CN113339208B - Method for selecting abnormal vibration segments of wind turbine generator - Google Patents
Method for selecting abnormal vibration segments of wind turbine generator Download PDFInfo
- Publication number
- CN113339208B CN113339208B CN202110824280.8A CN202110824280A CN113339208B CN 113339208 B CN113339208 B CN 113339208B CN 202110824280 A CN202110824280 A CN 202110824280A CN 113339208 B CN113339208 B CN 113339208B
- Authority
- CN
- China
- Prior art keywords
- vibration
- sequence
- vibrator
- point
- overrun
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- 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
-
- 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
- F03D80/00—Details, components or accessories not provided for in groups F03D1/00 - F03D17/00
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Sustainable Energy (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to a method for selecting a vibration abnormal segment of a wind turbine, which comprises the following steps: loading a vibration time sequence, and finding out a vibration overrun point with vibration larger than a threshold value in the vibration time sequence; intercepting a plurality of vibrator sequences for each vibration overrun point, generating a plurality of fragments for each vibrator sequence through variable point detection, wherein the final vibrator sequence of the vibration overrun point is the most suitable vibrator sequence; a. the number of fragments is closest to 3; b. the statistics of the segments where the vibration overrun points are located in the vibrator sequence are larger than those of other segments; c. the segment of the vibrator sequence where the vibration overrun point is located at the middle section. The method can automatically screen vibrator sequences (local abnormal data) from vibration time sequences with various resolutions, and provides a data basis for subsequent detailed analysis.
Description
Technical Field
The invention belongs to the technical field of wind power generation, and particularly relates to a method for selecting a vibration abnormal segment of a wind turbine.
Background
In the running process of the wind turbine generator, sudden vibration abnormality possibly occurs due to various reasons, so that vibration signals of equipment are required to be analyzed, abnormal fragments are extracted, fault reasons are further analyzed, preventive maintenance is realized, the sudden fault rate is reduced, and the equipment is enabled to run safely and reliably.
The current analysis method for the vibration abnormality of the wind turbine generator comprises the following steps: for example, spectrum analysis is carried out based on modal test and fast Fourier transform, and the reasons of equipment faults are found through vibration test comparison analysis between a normal wind field unit and an abnormal wind field unit and finite element analysis; analyzing the time domain and frequency domain data, performing time-frequency comparison research, and combining a cepstrum, an envelope spectrum and a wavelet transformation analysis method; and acquiring the rotating speed and vibration data of the fan, and judging whether vibration abnormality occurs or not according to the number of extreme acceleration in each rotating speed interval, the corresponding vibration amplitude and the aggregation degree after conversion treatment. The method for detecting the time sequence mutation points comprises the following steps: and performing variable point detection on the ARMA model and the ARCH model by a Bayesian method. Measuring the change point detection effect through an ROC curve; the regression prediction method is adopted, the predicted value is used for replacing the abnormal value, the predicted deviation is reduced as much as possible, and the detection accuracy is improved; respectively constructing quantitative indexes aiming at four performance degradation characteristic evaluation criteria, reconstructing a linear regression equation of a state variable and a time sequence, solving a coefficient of the linear regression equation, and extracting turning points of the time sequence through F test; also STL decomposition, classification and regression trees, ARIMA, exponential smoothing, neural networks, and the like.
Because the unit is operated under variable working conditions for a long time, the vibration condition of the unit is closely related to the design of the unit and the external wind resource condition.
Typically, a data preservation process exists prior to vibration analysis. The fan control system will save the data of the time period before and after the fault, including the vibration data. For fans with CMS system vibration sensors, the vibration data is also corresponding high frequency, medium frequency and low frequency data stored at intervals of a time period. If a fault occurs, the system automatically saves the data segments of a plurality of ms before and after the fault. In practical situations, no matter the spectrum analysis of vibration is carried out, or other indexes (such as power, rotation speed, wind direction and the like) of the fan are analyzed based on the vibration data in a combined mode, local data with larger vibration needs to be found for the vibration data which are acquired and stored. The work (searching local data) often depends on personal experience of vibration analysis engineers to select proper data segments for analysis, and no proper automatic and intelligent method has been available to directly give proper vibrator sequences (local abnormal data). Particularly, when vibration data are continuously collected, how to accurately find out the vibration data (vibrator sequences) which are interested by engineers before and after the vibration overrun depends more on the engineering experience of the engineers.
