CN112032003B - Method for monitoring operation performance of large wind turbine generator - Google Patents

Method for monitoring operation performance of large wind turbine generator Download PDF

Info

Publication number
CN112032003B
CN112032003B CN202010903361.2A CN202010903361A CN112032003B CN 112032003 B CN112032003 B CN 112032003B CN 202010903361 A CN202010903361 A CN 202010903361A CN 112032003 B CN112032003 B CN 112032003B
Authority
CN
China
Prior art keywords
data
wind speed
low
eps
speed section
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
Application number
CN202010903361.2A
Other languages
Chinese (zh)
Other versions
CN112032003A (en
Inventor
王欣
刘伟江
周民强
潘东浩
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.)
Zhejiang Windey Co Ltd
Original Assignee
Zhejiang Windey 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 Zhejiang Windey Co Ltd filed Critical Zhejiang Windey Co Ltd
Priority to CN202010903361.2A priority Critical patent/CN112032003B/en
Publication of CN112032003A publication Critical patent/CN112032003A/en
Application granted granted Critical
Publication of CN112032003B publication Critical patent/CN112032003B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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

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)
  • Wind Motors (AREA)
  • Control Of Eletrric Generators (AREA)

Abstract

The invention discloses a method for monitoring the operation performance of a large-scale wind turbine generator, which comprises the following steps: running data acquisition and data preprocessing; dividing a wind speed interval; dividing an operation area; and (5) judging the performance abnormity. According to the technical scheme, the operation area of the data scatter diagram shape of the wind speed interval in which the low-efficiency data points are easy to accumulate is divided, the whole operation space is divided into a normal operation area and a low-efficiency operation area, and the abnormal operation performance state of the unit is judged based on two quantitative indexes of the low-efficiency area and the low-efficiency data ratio.

