CN115263680A - Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model - Google Patents
Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model Download PDFInfo
- Publication number
- CN115263680A CN115263680A CN202210787537.1A CN202210787537A CN115263680A CN 115263680 A CN115263680 A CN 115263680A CN 202210787537 A CN202210787537 A CN 202210787537A CN 115263680 A CN115263680 A CN 115263680A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- abnormal
- pitch motor
- early warning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 57
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012549 training Methods 0.000 claims abstract description 20
- 238000004140 cleaning Methods 0.000 claims abstract description 4
- 238000013075 data extraction Methods 0.000 claims abstract description 4
- 230000008859 change Effects 0.000 claims abstract description 3
- 239000013598 vector Substances 0.000 claims description 20
- 230000008030 elimination Effects 0.000 claims description 12
- 238000003379 elimination reaction Methods 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000010248 power generation Methods 0.000 claims description 5
- 238000012544 monitoring process Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000003058 natural language processing Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 abstract description 2
- 230000007246 mechanism Effects 0.000 description 4
- 230000017525 heat dissipation Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
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
-
- 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)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models, which comprises the following steps: data extraction: extracting corresponding measuring point data from a fan SCADA historical database; data cleaning: removing data of each measuring point according to a corresponding principle; the working conditions of the fan are divided: when the model is trained, different models are trained according to different loads to obtain different baseline values; a TF-IDF model is adopted for training and an LSTM model is adopted for training, and a functional relation is trained to describe the temperature change of the variable pitch motor; the early warning model is operated: the model is operated once every 1 hour and is finally mapped into an early warning level. The abnormal temperature rise of the variable pitch motor is detected by combining an abnormal detection model constructed by a symbolic representation technology based on TF-IDF and a temperature prediction model of the variable pitch motor constructed by LSTM.
Description
Technical Field
The invention relates to the field of algorithms of machine learning, in particular to a pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models.
Background
The variable pitch system is a mechanical mechanism used for controlling the position angle of a blade relative to a rotating plane, a wind generating set controls the start and stop of the set, controls the output power and other tasks through variable pitch, the variable pitch control mode can be generally divided into two modes, one mode is a motor executing mechanism, the other mode is a hydraulic executing mechanism, the motor variable pitch executing mechanism utilizes a motor to independently control the blade, the existing mainstream wind generating set is a synchronous variable pitch system, so that three independent variable pitch motors execute the same variable pitch control instruction, if the motor continuously and frequently adjusts the blade, excessive heat load is generated to damage the motor, and the adjusting process of the variable pitch system is shown as follows.
Large wind speed- > 90 degrees of blade angle- > reducing rotating speed or stopping- > ensuring safety
The wind speed is small- > the blade angle is 0 degree- > the pursuit of the maximum rotating speed- > the maximum generating power
The variable pitch system is an important transmission chain control device of the wind generating set, and is an important part influencing the generating capacity of a fan, energy control and even unit safety, the variable pitch motor has the defects of asynchronous variable pitch control and even more serious potential safety hazard due to the fact that the motor is further cracked due to abnormal heat dissipation and the faults of overheating and locking of the motor are caused because the performance of the variable pitch motor is aged and degenerated and the working load is increased, the monitoring and early warning of the variable pitch motor and the heat dissipation of the variable pitch motor are facilitated to find the abnormal variable pitch motor and the heat dissipation in time, and the variable pitch system is maintained in time, so that the greater performance loss and the safety problem are avoided.
Disclosure of Invention
The invention aims to provide a pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models, which comprises the following steps:
step one, data extraction: extracting measuring point data from a fan SCADA historical database;
step two, data cleaning: removing the data of each measuring point according to two principles of null value removal and abnormal value removal;
thirdly, dividing the working conditions of the fan: when the model of abnormal working condition data is trained, different models are trained according to different loads to obtain different baseline values;
step four, model training: adopting two training models of TF-IDF and LSTM to train data;
step five, the early warning model operates: the model runs once every 1 hour, the data of the last 2 days are used every time, the data size of one window is collected, the distance between the TF-IDF vector of each temperature difference and the baseline model is calculated, the distance between the predicted value of the temperature of each variable pitch motor and the actual parameter of the equipment is calculated, the risk value is calculated, and finally the risk value is mapped into the early warning level.
