institute Co Ltd
Original Assignee
CCCC Shanghai Third Harbor Engineering Science and Technology Research Institute Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by CCCC Shanghai Third Harbor Engineering Science and Technology Research Institute Co Ltd filed Critical CCCC Shanghai Third Harbor Engineering Science and Technology Research Institute Co Ltd
Priority to CN202511494495.2A priority Critical patent/CN120969089B/en
Publication of CN120969089A publication Critical patent/CN120969089A/en
Application granted granted Critical
Publication of CN120969089B publication Critical patent/CN120969089B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics
    • F03D17/005Monitoring or testing of wind motors, e.g. diagnostics using computation methods, e.g. neural networks
    • F03D17/0065Monitoring or testing of wind motors, e.g. diagnostics using computation methods, e.g. neural networks for diagnostics
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics
    • F03D17/009Monitoring or testing of wind motors, e.g. diagnostics characterised by the purpose
    • F03D17/013Monitoring or testing of wind motors, e.g. diagnostics characterised by the purpose for detecting abnormalities or damage
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics
    • F03D17/009Monitoring or testing of wind motors, e.g. diagnostics characterised by the purpose
    • F03D17/015Monitoring or testing of wind motors, e.g. diagnostics characterised by the purpose for monitoring vibrations
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics
    • F03D17/027Monitoring or testing of wind motors, e.g. diagnostics characterised by the component being monitored or tested
    • F03D17/028Blades
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/10Pre-processing; Data cleansing
    • G06F18/15Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/22Matching criteria, e.g. proximity measures
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02Preprocessing

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Mechanical Engineering (AREA)
  • Combustion & Propulsion (AREA)
  • Chemical & Material Sciences (AREA)
  • Sustainable Energy (AREA)
  • Sustainable Development (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Health & Medical Sciences (AREA)
  • Toxicology (AREA)
  • Wind Motors (AREA)

Abstract

The invention provides a blade fault diagnosis method and device based on wind generating set vibration, which relates to the technical field of blade fault diagnosis of wind generating set vibration, the invention synchronously collects vibration signals at the cabin foundation and the blade root, after filtering and steady-state working condition screening, the first-order modal frequency and damping ratio of the blade under different working conditions are accurately identified through phase compensation and frequency response function calculation, and sequentially connecting modal parameters of a plurality of working points to construct a dynamic response track, forming a local mode feature vector by calculating relative vectors of each point and other points in the track, integrating the local mode feature vector into a global mode feature matrix, obtaining a track mode deviation index by calculating Frobenius norms of a matrix to be monitored and a reference matrix, and judging structural damage according to a statistics rule of continuous super-threshold of the index.