Description

Method for monitoring operation performance of large wind turbine generator
Technical Field
The invention relates to the technical field of wind power generation, in particular to a method for monitoring the operation performance of a large-scale wind turbine generator.
Background
Data show that a power curve is generally used as a main basis for evaluating the output performance of a unit in the wind power field, but a power difference curve obtained by a method of averaging by dividing data bin according to wind speed in the IEC61400-12-1 standard only represents the average output level of the unit in an evaluation period, and due to the fact that the data occupation ratio is relatively small in some abnormal working conditions, the abnormal working conditions are generally neutralized by a large amount of normal data in the power curve fitting process and even are forcibly eliminated in the data preprocessing stage, and running information is lost. Therefore, the traditional power curve monitoring method cannot effectively identify the abnormal or inefficient working condition of the unit in the actual operation process.
According to the feedback of a wind power engineering site, due to the reasons of the reduction of the output performance of a unit, the improper setting of control parameters, the instability of wind conditions, the change of air density and the like, the 'wind speed-active power' scatter diagram is easy to generate the accumulation of low-efficiency data points near the starting wind speed and the near rated wind speed. In order to monitor the abnormal output state of the unit more accurately, it is necessary to perform special monitoring on abnormal data points in the original operation data.
Chinese patent document CN110005580B discloses a "wind turbine generator running state monitoring method". The method comprises the following steps: acquiring sensor data of historical operation of a wind turbine generator, wherein the sensor data comprises a wind speed signal, a rotating speed signal, a variable pitch angle signal, a power signal, a strain signal and a vibration signal; respectively constructing a wind speed, rotating speed, variable pitch angle and power matching monitoring model and a rotating speed, power, strain and vibration frequency spectrum monitoring model according to the wind speed signal, the rotating speed signal, the variable pitch angle signal, the power signal, the strain signal and the vibration signal; and acquiring real-time operation data of the wind turbine generator, and comparing the real-time operation data of the wind turbine generator with the wind speed, the rotating speed, the variable pitch angle, the power matching monitoring model, the rotating speed, the power, the strain and the vibration spectrum monitoring model respectively to determine the current operation state of the wind turbine generator. The power difference curve obtained by the method for averaging in the technical scheme only represents the average output level of the unit in the evaluation period, and due to the fact that the data occupation ratio of some abnormal working conditions is relatively small, in the power curve fitting process, the abnormal working conditions are usually neutralized by a large amount of normal data and even are forcibly eliminated in the data preprocessing stage, and running information is lost. Therefore, the traditional power curve monitoring method cannot effectively identify the abnormal or inefficient working condition of the unit in the actual operation process.
Disclosure of Invention
The invention mainly solves the technical problem that the original power curve monitoring method cannot effectively identify the abnormal or inefficient working condition of the unit in the actual operation process, and provides the large-scale wind turbine unit operation performance monitoring method.
The technical problem of the invention is mainly solved by the following technical scheme: the invention comprises the following steps:
(1) running data acquisition and data preprocessing; and deep cleaning is carried out on the unit operation data by adopting a clustering method on the basis of conventional data screening.
(2) Dividing a wind speed interval; and dividing the wind speed interval of the unit into a low wind speed section and a high wind speed section according to the wind speed, and respectively monitoring the performance of the two wind speed intervals.
(3) Dividing an operation area; and dividing the operation area based on the shape of the data scatter diagram, and dividing the whole operation space into a normal operation area and an inefficient operation area.
(4) And (5) judging the performance abnormity. And judging the abnormal state of the unit operation performance based on the area size of the low-efficiency area and the low-efficiency data occupation ratio.