Preferably, the measuring point data adopted in the first step are the temperature of the variable pitch motor, the current of the variable pitch motor, the angle of the blade, the wind speed, the active power, the rotating speed of the hub, the state of the fan, the ambient temperature and the temperature of the engine room respectively, the time length is 1 year, and the time interval is 1 minute.
Preferably, in the second step, in each time section, as long as one measured point data is abnormal, all the time section data are removed;
in the null value elimination, when the unit is in a bad working condition or the operation state is unstable, the null value data in the SCADA system lasts for a period of time and can be preprocessed in a direct elimination mode, the operation unstable state means that the unit is frequently intermittently started and stopped or communication interruption occurs, and the SCADA data has null value records;
in the elimination of the abnormal values, for other abnormal values possibly existing, a boxplot method is adopted for preprocessing through statistical analysis of historical values of the wind turbine generator, data points obviously deviating from normal values are eliminated, the abnormal values in the SCADA data mainly mean that the measured value of the sensor exceeds the upper and lower physical limits of the sensor, or the measured result exceeds the current possible normal range, such as wind speed, environmental temperature and the like, exceeds the normal range, the data with the measured value of the sensor exceeding the upper and lower physical limits can be directly processed in a deleting mode, after extreme value judgment is carried out on the measured result, the data obviously exceeds the current normal range, such as the wind speed is less than 0 or greater than the upper historical limit, the rotating speed is less than 0 and the like, and the record items where the data are located are also deleted.
Preferably, in the removing of the abnormal value in the second step, when the preprocessing is performed by using a box chart method, the distribution of the whole data is observed, the whole distribution of the data is described by using statistics such as a median, a 25% quantile, a 75% quantile, an upper boundary and a lower boundary, and a box chart is generated by calculating the statistics, wherein the box contains most of normal data, and the data outside the upper boundary and the lower boundary of the box is regarded as abnormal data.
Preferably, in the elimination of the abnormal value in the second step, the calculation formula of the upper and lower boundaries is as follows:
where Q1 represents the 25% quantile of data, Q3 represents the 75% quantile of data, lowerLimit represents the lower limit of the data set, and UpperLimit represents the upper limit of the data set.
Preferably, in the third step, the abnormal working condition data mainly refers to data generated by the condition that equipment is not operated, equipment fails, and problems of wind abandon and electricity limiting or logical contradiction among multiple parameters, such as grid-connected power generation when the wind speed is lower than that of a wind cut-in speed unit, occur.
Preferably, in the third step, an early warning model is established for abnormal temperature rise of the variable pitch motor, and the working conditions of the fan are divided according to the following principle:
(1) Selecting data of which the wind speed is higher than the cut-in wind speed, the blade angle is more than 2 degrees and less than 50 degrees, and the state of the fan is a power generation state;
(2) Clustering by adopting a density-based clustering analysis method, and when the distance from an object to a clustering group needs to be calculated, using the vertical mapping distance from the object to a subspace;
(3) The clustered data set can be defined as different working conditions, generally a full-load working condition and different electricity-limiting working conditions.
Preferably, in the fourth step, the first step,
the baseline value was calculated using the TF-IDF model: under a stable working condition, the temperatures of the variable pitch motors of the three blades keep a certain relation, firstly, the temperatures of the three variable pitch motors are subtracted in pairs to obtain a characteristic vector, then, discretization is carried out on characteristic vector data according to the precision of 0.1 ℃, and then, historical data of the discretized characteristic vector are used for training cluster horizontal baseline vectors of temperature reference of the three variable pitch motors in pairs;
training by adopting an LSTM model: baseline values were calculated using the LSTM model: training a functional relation to describe the temperature change of the variable pitch motor based on historical data, establishing a model to predict the temperature, establishing the model as a base line, and then calling the trained model with the same working condition data to predict the temperature of the variable pitch motor;
after discretization processing is carried out on the feature vector data, the distribution of the data can be described by using a histogram method, the weighting method TF-IDF widely used in the field of natural language processing is adopted for processing, actual production data are used after processing, the space vector distance among the feature vectors is calculated and used as a risk value for early warning, and an early warning event can be output when any group of variable pitch motors are abnormal in temperature.