Disclosure of Invention
The invention aims to provide a method for selecting a vibration abnormal segment of a wind turbine, which can automatically screen vibrator sequences (local abnormal data) from vibration time sequences with various resolutions and provide a data basis for subsequent detailed analysis.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a method for selecting a vibration abnormal segment of a wind turbine generator is characterized by comprising the following steps of:
loading a vibration time sequence, and finding out a vibration overrun point with vibration larger than a threshold value in the vibration time sequence;
traversing all the vibration overrun points, intercepting a plurality of vibrator sequences for each vibration overrun point, detecting each vibrator sequence through a variable point to generate a plurality of fragments, wherein the final vibrator sequence of the vibration overrun point is the most suitable for the following conditions in the plurality of vibrator sequences;
a. the number of fragments is closest to 3;
b. the statistics of the segments where the vibration overrun points are located in the vibrator sequence are larger than those of other segments;
c. the segment of the vibrator sequence where the vibration overrun point is located at the middle section.
Further, the vibrator sequence is a segment that generates mean or variance statistic based by variable point detection.
Further, based on the conditions a, b, c, an evaluation function is defined for any vibration overrun point i
f(i,l,r,acc)
Where i is the position where the vibration overrun point is located, l is the position amount where the starting point of the vibrator sequence is reduced relative to i, r is the position amount where the end point of the vibrator sequence is increased relative to i, and acc is the vibration time sequence;
after the conditions a, b and c are quantized, an evaluation function is combined to obtain a function value f based on the position sequence from i-l to i+r, and the return value of the evaluation function is configured to be smaller and more optimal, and the quantized comprehensive value of the conditions a, b and c is taken as the return value.
Further, the quantization of the conditions a, b, c includes:
condition a quantification: the number of fragments after the change point detection is recorded as n, and the fragments are quantized into (n-3) 2;
condition b quantification: if the fragment statistics after the detection and segmentation of the variable point at the moment i is the maximum value of 0, otherwise, the fragment statistics is 1;
condition c quantification: whether the moment i is at the middle position or not, the middle position is marked as 0, otherwise, the moment i is marked as 1;
quantizing and synthesizing the conditions a, b and c into a formula a+m (b+c) by adopting a penalty function outlier method, and taking the formula a+m (b+c) as a return value of f (i, l, r, acc), wherein m=1;
and (3) solving the minimum value min [ f (i, l, r, acc) ] of the evaluation function through an optimization algorithm, and combining the i index position and the r index position in the optimization result with the i index position and the acc index position after optimization to obtain an optimized vibrator sequence, namely a final vibrator sequence.
Further, the optimization algorithm adopts a genetic algorithm based on integer optimization to calculate the minimum value min [ f (i, l, r, acc) ] of the evaluation function, and the genetic algorithm comprises the steps of: and (3) randomly generating an integer between the initial parameters of l and r and Smin-Smax, wherein the total population is P, and the total iteration number is t.
Further, the crossing method in the genetic algorithm is as follows: internally, l and r are interchanged.
Further, the lower the fitness of the individual in the crossing method is, the higher the probability of crossing;
the probability of each individual crossing is calculated by adopting a standard normal distribution probability density function D, the rank order of the fitness of the individuals needing crossing is calculated and is marked as mu, the crossing probability of the individuals is D (mu),
further, the mutation method in the genetic algorithm comprises the following steps: trigger variation is selected from individuals who have crossed.
Further, the mutation adopts real-value mutation, wherein the mutation amplitude is reduced as the population algebra increases;
the variation amplitude is v+ [ Smax- (1-q) Smin ]. D #, wherein:
q is the ratio of individuals triggering variation in individuals who have crossed,
v is the actual value of the individual,
d is the mutation direction, D of single mutation is a random number of-1 or 1,
e is the coefficient of variation, and the current population coefficient is denoted as g, e= (t-g)/t.
Further, if l or r after mutation exceeds Smax or falls below Smin, it is set as new Smax or Smin.
Compared with the prior art, the invention has the following beneficial effects: the method can automatically screen vibrator sequences (local abnormal data) from vibration time sequences with various resolutions, provides a data base for subsequent detailed analysis (vibration abnormal analysis of the wind turbine generator), does not need manual participation, improves accuracy and reduces workload.
Drawings
Fig. 1 is a general flowchart of a method for selecting abnormal vibration segments of a wind turbine generator system according to an embodiment.
FIG. 2 is an overall flow chart of vibration overrun sub-sequence selection in an embodiment.