Preferably, the operation data acquisition in the step 1 includes extracting data samples from a historical database of the unit according to an evaluation period, wherein the evaluation period is preferably 1-6 months, and the data channel includes a real-time wind speed, an active power, a generator rotating speed, a pitch angle, a unit operation state flag bit and a power limiting flag bit; the data preprocessing is specifically to carry out deep cleaning on outliers and edge points in a scatter diagram of the operating data by adopting a parameter self-tuning DBSCAN clustering algorithm on the basis of conventional data screening.
Preferably, the conventional data screening in step 1 comprises: rejecting shutdown data and faulty operation data according to the unit operation state zone bit; rejecting power-limiting data according to the power-limiting zone bit; and eliminating data beyond the normal operation range according to the wind speed, the active power and the rotating speed of the generator.
Preferably, the wind speed interval in the step 2 is divided into a low wind speed section and a high wind speed section, the low wind speed section is a starting wind speed section of 4-6m/s, and the high wind speed section is a near wind speed section of 8-12 m/s.
Preferably, the step 3 operation region division calculates the boundaries of the normal region and the inefficient region based on the spatial distribution characteristics of the scattered point data of the wind speed-active power.
Preferably, the calculation of the boundary of the operating region divides the data bin according to the wind speed, wherein the bin interval is 0.5m/s, and the specific calculation method comprises the following steps:
a low wind speed section:
edge_Lv＝median_Lv-W，v＝4，4.5，…6
W＝median_L4-(Q14-1.5IQR4)
IQR4＝Q34-Q14
wherein, Q34、Q14、median_L4Upper and lower quartile values and median, mean _ L, of data bin at wind speed 4m/s, respectivelyvThe median of the data bin with the wind speed v;
a high wind speed section:
edge_Hv＝Q1v-1.5(Q3v-Q1v)，v＝8，8.5，…12
wherein, Q3v、Q1vRespectively the upper and lower quartile values of the data bin of the wind speed v.
Preferably, in the step 4, the data occupation ratio of the low-efficiency area and the low-efficiency area is used as an evaluation index of the abnormal degree of the unit operation state, a threshold value is set according to the probability distribution condition of the index of the quantity of the same machine type, and the unit operation performance is considered to be abnormal when the index exceeds the threshold value.
Preferably, the calculation of the area of the low-efficiency area firstly performs 0-1 normalization on the raw data of the low wind speed section and the high wind speed section and the calculation result of the boundary, and the calculation method of the low-efficiency area comprises the following steps:
Figure BDA0002660534180000041
Figure BDA0002660534180000042
wherein, min _ Lv、min_HvIs the minimum value of the data bin for wind speed v.
Preferably, the accuracy of the DBSCAN algorithm is related to the selection of two parameters of Eps and Minpts, and the method for self-tuning the two parameters of Eps and Minpts comprises the following specific steps:
1) carrying out standardization processing on the data samples;
2) initializing, namely selecting an acceptable proportion epsilon-noise of the noise data to be 5%, and enabling k to be 2;
3) calculating the geometric distance k-dis between each object in the data sample and its k-th nearest objectAnd n object k-dis sets in the data set are marked as Distk＝{k-dis1,k-dis2,…,k-disn}；
4) For DistkThe elements in the set are subjected to probability statistics, and m elements within the probability range of 5% -95% are selected to form a new set Dist according to the cumulative probability distribution resultk *To Distk *Is taken as the mathematical expected value of Epsk：
Figure BDA0002660534180000043
5) Eps of all objects in the datasetkThe number of points in the domain is set Pk＝{p1,p2,…,pnIs PkAccording to the accumulated probability distribution result, selecting one element in the probability range of 5% -95% to form a new set Pk *A 1 is to Pk *Reassigns the mathematical expectation value of (C) to Minptsk：
Figure BDA0002660534180000051
6) In MinptskAnd EpskClustering the data set, and calculating the noise data Ratio according to the clustering resultk；
7) If k is>2, and Ratio-noisekSatisfies the following conditions: i Ratio-noisek-1-Ratio-noisek| ≦ ε -noise, then Minpts ≦ Minptsk，Eps＝EpskAnd ending; otherwise, let k be k +1, return to 3) to be executed again.
The invention has the beneficial effects that:
1. the method has the advantages that the averaging processing is not needed, all working condition information of the unit can be reserved, the abnormal performance state of the unit is more sensitive, and the monitoring precision is higher.
2. In the data preprocessing step, the clustering method is adopted to deeply clean the unit operation data on the basis of conventional data screening, so that the effectiveness of a data sample is ensured.
3. Aiming at the characteristic that the operation rules of the wind turbine generator set are different in different wind speed intervals, the whole operation space of the wind turbine generator set is divided into a low wind speed section and a high wind speed section for monitoring respectively, and the boundary of the operation area of the two sections adopts different calculation methods, so that the monitoring precision is further improved.