Description

Blade fault diagnosis method and device based on vibration of wind generating set
Technical Field
The invention relates to the technical field of blade fault diagnosis of vibration of a wind generating set, in particular to a blade fault diagnosis method and device based on vibration of the wind generating set.
Background
The wind turbine generator system blade is used as a core component for capturing wind energy, is subjected to complex alternating load for a long time, is easy to generate structural damage such as cracks, layering and the like, and can cause huge economic loss and safety accidents if not detected in time, so that vibration-based blade state monitoring and fault diagnosis are carried out, and the wind turbine generator system blade has important engineering value for guaranteeing safe operation and realizing predictive maintenance.
In the prior art, vibration-based blade diagnosis methods mainly rely on identifying the modal parameters of a blade, such as the natural frequency and damping ratio, under specific operation conditions, and implementing fault identification by monitoring the deviation of the parameters relative to a healthy reference, such methods generally assume that the change of the modal parameters is isolated, namely only whether the absolute deviation of the frequency or damping ratio exceeds a threshold value or not, however, the modal parameters of the blade can show inherent fluctuation due to load change under different operation conditions such as wind speed, power and the like, and the prior methods treat the data points under different operation conditions as isolated or attempt to establish a fixed threshold range allowing fluctuation, which essentially neglects the continuity and the integrity of the dynamic characteristics of the blade along with the evolution of the operation conditions.
The defects of the prior art are that natural fluctuation of modal parameters caused by normal operation load change and abnormal change caused by structural damage are difficult to distinguish effectively, and for early and tiny damage, parameter change signals caused by the natural fluctuation are weak and are very easy to submerge in normal working condition fluctuation, so that diagnosis sensitivity is low, early warning cannot be realized, and in addition, the inherent relevance among modal parameters cannot be captured, the directionality of the damage is not strong, the damage is easy to be interfered by environmental noise, and the false alarm rate is high.
The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it may include information that does not form the prior art that is already known to a person of ordinary skill in the art.
Disclosure of Invention
The invention aims to provide a blade fault diagnosis method and device based on vibration of a wind generating set, so as to solve the problems in the background technology.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a blade fault diagnosis method based on vibration of a wind generating set comprises the following specific steps:
Step 1, arranging sensors at the basic position of a cabin and the root of a blade of a wind generating set, synchronously acquiring cabin reference vibration signals and blade response vibration signals, and screening steady-state effective data segments through a preset time window;
Step 2, analyzing time delay between a cabin reference signal and a blade response signal according to the effective data segments, and identifying modal parameters of the first order of the blade through phase compensation and frequency response function calculation, wherein the modal parameters comprise modal frequencies and modal damping ratios, and repeating the operations on a plurality of different effective data segments to obtain a series of modal parameters;
step 3, sequentially connecting a series of modal parameters in a coordinate system with the modal frequency as an abscissa and the modal damping ratio as an ordinate, constructing a track of dynamic response, calculating relative vectors of each modal parameter on the track and all other modal parameters in the track to form local feature vectors, integrating all the local feature vectors, and constructing a global feature matrix representing the overall track morphology;
And 4, calculating a reference global feature matrix when the blade is healthy and lossless according to the steps 1 to 3, calculating a Frobenius norm between the global feature matrix and the reference global feature matrix, defining the norm value as a track deviation degree, and judging whether the wind generating set blade has structural damage according to a track deviation construction state transition probability matrix.
Further, a time window with the current time T as an end point and a fixed length T is traced backWherein,For the time length of 10 complete rotation periods, synchronous signal acquisition is carried out through a reference acceleration sensor positioned on a main frame of the engine room and a response acceleration sensor positioned on the inner side of a blade root flange in a time window so as to acquire an engine room reference vibration signal and a blade response vibration signal;
The method comprises the steps of filtering signals by band-pass filtering, wherein the band-pass frequency range covers the first-order waving modal frequency of the blade, namely, the vibration signals are band-pass filtered to retain the main modal frequency components of the blade, and the wind speed and output power signals are monitored in real time and set to be the length in the running process of the unit The sliding window is made to be in the time windowAnd sliding in the range, when the fluctuation ranges of the wind speed and the output power are respectively smaller than the respective preset thresholds in the sliding window, judging that the unit enters a steady-state working condition, and taking the time interval corresponding to the sliding window as an effective data segment.
Further, identifying modal parameters of the first order of the blade based on each valid data segment specifically includes calculating a nacelle reference signal under each valid data segmentResponsive to blade signalsCross-power spectrum of (a)Self-power spectrum of cabin reference signalWhereinAndRespectively isAndIs used for the fourier transform of (a),Wherein for the desired operation(s),Is complex conjugated;
By aligning Performing inverse Fourier transform to obtain a cross-correlation function: Wherein Is an inverse fourier transform;
Searching for a lead Initial time delay to take maximum valueBy usingParabolic interpolation of cross-correlation function values at three points to accurately determine time delayThe interpolation formula is as follows:
;
wherein, the Is the sampling time interval.
Further, to eliminate time delayInfluence on phase information, cross-power spectrum of both in frequency domainAnd (3) compensating, wherein the compensation formula is as follows:
;
wherein, the In units of imaginary numbers,For frequency, the compensated positive sign is used to counteract the time delay in the blade response signal relative to the nacelle reference signalThereby correcting the phase of the cross power spectrum;
And uses the compensated cross power spectrum Self-power spectrum with cabin reference signalCalculating a frequency response function from the nacelle to the blade: in the following And (3) identifying the highest peak value in the theoretical frequency range of the first-order waving mode of the blade, determining the frequency corresponding to the peak value as the mode frequency, and calculating the mode damping ratio of the first-order mode by using a half-power bandwidth method.
Further, when constructing the locus of dynamic response, the modal frequency and modal damping ratio corresponding to each effective data segment are taken as one data pointN data points are obtained in total and are sequentially connected according to the starting time sequence of the effective data segments to form a track, whereinIs the modal frequency of the ith data point,Modal damping ratio for the i-th data point;
for the ith data point on the trace Its local feature vectorIs oneThe column vector of the dimension has the mathematical expression:
;
wherein, the I is the index of the data point, and
Further, performing Z-score normalization processing on the modal frequencies and modal damping ratios in all effective data segments in a time window, and performing normalization on all N local feature vectors according to the corresponding data pointsSequentially stacked as row vectors to form a global feature matrixProcessing the signals of the same unit blade in the healthy and lossless state by the same method to obtain a reference global feature matrix;
Defining track departure indexThe Frobenius norm, which is the difference between the two matrices, is calculated as follows:
;
wherein, the As a reference global feature matrix,As a global feature matrix of the system,Representing the Frobenius norm of the matrix.
Further, under the condition that the same unit blade is healthy and lossless, historical data of track deviation indexes are obtained through a plurality of historical time windows, and the average value of the historical data is calculatedAnd standard deviationReal-time monitoring the obtained track deviation indexAbnormal probability of (2)Calculated from the following formula:
;
wherein, the A cumulative distribution function that is a standard normal distribution;
and when the abnormal probability is continuously higher than a set threshold value in K continuous time windows, judging that the blade has structural damage, wherein K is more than or equal to 3.
The invention further provides a blade fault diagnosis device based on the vibration of the wind generating set, which is used for executing the blade fault diagnosis method based on the vibration of the wind generating set, and comprises the following steps:
The signal acquisition module is used for arranging sensors at the basic position of a cabin and the root of a blade of the wind generating set, synchronously acquiring cabin reference vibration signals and blade response vibration signals, and screening steady-state effective data segments through a preset time window;
The modal calculation module is used for analyzing the time delay between the cabin reference signal and the blade response signal according to the effective data segments, and identifying the modal parameters of the first order of the blade through phase compensation and frequency response function calculation, wherein the modal parameters comprise modal frequencies and modal damping ratios, and repeating the operations on a plurality of different effective data segments to obtain a series of modal parameters;