Preferably, in the training by using the LSTM model in the fourth step, the model is constructed by using the following method:
the original data are normalized, the purpose of normalizing the original data is that the monitoring data stored in the SCADA data are various, the physical meanings of different types of data are different, the units are different, and the data in different units and different magnitudes can be compared and analyzed conveniently in order to eliminate the influence caused by different dimensions.
Preferably, when the LSTM model is adopted for training and the LSTM model is utilized for calculating the baseline value in the fourth step, the normalization formula is
The invention has the technical effects and advantages that:
(1) According to the method, the abnormal temperature rise of the variable pitch motor is detected by combining an abnormal detection model constructed by a symbolic representation technology based on TF-IDF and a variable pitch motor temperature prediction model constructed by LSTM, so that the transverse and longitudinal comparison of the variable pitch motor is effectively realized, namely the temperature of the variable pitch motor is compared with the temperature of other variable pitch motors at the same time and the historical near working condition of the variable pitch motor, and the early warning accuracy is improved;
(2) Meanwhile, the invention also adopts TF-IDF symbolization characterization technology to reduce the measurement error caused by sensor noise and improve the signal-to-noise ratio;
(3) When the LSTM is used for modeling, the working conditions are divided firstly, the data set under the relatively stable working conditions is obtained, and the data set is used for model training, so that the problem that the model fails or is too complex due to severe fluctuation of the working conditions is solved.
Drawings
FIG. 1 is a schematic flow chart of the present invention.
FIG. 2 is a box type diagram of the invention for eliminating abnormal data.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
The invention provides a pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models, which comprises the following steps:
step one, data extraction: extracting measuring point data from a fan SCADA historical database;
step two, data cleaning: removing data of each measuring point according to two principles of null value removal and abnormal value removal;
thirdly, dividing the working conditions of the fan: when the model of abnormal working condition data is trained, different models are trained according to different loads to obtain different baseline values.
Step four: model training: adopting two training models of TF-IDF and LSTM to train data;
step five: the early warning model operates: the model runs once every 1 hour, the data of the last 2 days are used every time, the data size of one window is collected, the distance between the TF-IDF vector of each temperature difference and the baseline model is calculated, the distance between the predicted value of the temperature of each variable pitch motor and the actual parameter of the equipment is calculated, the risk value is calculated, and finally the risk value is mapped into the early warning level.