4. The method has low requirement on the quality of the operation data, does not require the completeness of a time sequence, has small related calculated amount, does not need highly configured calculation resources, and can be easily applied to the field of a wind power plant to monitor the operation performance state of a unit on line and give an alarm for performance abnormality.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
FIG. 2 is a flow chart of an improved DBSCAN clustering algorithm based on a parameter self-tuning method of the present invention.
FIG. 3 is a diagram illustrating an exemplary data preprocessing effect of the present invention.
FIG. 4 is an exemplary diagram of the division result of the low wind speed operation region according to the present invention.
FIG. 5 is an exemplary graph of the high wind speed segment operation region division result of the present invention.
FIG. 6 is an exemplary illustration of the low efficiency area of a low wind speed section of the present invention.
FIG. 7 is an exemplary plot of the area of the low efficiency region of a high wind speed segment of the present invention.
FIG. 8 is a probability distribution plot for the area of an inefficient region of the present invention.
FIG. 9 is a data specific probability distribution diagram of a certain model of the invention for low efficiency region.
Detailed Description
The technical scheme of the invention is further specifically described by the following embodiments and the accompanying drawings.
Example (b): as shown in fig. 1, the method for monitoring the operation performance of a large wind turbine generator according to this embodiment includes 4 basic steps of obtaining and preprocessing operation data, dividing wind speed intervals, dividing operation areas of different wind speed segments, and determining operation states of different wind speed segments, where each step is described in detail as follows:
s1, operation data acquisition and preprocessing
Firstly, extracting 5 data channels including real-time wind speed, active power, generator rotating speed, unit running state zone bits and electricity limiting zone bits from a unit historical database according to an evaluation period, wherein the evaluation period is preferably 1-6 months. The data preprocessing comprises 2 links of conventional data screening and data deep cleaning.
The conventional data screening is used for sequentially screening original sample data according to the following conditions:
1) removing shutdown data and faulty operation data according to the unit operation state zone bit (the unit operation state zone bit is 1);
2) removing power-limiting data according to the power-limiting flag bit (the power-limiting flag bit is equal to 0);
3) and (4) eliminating data beyond a normal operation range according to the wind speed, the active power and the rotating speed of the generator (the lower limit of the wind speed is less than the upper limit of the wind speed, the lower limit of the active power is less than the upper limit of the active power, the lower limit of the rotating speed is less than the rotating speed of the generator and less than the upper limit of the rotating speed).
And (4) deeply cleaning data, performing cluster analysis on the data after conventional screening by adopting a DBSCAN algorithm with self-tuning parameters, and identifying and removing outlier data and edge burr data. Fig. 2 is a flow chart of an improved DBSCAN clustering algorithm based on a parameter self-tuning method, and shows a specific implementation process of clustering data samples after conventional screening.
The DBSCAN algorithm is a classic clustering algorithm based on density, the accuracy of the DBSCAN algorithm is related to the selection of two parameters of Eps and Minpts, the invention provides a self-tuning method of the two parameters of Eps and Minpts, as shown in figure 2, the specific steps are as follows:
1) carrying out standardization processing on the data samples;
2) initializing, namely selecting an acceptable proportion epsilon-noise of the noise data to be 5%, and enabling k to be 2;
3) calculating the geometric distance between each object in the data sample and its k-th nearest objectk-dis, the set of n objects in the dataset k-dis is denoted as Distk＝{k-dis1,k-dis2,…,k-disn}。
4) For DistkThe elements in the list are subjected to probability statistics, and m elements within the range of 5% -95% of probability are selected to form a new set according to the cumulative probability distribution result
Figure BDA0002660534180000071
To be provided with
Figure BDA0002660534180000072
Is taken as the mathematical expected value of Epsk：
Figure BDA0002660534180000073
5) Eps of all objects in the datasetkThe number of points in the domain is set Pk＝{p1,p2,…,pnIs PkThe elements in the list are subjected to probability statistics, and according to the accumulated probability distribution result, the elements in the range of 5% -95% probability are selected to form a new set
Figure BDA0002660534180000074