the characteristic construction module is used for sequentially connecting a series of modal parameters in a coordinate system taking modal frequency as an abscissa and modal damping ratio as an ordinate to construct a track of dynamic response, calculating relative vectors of each modal parameter on the track and all other modal parameters in the track to form local feature vectors, integrating all the local feature vectors, and constructing a global feature matrix representing the overall track shape;
the result judging module is used for calculating a reference global feature matrix when the blade is healthy and lossless according to the steps 1 to 3, calculating a Frobenius norm between the global feature matrix and the reference global feature matrix, defining the norm value as a track deviation degree, and judging whether the wind turbine blade has structural damage according to a track deviation construction state transition probability matrix.
Compared with the prior art, the invention has the beneficial effects that:
According to the invention, through accurate phase compensation and modal parameter identification, the accuracy of basic data points of the constructed track is ensured; the technical characteristic of the method enables the method to capture track form distortion caused by structural damage sharply, and the change of the integral form is incomparable by weak deviation of single parameter points, thereby realizing high-sensitivity detection of early and tiny damage;
The diagnosis basis of the invention is the morphological similarity of the whole track, namely, the method is realized by calculating the Frobenius norms of two global mode feature matrixes, so that the method is insensitive to fluctuation of single data points due to random noise or instantaneous disturbance, the track form of healthy blades keeps intrinsic consistency even under different normal running loads, once damage occurs, the consistency is destroyed, the deviation index is obviously increased, and the diagnosis result is not easily influenced by normal working condition fluctuation due to the judgment logic based on the mode similarity rather than the absolute value of the parameters;
According to the invention, through setting the continuous monitoring judgment conditions based on statistics, the accidental interference signals can be effectively filtered while the real fault trend is accurately captured, and highly reliable early warning information is provided for wind farm operation and maintenance personnel, so that the real and effective predictive maintenance is realized.
Drawings
FIG. 1 is a schematic flow chart of the overall method of the present invention;
FIG. 2 is a histogram of trace deviation index versus anomaly probability in accordance with the present invention;
FIG. 3 is a plot of modal frequencies versus modal damping ratios for the present invention;
FIG. 4 is a graph of trace deviation index versus anomaly probability fit according to the present invention;
FIG. 5 is a normal probability map of anomaly probabilities in accordance with the present invention;
FIG. 6 is a flow chart of the overall device structure of the present invention.
Detailed Description
The present invention will be further described in detail with reference to specific embodiments in order to make the objects, technical solutions and advantages of the present invention more apparent.
It is to be noted that unless otherwise defined, technical or scientific terms used herein should be taken in a general sense as understood by one of ordinary skill in the art to which the present invention belongs. The terms "first," "second," and the like, as used herein, do not denote any order, quantity, or importance, but rather are used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that elements or items preceding the word are included in the element or item listed after the word and equivalents thereof, but does not exclude other elements or items. The terms "connected" or "connected," and the like, are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "up", "down", "left", "right" and the like are used only to indicate a relative positional relationship, and when the absolute position of the object to be described is changed, the relative positional relationship may be changed accordingly.
Examples:
Referring to fig. 1-5, the present invention provides a technical solution:
a blade fault diagnosis method based on vibration of a wind generating set comprises the following specific steps:
Step 1, arranging sensors at the basic position of a cabin and the root of a blade of a wind generating set, synchronously acquiring cabin reference vibration signals and blade response vibration signals, and screening steady-state effective data segments through a preset time window;
For the cabin foundation, the position sensing is the comprehensive vibration transmitted to the cabin by the whole transmission chain (comprising a wind wheel, a main shaft, a gear box and a generator), and the comprehensive vibration is taken as a reference signal and represents the combined action of an excitation source and a transmission path; for the root of the blade, the position directly senses the local dynamic response of the blade, and because the blade is a flexible body, the vibration of the blade contains rich structural dynamics information, and the root response is combined with a cabin reference signal to separate the dynamic characteristics of the blade, the synchronous acquisition is used for ensuring that two signals are strictly aligned in time, which is the basis for carrying out time delay estimation, cross-correlation analysis and frequency response function calculation subsequently, and if the two signals are not synchronous, systematic errors exist in all subsequent analysis;
the time window refers to a total duration frame of data acquisition, the effective data segment refers to a high-quality short data segment which is screened from total acquired data and is suitable for subsequent precise modal analysis, and the screening standard is that the unit is in a steady-state working condition;
Setting a time window with the current time T as an end point and tracing back a fixed length T Wherein,For the time length of 10 complete rotation periods, synchronous signal acquisition is carried out through a reference acceleration sensor positioned on a main frame of the engine room and a response acceleration sensor positioned on the inner side of a blade root flange in a time window to acquire an engine room reference vibration signal and a blade response vibration signal, wherein the sampling frequency of the reference acceleration sensor and the response acceleration sensor is set to be 100Hz, namely sampling time intervalSecond, this arrangement satisfies the Nyquist sampling theorem, effectively covering the principal frequency component of blade vibration (below 50 Hz), and providing sufficient resolution for subsequent cross-correlation calculations and time delay interpolation, over a time windowIn, synchronize signal acquisitionTo be performed at intervals, ensuring that the nacelle reference vibration signal has a consistent time base with the time series of blade response vibration signals
WhereinFor the current rotational speed of the wind turbine blades,Ensuring that the data buffer is large enough, e.g. ifSecond (about 0.33Hz for a typical fan, 10 turns, 30 seconds), then T needs at least more than 90 seconds; the 10 rotation periods are selected to be representative in statistics, namely, the stable vibration period cannot be captured when the time is too short and is easily influenced by random fluctuation, the response speed of the system to the working condition change is reduced when the time is too long, and the 10 periods are empirical values for balancing the stability and the timeliness;
The band-pass filtering is adopted to filter the signals, the passband frequency range covers the waving modal frequency of the blade, namely the vibration signals are subjected to the band-pass filtering to keep the main modal frequency components of the blade, the wind speed and the output power signals are monitored in real time in the running process of the unit, and a length is set as The sliding window is made to be in the time windowSliding in the range, judging that the unit enters a steady-state working condition when the fluctuation ranges of the wind speed and the output power are respectively smaller than the respective preset thresholds in the sliding window continuously, and taking the time interval corresponding to the sliding window as an effective data segment;
The filter uses a zero-phase digital filter to avoid phase distortion, and removes high-frequency noise such as vibration of a gear box and a generator and low-frequency interference such as tower swing, so that vibration components related to a blade waving mode are highlighted, and the signal to noise ratio is improved;
the blade is most easily excited and is most sensitive to damage, and is usually a low-order mode, particularly a first-order waving mode, firstly, the frequency range of the first-order waving mode of the blade is estimated through simulation calculation or unit design data, for example, for a large-scale wind turbine blade, the range is between 0.8Hz and 1.5Hz, then the passband of a bandpass filter is set in the range, for example, 0.5Hz-2.0Hz, so that the signal-to-noise ratio can be remarkably improved, and the subsequent mode identification is more accurate;
Length of Is the execution unit of steady state judgment toFor step length, atAnd (3) sliding the window and checking the data in each window to evaluate the steady state characteristics, wherein the steady state quantization standard is based on the statistical fluctuation of the data in the window, the fluctuation range is required to be smaller than a preset threshold value, and when the method is concretely implemented, for wind speed data, the variation coefficient (the ratio of standard deviation to mean value) of the wind speed value in the window is calculated, the variation coefficient is required to be smaller than the preset threshold value (such as not more than 0.1), for power data, the variation coefficient (the ratio of standard deviation to mean value) of the power value in the window is calculated, the variation coefficient is required to be smaller than the preset threshold value (such as not more than 0.03), wherein the variation coefficient is defined as the ratio of standard deviation to mean value, and the time window is judged to be in steady state working condition only when the fluctuation statistics of all monitoring parameters in the window meet the respective preset threshold value requirements.
Step 2, analyzing time delay between a cabin reference signal and a blade response signal according to the effective data segments, and identifying modal parameters of the first order of the blade through phase compensation and frequency response function calculation, wherein the modal parameters comprise modal frequencies and modal damping ratios, and repeating the operations on a plurality of different effective data segments to obtain a series of modal parameters;