In the first step, the adopted measuring point data are respectively the temperature of a variable pitch motor, the current of the variable pitch motor, the angle of a blade, the wind speed, the active power, the rotating speed of a hub, the state of a fan, the ambient temperature and the temperature of an engine room, the time length is 1 year, the time interval is 1 minute, and the extracted measuring point data are shown in the following table 1;
| pitch_mot_tmp1 | temperature of variable pitch motor 1 |
| pitch_mot_tmp2 | Temperature of variable pitch motor 2 |
| pitch_mot_tmp3 | Variable pitch motor 3 temperature |
| pitch_mot_curr1 | Variable pitch motor 1 current |
| pitch_mot_curr2 | Current of variable pitch motor 2 |
| pitch_mot_curr3 | Variable pitch motor 3 current |
| blade_ang1 | Angle of blade 1 |
| blade_ang2 | 2 angle of blade |
| blade_ang3 | Angle of blade 3 |
| wd_spd | Wind speed |
| atpwr | Active power |
| hub_spd | Rotational speed of hub |
| wd_tur_st | Fan status |
| hub_tmp | Ambient temperature |
| nac_tmp | Cabin temperature |
TABLE 1
In the second step, in each time section, as long as one measuring point data is abnormal, all the time section data are removed;
in the null value elimination, when the unit is in a bad working condition or the running state is unstable, the null value data in the SCADA system lasts for a period of time and can be preprocessed in a direct elimination mode, the running unstable state refers to the condition that the unit is frequently intermittently started and stopped or communication is interrupted, and the SCADA data has null value records;
in the elimination of the abnormal values, for other abnormal values possibly existing, a boxplot method is adopted for preprocessing through statistical analysis of historical values of the wind turbine generator, data points obviously deviating from normal values are eliminated, the abnormal values in the SCADA data mainly mean that the measured value of the sensor exceeds the upper and lower physical limits of the sensor, or the measured result exceeds the current possible normal range, such as wind speed, environmental temperature and the like, exceeds the normal range, the data with the measured value of the sensor exceeding the upper and lower physical limits can be directly processed in a deleting mode, after extreme value judgment is carried out on the measured result, the data obviously exceeds the current normal range, for example, the wind speed is less than 0 or greater than the upper historical limit, the rotating speed is less than 0 and the like, and the record items where the data are located are also deleted;
in the abnormal value elimination, when the preprocessing is carried out by adopting a box diagram method, the overall distribution condition of the data is observed, the overall distribution condition of the data is described by using statistics such as a median, a 25% quantile, a 75% quantile, an upper boundary and a lower boundary, and the like, a box diagram is generated by calculating the statistics, the box contains most of normal data, the abnormal data is considered to be outside the upper boundary and the lower boundary of the box, and the upper boundary and the lower boundary calculation formula are shown as follows:
where Q1 represents the 25% quantile of data, Q3 represents the 75% quantile of data, lowerLimit represents the lower limit of the data set, and UpperLimit represents the upper limit of the data set.
In the third step, an early warning model is established for abnormal temperature rise of the variable pitch motor, and the working conditions of the fan are divided according to the following principle:
(1) Selecting data of which the wind speed is higher than the cut-in wind speed, the blade angle is more than 2 degrees and less than 50 degrees, and the state of the fan is a power generation state;
(2) Clustering by adopting a density-based clustering analysis method, and when the distance from an object to a clustering group needs to be calculated, using the vertical mapping distance from the object to a subspace;
(3) The clustered data set can be defined as different working conditions, generally a full-load working condition and different electricity-limiting working conditions.
In the fourth step, after discretization processing is carried out on the feature vector data, a histogram method is used for describing the distribution of the data, a weighting method TF-IDF widely used in the field of natural language processing is adopted for processing, actual production data are used after processing, the space vector distance between the feature vectors is calculated to serve as a risk value for early warning, an early warning event can be output when any group of variable pitch motors are abnormal in temperature, and in LSTM model training, a model is constructed by adopting the following method:
firstly, normalizing original data, aiming at the purpose of normalizing the original data that monitoring data stored in SCADA data are various, different types of data have different physical meanings and different units, and in order to eliminate the influence caused by different dimensions, the data of different units and different magnitudes are convenient to compare and analyze, an LSTM model is adopted to establish a model for predicting the temperature of a variable pitch motor, the detailed parameters of the model are shown in the following table 2, when the model runs, the working conditions are divided according to the working condition division principle, then the model trained by the same working condition data is called to predict the temperature of the variable pitch motor, and when the LSTM model is adopted to train and the LSTM model is utilized to calculate a baseline value, the normalization formula is
TABLE 2
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still make modifications to the technical solutions described in the foregoing embodiments, or make equivalent substitutions and improvements to part of the technical features of the foregoing embodiments, and any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A variable pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models is characterized by comprising the following steps:
step one, data extraction: extracting measuring point data from a fan SCADA historical database;
step two, data cleaning: removing the data of each measuring point according to two principles of null value removal and abnormal value removal;
thirdly, dividing the working conditions of the fan: when the model of abnormal working condition data is trained, different models are trained according to different loads to obtain different baseline values;
step four, training a model: adopting two training models of TF-IDF and LSTM to train data;
step five, the early warning model operates: the model runs once every 1 hour, data of the last 2 days are used each time, data amount of one window is collected, the distance between the TF-IDF vector of each temperature difference and the baseline model is calculated, the distance between the predicted value of each variable pitch motor temperature and the actual parameter of the equipment is calculated, a risk value is calculated, and finally the risk value is mapped to be an early warning level.