Will be provided with
Figure BDA0002660534180000075
Reassigns the mathematical expectation value of (C) to Minptsk：
Figure BDA0002660534180000076
6) In MinptskAnd EpskClustering the data set, and calculating the noise data Ratio according to the clustering resultk。
7) If k is>2, and Ratio-noisekSatisfies the following conditions: i Ratio-noisek-1-Ratio-noisek| ≦ ε -noise, then Minpts ≦ Minptsk，Eps＝EpskAnd ending; otherwise, let k be k +1, return to 3) to be executed again.
Fig. 3 is an exemplary diagram of data preprocessing effect, which shows the variation of scattered points of data after the original data is subjected to conventional data screening and deep data cleaning.
S2, dividing wind speed interval
Selecting a starting wind speed section of 4-6m/s as a low wind speed section, and selecting a wind speed section near a full wind speed of 8-12m/s as a high wind speed section.
S3, dividing operation areas of different wind speed sections
The running region boundary calculation divides the data bin according to the wind speed, and the bin interval is 0.5 m/s. The data bin interval boundaries of the low wind speed segments are [3.75, 4.25], [4.25, 4.75], …, [5.75, 6.25 ]. The data bin interval boundaries of the high wind speed section are [7.75, 8.25], [8.25, 8.75], …, [11.75, 12.25 ].
FIG. 4 is an exemplary graph of the division result of the low wind speed segment operation region, which shows the boundary position of the operation region of a certain 2MW unit in the 4-6m/s wind speed segment. The specific operation area boundary calculation method of the low wind speed section comprises the following steps:
edge_Lv＝median_Lv-W，v＝4，4.5，…6
W＝median_L4-(Q14-1.5IQR4)
IQR4＝Q34-Q14
wherein, Q34、Q14、median_L4Upper and lower quartile values and median, mean _ L, of data bin at wind speed 4m/s, respectivelyvThe median of the bin for wind speed v.
FIG. 5 is an exemplary graph of the division result of the high wind speed segment operation region, showing the boundary position of the operation region of a certain 2MW unit in the 8-12m/s wind speed segment. The specific operation area boundary calculation method of the high wind speed section comprises the following steps:
edge_Hv＝Q1v-1.5(Q3v-Q1v)，v＝8，8.5，…12
wherein, Q3v、Q1vRespectively the upper and lower quartile values of the data bin of the wind speed v.
S4, judging the running states of different wind speed sections
And comparing the area of the low-efficiency area with the data of the low-efficiency area to obtain the abnormal degree evaluation index of the running state of the unit.
Fig. 6 and 7 are calculation example diagrams of the low-efficiency area, which show the low-efficiency area of the low wind speed section and the high wind speed section of a certain 2MW unit. Firstly, respectively carrying out 0-1 normalization processing on the original data of a low wind speed section and a high wind speed section and the calculation result of the boundary of an operating area, wherein the calculation method of the area of the low-efficiency area comprises the following steps:
Figure BDA0002660534180000091
Figure BDA0002660534180000092
wherein, min _ Lv、min_HvIs the minimum value of the data bin for wind speed v.
The method for calculating the data occupation ratio of the low-efficiency area comprises the following steps:
r _ L is the data volume of the low-efficiency area of the low wind speed section/the total data volume of the low wind speed section multiplied by 100 percent
R _ H is the data volume of the low-efficiency area of the high wind speed section/the total data volume of the high wind speed section multiplied by 100 percent
And setting a threshold value according to the probability distribution condition of the two indexes of the same machine type, and considering that the running performance of the machine set is abnormal when the indexes exceed the threshold value.
Fig. 8 is a 2MW model low-efficiency region area probability distribution diagram, with a 95% fractional value as the low-efficiency region area threshold, in this example, the low-wind speed segment and the high-wind speed segment model low-efficiency region area thresholds are S _ L _ threshold of 0.022 and S _ H _ threshold of 0.058, respectively.
Fig. 9 is a 2MW model inefficiency data ratio probability distribution diagram, which uses a 95% quantile value as a inefficiency data ratio threshold, and the inefficiency data ratio thresholds of the models in the low wind speed segment and the high wind speed segment are respectively R _ L _ threshold 0.078 and R _ H _ threshold 0.116 in this example.
The unit running state judgment principle is as follows:
Figure BDA0002660534180000093
Figure BDA0002660534180000101
the specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Although the terms clustering algorithm, wind speed interval, operating area, etc. are used more herein, the possibility of using other terms is not excluded. These terms are used merely to more conveniently describe and explain the nature of the present invention; they are to be construed as being without limitation to any additional limitations that may be imposed by the spirit of the present invention.

