CN112733283A - Wind turbine generator component fault prediction method - Google Patents

Wind turbine generator component fault prediction method Download PDF

Info

Publication number
CN112733283A
CN112733283A CN202011520719.XA CN202011520719A CN112733283A CN 112733283 A CN112733283 A CN 112733283A CN 202011520719 A CN202011520719 A CN 202011520719A CN 112733283 A CN112733283 A CN 112733283A
Authority
CN
China
Prior art keywords
wind turbine
neural network
data
component
failure
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
Application number
CN202011520719.XA
Other languages
Chinese (zh)
Inventor
杨政厚
田宏哲
孙新佳
曹利蒲
刘鹏飞
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.)
Beijing Huaneng Xinrui Control Technology Co Ltd
Original Assignee
Beijing Huaneng Xinrui Control Technology 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 Beijing Huaneng Xinrui Control Technology Co Ltd filed Critical Beijing Huaneng Xinrui Control Technology Co Ltd
Priority to CN202011520719.XA priority Critical patent/CN112733283A/en
Publication of CN112733283A publication Critical patent/CN112733283A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/17—Mechanical parametric or variational design
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/06—Wind turbines or wind farms
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

The invention provides a wind turbine component fault prediction method, which comprises the following steps: acquiring operation state data of monitored wind turbine components; analyzing faults of the wind turbine generator component, and acquiring historical data and real-time data of operation of the wind turbine generator component; preprocessing the historical data and the real-time data; training a neural network by utilizing the preprocessed historical data to obtain a neural network prediction model; and comparing the real-time data with the fault early warning criterion by combining a neural network and an Euclidean distance algorithm to realize fault early warning. The failure prediction method provided by the invention can effectively win spare parts and maintenance plan optimization time, and can also eliminate false alarms of the SCADA system and reduce the maintenance amount, thereby reducing the maintenance cost of the wind power plant.