time delay analysis physically takes time for vibrations to travel from the nacelle to the blade root, this delay containing information about the propagation of the signal in the structural path, accurate estimation of it, being the first step to ensure accuracy of subsequent analysis, the blade responding to the signal due to the time delay Will lag in phase behind the nacelle reference signalThis phase lag distorts the calculated frequency response function, and phase compensation is an alignment operation aimed at eliminating this propagation effect, allowing us to observe pure blade dynamics subsystem characteristics;
The frequency response function describes the relation between the output (response) and the input (excitation) of the unit blade on the frequency domain, the unit blade resonates at the modal frequency, the response is greatly amplified, the frequency response function is represented as a prominent peak value, and the key modal frequency and damping ratio can be extracted by analyzing the peak value;
the method for identifying the modal parameters of the first order of the blade based on each effective data segment comprises the steps of calculating a nacelle reference signal under each effective data segment Responsive to blade signalsCross-power spectrum of (a)Self-power spectrum of cabin reference signalWhereinAndRespectively isAndIs used for the fourier transform of (a),Wherein for the desired operation(s),Is complex conjugated;
Is itself a complex number, representing the frequency content in both signals Common power and phase relationship on it, which is reflected in frequencyWhere the magnitude of the linear correlation of the blade response signal with the nacelle reference signal is greater, the frequency component is representedFrom the slaveTo transfer toThe stronger the energy, the frequency is the natural frequency of the system,Is directly dependent on the size of (2)AndIf input is made of the amplitude of (2)At which the energy is present and at which the system is readily excited (i.e., near the modal frequency), then the outputWill be large in amplitude, resulting inThe amplitude of (2) is also very large;
is a real number representing the blade reference signal The distribution of its own power in the frequency domain, which reflects the individual frequency components in the excitation signalThe amount of energy involved, which indicates at which frequencies the system is excited,Directly byIs determined by the square of the magnitude of (a),The larger the amplitude of (a) is,The greater the value at that frequency;
By aligning Performing inverse Fourier transform to obtain a cross-correlation function: Wherein Is an inverse fourier transform;
Representing delaying two signals at different times The following similarity, which reflects the signalWith a time delayTo a similar extent whenEqual to the real time delay between the two signals,A maximum value is reached and is therefore a critical function indicating the time delay between signals; the larger the value of (2) is, the time delay is represented The more similar the waveforms of the two signals are, the corresponding global maximumIs the primary time delay to be found;
Searching for a lead Initial time delay to take maximum valueBy usingParabolic interpolation of cross-correlation function values at three points to accurately determine time delayThe interpolation formula is as follows:
;
wherein, the Is a sampling time interval;
Since the signal is discretely sampled, the maximum point is initially found Can only be accurate to the sampling intervalThe real maximum point is positioned between two sampling points, parabolic interpolation utilizes the maximum point and the left and right adjacent points thereof to fit a parabola and find the vertex of the parabola, thereby delaying the timeThe accuracy of (2) is improved to a sub-sampling level;
The above formula passes through the three function values Calculating the vertex abscissa of the parabola ifThe larger the value relative to the left and right points (i.e., the sharper the peak), then the calculatedThe closer the approach isIf the peak is asymmetric, the peak is not symmetric,Will shift to the higher function side;
to eliminate time delay Influence on phase information, cross-power spectrum of both in frequency domainAnd (3) compensating, wherein the compensation formula is as follows:
;
wherein, the In units of imaginary numbers,For frequency, the compensated positive sign is used to counteract the time delay in the blade response signal relative to the nacelle reference signalThereby correcting the phase of the cross-power spectrum, the formula being multiplied byTo compensate (cancel) for time delays in response signalsWith resultant phase lagThis positive sign compensates for the delay itself, which has the effect of advancing the response signal in phase so as to align with the reference signal;
And uses the compensated cross power spectrum Self-power spectrum with cabin reference signalCalculating a frequency response function from the nacelle to the blade: in the following Identifying the highest peak value in the theoretical frequency range of the first-order waving mode of the blade, determining the frequency corresponding to the peak value as the mode frequency, and calculating the mode damping ratio of the first-order mode by using a half-power bandwidth method;
Is a complex number containing amplitude and phase information describing the blade as a system whose vibrational response is related to the frequency domain of the excited, at the modal frequencies of the blade, A significant peak occurs in the amplitude of (c) because the system resonates at this frequency, a small excitation can produce a large response,The larger the amplitude of (a) is, the more the system is at that frequencyThe stronger the ability of the vibration response to amplify, which strongly suggests that this frequency is a natural frequency of the system;
within a predicted theoretical frequency range, e.g. obtained by design drawing or simulation, find Peak in the amplitude spectrum, the peak corresponding to the frequencyI.e. the identified first order flapping modal frequency, which is a fundamental inherent property of the blade;
at the identified modal frequencies Find the amplitude at the peak of (2)Calculating the-3 d point, i.e. the amplitude drops toTwo frequencies corresponding to each otherAnd,Calculating half power bandwidthModal damping ratioThe calculation formula of (2) is as follows: The damping ratio reflects the capacity of the structural system to consume vibration energy, the larger the damping ratio is, the stronger the capacity of the system to dissipate energy is, the vibration is attenuated faster, the damping ratio is generally in a stable and smaller range for a healthy blade, if the damping ratio is obviously increased, the damping ratio means that friction type damages exist in the structure, such as the mutual friction of crack surfaces to consume energy, if the damping ratio is obviously reduced, the damping ratio means that the structural connection is loosened, and the energy consumption mechanism is weakened, so that the damping ratio is the same as the modal frequency and is a key index for diagnosing the health state of the structure.
Step 3, sequentially connecting a series of modal parameters in a coordinate system with the modal frequency as an abscissa and the modal damping ratio as an ordinate, constructing a track of dynamic response, calculating relative vectors of each modal parameter on the track and all other modal parameters in the track to form local feature vectors, integrating all the local feature vectors, and constructing a global feature matrix representing the overall track morphology;
Each data point represents the 'dynamic fingerprint' of the blade under a specific steady-state working condition, the starting time sequence is essentially working condition sequence, and the sequence implies the change of the load from small to large because the data are collected under the steady-state working condition with different wind speeds (i.e. different loads), so the track reveals the inherent rule of the change of the modal parameter (frequency and damping) of the blade along with the wind load, and a healthy blade has stable and repeatable track form;
when constructing the track of dynamic response, the modal frequency and modal damping ratio corresponding to each effective data segment are taken as one data point N data points are obtained in total and are sequentially connected according to the starting time sequence of the effective data segments to form a track, whereinIs the modal frequency of the ith data point,Modal damping ratio for the i-th data point;
for the ith data point on the trace Its local feature vectorIs oneThe column vector of the dimension has the mathematical expression:
;
wherein, the I is the index of the data point, and;
In order to quantify the geometry of the entire trajectory, a mathematical tool is required that describes the global relationship of each data point, local feature vectorIt is to create a "global context description" for each data point on the trace;
Itself a high-dimensional vector whose elements are the current data points The difference in coordinates from all other data points on the track,Reflecting data pointsThe unique position and role in the overall track shape is that it is not just a data point information, but encodesRelative to all other data points on the trace ifIs an abnormal point (e.g., due to injury), its relative relationship to most points on a healthy track changes dramatically, resulting in itA significant change occurs;
The size of each element in (a) represents The Euclidean distance component in the modal parameter space from other data points on each axis (frequency, damping) ifThese differences are then severely deviated from the healthy track position in which it should be due to injuryThe absolute value of (2) will generally increase;
Directly dependent on these coordinate values, if the unit blade is healthy, the track form constituted by all data points is stable, then each Is also stable, if damage occurs, the coordinates of certain data points will change, which will result in that pointAlmost all elements in the data are changed, and other data points are slightly influencedThis general nature of traction makes the feature extremely sensitive to local injury;
Performing Z-score normalization processing on the modal frequencies and modal damping ratios in all effective data segments in a time window, and performing normalization on all N local feature vectors according to the corresponding data points Sequentially stacked as row vectors to form a global feature matrixProcessing the signals of the same unit blade in the healthy and lossless state by the same method to obtain a reference global feature matrix;
The global feature matrix integrates N vectors describing the local feature vector of each point into a mathematical object capable of describing the complete shape of the whole track;