Claims (4)

1. A method for monitoring the operation performance of a large-scale wind turbine generator is characterized by comprising the following steps:
(1) acquiring operation data and preprocessing the data, and deeply cleaning the unit operation data by adopting a clustering method on the basis of conventional data screening;
(2) the wind speed interval is divided into a low wind speed section and a high wind speed section, wherein the low wind speed section is selected to be a 4-6m/s starting wind speed section, and the high wind speed section is selected to be a near full wind speed section of 8-12 m/s;
(3) dividing an operation region, calculating the boundary of a normal region and an inefficient region based on the spatial distribution characteristics of scattered point data of wind speed-active power, dividing a data bin according to the wind speed in the operation region boundary calculation, wherein the bin interval is 0.5m/s, and the specific calculation method comprises the following steps:
a low wind speed section:
edge_Lv＝median_Lv-W，v＝4，4.5，…6
W＝median_L4-(Q14-1.5IQR4)
IQR4＝Q34-Q14
wherein, Q34、Q14、median_L4Upper and lower quartile values and median, mean _ L, of data bin at wind speed 4m/s, respectivelyvThe median of the data bin with the wind speed v;
a high wind speed section:
edge_Hv＝Q1v-1.5(Q3v-Q1v)，v＝8，8.5，…12
wherein, Q3v、Q1vThe upper and lower quartile values of the data bin of the wind speed v are respectively;
(4) and (2) judging performance abnormity, namely taking the data occupation ratio of the low-efficiency area and the low-efficiency area as the abnormal degree evaluation indexes of the unit running state, setting a threshold value according to the probability distribution condition of the two indexes of the same machine type, when the indexes exceed the threshold value, judging that the unit running performance is abnormal, calculating the area of the low-efficiency area, firstly, respectively carrying out 0-1 normalization on the original data and the boundary calculation result of the low wind speed section and the high wind speed section, wherein the calculation method of the low-efficiency area comprises the following steps:
Figure FDA0003030587680000011
Figure FDA0003030587680000021
wherein, min _ Lv、min_HvIs the minimum value of the data bin for wind speed v.
2. The method for monitoring the running performance of the large-scale wind turbine generator set according to claim 1, wherein the step 1 of obtaining the running data comprises extracting data samples from a historical database of the large-scale wind turbine generator set according to an evaluation period, wherein the evaluation period is preferably 1-6 months, and a data channel comprises a real-time wind speed, an active power, a generator rotating speed, a pitch angle, a unit running state flag bit and a power limiting flag bit; the data preprocessing is specifically to carry out deep cleaning on outliers and edge points in a scatter diagram of the operating data by adopting a parameter self-tuning DBSCAN clustering algorithm on the basis of conventional data screening.
3. The method for monitoring the operation performance of the large-scale wind turbine generator according to claim 2, wherein the conventional data screening in the step 1 comprises the following steps: rejecting shutdown data and faulty operation data according to the unit operation state zone bit; rejecting power-limiting data according to the power-limiting zone bit; and eliminating data beyond the normal operation range according to the wind speed, the active power and the rotating speed of the generator.
4. The method for monitoring the operating performance of the large-scale wind turbine generator set according to claim 2, wherein the accuracy of the DBSCAN clustering algorithm is related to the selection of two parameters, namely Eps and Minpts, and a self-tuning method for the two parameters, namely Eps and Minpts, comprises the following specific steps:
1) carrying out standardization processing on the data samples;
2) initializing, namely selecting an acceptable proportion epsilon-noise of the noise data to be 5%, and enabling k to be 2;
3) calculating the geometric distance k-dis between each object in the data sample and the k-th nearest object thereof, and recording n object k-dis sets in the data set as Distk＝{k-dis1,k-dis2,…,k-disn}；
4) For DistkThe elements in the set are subjected to probability statistics, and m elements within the probability range of 5% -95% are selected to form a new set Dist according to the cumulative probability distribution resultk *To Distk *Is taken as the mathematical expected value of Epsk：
Figure FDA0003030587680000022
5) Eps of all objects in the datasetkThe number of points in the domain is set Pk＝{p1,p2,…,pnIs PkAccording to the accumulated probability distribution result, selecting one element in the probability range of 5% -95% to form a new set Pk *A 1 is to Pk *Reassigns the mathematical expectation value of (C) to Minptsk：
Figure FDA0003030587680000031
6) In MinptskAnd EpskClustering the data set, and calculating the noise data Ratio according to the clustering resultk；
7) If k is>2, and Ratio-noisekSatisfies the following conditions: i Ratio-noisek-1-Ratio-noisek| ≦ ε -noise, then Minpts ≦ Minptsk，Eps＝EpskAnd ending; otherwise, let k be k +1, return to 3) to be executed again.
CN202010903361.2A 2020-09-01 2020-09-01 Method for monitoring operation performance of large wind turbine generator Active CN112032003B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010903361.2A CN112032003B (en) 2020-09-01 2020-09-01 Method for monitoring operation performance of large wind turbine generator