Description

Wind turbine generator component fault prediction method
Technical Field
The invention relates to the field of wind power, in particular to a wind turbine component fault prediction method.
Background
The failure prediction means that the future failure trend or the remaining effective life is predicted by analyzing and processing historical data and then by current data. With the development of technology and the need of production life, equipment overhaul is gradually developed from the initial accident overhaul to the current state overhaul. The fault prediction technology is the basis for the condition maintenance of the equipment. The condition maintenance of the wind turbine generator equipment is to analyze the running condition of the equipment and predict the development trend of the equipment on the basis of on-line monitoring of the equipment, and then timely maintain the equipment which possibly breaks down. Compared with the traditional fault monitoring mode, the fault prediction method has the greatest advantage that the possible faults can be predicted before the faults occur, if the prediction lead is long enough and the prediction result is accurate enough, field operators can immediately take corresponding measures to prevent the faults from occurring, and the safety, the stability and the economy of the operation of the wind turbine generator are improved.
The current mainstream wind turbine generator fault prediction method mainly comprises the following steps: vibration monitoring, oil monitoring, acoustic emission monitoring, strain measurement, mode identification method and intelligent early warning method. The vibration monitoring method is to identify the fault of the equipment by analyzing a time domain and a frequency domain of a vibration signal recorded by an equipment sensor, and mainly comprises methods such as frequency spectrum analysis, cepstrum analysis, envelope analysis, time waveform analysis and the like. The oil monitoring is to obtain the lubricating and abrasion state information of the machine by analyzing the performance change of the lubricating oil of the monitored equipment and the condition of the carried abrasion particles, thereby evaluating the working condition of the equipment and determining the fault reason. Acoustic emission monitoring is a non-destructive inspection method for assessing the integrity of a material's properties or structure by receiving and analyzing the material's acoustic emission signals. The strain measurement is to apply a strain gauge and a strain gauge to measure the surface strain of a device component, determine the stress state of the component according to the relationship between the strain and the stress, and judge the damage condition of an element, and mainly comprises an electrical measurement method, a light measurement method, an acoustic measurement method, a brittle coating method, a strain mechanical measurement method and the like. The pattern recognition method is a basic intelligent method, and mainly utilizes the characteristics or attributes provided in the information acquired from the diagnostic object to perform classification and recognition according to some pattern recognition algorithms. The intelligent early warning method is based on data information of online monitoring, realizes evaluation on wind power running conditions by using an artificial intelligence method, and does not completely depend on experimental data of a wind turbine generator.
The prior art proposes that a fan assembly normal model is built by utilizing a neural network to carry out fault early warning on a fan gear box, the fault early warning is realized by observing the error between a predicted value and real-time data of the neural network and the increase of error frequency, the neural network model is improved, and an optimal performance model is selected to carry out fault prediction on a fan generator bearing.
However, the precision of the early warning methods is not high, and the fault signals cannot be accurately identified, so that fault early warning is realized.
Disclosure of Invention
Aiming at the problems in the prior art, the invention proposes a wind turbine component fault prediction method, so that spare parts and maintenance plan optimization time can be effectively won, false alarms of an SCADA system can be eliminated, the maintenance amount is reduced, and the maintenance cost of a wind power plant is reduced.
Specifically, the embodiment of the invention provides a wind turbine component fault prediction method, which comprises the following steps:
acquiring operation state data of monitored wind turbine components;
analyzing faults of the wind turbine generator component, and acquiring historical data and real-time data of operation of the wind turbine generator component;
preprocessing the historical data and the real-time data;
training a neural network by utilizing the preprocessed historical data to obtain a neural network prediction model;
and comparing the real-time data with the fault early warning criterion by combining a neural network and an Euclidean distance algorithm to realize fault early warning.
Optionally, the wind turbine component comprises: the wind wheel, the yaw system, the variable pitch system, the hydraulic system, the generator, the main bearing and the SCADA system.
Optionally, simulation comparison is performed by using the neural network before and after optimization.
Optionally, the pre-treatment employs Z-simplification.
Optionally, the obtaining the neural network prediction model includes optimizing the neural network initial threshold and the weight by using a genetic algorithm to establish the neural network prediction model.
Optionally, the obtaining the neural network prediction model includes a loop operation until a training target reaches a set requirement or the number of iterations reaches a set target.
Optionally, the fault pre-warning comprises fault pre-warning a single critical component.
Optionally, the number of neurons in the input layer of the fault early warning is 6, and the output parameter is a fault early warning component.
The invention also proposes an electronic device comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for wind turbine component failure prediction according to the present invention.
The invention further provides a computer-readable storage medium storing a computer program, which is characterized in that when the computer program is executed by a processor, the method for predicting the fault of the wind turbine component provided by the invention is realized.
The invention designs a wind turbine generator fault early warning method by taking SCADA data of a power plant as a basis and combining a GA-BP neural network algorithm and a Euclidean distance algorithm. Compared with the traditional fault early warning scheme, the method solves the contradiction between the fault early warning time and the early warning precision, can adapt to the complex and variable conditions of the operation condition of the wind turbine generator, can effectively win the spare parts and the maintenance plan optimization time, and can also reject the false alarm of the SCADA system and reduce the maintenance amount, thereby reducing the maintenance cost of the wind power plant.
Drawings
One or more embodiments are illustrated by way of example in the accompanying drawings, which correspond to the figures in which like reference numerals refer to similar elements and which are not to scale unless otherwise specified.
Fig. 1 is a flow chart of a failure prediction method according to an embodiment of the invention.
FIG. 2 is a GA-BP neural network prediction model according to an embodiment of the present invention.
FIG. 3 is a schematic flow chart of passing fault warning criteria and real-time SCADA data according to an embodiment of the present invention.
FIG. 4 is a BP neural network performance curve according to an embodiment of the present invention.
FIG. 5 is a graph of GA-BP neural network performance according to one embodiment of the present invention.