Fig. 3 is a schematic diagram of acceleration corresponding to a vibration time sequence of a certain period in the embodiment.
FIG. 4 is a schematic view of a portion of a segment selected from the acceleration effective values of FIG. 3.
Fig. 5 is a schematic diagram of acceleration corresponding to a vibration time series for a certain period of time in example 1.
FIG. 6 is a schematic view of a portion of a segment selected from the acceleration effective values of FIG. 5.
Fig. 7 is a schematic diagram of a segment of the vibrator sequence for a certain period of time in example 1.
Fig. 8 is a schematic diagram II of a section of the vibrator sequence for a certain period of time in example 1.
Fig. 9 is a schematic diagram III of a segment of the vibrator sequence for a certain period of time in example 1.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below in connection with specific embodiments. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 3, an acceleration diagram corresponding to a vibration time sequence in a certain time period includes X, Y directions of acceleration and an effective acceleration value, 100 data points before and after a vibration overrun point (effective acceleration value >0.12 g) are selected, and a vertical line is a vibration overrun point mark. By observing the vibration time sequence, the vibration time sequence is found to have a process of changing from low vibration to high vibration suddenly, and after a period of time, the vibration overruns and then the vibration is reduced again. Through analysis of a plurality of vibration time sequences, the vibrator sequence capable of finding local abnormality comprises square frame parts in the figure, and is divided into three parts of low, overrun and low, namely the main part of engineers needing to carry out subsequent analysis. The vibrator sequence corresponding to the overrun point comprises a segment which is positioned in the middle and comprises the overrun point and segments with lower vibration amplitudes which are positioned at two sides. Based on the found rule, the method screens out vibrator sequences with local anomalies of low vibration, overrun and low vibration from the vibration time sequences with various resolutions, and provides a data basis for subsequent detailed analysis.
As shown in FIG. 1, a method for selecting a vibration anomaly segment of a wind turbine generator includes:
s100, loading a vibration time sequence acc;
s200, finding out a vibration overrun point with vibration larger than a threshold value (generally 0.12-0.14 g) in a vibration time sequence;
s300, traversing all the vibration overrun points, and finding out a vibrator sequence corresponding to each vibration overrun point, as shown in fig. 2.
As shown in fig. 4, the vibrator sequence to be intercepted should theoretically satisfy the following condition for each vibration overrun point by the above analysis:
the vibration value Yi of the vibrator sequence at the moment i exceeds the set threshold value (generally 0.12-0.14 g), the Yi is positioned in the corresponding segment of the vibrator sequence, the vibrator sequence is marked as Yi, a plurality of segments based on mean value or variance statistics can be generated by using moving point detection based on the Yi sequence, the mean value statistics is adopted for the effective acceleration value, and the variance statistics is adopted for the XY acceleration.
a. For an ideal vibrator sequence, the number of fragments cut out from the moving point detection result is 3.
In the moving point detection result of the Yi sequence, the segment at the moment i is s-e moment, and the vibration is maximum compared with other segments, for example, the statistic (mean value) of the yi in the result of fig. 4 is larger than that of the other segments.
And c, in the moving point detection result of the Yi sequence, the segment at the moment i is positioned in the middle segment to be optimal.
S310, based on the analysis conditions a, b and c, defining an evaluation function f (i, l, r and acc) for any vibration overrun point i, wherein i is the position where the vibration overrun point is located (as a fixed parameter), l is the position where the starting point of the vibrator sequence is reduced relative to i (as a variable parameter), r is the position where the end point of the vibrator sequence is increased relative to i (as a variable parameter), and acc is the vibration time sequence (as a fixed parameter).
After the conditions a, b and c are quantized, an evaluation function is combined to obtain a function value f based on the i-l to i+r position sequence, the return value of the evaluation function is configured to be smaller and more optimal, and the quantized comprehensive values of the conditions a, b and c are used as optimization targets.
The quantification of the conditions a, b, c includes:
condition a quantification: the number of fragments after the change point detection is recorded as n, and then the fragments are quantized into (n-3) 2 Denoted by a.
Condition b quantification: if the fragment statistics after the detection and segmentation of the variable point at the moment i is the maximum value of 0, otherwise, the fragment statistics are marked as 1, and b is used for representing the fragment statistics.
Condition c quantification: and (3) judging whether the moment i is at the middle position, wherein the middle position is marked as 0, otherwise, the moment i is marked as 1, and the moment i is marked as c.