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010903361.2A CN112032003B (en) 2020-09-01 2020-09-01 Method for monitoring operation performance of large wind turbine generator

Publications (2)

Publication Number Publication Date
CN112032003A CN112032003A (en) 2020-12-04
CN112032003B true CN112032003B (en) 2021-08-17

Family

ID=73590455

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010903361.2A Active CN112032003B (en) 2020-09-01 2020-09-01 Method for monitoring operation performance of large wind turbine generator

Country Status (1)

Country Link
CN (1) CN112032003B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112883019A (en) * 2021-02-20 2021-06-01 上海电气风电集团股份有限公司 Data processing method and system
CN113236508B (en) * 2021-05-31 2022-04-05 浙江运达风电股份有限公司 Method for detecting wind speed-power abnormal data of wind generating set
CN113339207B (en) * 2021-06-29 2023-05-26 浙江理工大学 Method and system for dividing operation conditions of main transmission system of wind turbine generator
CN114033631B (en) * 2021-11-08 2023-08-01 浙江运达风电股份有限公司 Online identification method for wind energy utilization coefficient of wind turbine generator
CN116106744B (en) * 2023-04-13 2023-06-27 合肥荣叙科技有限公司 Permanent magnet synchronous motor testing system and testing method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014046355A1 (en) * 2012-09-18 2014-03-27 한국전력공사 Method of automatically calculating power curve limit for power curve monitoring of wind turbine
CN105134485A (en) * 2015-08-18 2015-12-09 山东大学 Inertial frequency-modulation drive rotating speed protection control system and method for doubly-fed wind turbine
CN106014858A (en) * 2016-07-21 2016-10-12 浙江运达风电股份有限公司 Automatic calibration method and device for air aligning errors of wind generation set
CN108412710A (en) * 2018-01-30 2018-08-17 同济大学 A kind of Wind turbines wind power data cleaning method
CN111275570A (en) * 2020-01-13 2020-06-12 浙江大学 Wind turbine generator set power abnormal value detection method based on iterative statistics and hypothesis test

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014046355A1 (en) * 2012-09-18 2014-03-27 한국전력공사 Method of automatically calculating power curve limit for power curve monitoring of wind turbine
CN105134485A (en) * 2015-08-18 2015-12-09 山东大学 Inertial frequency-modulation drive rotating speed protection control system and method for doubly-fed wind turbine
CN106014858A (en) * 2016-07-21 2016-10-12 浙江运达风电股份有限公司 Automatic calibration method and device for air aligning errors of wind generation set
CN108412710A (en) * 2018-01-30 2018-08-17 同济大学 A kind of Wind turbines wind power data cleaning method
CN111275570A (en) * 2020-01-13 2020-06-12 浙江大学 Wind turbine generator set power abnormal value detection method based on iterative statistics and hypothesis test

Also Published As

Publication number Publication date
CN112032003A (en) 2020-12-04

Similar Documents

Publication Publication Date Title
CN112032003B (en) Method for monitoring operation performance of large wind turbine generator
CN105631596B (en) Equipment fault diagnosis method based on multi-dimensional piecewise fitting
CN107885959B (en) Wind power modeling and performance evaluation method based on confidence equivalent power curve band
CN109740175B (en) Outlier discrimination method for power curve data of wind turbine generator
WO2022001379A1 (en) Data processing method, and performance evaluation method and system of wind generating set
CN108443088B (en) Wind turbine generator system state judging method based on cumulative probability distribution
CN110685868A (en) Wind turbine generator fault detection method and device based on improved gradient elevator
CN110362045B (en) Marine doubly-fed wind turbine generator fault discrimination method considering marine meteorological factors
CN111931834B (en) Method, equipment and storage medium for detecting abnormal flow data in extrusion process of aluminum profile based on isolated forest algorithm
CN110795690A (en) Wind power plant operation abnormal data detection method
CN111767951A (en) Method for discovering abnormal data by applying isolated forest algorithm in residential electricity safety analysis
CN111209934A (en) Fan fault prediction and alarm method and system
CN114417971A (en) Electric power data abnormal value detection algorithm based on K nearest neighbor density peak clustering
CN107808209B (en) Wind power plant abnormal data identification method based on weighted kNN distance
CN117421994A (en) Edge application health monitoring method and system
Zhang et al. An improved association rule mining-based method for discovering abnormal operation patterns of HVAC systems
CN115326393A (en) Wind turbine generator bearing pair fault diagnosis method based on temperature information
CN110083804B (en) Wind power plant SCADA data missing intelligent repairing method based on condition distribution regression
Nashrullah et al. Performance Evaluation of Superstate HMM with Median Filter For Appliance Energy Disaggregation
Mascali et al. Synthetic ground truth generation of an electricity consumption dataset
Yang et al. Change-point estimation for repairable systems combining bootstrap control charts and clustering analysis: Performance analysis and a case study
CN117312617B (en) Real-time sewage treatment method and system based on sewage data monitoring
CN116738158B (en) Intelligent evaluation method for loss of distribution box system
CN113447813B (en) Fault diagnosis method and equipment for offshore wind generating set
CN112459970B (en) Wind turbine generator anomaly detection method using data counterstudy

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