Z-score normalization is used for eliminating deviation caused by different dimension and magnitude of two parameters of modal frequency and damping ratio, the frequency value is usually about 1Hz, the damping ratio is only 0.01, if the frequency is not processed, tiny fluctuation of the frequency can be dominant in a characteristic vector to cover the change of the damping ratio;
The mean value and standard deviation used for normalization must come from data monitored for a long time under the health state of the blade, the data to be monitored is normalized by using the fixed health reference statistic, and the new mean value and standard deviation are calculated by using the data of the data to be monitored, so that all the data can be compared under the same reference;
Rather than just a "digitized template" of the health track morphology, it is more precisely a relative spatial relationship-based health track behavior code embedded with temporal sequence information, which contains the relative vector relationships (spatial information) between all data points, while by the exact order of the rows, The time stamp sequence (time information) when these data points were confirmed is retained in strong association; Defining the specific spatial configuration formed by the dynamic parameters of healthy leaves under the time sequence, and any factor (such as structural damage) destroying the spatial-temporal configuration can lead to the currently constructed global feature matrix And (3) withCreating a large difference.
Step 4, calculating a reference global feature matrix when the blade is healthy and lossless according to the steps 1 to 3, calculating a Frobenius norm between the global feature matrix and the reference global feature matrix, defining the norm value as a track deviation degree, and judging whether the wind generating set blade has structural damage according to a track deviation construction state transition probability matrix;
Defining track departure index The Frobenius norm, which is the difference between the two matrices, is calculated as follows:
;
wherein, the As a reference global feature matrix,As a global feature matrix of the system,The Frobenius norm of the matrix;
compressing the overall difference between the two complex matrices into a single, comparable scalar value, the Frobenius norm is the mathematical tool to achieve this goal; Quantitatively reflecting the comprehensive deviation degree between the overall dynamic behavior mode of the blade to be monitored and the health reference mode thereof, and calculating the overall level of the difference value of all corresponding position elements in the two global feature matrixes; The larger the value of (i) is, the larger the difference between the current dynamic response track and the healthy track of the blade to be monitored in the overall shape is, the difference is caused by the change of the structural characteristics of the blade (namely, the occurrence of damage), and the difference is approximately 0 The value means that the current behavior is almost consistent with the healthy behavior;
Is the square root of the sum of these difference squares, and therefore, AndThe difference between any corresponding elements increases, which results inIs increased by a value of (a),Sensitive to any tiny change in the matrix, and can comprehensively and comprehensively capture any local or global morphological distortion caused by damage;
Under the same unit blade health lossless state, acquiring historical data of track deviation index through a plurality of historical time windows, and calculating the average value of the historical data And standard deviationReal-time monitoring the obtained track deviation indexAbnormal probability of (2)Calculated from the following formula:
;
wherein, the A cumulative distribution function that is a standard normal distribution;
the partial time window and the specific data of the anomaly probability are shown in table 1.
TABLE 1 statistical graph of anomaly probabilities
Through analysis of the data, a certain correlation is observed between different characteristic parameters, for example, as can be seen from the data, a certain negative correlation trend exists between the modal frequency and the track deviation index, the modal frequency gradually decreases (from 0.5995 to 0.5858) along with the increase of a time window, and the track deviation index increases (from 0.027 to 0.064). This indicates that the vibration characteristics of the blade change under the condition of low modal frequency, resulting in an increase in the degree of trajectory deviation, reflecting deterioration of the structural state;
When the relation between the modal frequency and the anomaly probability is analyzed, the anomaly probability shows a gradually increasing trend along with the reduction of the modal frequency, and the anomaly probability is increased due to the fact that the reduction of the modal frequency is related to damage of the blade structure, so that the sensitivity of anomaly detection is improved, for example, the modal frequency of a time window 1 is 0.5995, the corresponding anomaly probability is 0.556, the modal frequency of a time window 15 is 0.5858, the anomaly probability is increased to 0.811, and the comprehensive influence of the modal frequency and the track deviation degree is considered to optimize the accuracy and the reliability of blade health monitoring.
When the anomaly probability continues to be higher than a set threshold for a succession of K time windows, then a structural damage is determined to exist in the leaf, where K is 3 or more, this determination strategy perfectly balances sensitivity and reliability, which is sufficiently sensitive to early damage (byCapture), has strong anti-interference capability through persistence requirements, effectively avoids false alarm, and makes a final diagnosis conclusion very stable enough to be used as a decision basis for shutdown maintenance;
when the blade is healthy, repeatedly executing the steps 1-4 for a long time to calculate a large number of Values that constitute a state of healthDistribution of samples by means of their averageAnd standard deviationTo fully describe, meanRepresenting healthy leavesThe normal average level of the value, which should ideally be close to 0, is usually a value slightly greater than 0 due to various random factors; representing healthy leaves The normal fluctuation range of the value is set,The larger the indication of healthUnstable and large fluctuation; And Together define a reference for health status, e.g.Is the limit range of normal fluctuations;
Will be current The value is converted into a more visual and explanatory index-which is the probability of an outlier-which is compared with the direct useThe absolute value of (2) is more finely judged by threshold value, and the abnormality probability is higherIs a value between 0 and 1,Indicating the deviation of the track currently observedPossibility of belonging to health distribution; The larger the value of (2) is, the current is represented The value comes from an abnormal state, so that the blade state is extremely likely to be abnormal (damaged), e.gIndicating that in a healthy state, only a 5% probability of the current situation is observedValues or greater, and therefore with high confidence, the unit blades are considered abnormal;
is normalized Values, i.e.It represents the currentBy how many standard deviations from the healthy mean,The larger the value of (c) is,The closer to 1 the value is (i.e., the cumulative probability of the value in a normal distribution), and thereforeThe closer to 1, which is a monotonically increasing relationship, normalizedThe greater the value, the probability of abnormalityThe larger the size;
When (when) Continuously below the set threshold for a continuous K time windowsFor exampleWhen the blade is judged to have structural damage, wherein K is more than or equal to 3, the judgment strategy balances sensitivity and specificity, and requires that an abnormal signal is continuous, false alarms caused by instantaneous interference are avoided, for example, 3 continuous time windows are all providedThere is a high degree of confidence in determining the impairment.
Referring to fig. 6, the present invention further provides a device for diagnosing a blade fault based on vibration of a wind turbine generator, where the device is configured to perform the above method for diagnosing a blade fault based on vibration of a wind turbine generator, and the method includes:
The signal acquisition module is used for arranging sensors at the basic position of a cabin and the root of a blade of the wind generating set, synchronously acquiring cabin reference vibration signals and blade response vibration signals, and screening steady-state effective data segments through a preset time window;
The modal calculation module is used for analyzing the time delay between the cabin reference signal and the blade response signal according to the effective data segments, and identifying the modal parameters of the first order of the blade through phase compensation and frequency response function calculation, wherein the modal parameters comprise modal frequencies and modal damping ratios, and repeating the operations on a plurality of different effective data segments to obtain a series of modal parameters;
the characteristic construction module is used for sequentially connecting a series of modal parameters in a coordinate system taking modal frequency as an abscissa and modal damping ratio as an ordinate to construct a track of dynamic response, calculating relative vectors of each modal parameter on the track and all other modal parameters in the track to form local feature vectors, integrating all the local feature vectors, and constructing a global feature matrix representing the overall track shape;
the result judging module is used for calculating a reference global feature matrix when the blade is healthy and lossless according to the steps 1 to 3, calculating a Frobenius norm between the global feature matrix and the reference global feature matrix, defining the norm value as a track deviation degree, and judging whether the wind turbine blade has structural damage according to a track deviation construction state transition probability matrix.
The above formulas are all formulas with dimensions removed and numerical values calculated, the formulas are formulas with a large amount of data collected for software simulation to obtain the latest real situation, and preset parameters in the formulas are set by those skilled in the art according to the actual situation.
The above embodiments may be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, the above-described embodiments may be implemented in whole or in part in the form of a computer program product. Those of skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application.