And (3) quantizing and synthesizing the conditions a, b and c into a formula a+m (b+c) by adopting a penalty function outlier method, and taking the formula a, b and c as a return value of f (i, l, r and acc) to be an optimization problem, namely obtaining the minimum value of the functions, wherein m=1, a= (n-3) 2, b=0 or 1, and c=0 or 1.
S320, calculating a minimum value min [ f (i, l, r, acc) ] of an evaluation function through a genetic algorithm based on integer optimization, combining the l index position and the r index position in an optimization result with i and acc after optimization, and obtaining an optimized vibrator sequence, namely a final vibrator sequence (storing the vibrator sequence to a designated position and providing input for subsequent post-processing).
The specific optimization process of the genetic algorithm comprises the following steps: generating an initial population, calculating fitness, selecting, crossing, mutating, and ending the iteration. The method specifically comprises the following steps:
s321, generating an initial population: the initial parameters of l and r are randomly generated into integers from Smin (default 3) to Smax (default 200), the total population is P (default 200), and the total iteration number is t (default 200).
S322, calculating fitness: the fitness of each individual is calculated by bringing the individual into the objective function f, wherein the smaller the value of a+m (b+c) the greater the probability that individual is selected.
S323, selecting: and selecting according to the probability of selecting the individuals, and eliminating the individuals with low fitness.
S324, crossing: the crossover method is the interchange of l and r inside the individual. The crossover rate was 25% of the total population P. The lower the fitness of the individual in the crossing method, the higher the probability of crossing. The probability of each individual crossing is calculated by adopting a standard normal distribution probability density function D, the rank order of the fitness (f function value) of the individual needing crossing is calculated and is recorded as mu, the crossing probability of the individual is D (mu),
s325, mutation: the mutation is triggered by selecting 33% of individuals from the individuals with crossover, and the mutation adopts real-value mutation, wherein the mutation amplitude is reduced with the increase of population algebra.
The variation amplitude is v+ [ Smax-0.67Smin ]. D, wherein: q is the ratio of individuals triggering mutation in individuals who cross, v is the actual value of the individuals (on l or r), D is the mutation direction, D of single mutation is a random number of-1 or 1, e is the mutation coefficient, and the current population coefficient is recorded as g, e= (t-g)/t. If l or r after mutation exceeds Smax or is lower than Smin, then it is set as new Smax or Smin.
S324, iteration: the fitness of new individuals in the population (i.e., individuals who have undergone mutation and crossover) is re-evaluated using the f-function. The iterative range and the initializing range of the optimization algorithm for l and r are the same.
And (3) obtaining the minimum value min [ f (i, l, r, acc) ] of the evaluation function through the optimization, combining the i index position and the r index position in the optimization result with the i index position and the acc index position after the optimization, obtaining an optimized vibrator sequence, namely a final vibrator sequence, and storing the subsequence to a designated position to provide input for subsequent post-processing.
Example 1
The practical operation wind farm is adopted in the example, the installed capacity of the wind farm is 5 ten thousand kW, and the data recording resolution is 5s. In the middle 3 th year of 2020, the wind farm 10# unit frequently generates fault alarm of vibration overrun, and the unit is taken as an example to select a vibrator sequence.
Fig. 5 is a vibration time series chart of about 13min before and after the overrun of the vibration for a certain time, in which the vertical line portion is the time of the overrun point of the vibration.
After the abnormal segment selection by the method, the vibrator sequence result of fig. 6 is obtained. The result shows that the optimization function can accurately select the result of meeting the variable point detection of low, high and low, the point corresponding to the vertical line in the middle section is the position of the vibration overrun point, and the vibration mean value (refer to the horizontal line in the figure) of the section is larger than that of the rest 2 sections.