FIG. 6 is a comparison of a fit of a predicted output to an expected output for one embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, it will be appreciated by those of ordinary skill in the art that numerous technical details are set forth in order to provide a better understanding of the present application in various embodiments of the present invention. However, the technical solution claimed in the present application can be implemented without these technical details and various changes and modifications based on the following embodiments. The following embodiments are divided for convenience of description, and should not constitute any limitation to the specific implementation manner of the present invention, and the embodiments may be mutually incorporated and referred to without contradiction.
The wind turbine generator comprises the following components and principle functions:
the wind wheel mainly comprises blades and a profile, is designed by adopting an air principle, and is a main part for capturing wind energy by a wind turbine generator. The blades are fixed on a wheel shaft with the profile, the wind wheel transmits torque generated by rotation to a main transmission shaft, and mechanical equipment and electrical equipment for changing the pitch are further installed in the wind wheel.
The direct-drive generator is a generator directly driven by wind power, also called a gearless wind motor, and adopts a mode that a multi-pole motor is directly connected with an impeller to drive, so that the traditional component of a gear box is omitted. The direct-drive wind driven generator has the advantages that: the power generation efficiency is high, the reliability is high, the operation and maintenance cost is low, and the power grid access performance is excellent. The direct-drive wind turbine generator set is not provided with a gear box, the low-speed wind wheel is directly connected with the generator, various harmful impact loads are completely born by the generator system, and the requirement on the generator is high. Meanwhile, in order to improve the power generation efficiency, the number of poles of the generator is very large, usually about 100 poles, the structure of the generator becomes very complex and large, and the hoisting maintenance of the whole generator is required. And the use of permanent magnetic materials and rare earths adds some uncertainty factors.
Pitch systems mainly comprise two main types: hydraulic and electric pitch systems. The hydraulic pitch variation is realized by applying pressure to liquid to drive an actuating mechanism to adjust the pitch variation; the electric variable pitch is realized by adjusting the variable pitch through a servo motor driving gear mechanism. The electric pitch system comprises two parts, namely a pitch drive and a pitch bearing. The main working principle of the electric variable pitch system is that when the wind speed which changes constantly is responded, the servo motor drives the speed reducing device to drive the pinion to rotate, and the pinion is meshed with the inner ring of the variable pitch bearing, so that the variable pitch bearing and the blade group rotate simultaneously, and the purpose of changing the pitch angle of the fan is achieved.
Yaw refers to the process by which the rotor of a wind turbine rotates about an axis with respect to the wind. The yaw system has the function of capturing the wind direction to control the accurate, stable and reliable wind alignment of the engine room, and can obtain wind energy to the maximum extent. Yawing systems can be divided into two main categories according to different yawing modes: passive yaw systems and active yaw systems. The passive yaw system is characterized in that a fan is made to yaw under the action of wind, and three common direction adjusting devices are provided, namely a rudder wheel side wind wheel, a wind direction tracking device and a tail vane direction adjusting device; the active yawing system mainly uses a motor to drag a gear mechanism to enable a wind wheel to always keep a front face facing the wind, so that the maximum wind energy is obtained.
The hydraulic system functions to increase the force by changing the pressure of the fluid, transmitting power to the mechanical unit. The liquid has the advantages of high density, stepless speed regulation, stable and reliable rotation, relative easiness in part replacement and the like, so that the liquid is widely applied to large-scale wind turbines. In a generator set, a hydraulic system is mainly used for a mechanical brake mechanism, a pitch change mechanism, a yaw system drive and yaw brake of a fan. Furthermore, hydraulic systems are used in generator cooling, converter temperature control and cooling of gearbox lubrication oil, all of which play irreplaceable important roles.
The main bearing is used as an important component in a rotating mechanical system, generally is a self-aligning roller bearing, and also adopts a large-cone-angle double-row tapered roller bearing, and most impeller main shafts are supported by two self-aligning roller bearings. Since the main shaft of the impeller is subjected to a very large load, and the shaft is long and easily deformed, the bearing must have good aligning performance.
The basic functions of the SCADA include remote control, real-time monitoring, historical data browsing and the like of a wind power plant, and normal operation and safety of the wind turbine generator are guaranteed. In the operation process of the fan, hundreds of operation parameters need to be monitored by the SCADA system, the operation state of the fan is observed according to the operation data, and meanwhile, the functions of protection, monitoring and the like are also achieved. The real-time tracking parameters are important basis for examining the comprehensive performance of the wind turbine generator and judging whether the wind turbine generator fails or not. If the real-time monitoring data are deeply analyzed, more valuable information can be researched, and the abnormal condition of the fan can be accurately and timely seen, along with the change of the data. The change of the data helps the operator to improve the safe and reliable performance of the fan unit and greatly reduce the cost for maintenance.
The invention provides a wind turbine component fault prediction method based on a neural network and an Euclidean distance algorithm, aiming at improving the operation safety and stability of a wind turbine and the remote monitoring automation level.
Specifically, the method is realized by adopting the following steps:
and acquiring state data of monitored wind turbine components such as a wind wheel, a yaw system, a pitch system, a hydraulic system, a generator, a main bearing, an SCADA system and the like.
And analyzing faults of main components of the fan, and acquiring historical SCADA data and real-time SCADA data of fan operation through an SCADA system.
The main failure modes and reasons of the wind wheel are shown in table 1:
TABLE 1
Figure BDA0002849384300000051
The main failure modes and reasons of the generator are shown in table 2:
TABLE 2
Figure BDA0002849384300000052
Figure BDA0002849384300000061
The failure modes and reasons of the pitch system are shown in table 3:
TABLE 3
Figure BDA0002849384300000062
Figure BDA0002849384300000071
Yaw system failure modes and causes are shown in table 4:
TABLE 4
Figure BDA0002849384300000072
The failure modes and reasons of the hydraulic system are shown in table 5:
TABLE 5
Figure BDA0002849384300000073
Figure BDA0002849384300000081
The main bearing failure modes and causes are shown in table 6:
TABLE 6
Figure BDA0002849384300000082
Preprocessing data:
a simple filtering of the data is performed. The wind turbine generator system fault early warning based on the SCADA data needs to ensure that a partial data set of the SCADA data is normal and complete, but a complete and normal data set is difficult to find out under normal conditions of the SCADA data. In general, data acquired by the SCADA system is discontinuous and inconsistent, and the SCADA data needs to be simply filtered in a first step to filter out lost and abnormal data.
And carrying out standardization processing on the data. The data normalization process is to scale the data according to the size of intervals, wherein the classical methods are Z normalization, 0-1 normalization and Z simplification. The data normalization method adopted by the invention is Z simplification.
0-1 normalization, also known as dispersion normalization, is a linear transformation of the raw data such that the resulting values map between [0, -1 ]. The transfer function is as follows:
x*＝(x-xmin)/(xmax-xmin)……………………(1)
in the formula (1), x*For the processed data result, x is the sample data to be processed, xminIs the minimum value of sample data, xmaxIs the maximum value of the sample data.
Z normalization, the method can carry out data normalization on the mean value and the standard deviation of the sample data. The processed data were in accordance with a standard normal distribution, i.e. the mean was 0 and the standard deviation was 1, the transfer function was as follows:
x*＝(x-xμ)/xδ......................................(2)
in the formula (2), xμIs the mean of all sample data, xбIs the standard deviation of all sample data.
Z simplification indicates that the larger x is*The smaller the number, the more a number can be normalized to [0, 1]]Meanwhile, compared with 0-1 normalization, Z normalization needs to depend on all data of a sample, and Z simplification only depends on current data, so that the method can be used dynamically and is well understood. The simplified transfer function for Z is as follows:
x*＝1/(x+1)...........................................(3)
and filtering abnormal data. In order to reduce data analysis errors, the data needs to be processed to filter out abnormal data points. Wind power sample data xnBy obtaining an average value of sample data
Figure BDA0002849384300000091
And the mean square error sigma, to determine the sample outliers, since wind speed has uncertainty,
Figure BDA0002849384300000092
can be obtained by the following exponential smoothing formula:
Figure BDA0002849384300000093
in the formula (4), the first and second groups,
Figure BDA0002849384300000094
is the average value at time t, xtIs the true value for the time instant t,
Figure BDA0002849384300000095
is the average of the previous time instant.
The normal data judgment formula can be obtained according to the formula (4), and the residual data is judged to be abnormal data for filtering.
Figure BDA0002849384300000096
In formula (5), k is determined by counting small probability events, and abnormal judgment is carried out on data by setting k and a, when x istWhen the formula (5) is satisfied, the current data can be judged to be normal values, otherwise, the current data are filtered for abnormity.
And selecting the processed normal historical data to train the neural network, and obtaining a GA-BP neural network prediction model through testing.
Initializing population P, cross probability, cross scale PcProbability of mutation PmAnd each network layer weight and threshold W1、B1And W2、B2And initializing, and adopting real number coding in a recoding process.
The fitness function is set as the reciprocal of the sum of squares of the errors of the neural network:
Figure BDA0002849384300000101
in equation (6), SE is the sum of the squares of the errors between the predicted output and the desired output via the network and the neural network.
Calculating and ordering each individual evaluation function, and selecting network individuals according to the following probability values:
Figure BDA0002849384300000102
in the formula (7), n is the number of all individuals, fiIs the fitness value of the ith individual.
Crossover and mutation operations: the optimal individuals do not do crossover operations but are directly copied into the next generation. While other individuals use the cross probability PcTwo individuals are operated in a crossover operation, resulting in two additional new individuals. Also, the optimal individuals do not perform cheap operations and are directly copied to the next generation. While other individuals use the mutation probability PmMutation operations were performed to generate two additional new individuals.
And circulating the operation until the training target reaches the set requirement or the iteration number reaches the set target.
And comparing the fault early warning criterion with the real-time SCADA data to realize fault early warning.
The block diagram of the GA-BP neural network and Euclidean distance fault early warning is shown in figure 3.
The invention optimizes the initial threshold and weight of GA-BP neural network by genetic algorithm to establish neural network prediction model, selects 6 characteristic parameters as the input of BP neural network, and the number of neurons in input layer is 6. The output parameters are fault early warning component parameters.
The invention carries out fault early warning aiming at a single key component, so the number of neurons in an output layer is 1, and an implicit layer empirical formula is shown as follows:
Figure BDA0002849384300000103
in equation (8), a is [1,2,3, …,10], m is the input layer neuron number, n is the output layer neuron number, and the number of the obtained optimal hidden layer neurons is 8. The transfer function of the hidden layer is an S-type tangent function tan sig, the transfer function of the output layer is an S-type logarithmic function log sig, the training algorithm is a rainlm algorithm, the performance index is MSE, the training target is 0.0001, the running population scale of the genetic algorithm is 50, the maximum genetic algebra is 100, the variation rate is 0.09, and the cross probability is 0.7.
Step 6: and carrying out simulation comparison by using the neural networks before and after optimization.
The resulting predicted mean square deviation values for the main bearing temperatures are shown in table 7:
TABLE 7
Figure BDA0002849384300000111
According to the simulation comparison result, the fault early warning method can identify the fault signal more accurately and realize fault early warning.
The invention also relates to an electronic device comprising: at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for wind turbine component failure prediction according to the present invention.
Where the memory and processor are connected by a bus, the bus may comprise any number of interconnected buses and bridges, the buses connecting together one or more of the various circuits of the processor and the memory. The bus may also connect various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be one element or a plurality of elements, such as a plurality of receivers and transmitters, providing a means for communicating with various other apparatus over a transmission medium. The data processed by the processor is transmitted over a wireless medium via an antenna, which further receives the data and transmits the data to the processor.
The processor is responsible for managing the bus and general processing and may also provide various functions including timing, peripheral interfaces, voltage regulation, power management, and other control functions. And the memory may be used to store data used by the processor in performing operations.
The invention also provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the wind turbine component fault prediction method provided by the invention.
That is, as can be understood by those skilled in the art, all or part of the steps in the method for implementing the embodiments described above may be implemented by a program instructing related hardware, where the program is stored in a storage medium and includes several instructions to enable a device (which may be a single chip, a chip, or the like) or a processor (processor) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific examples for carrying out the invention, and that various changes in form and details may be made therein without departing from the spirit and scope of the invention in practice.

Claims (10)

1. A wind turbine component fault prediction method is characterized by comprising the following steps:
acquiring operation state data of monitored wind turbine components;
analyzing faults of the wind turbine generator component, and acquiring historical data and real-time data of operation of the wind turbine generator component;
preprocessing the historical data and the real-time data;
training a neural network by utilizing the preprocessed historical data to obtain a neural network prediction model;
and comparing the real-time data with the fault early warning criterion by combining a neural network and an Euclidean distance algorithm to realize fault early warning.
2. The method of predicting a failure of a wind turbine component of claim 1, wherein the wind turbine component comprises: the wind wheel, the yaw system, the variable pitch system, the hydraulic system, the generator, the main bearing and the SCADA system.
3. The method of predicting the failure of the wind turbine component according to claim 1 or 2, further comprising performing simulation comparison by using neural networks before and after optimization.
4. The method of predicting a failure of a wind turbine component of claim 1, wherein the preprocessing employs Z-simplification.
5. The method of claim 1, wherein the obtaining the neural network prediction model comprises optimizing the neural network initial threshold and weight using a genetic algorithm to build the neural network prediction model.
6. The method of claim 1, wherein the obtaining the neural network prediction model comprises a loop operation until a training target meets a set requirement or a number of iterations reaches a set target.
7. The method of claim 1, wherein the fault pre-warning comprises fault pre-warning a single critical component.
8. The method for predicting the failure of the wind turbine generator component according to claim 7, wherein the number of the neurons of the input layer of the failure early warning is 6, and the output parameter is a failure early warning component.
9. An electronic device, comprising:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the wind turbine component failure prediction method of any of claims 1 to 8.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the method for wind turbine component failure prediction according to any one of claims 1 to 8.
CN202011520719.XA 2020-12-21 2020-12-21 Wind turbine generator component fault prediction method Pending CN112733283A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011520719.XA CN112733283A (en) 2020-12-21 2020-12-21 Wind turbine generator component fault prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011520719.XA CN112733283A (en) 2020-12-21 2020-12-21 Wind turbine generator component fault prediction method

Publications (1)

Publication Number Publication Date
CN112733283A true CN112733283A (en) 2021-04-30

Family

ID=75605000

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011520719.XA Pending CN112733283A (en) 2020-12-21 2020-12-21 Wind turbine generator component fault prediction method

Country Status (1)

Country Link
CN (1) CN112733283A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113339208A (en) * 2021-07-21 2021-09-03 申能新能源(青海)有限公司 Method for selecting abnormal vibration fragments of wind turbine generator
CN113468728A (en) * 2021-06-11 2021-10-01 国电南京自动化股份有限公司 Variable pitch system fault prediction method based on neural network
CN114215706A (en) * 2021-12-27 2022-03-22 南京邮电大学 Wind turbine generator blade cracking fault early warning method and device
CN114320773A (en) * 2021-12-22 2022-04-12 中国大唐集团新能源科学技术研究院有限公司 Wind turbine generator fault early warning method based on power curve analysis and neural network
CN115293057A (en) * 2022-10-10 2022-11-04 深圳先进技术研究院 Wind driven generator fault prediction method based on multi-source heterogeneous data
CN115596620A (en) * 2022-12-15 2023-01-13 深圳鹏锐信息技术股份有限公司（Cn） Wind generating set fault intelligent analysis method and system based on artificial intelligence
CN117171590A (en) * 2023-11-02 2023-12-05 默拓(江苏)电气驱动技术有限公司 Intelligent driving optimization method and system for motor

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110410282A (en) * 2019-07-24 2019-11-05 河北工业大学 Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM
CN111581597A (en) * 2020-03-17 2020-08-25 华电电力科学研究院有限公司 Wind turbine generator gearbox bearing temperature state monitoring method based on self-organizing kernel regression model

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110410282A (en) * 2019-07-24 2019-11-05 河北工业大学 Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM
CN111581597A (en) * 2020-03-17 2020-08-25 华电电力科学研究院有限公司 Wind turbine generator gearbox bearing temperature state monitoring method based on self-organizing kernel regression model

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
杨晓: "风力发电机组传动系统运行状态评价方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》, no. 03, pages 042 - 221 *
梁坤鑫: "数据驱动的风电机组部件故障预警", 中国优秀硕士学位论文全文数据库工程科技Ⅱ辑, no. 02, pages 042 - 508 *

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113468728A (en) * 2021-06-11 2021-10-01 国电南京自动化股份有限公司 Variable pitch system fault prediction method based on neural network
CN113468728B (en) * 2021-06-11 2024-02-02 国电南京自动化股份有限公司 Variable pitch system fault prediction method based on neural network
CN113339208B (en) * 2021-07-21 2023-08-22 申能新能源(青海)有限公司 Method for selecting abnormal vibration segments of wind turbine generator
CN113339208A (en) * 2021-07-21 2021-09-03 申能新能源(青海)有限公司 Method for selecting abnormal vibration fragments of wind turbine generator
CN114320773B (en) * 2021-12-22 2023-09-22 大唐可再生能源试验研究院有限公司 Wind turbine generator system fault early warning method based on power curve analysis and neural network
CN114320773A (en) * 2021-12-22 2022-04-12 中国大唐集团新能源科学技术研究院有限公司 Wind turbine generator fault early warning method based on power curve analysis and neural network
CN114215706A (en) * 2021-12-27 2022-03-22 南京邮电大学 Wind turbine generator blade cracking fault early warning method and device
CN114215706B (en) * 2021-12-27 2024-02-20 南京邮电大学 Early warning method and device for cracking faults of wind turbine generator blades
CN115293057B (en) * 2022-10-10 2022-12-20 深圳先进技术研究院 Wind driven generator fault prediction method based on multi-source heterogeneous data
CN115293057A (en) * 2022-10-10 2022-11-04 深圳先进技术研究院 Wind driven generator fault prediction method based on multi-source heterogeneous data
CN115596620A (en) * 2022-12-15 2023-01-13 深圳鹏锐信息技术股份有限公司（Cn） Wind generating set fault intelligent analysis method and system based on artificial intelligence
CN115596620B (en) * 2022-12-15 2023-03-07 深圳鹏锐信息技术股份有限公司 Wind generating set fault intelligent analysis method and system based on artificial intelligence
CN117171590A (en) * 2023-11-02 2023-12-05 默拓(江苏)电气驱动技术有限公司 Intelligent driving optimization method and system for motor
CN117171590B (en) * 2023-11-02 2024-01-26 默拓(江苏)电气驱动技术有限公司 Intelligent driving optimization method and system for motor

Similar Documents

Publication Publication Date Title
CN112733283A (en) Wind turbine generator component fault prediction method
CN110410282B (en) SOM-MQE and SFCM (Small form-factor pluggable) based wind turbine generator health state online monitoring and fault diagnosis method
Wang et al. SCADA data based condition monitoring of wind turbines
CN102022264B (en) System and method for wind turbine health management
CN106870298B (en) Blade root bolt fracture fault detection method based on machine learning
CN110926809B (en) Big data analysis-based wind turbine generator transmission chain fault early warning method
CN107291991B (en) Early defect early warning method for wind turbine generator based on dynamic network sign
US20110020122A1 (en) Integrated condition based maintenance system for wind turbines
Zhan et al. Abnormal vibration detection of wind turbine based on temporal convolution network and multivariate coefficient of variation
CN115013261B (en) State monitoring method and system for offshore wind farm
CN111415070A (en) Wind turbine generator gearbox oil temperature over-temperature fault early warning method based on SCADA data
CN109583075B (en) Permanent magnet direct-drive wind turbine service quality evaluation method based on temperature parameter prediction
CN116484751B (en) Fatigue life assessment method and device for wind turbine generator components
CN103925155A (en) Self-adaptive detection method for abnormal wind turbine output power
CN116067432A (en) Escalator variable working condition fault diagnosis method
CN116771610A (en) Method for adjusting fault evaluation value of variable pitch system of wind turbine
CN115456041A (en) Equipment fault early warning method and device, computing equipment and storage medium
CN114320773B (en) Wind turbine generator system fault early warning method based on power curve analysis and neural network
Lu et al. Gear Fault Diagnosis and Life Prediction of Petroleum Drilling Equipment Based on SOM Neural Network
Zhou et al. A fatigue life prediction method for the drive system of wind turbine using internet of things
Zhang Comparison of data-driven and model-based methodologies of wind turbine fault detection with SCADA data
CN114352483A (en) Fatigue loading in wind turbines and use of operational metadata
CN114295367A (en) Wind turbine generator gearbox working condition online monitoring method
Yang et al. Intelligent fan system based on big data and artificial intelligence
Ou et al. Fault Prediction Model of Wind Power Pitch System Based on BP Neural Network

Legal Events

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