Claims (8)

1.一种基于风力发电机组振动的叶片故障诊断方法,其特征在于,具体步骤包括:1. A method for diagnosing blade faults based on the vibration of a wind turbine generator set, characterized in that the specific steps include: 步骤1:在风力发电机组的机舱基础位置和叶片根部布设传感器,同步采集机舱参考振动信号与叶片响应振动信号,并通过预设时间窗口筛选出稳态的有效数据段;Step 1: Install sensors at the nacelle foundation and blade root of the wind turbine generator set to synchronously collect the nacelle reference vibration signal and the blade response vibration signal, and filter out the effective steady-state data segment through a preset time window; 步骤2:根据有效数据段,分析机舱参考信号与叶片响应信号间的时间延迟,通过相位补偿和频率响应函数计算,识别该叶片一阶的模态参数,其包括模态频率和模态阻尼比,在多个不同的有效数据段重复上述操作,获得一系列的模态参数;Step 2: Based on the valid data segments, analyze the time delay between the nacelle reference signal and the blade response signal. Through phase compensation and frequency response function calculation, identify the first-order modal parameters of the blade, including modal frequency and modal damping ratio. Repeat the above operation in multiple different valid data segments to obtain a series of modal parameters. 步骤3:将一系列模态参数在以模态频率为横坐标、模态阻尼比为纵坐标的坐标系中依次连接,构建动力学响应的轨迹,针对该轨迹上的每个模态参数,计算其与轨迹中所有其他模态参数的相对矢量以形成局部特征向量,整合所有局部特征向量,构建表征整体轨迹形态的全局特征矩阵;Step 3: Connect a series of modal parameters in a coordinate system with modal frequency as the abscissa and modal damping ratio as the ordinate to construct the trajectory of the dynamic response. For each modal parameter on the trajectory, calculate its relative vector with all other modal parameters in the trajectory to form a local feature vector. Integrate all local feature vectors to construct a global feature matrix that characterizes the overall trajectory shape. 步骤4:按照步骤1至3计算该叶片健康无损时的基准全局特征矩阵,计算全局特征矩阵与基准全局特征矩阵间的Frobenius范数,将该范数值定义为轨迹偏离度,根据轨迹偏离构建状态转移概率矩阵判断风力发电机组叶片是否存在结构损伤。Step 4: Calculate the baseline global feature matrix when the blade is healthy and undamaged, following steps 1 to 3. Calculate the Frobenius norm between the global feature matrix and the baseline global feature matrix. Define the norm value as the trajectory deviation. Construct a state transition probability matrix based on the trajectory deviation to determine whether there is structural damage to the wind turbine blade. 2.根据权利要求1所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:设定一个以当前时刻t为终点,往前追溯固定长度为T的时间窗口,其中为10个完整旋转周期的时间长度,在时间窗口内通过位于机舱主机架的参考加速度传感器与位于叶片根部法兰内侧的响应加速度传感器进行同步信号采集,以获取机舱参考振动信号与叶片响应振动信号;2. The blade fault diagnosis method based on wind turbine generator vibration according to claim 1, characterized in that: a time window with a fixed length T tracing back from the current time t is set. ,in , The time length is 10 complete rotation cycles. Within the time window, synchronous signal acquisition is performed by a reference acceleration sensor located on the main frame of the nacelle and a response acceleration sensor located inside the blade root flange to obtain the nacelle reference vibration signal and the blade response vibration signal. 采用带通滤波对信号进行滤波处理,其通带频率范围覆盖叶片的一阶挥舞模态频率,即对振动信号进行带通滤波以保留叶片主要模态频率成分;在机组运行过程中,实时监测风速与输出功率信号,设定一个长度为的滑动窗口,使该滑动窗口在时间窗口范围内滑动,当风速和输出功率的波动范围在滑动窗口内均分别小于其各自的预设阈值时,则判定机组进入稳态工况,把这一滑动窗口对应的时间区间作为一个有效数据段。A bandpass filter is used to filter the signal, with its passband frequency range covering the first-order flapping mode frequency of the blades. This bandpass filtering preserves the main modal frequency components of the blades. During unit operation, wind speed and output power signals are monitored in real time, and a range of lengths is set. A sliding window, making the sliding window appear within the time window. Within the sliding window, when the fluctuation range of wind speed and output power is less than their respective preset thresholds, the unit is determined to have entered steady-state operation, and the time interval corresponding to this sliding window is taken as a valid data segment. 3.根据权利要求2所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:基于每一个有效数据段识别叶片一阶的模态参数,具体包括:计算每一有效数据段下机舱参考信号与叶片响应信号的互功率谱以及机舱参考信号的自功率谱,其中分别为的傅里叶变换,其中为期望运算,为复共轭;3. The blade fault diagnosis method based on wind turbine generator vibration according to claim 2, characterized in that: the first-order modal parameters of the blade are identified based on each effective data segment, specifically including: calculating the nacelle reference signal under each effective data segment. With blade response signal cross power spectrum and the self-power spectrum of the cabin reference signal ,in and They are respectively and Fourier transform, Where is the expectation operation, For complex conjugate; 通过对进行逆傅里叶变换得到互相关函数:,其中为逆傅里叶变换;Through the The cross-correlation function is obtained by performing an inverse Fourier transform: ,in This is the inverse Fourier transform; 寻找使取得最大值的初始时间延迟,利用三点处的互相关函数值进行抛物线插值,以精确确定时间延迟,插值公式如下:Searching for Initial time delay to reach the maximum value ,use , , Parabolic interpolation of the cross-correlation function values at the three points is performed to accurately determine the time delay. The interpolation formula is as follows: ; 其中,为采样时间间隔。in, This represents the sampling time interval. 4.根据权利要求3所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:为消除时间延迟对相位信息的影响,在频域对两者的互功率谱进行补偿,补偿公式为:4. The blade fault diagnosis method based on wind turbine generator vibration according to claim 3, characterized in that: to eliminate time delay The impact on phase information, in the frequency domain, on the cross-power spectrum of the two. Compensation will be provided, and the compensation formula is as follows: ; 其中,为虚数单位,为频率,补偿的正号用于抵消叶片响应信号中相对于机舱参考信号的时间延迟,从而校正互功率谱的相位;in, The imaginary unit, The positive sign of the compensation, representing the frequency, is used to offset the time delay in the blade response signal relative to the nacelle reference signal. This corrects the phase of the cross power spectrum; 并利用补偿后的互功率谱与机舱参考信号的自功率谱,计算从机舱到叶片的频率响应函数:,在的幅值谱中,识别出位于叶片的一阶挥舞模态理论频率范围内的最高峰值,将该峰值对应的频率确定为模态频率,并利用半功率带宽法计算该一阶模态的模态阻尼比。