In the period from 16 days in 3 months in 2020 to 18 days in 3 months in 2020, 4 times of vibration overrun occur in total, and the vibrator sequence of the vibration overrun point can be accurately selected by using the method, so that engineers can analyze deep-level reasons of the vibration overrun. Fig. 7, 8, 9 show the selection result of the remaining 3 vibration anomalies in this period.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (8)
1. A method for selecting a vibration abnormal segment of a wind turbine generator is characterized by comprising the following steps of:
loading a vibration time sequence, and finding out a vibration overrun point with vibration larger than a threshold value in the vibration time sequence;
traversing all the vibration overrun points, intercepting a plurality of vibrator sequences for each vibration overrun point, detecting each vibrator sequence through a variable point to generate a plurality of fragments, wherein the final vibrator sequence of the vibration overrun point is the most suitable for the following conditions in the plurality of vibrator sequences;
a. the number of fragments is closest to 3;
b. the statistics of the segments where the vibration overrun points are located in the vibrator sequence are larger than those of other segments;
c. the segment where the vibration overrun point is located in the vibrator sequence is located in the middle section;
the selecting of the final vibrator sequence includes: based on the conditions a, b and c, defining an evaluation function f (i, l, r and acc) for any vibration overrun point i, wherein i is the position of the vibration overrun point, l is the position quantity of the starting point of the vibrator sequence reduced relative to i, r is the position quantity of the end point of the vibrator sequence increased relative to i, and acc is the vibration time sequence; after the conditions a, b and c are quantized, combining an evaluation function to obtain a function value f based on the position sequence from i-l to i+r, and constructing the return value of the evaluation function into a smaller and more optimal form, wherein the quantized comprehensive values of the conditions a, b and c are taken as the return value;
the quantification of the conditions a, b, c includes:
condition a quantification: the number of fragments after the change point detection is recorded as n, and the fragments are quantized into (n-3) 2;
condition b quantification: if the fragment statistics after the detection and segmentation of the variable point at the moment i is the maximum value of 0, otherwise, the fragment statistics is 1;
condition c quantification: whether the moment i is at the middle position or not, the middle position is marked as 0, otherwise, the moment i is marked as 1;
quantizing and synthesizing the conditions a, b and c into a formula a+m (b+c) by adopting a penalty function outlier method, and taking the formula a+m (b+c) as a return value of f (i, l, r, acc), wherein m=1;
and (3) solving the minimum value min [ f (i, l, r, acc) ] of the evaluation function through an optimization algorithm, and combining the i index position and the r index position in the optimization result with the i index position and the acc index position after optimization to obtain an optimized vibrator sequence, namely a final vibrator sequence.
2. The method according to claim 1, characterized in that: the vibrator sequence is a segment that generates mean or variance statistic based by variable point detection.
3. The method according to claim 1, characterized in that: the optimization algorithm adopts a genetic algorithm based on integer optimization to calculate the minimum value min [ f (i, l, r, acc) ] of an evaluation function, and the genetic algorithm generates an initial population comprising: and (3) randomly generating an integer between the initial parameters of l and r and Smin-Smax, wherein the total population is P, and the total iteration number is t.
4. A method according to claim 3, characterized in that: the crossing method in the genetic algorithm comprises the following steps: internally, l and r are interchanged.
5. The method according to claim 4, wherein: the lower the fitness of the individual in the crossing method is, the higher the probability of crossing is;
the probability of each individual crossing is calculated by adopting a standard normal distribution probability density function D, the rank order of the fitness of the individuals needing crossing is calculated and is marked as mu, the crossing probability of the individuals is D (mu),
6. a method according to claim 3, characterized in that: the mutation method in the genetic algorithm comprises the following steps: trigger variation is selected from individuals who have crossed.
7. The method according to claim 6, wherein: the mutation adopts real-value mutation, wherein the mutation amplitude is reduced along with the increase of population algebra;
the variation amplitude is v+ [ Smax- (1-q) Smin ]. D #, wherein:
q is the ratio of individuals triggering variation in individuals who have crossed,
v is the actual value of the individual,
d is the mutation direction, D of single mutation is a random number of-1 or 1,
e is the coefficient of variation, and the current population coefficient is denoted as g, e= (t-g)/t.