2. The abnormal temperature rise early warning method for the pitch motor combined with the TF-IDF and LSTM models according to claim 1, wherein the measured point data adopted in the first step are the temperature of the pitch motor, the current of the pitch motor, the angle of the blade, the wind speed, the active power, the rotating speed of the hub, the state of the fan, the ambient temperature and the temperature of the cabin respectively, the time length is 1 year, and the time interval is 1 minute.
3. The abnormal temperature rise early warning method of the pitch motor combined with the TF-IDF and LSTM models is characterized in that in the second step, as long as one measured point data is abnormal in each time section, all the time section data are removed;
in the null value elimination, when the unit is in a bad working condition or the running state is unstable, the null value data in the SCADA system lasts for a period of time and can be preprocessed in a direct elimination mode, the running unstable state refers to the condition that the unit is frequently intermittently started and stopped or communication is interrupted, and the SCADA data has null value records;
in the process of removing the abnormal values, for other abnormal values which may exist, the historical values of the wind turbine generator are statistically analyzed, a box diagram method is adopted for preprocessing, and data points which obviously deviate from normal values are deleted.
4. The abnormal temperature rise early warning method of the pitch motor combining the TF-IDF model and the LSTM model is characterized in that in the abnormal value elimination in the second step, when the preprocessing is carried out by adopting the box diagram method, the overall distribution situation of the data is observed, the overall distribution situation of the data is described by using statistics such as a median, a 25% quantile, a 75% quantile, an upper boundary and a lower boundary, and a box diagram is generated by calculating the statistics, wherein the box contains most of normal data, and the abnormal data is considered to be abnormal data outside the upper boundary and the lower boundary of the box.
5. Variable pitch motor abnormal temperature rise early warning method combining TF-IDF and LSTM models according to claim 1In the elimination of the abnormal value in the second step, the calculation formula of the upper and lower boundaries is as follows:
where Q1 represents the 25% quantile of data, Q3 represents the 75% quantile of data, lowerLimit represents the lower limit of the data set, and UpperLimit represents the upper limit of the data set.
6. The abnormal temperature rise early warning method for the pitch motor combined with the TF-IDF and LSTM models according to claim 1, wherein in the third step, the abnormal working condition data mainly refer to data generated by the fact that equipment does not operate, equipment fails, and problems such as wind abandonment and power limitation occur or logical contradictions among multiple parameters occur, for example, the wind speed is lower than a cut-in wind speed unit and grid-connected power generation is carried out.
7. The abnormal temperature rise early warning method of the pitch motor combined with the TF-IDF and LSTM models according to claim 1, wherein in the third step, the early warning model is established for the abnormal temperature rise of the pitch motor, and the working conditions of the fan are divided according to the following principle:
(1) Selecting data of which the wind speed is higher than the cut-in wind speed, the blade angle is more than 2 degrees and less than 50 degrees, and the state of the fan is a power generation state;
(2) Clustering by adopting a density-based clustering analysis method, and when the distance from an object to a clustering group needs to be calculated, using the vertical mapping distance from the object to a subspace;
(3) The clustered data set can be defined as different working conditions, generally a full-load working condition and different power limiting working conditions.
8. The abnormal temperature rise early warning method for the pitch motor combining the TF-IDF and LSTM models according to claim 1, wherein in the fourth step,
the baseline values were calculated using the TF-IDF model: under a stable working condition, the temperatures of the variable pitch motors of the three blades keep a certain relation, firstly, the temperatures of the three variable pitch motors are subtracted from each other in pairs to obtain a characteristic vector, then, discretization is carried out on characteristic vector data according to the precision of 0.1 ℃, and then, historical data of the discretized characteristic vector are used for training cluster horizontal baseline vectors of the three variable pitch motor temperatures which are referred to in pairs;
training by adopting an LSTM model: baseline values were calculated using the LSTM model: based on historical data, training a functional relation to describe the temperature change of the variable pitch motor, establishing a model to predict the temperature, establishing the model as a base line, and calling the trained model with the same working condition data to predict the temperature of the variable pitch motor;
after discretization processing is carried out on the feature vector data, the distribution of the data can be described by using a histogram method, the weighting method TF-IDF widely used in the field of natural language processing is adopted for processing, actual production data are used after processing, the space vector distance among the feature vectors is calculated and used as a risk value for early warning, and an early warning event can be output when any group of variable pitch motors are abnormal in temperature.
9. The abnormal temperature rise early warning method for the pitch motor combined with the TF-IDF and LSTM models according to claim 1, wherein in the LSTM model training in the fourth step, the model is constructed by adopting the following method:
the original data are normalized, the purpose of normalizing the original data is that the monitoring data stored in the SCADA data are various, the physical meanings of different types of data are different, the units are different, and the data with different units and different magnitudes are convenient to compare and analyze in order to eliminate the influence brought by different dimensions.
10. The abnormal temperature rise early warning method of the pitch motor combining the TF-IDF model and the LSTM model as claimed in claim 1, wherein in the fourth step, when the LSTM model is adopted for training and the LSTM model is utilized for calculating the baseline value, the normalization formula is
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210787537.1A CN115263680A (en) | 2022-07-04 | 2022-07-04 | Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210787537.1A CN115263680A (en) | 2022-07-04 | 2022-07-04 | Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115263680A true CN115263680A (en) | 2022-11-01 |
Family
ID=83763879
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210787537.1A Pending CN115263680A (en) | 2022-07-04 | 2022-07-04 | Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115263680A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842322A (en) * | 2023-07-19 | 2023-10-03 | 深圳市精微康投资发展有限公司 | Electric motor operation optimization method and system based on data processing |
| CN117212074A (en) * | 2023-09-25 | 2023-12-12 | 武汉盈风能源科技有限公司 | Wind power generation system, temperature rise early warning method of variable pitch motor of wind power generation system and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180285391A1 (en) * | 2017-04-02 | 2018-10-04 | Sas Institute Inc. | Methods, Mediums, and Systems for Data Harmonization and Data Mapping in Specified Domains |
| EP3620983A1 (en) * | 2018-09-05 | 2020-03-11 | Sartorius Stedim Data Analytics AB | Computer-implemented method, computer program product and system for data analysis |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111985711A (en) * | 2020-08-19 | 2020-11-24 | 华北电力大学（保定） | Wind power probability prediction model establishing method |
| CN112052426A (en) * | 2020-09-01 | 2020-12-08 | 国家电投集团江西电力有限公司 | Temperature rise fault early warning method for fan variable pitch motor |
| CN113420509A (en) * | 2021-07-07 | 2021-09-21 | 华能(浙江)能源开发有限公司清洁能源分公司 | Wind turbine state evaluation method and device and storage medium |
| CN114382662A (en) * | 2022-01-21 | 2022-04-22 | 华电安诺(北京)信息科技有限公司 | Fan state early warning method based on digital twinning |
-
2022
- 2022-07-04 CN CN202210787537.1A patent/CN115263680A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180285391A1 (en) * | 2017-04-02 | 2018-10-04 | Sas Institute Inc. | Methods, Mediums, and Systems for Data Harmonization and Data Mapping in Specified Domains |
| EP3620983A1 (en) * | 2018-09-05 | 2020-03-11 | Sartorius Stedim Data Analytics AB | Computer-implemented method, computer program product and system for data analysis |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111985711A (en) * | 2020-08-19 | 2020-11-24 | 华北电力大学（保定） | Wind power probability prediction model establishing method |
| CN112052426A (en) * | 2020-09-01 | 2020-12-08 | 国家电投集团江西电力有限公司 | Temperature rise fault early warning method for fan variable pitch motor |
| CN113420509A (en) * | 2021-07-07 | 2021-09-21 | 华能(浙江)能源开发有限公司清洁能源分公司 | Wind turbine state evaluation method and device and storage medium |
| CN114382662A (en) * | 2022-01-21 | 2022-04-22 | 华电安诺(北京)信息科技有限公司 | Fan state early warning method based on digital twinning |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842322A (en) * | 2023-07-19 | 2023-10-03 | 深圳市精微康投资发展有限公司 | Electric motor operation optimization method and system based on data processing |
| CN116842322B (en) * | 2023-07-19 | 2024-02-23 | 深圳市精微康投资发展有限公司 | Electric motor operation optimization method and system based on data processing |
| CN117212074A (en) * | 2023-09-25 | 2023-12-12 | 武汉盈风能源科技有限公司 | Wind power generation system, temperature rise early warning method of variable pitch motor of wind power generation system and storage medium |
| CN117212074B (en) * | 2023-09-25 | 2024-03-12 | 武汉盈风能源科技有限公司 | Wind power generation system, temperature rise early warning method of variable pitch motor of wind power generation system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105275742B (en) | A kind of control method of Wind turbines adaptive environment | |
| CN115263680A (en) | Abnormal temperature rise early warning method for variable pitch motor by combining TF-IDF model and LSTM model | |
| CN108932580A (en) | Wind turbines pitch variable bearings wear monitoring and method for early warning based on data modeling | |
| CN105257471B (en) | Wind power generating set award setting method, apparatus and system | |
| CN110412966B (en) | Method and device for monitoring temperature abnormity of variable pitch motor | |
| CN110067708B (en) | Method for identifying yaw wind disharmony by using power curve | |
| CN110907170B (en) | Wind turbine generator gearbox bearing temperature state monitoring and fault diagnosis method | |
| CN109973329B (en) | Method for judging temperature abnormity of frequency converter cabinet of wind driven generator cabin | |
| CN110094310B (en) | Method for identifying wind power generator yaw wind disharmony | |
| CN110939550B (en) | Monitoring method and early warning method for temperature abnormity of variable pitch motor | |
| CN103925155A (en) | Self-adaptive detection method for abnormal wind turbine output power | |
| CN111927717B (en) | System and method for online monitoring noise of transmission chain of fan engine room | |
| CN116624341A (en) | Wind turbine generator power abnormality diagnosis method, system and equipment | |
| CN117365869A (en) | Self-adaptive early warning strategy design method for wind turbine blade tower sweeping faults | |
| CN116879735A (en) | Temperature fault identification method and system for variable pitch motor of wind turbine generator | |
| CN116771610A (en) | Method for adjusting fault evaluation value of variable pitch system of wind turbine | |
| CN110578659B (en) | System and method for processing SCADA data of wind turbine generator | |
| CN114753980B (en) | Method and system for monitoring icing of fan blade | |
| CN110222393A (en) | The fan blade icing method for monitoring abnormality divided based on fine granularity wind-powered electricity generation generating state | |
| CN114623051A (en) | Intelligent identification and early warning method for icing state of wind power blade | |
| CN111794921B (en) | Onshore wind turbine blade icing diagnosis method based on migration component analysis | |
| EP3526471A1 (en) | Determining loads on a wind turbine | |
| CN115929569A (en) | Fault diagnosis method for variable pitch system of wind turbine generator | |
| CN114444291B (en) | Method, system, equipment and medium for finely measuring and calculating power generation loss of fan | |
| CN115062007A (en) | Wind turbine generator set wind speed and power data cleaning method based on isolated forest algorithm |
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 |