And using the compensated cross power spectrum Self-power spectrum of the cabin reference signal Calculate the frequency response function from the nacelle to the blades: ,exist In the amplitude spectrum, the highest peak value located within the theoretical frequency range of the first-order flapping mode of the blade is identified, the frequency corresponding to the peak value is determined as the modal frequency, and the modal damping ratio of the first-order mode is calculated using the half-power bandwidth method. 5.根据权利要求4所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:构建动力学响应的轨迹时,将各有效数据段对应的模态频率和模态阻尼比作为一个数据点,共得到N个数据点并按照有效数据段的起始时刻先后顺序,依次连接各数据点形成轨迹,其中为第i个数据点的模态频率,为第i个数据点的模态阻尼比;5. The blade fault diagnosis method based on wind turbine generator vibration according to claim 4, characterized in that: when constructing the trajectory of the dynamic response, the modal frequency and modal damping ratio corresponding to each effective data segment are taken as a data point. A total of N data points were obtained, and the data points were connected sequentially according to the starting time of the valid data segments to form a trajectory. Let i be the modal frequency of the i-th data point. Let i be the modal damping ratio of the i-th data point; 对轨迹上的第i个数据点,其局部特征向量是一个维的列向量,其数学表达式如下:For the i-th data point on the trajectory Its local feature vector It is A column vector of dimension 1 has the following mathematical expression: ; 其中,为转置符,i为数据点的索引,且in, `i` is the transpose operator, where `i` is the index of the data point, and . 6.根据权利要求5所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:对时间窗口内所有有效数据段内的模态频率和模态阻尼比进行Z-score归一化处理,将归一化后的所有N个局部特征向量,按照其对应的数据点的顺序,依次作为行向量进行堆叠,从而形成全局特征矩阵;对同一机组叶片在健康无损状态下的信号以相同的方法处理,以得到基准全局特征矩阵6. The blade fault diagnosis method based on wind turbine generator vibration according to claim 5, characterized in that: the modal frequencies and modal damping ratios in all effective data segments within the time window are Z-score normalized, and all N normalized local feature vectors are sorted according to their corresponding data points. The elements are stacked sequentially as row vectors to form the global feature matrix. Signals from the same unit's blades under healthy and undamaged conditions are processed using the same method to obtain a baseline global feature matrix. ; 定义轨迹偏离度指标为两矩阵之差的Frobenius范数,计算方法如下:Define trajectory deviation index The Frobenius norm of the difference between two matrices is calculated as follows: ; 其中,为基准全局特征矩阵,为全局特征矩阵,表示矩阵的Frobenius范数。in, As the baseline global feature matrix, The global feature matrix, This represents the Frobenius norm of the matrix. 7.根据权利要求6所述的一种基于风力发电机组振动的叶片故障诊断方法,其特征在于:在同一机组叶片健康无损状态下,通过多个历史的时间窗口获取轨迹偏离度指标的历史数据,并计算其均值和标准差,实时监测得到的轨迹偏离度指标的异常概率由下式计算:7. The method for blade fault diagnosis based on wind turbine vibration according to claim 6, characterized in that: under the same healthy and undamaged blade condition, historical data of trajectory deviation index are obtained through multiple historical time windows, and their average value is calculated. and standard deviation The trajectory deviation index obtained through real-time monitoring abnormal probability Calculated by the following formula: ; 其中,为标准正态分布的累积分布函数;in, The cumulative distribution function of the standard normal distribution; 当所述异常概率在连续K个时间窗口内持续高于设定阈值时,则判定叶片存在结构损伤,其中K大于等于3。When the abnormal probability is continuously higher than a set threshold for K consecutive time windows, it is determined that the blade has structural damage, where K is greater than or equal to 3. 8.一种基于风力发电机组振动的叶片故障诊断装置,其特征在于:所述装置用于执行权利要求1-7任一项所述的一种基于风力发电机组振动的叶片故障诊断方法,包括:8. A blade fault diagnosis device based on wind turbine generator vibration, characterized in that: the device is used to execute the blade fault diagnosis method based on wind turbine generator vibration as described in any one of claims 1-7, comprising: 信号采集模块,用于在风力发电机组的机舱基础位置和叶片根部布设传感器,同步采集机舱参考振动信号与叶片响应振动信号,并通过预设时间窗口筛选出稳态的有效数据段;The signal acquisition module is used to deploy sensors at the nacelle foundation and blade root of the wind turbine generator set to simultaneously acquire the nacelle reference vibration signal and the blade response vibration signal, and to filter out the effective steady-state data segments through a preset time window. 模态计算模块,用于根据有效数据段,分析机舱参考信号与叶片响应信号间的时间延迟,通过相位补偿和频率响应函数计算,识别该叶片一阶的模态参数,其包括模态频率和模态阻尼比,在多个不同的有效数据段重复上述操作,获得一系列的模态参数;The modal calculation module is used to analyze the time delay between the nacelle reference signal and the blade response signal based on the effective data segment. Through phase compensation and frequency response function calculation, it identifies the first-order modal parameters of the blade, including the modal frequency and modal damping ratio. The above operation is repeated in multiple different effective data segments to obtain a series of modal parameters. 特征构建模块,用于将一系列模态参数在以模态频率为横坐标、模态阻尼比为纵坐标的坐标系中依次连接,构建动力学响应的轨迹,针对该轨迹上的每个模态参数,计算其与轨迹中所有其他模态参数的相对矢量以形成局部特征向量,整合所有局部特征向量,构建表征整体轨迹形态的全局特征矩阵;The feature construction module is used to connect a series of modal parameters in a coordinate system with modal frequency as the abscissa and modal damping ratio as the ordinate to construct the trajectory of the dynamic response. For each modal parameter on the trajectory, its relative vector with all other modal parameters in the trajectory is calculated to form a local feature vector. All local feature vectors are integrated to construct a global feature matrix that characterizes the overall trajectory shape. 结果判断模块,用于按照步骤1至3计算该叶片健康无损时的基准全局特征矩阵,计算全局特征矩阵与基准全局特征矩阵间的Frobenius范数,将该范数值定义为轨迹偏离度,根据轨迹偏离构建状态转移概率矩阵判断风力发电机组叶片是否存在结构损伤。The result judgment module is used to calculate the baseline global feature matrix when the blade is healthy and undamaged according to steps 1 to 3, calculate the Frobenius norm between the global feature matrix and the baseline global feature matrix, define the norm value as the trajectory deviation, and construct a state transition probability matrix based on the trajectory deviation to determine whether there is structural damage to the wind turbine blade.
CN202511494495.2A 2025-10-20 2025-10-20 Blade Fault Diagnosis Method and Device Based on Wind Turbine Generator Vibration Active CN120969089B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202511494495.2A CN120969089B (en) 2025-10-20 2025-10-20 Blade Fault Diagnosis Method and Device Based on Wind Turbine Generator Vibration

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202511494495.2A CN120969089B (en) 2025-10-20 2025-10-20 Blade Fault Diagnosis Method and Device Based on Wind Turbine Generator Vibration

Publications (2)

Publication Number Publication Date
CN120969089A true CN120969089A (en) 2025-11-18
CN120969089B CN120969089B (en) 2026-01-30

Family

ID=97641932

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202511494495.2A Active CN120969089B (en) 2025-10-20 2025-10-20 Blade Fault Diagnosis Method and Device Based on Wind Turbine Generator Vibration

Country Status (1)

Country Link
CN (1) CN120969089B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB745421A (en) * 1950-12-05 1956-02-29 British Thomson Houston Co Ltd Improvements in or relating to electron discharge devices
CN115496102A (en) * 2022-09-23 2022-12-20 西安热工研究院有限公司 Wind turbine generator blade fault diagnosis method and device, equipment and storage medium
CN116542101A (en) * 2023-05-09 2023-08-04 重庆大学 Wind turbine generator blade load dynamic estimation method
CN120369093A (en) * 2025-04-15 2025-07-25 江苏国电南自海吉科技有限公司 Wind turbine generator blade vibration detection method based on optical fiber whispering gallery mode singular point microcavity model
CN120742003A (en) * 2025-08-29 2025-10-03 北京华科同安监控技术有限公司 Power generation equipment state fault diagnosis method and system based on artificial intelligence

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB745421A (en) * 1950-12-05 1956-02-29 British Thomson Houston Co Ltd Improvements in or relating to electron discharge devices
CN115496102A (en) * 2022-09-23 2022-12-20 西安热工研究院有限公司 Wind turbine generator blade fault diagnosis method and device, equipment and storage medium
CN116542101A (en) * 2023-05-09 2023-08-04 重庆大学 Wind turbine generator blade load dynamic estimation method
CN120369093A (en) * 2025-04-15 2025-07-25 江苏国电南自海吉科技有限公司 Wind turbine generator blade vibration detection method based on optical fiber whispering gallery mode singular point microcavity model
CN120742003A (en) * 2025-08-29 2025-10-03 北京华科同安监控技术有限公司 Power generation equipment state fault diagnosis method and system based on artificial intelligence

Also Published As

Publication number Publication date
CN120969089B (en) 2026-01-30

Similar Documents

Publication Publication Date Title
CA2868643C (en) Wind turbine and method for evaluating the health state of the blades thereof
CN114708885A (en) Fan fault early warning method based on sound signals
Lu et al. Early fault warning and identification in condition monitoring of bearing via wavelet packet decomposition coupled with graph
CN113567162A (en) Fan fault intelligent diagnosis device and method based on acoustic sensor
CN120578897A (en) A method for fault analysis of hydropower station equipment based on state data mining
Cui et al. An anomaly detection approach using wavelet transform and artificial neural networks for condition monitoring of wind turbines' gearboxes
CN111881594A (en) Non-stationary signal state monitoring method and system for nuclear power equipment
CN119532231A (en) Abnormal detection method for fan operation status
CN115220396A (en) Intelligent monitoring method and system for numerical control machine tool
Chen et al. Markov modeling of signal condition transitions for bearing diagnostics under external interference conditions
CN119688342A (en) Coal mill fault detection system and method based on intelligent deep learning technology
CN117365869A (en) An adaptive early warning strategy design method for wind turbine blade sweeping failure
CN119446186A (en) Multi-channel audio acquisition and intelligent fusion coal mill fault voiceprint recognition system and method
CN118705204A (en) A control system and method for a train variable frequency axial flow fan
CN118998004A (en) Intelligent detection supervision system and detection method for wind power generation equipment
CN118503882A (en) Method for collecting operating state parameters and diagnosing faults of three-phase horizontal decanter centrifuge
CN119641570A (en) A wind turbine condition monitoring method based on distributed optical fiber sensing
CN118673305A (en) Wind turbine generator vibration fault on-line diagnosis system based on feature extraction
CN112347944A (en) A mechanical transmission fault detection method based on zero-value proportional spectrum features
CN120969089B (en) Blade Fault Diagnosis Method and Device Based on Wind Turbine Generator Vibration
CN120819475A (en) Fault monitoring method and system for doubly-fed wind turbine generator system based on frequency monitoring
CN120761021A (en) A method for monitoring wear and fatigue of sluice gate opening and closing equipment
CN120404212A (en) A wire drawing machine fault detection method based on the Internet of Things
CN120145089A (en) Motor performance detection method and system based on data analysis
CN118815653A (en) A monitoring optimization method and system for high-head long-short blade turbines

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