8. The method according to claim 7, wherein: if l or r after mutation exceeds Smax or is lower than Smin, then it is set as new Smax or Smin.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110824280.8A CN113339208B (en) | 2021-07-21 | 2021-07-21 | Method for selecting abnormal vibration segments of wind turbine generator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110824280.8A CN113339208B (en) | 2021-07-21 | 2021-07-21 | Method for selecting abnormal vibration segments of wind turbine generator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113339208A CN113339208A (en) | 2021-09-03 |
| CN113339208B true CN113339208B (en) | 2023-08-22 |
Family
ID=77480263
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110824280.8A Active CN113339208B (en) | 2021-07-21 | 2021-07-21 | Method for selecting abnormal vibration segments of wind turbine generator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113339208B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117889945B (en) * | 2024-03-14 | 2024-05-31 | 河南鹏飞建设工程有限公司 | Highway bridge construction vibration testing method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005215881A (en) * | 2004-01-28 | 2005-08-11 | Fujikura Rubber Ltd | Vibration removing system, tuning method therefor, and computer-readable recording medium with program recorded thereon for executing tuning method |
| CN105606360A (en) * | 2015-11-24 | 2016-05-25 | 国网内蒙古东部电力有限公司电力科学研究院 | Fault diagnosis method for condition-variable planetary gear box based on multi-sensor information fusion |
| CN110685868A (en) * | 2019-10-25 | 2020-01-14 | 长沙理工大学 | Wind turbine generator fault detection method and device based on improved gradient elevator |
| CN110905732A (en) * | 2018-09-17 | 2020-03-24 | 中车株洲电力机车研究所有限公司 | Method and system for identifying unbalance of wind wheel of wind turbine generator and storage medium |
| CN112733283A (en) * | 2020-12-21 | 2021-04-30 | 北京华能新锐控制技术有限公司 | Wind turbine generator component fault prediction method |
-
2021
- 2021-07-21 CN CN202110824280.8A patent/CN113339208B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005215881A (en) * | 2004-01-28 | 2005-08-11 | Fujikura Rubber Ltd | Vibration removing system, tuning method therefor, and computer-readable recording medium with program recorded thereon for executing tuning method |
| CN105606360A (en) * | 2015-11-24 | 2016-05-25 | 国网内蒙古东部电力有限公司电力科学研究院 | Fault diagnosis method for condition-variable planetary gear box based on multi-sensor information fusion |
| CN110905732A (en) * | 2018-09-17 | 2020-03-24 | 中车株洲电力机车研究所有限公司 | Method and system for identifying unbalance of wind wheel of wind turbine generator and storage medium |
| CN110685868A (en) * | 2019-10-25 | 2020-01-14 | 长沙理工大学 | Wind turbine generator fault detection method and device based on improved gradient elevator |
| CN112733283A (en) * | 2020-12-21 | 2021-04-30 | 北京华能新锐控制技术有限公司 | Wind turbine generator component fault prediction method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113339208A (en) | 2021-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | Short-term wind speed forecasting based on long short-term memory and improved BP neural network | |
| CN109726503B (en) | Missing data filling method and device | |
| CN113420509A (en) | Wind turbine state evaluation method and device and storage medium | |
| CN110685868A (en) | Wind turbine generator fault detection method and device based on improved gradient elevator | |
| CN107292446B (en) | Hybrid wind speed prediction method based on component relevance wavelet decomposition | |
| CN111881594B (en) | Non-stationary signal state monitoring method and system for nuclear power equipment | |
| CN117458955A (en) | Operation control method and system of motor | |
| CN110533314A (en) | A kind of wind power plant exception unit recognition methods based on probability density distribution | |
| CN113339208B (en) | Method for selecting abnormal vibration segments of wind turbine generator | |
| Shi et al. | Study of wind turbine fault diagnosis and early warning based on SCADA data | |
| CN114429238A (en) | Wind turbine generator fault early warning method based on space-time feature extraction | |
| CN117410961A (en) | Wind power prediction method, device, equipment and storage medium | |
| CN116956120A (en) | Prediction method for water quality non-stationary time sequence based on improved TFT model | |
| CN114117887B (en) | Real-time evaluation method, system and medium for online power generation performance of wind turbine generator | |
| CN110766215B (en) | Wind power climbing event prediction method based on feature adaptive selection and WDNN | |
| CN113468473B (en) | Real-time evaluation method and system for running state of outdoor fixed large-scale mechanical equipment | |
| CN114548498A (en) | Wind speed prediction method and system for local area of overhead transmission line | |
| CN114298300A (en) | Uncertainty prediction method, uncertainty prediction device, electronic device, and storage medium | |
| CN117494573B (en) | Wind speed prediction method and system and electronic equipment | |
| Mu et al. | The NAO variability prediction and forecasting with multiple time scales driven by ENSO using machine learning approaches | |
| CN117310479A (en) | Motor fault prediction method and system based on time-frequency characteristic signal distribution optimization | |
| CN116522993A (en) | Chemical process fault detection method based on countermeasure self-coding network | |
| CN116070669A (en) | Workshop energy consumption prediction method and management system based on improved deep belief network | |
| CN115795999A (en) | Performance abnormity early warning method for long-term service pumped storage unit | |
| CN115901259A (en) | Rolling bearing weak fault diagnosis method based on two-dimensional image and CNN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |