CN117738851A - Wind turbine generator system fault diagnosis method based on dynamic load - Google Patents
Wind turbine generator system fault diagnosis method based on dynamic load Download PDFInfo
- Publication number
- CN117738851A CN117738851A CN202311525723.9A CN202311525723A CN117738851A CN 117738851 A CN117738851 A CN 117738851A CN 202311525723 A CN202311525723 A CN 202311525723A CN 117738851 A CN117738851 A CN 117738851A
- Authority
- CN
- China
- Prior art keywords
- evaluation value
- monitoring
- generating
- fault
- time sequence
- 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
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000003745 diagnosis Methods 0.000 title claims abstract description 24
- 238000011156 evaluation Methods 0.000 claims abstract description 203
- 238000012544 monitoring process Methods 0.000 claims abstract description 196
- 238000013210 evaluation model Methods 0.000 claims abstract description 47
- 238000012937 correction Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 9
- 238000013527 convolutional neural network Methods 0.000 abstract description 6
- 238000013528 artificial neural network Methods 0.000 abstract description 4
- 125000004122 cyclic group Chemical group 0.000 abstract description 4
- 238000005299 abrasion Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
Abstract
The application relates to the technical field of wind turbines, in particular to a wind turbine fault diagnosis method based on dynamic load. Comprising the following steps: establishing a space evaluation model and a time sequence evaluation model; generating a plurality of monitoring points according to the wind turbine generator equipment parameters, and generating a feedback time node according to the wind turbine generator equipment parameters and the time sequence evaluation model; and generating a space evaluation value and a time sequence evaluation value of each monitoring point according to a preset feedback time node, and judging whether to generate fault early warning according to the space evaluation value and the time sequence evaluation value. And extracting spatial characteristic data of wind turbine generator equipment by using a convolutional neural network technology, and establishing a spatial evaluation model. And extracting time sequence characteristics of each piece of sub-equipment of the wind turbine generator by using a cyclic neural network technology, and establishing a time sequence evaluation model. And performing fault diagnosis on each piece of sub equipment through the space evaluation model and the time sequence evaluation model, and improving the accuracy of fault diagnosis.
Description
Technical Field
The application relates to the technical field of wind turbines, in particular to a wind turbine fault diagnosis method based on dynamic load.
Background
Wind power generation is a renewable energy technology, converts wind energy into electric energy, has the advantages of environmental protection, economy, sustainability and the like, and becomes one of the most popular clean energy at present. With the worldwide increasing demand for renewable energy sources, the installed capacity of wind turbines is expanding. The wind power integration grid-connected installed capacity scale of China continuously rises, and particularly the wind power integration grid-connected installed capacity scale is in the offshore wind power field. However, the wind turbine generator sets face various fault challenges in the long-term operation process, and the reliability, the economic benefit and the service life of equipment are seriously affected.
In wind turbines, dynamic loading is an important fault influencing factor. Dynamic load refers to the continuous change of the stress state of the wind turbine caused by climate condition factors such as wind speed, wind direction and the like. Dynamic loads may cause fatigue damage, vibration enhancement, and other potential safety hazards to various components in the wind turbine. At present, fault diagnosis is carried out on the wind turbine based on dynamic load, fault feature extraction and fault classification are adopted traditionally, a certain effect is achieved, and certain limitation exists in the aspects of processing complex dynamic load data and modeling dependency relationship. The accuracy and efficiency of wind turbine generator system fault diagnosis are lower.
Disclosure of Invention
The purpose of the present application is: in order to solve the technical problems, the application provides a wind turbine generator fault diagnosis method based on dynamic load, which aims to improve the accuracy of wind turbine generator fault diagnosis and early warn the fault risk of the wind turbine generator in time.
In some embodiments of the application, spatial feature data of wind turbine generator equipment is extracted by using a convolutional neural network technology, and a spatial evaluation model is established. And extracting time sequence characteristics of each piece of sub-equipment of the wind turbine generator by using a cyclic neural network technology, and establishing a time sequence evaluation model. And performing fault diagnosis on each piece of sub equipment through the space evaluation model and the time sequence evaluation model, and improving the accuracy of fault diagnosis.
In some embodiments of the application, the association relation among all the sub-devices is established by generating a space evaluation model, and the accuracy of fault diagnosis of all the monitoring points is improved by dynamically capturing the space characteristic parameters in the dynamic load data of the wind turbine generator through a preset feedback time node.
In some embodiments of the present application, a time sequence evaluation model is generated, a time sequence evaluation period of each piece of sub-equipment is established, and accumulated risks caused by running wear are fully considered by extracting time sequence characteristic parameters in dynamic load data, so that the ability of identifying potential faults is improved, and the fault risks are early warned in time. And ensuring the safe operation of the wind turbine generator.
In some embodiments of the present application, a wind turbine generator fault diagnosis method based on dynamic load is provided, including:
establishing a space evaluation model and a time sequence evaluation model;
generating a plurality of monitoring points according to the wind turbine generator equipment parameters, and generating a feedback time node according to the wind turbine generator equipment parameters and the time sequence evaluation model;
and generating a space evaluation value and a time sequence evaluation value of each monitoring point according to a preset feedback time node, and judging whether fault early warning is generated or not according to the space evaluation value and the time sequence evaluation value.
In some embodiments of the present application, the generating the plurality of monitoring points includes:
generating a monitoring point number array A, A= (a 1, a2 … an), wherein n is the number of monitoring points, and ai is the ith monitoring point;
setting characteristic data of each monitoring point;
sequentially selecting target monitoring points ai according to the monitoring point number sequence A;
generating each target monitoring point and residual monitoring points according to the space evaluation modelThe correlation degree series Bi, bi= (b) i1 ，b i2 …b in ) Wherein b ii The degree of association between the ith monitoring point and the target monitoring point ai.
In some embodiments of the present application, when generating the plurality of monitoring points, the method further includes:
presetting a first association threshold C1 and a second association threshold C2, wherein C1 is smaller than C2;
if b ii When the relation between the first relation threshold C1 and the second relation threshold C2 is set, the ith monitoring point and the target monitoring point ai are set as primary relation monitoring points;
if b ii And when the relation degree is larger than the second relation degree threshold C2, setting the ith monitoring point and the target monitoring point ai as secondary relation monitoring points.
In some embodiments of the present application, when generating the feedback time node, the method includes:
acquiring historical operation parameters of the wind turbine generator, and generating historical operation time and fault frequency according to the historical operation parameters;
generating a first historical evaluation value D1 according to the historical operation time length;
generating a second historical evaluation value D2 according to the fault frequency;
generating a monitoring evaluation value f according to the first history evaluation value D1 and the second history evaluation value D2;
f= e1+e2×d2, wherein e1 is a preset first weight coefficient, e2 is a preset second weight coefficient, and e1+e2=1;
and setting a time interval t between adjacent feedback time nodes according to the monitoring evaluation value f.
In some embodiments of the present application, setting the time interval t includes:
presetting a first monitoring evaluation value interval (F1, F2), a second monitoring evaluation value interval (F2, F3) and a third monitoring evaluation value interval (F3, F4);
if the monitoring evaluation value f is in the preset first monitoring evaluation value interval, setting the time interval as a preset first time interval t1, namely t=t1; if the monitoring evaluation value f is in the preset second monitoring evaluation value interval, setting the time interval as a preset second time interval t2, namely t=t2; if the monitoring evaluation value f is in the preset third monitoring evaluation value interval, the set time interval is a preset third time interval t3, namely t=t3; and t1 > t2 > t3.
In some embodiments of the present application, generating the spatial evaluation value includes;
acquiring monitoring data packets of all monitoring points according to the feedback time node, and preprocessing the monitoring data packets to generate feedback data packets;
generating initial evaluation values of all monitoring points according to the feedback data packet and the space evaluation model;
sequentially selecting target monitoring points, and correcting initial evaluation values of the target monitoring points according to feedback data packets of primary associated monitoring points and secondary associated monitoring points of the target monitoring points;
generating a space evaluation value g of the target monitoring point according to the correction result;
a spatial evaluation value array G, g= (G1, G2 … gn) is established, where gi is the spatial evaluation value of the i-th monitoring point.
In some embodiments of the present application, when generating the timing evaluation value, the method includes:
establishing a time sequence evaluation period of the monitoring point according to the current feedback time node;
acquiring corresponding feedback data packets of each feedback time node in a time sequence evaluation period;
generating a characteristic value fluctuation curve according to all the feedback data packets;
generating a time sequence evaluation value h of a current feedback time node of the monitoring point according to the characteristic value fluctuation curve and the time sequence evaluation model;
a time series evaluation value sequence H, h= (H1, H2 … hn) is established, wherein hi is the time series evaluation value of the i-th monitoring point.
In some embodiments of the present application, when establishing the time sequence evaluation period of the monitoring point, the method includes:
presetting a first space evaluation value interval (G1, G2), a second space evaluation value interval (G2, G3) and a third space evaluation value interval (G3, G4);
acquiring a space evaluation value gi of an ith monitoring point of a current feedback time node, and setting a time sequence evaluation period duration j of the ith monitoring point according to the space evaluation value gi;
if the space evaluation value g is in the preset first space evaluation value interval, setting the time sequence evaluation period duration j as a preset first time sequence evaluation period duration j1, namely j=j1; if the space evaluation value g is in the preset second space evaluation value interval, setting the time sequence evaluation period duration j as a preset second time sequence evaluation period duration j2, namely j=j2; if the space rating value g is in the preset third space rating value interval, setting the time sequence rating period duration j as a preset third time sequence rating period duration j3, namely j=j3, and j1 is less than j2 and less than j3.
In some embodiments of the present application, when determining whether to generate the fault early warning according to the spatial evaluation value and the time sequence evaluation value, the method includes:
acquiring a space evaluation value gi of an ith monitoring point, and generating a first fault probability K of the ith monitoring point according to the space evaluation value gi i1 ；
Acquiring a time sequence evaluation value hi of an ith monitoring point, and generating a second fault probability K of the ith monitoring point according to the space evaluation value hi i2 ；
According to the first failure probability K i1 And a second failure probability K i2 Generating a fault evaluation value Pi of an ith monitoring point;
pi＝e3＊K i1 +e4＊K i2 wherein e3 is a third weight coefficient, e4 is a fourth weight coefficient, and e3+e4=1;
and judging whether to generate fault early warning according to the fault evaluation value pi.
In some embodiments of the present application, when determining whether to generate the fault early warning according to the fault evaluation value Pi, the method includes:
presetting a first fault evaluation value threshold value P1 and a second fault evaluation value threshold value P2, wherein P1 is smaller than P2;
if pi is in a preset first fault evaluation value threshold value P1 and a preset second fault evaluation value threshold value P2s, generating a first-level fault early warning instruction of an ith monitoring point;
if pi is larger than a preset second fault evaluation value threshold value P2, generating a second fault early warning value instruction of the ith monitoring point;
and generating an overhaul plan according to all the fault early warning instructions.
Compared with the prior art, the wind turbine generator fault diagnosis method based on dynamic load has the beneficial effects that:
and extracting spatial characteristic data of wind turbine generator equipment by using a convolutional neural network technology, and establishing a spatial evaluation model. And extracting time sequence characteristics of each piece of sub-equipment of the wind turbine generator by using a cyclic neural network technology, and establishing a time sequence evaluation model. And performing fault diagnosis on each piece of sub equipment through the space evaluation model and the time sequence evaluation model, and improving the accuracy of fault diagnosis.
And establishing an association relation among all the sub-equipment by generating a space evaluation model, and dynamically capturing space characteristic parameters in dynamic load data of the wind turbine generator by a preset feedback time node, thereby improving the accuracy of fault diagnosis of all the monitoring points.
By generating a time sequence evaluation model, establishing a time sequence evaluation period of each piece of sub-equipment, and extracting time sequence characteristic parameters in dynamic load data, the accumulated risk caused by running abrasion is fully considered, so that the identification capability of fault hidden danger is improved, and the fault risk is early warned in time. And ensuring the safe operation of the wind turbine generator.
Drawings
Fig. 1 is a schematic flow chart of a wind turbine generator fault diagnosis method based on dynamic load in a preferred embodiment of the application.
Detailed Description
The detailed description of the present application is further described in detail below with reference to the drawings and examples. The following examples are illustrative of the present application, but are not intended to limit the scope of the present application.
In the description of the present application, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate description of the present application and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present application.
The terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present application, unless otherwise indicated, the meaning of "a plurality" is two or more.
In the description of the present application, it should be noted that, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the terms in this application will be understood by those of ordinary skill in the art in a specific context.
As shown in fig. 1, a wind turbine generator fault diagnosis method based on dynamic load according to a preferred embodiment of the present application includes:
s101: establishing a space evaluation model and a time sequence evaluation model;
s102: generating a plurality of monitoring points according to the wind turbine generator equipment parameters, and generating a feedback time node according to the wind turbine generator equipment parameters and the time sequence evaluation model;
s103: and generating a space evaluation value and a time sequence evaluation value of each monitoring point according to a preset feedback time node, and judging whether to generate fault early warning according to the space evaluation value and the time sequence evaluation value.
Specifically, a space evaluation model is established by using a convolutional neural network in processing space characteristic parameters of the wind turbine, and a time sequence evaluation model is established by using a convolutional neural network in processing time sequence characteristics in wind turbine data.
Specifically, when generating a plurality of monitoring points, the method includes:
generating a monitoring point number array A, A= (a 1, a2 … an), wherein n is the number of monitoring points, and ai is the ith monitoring point;
setting characteristic data of each monitoring point;
sequentially selecting target monitoring points ai according to the monitoring point sequence A;
generating a correlation degree array Bi, bi= (b) between each target monitoring point and the rest monitoring points according to the space evaluation model i1 ，b i2 …b in ) Wherein b ii The degree of association between the ith monitoring point and the target monitoring point ai.
Presetting a first association threshold C1 and a second association threshold C2, wherein C1 is smaller than C2;
if b ii When the relation between the first relation threshold C1 and the second relation threshold C2 is set, the ith monitoring point and the target monitoring point ai are set as primary relation monitoring points;
if b ii And when the relation degree is larger than the second relation degree threshold C2, setting the ith monitoring point and the target monitoring point ai as secondary relation monitoring points.
Specifically, corresponding monitoring points and dynamic load parameters to be collected are set according to the type of the fault to be diagnosed, for example, the rotating speed data of the wind turbine, the vibration signal data of the wind turbine and the power data of the wind turbine are collected. And acquiring corresponding dynamic load data by establishing a plurality of monitoring points.
Specifically, the monitoring points include, but are not limited to, fan blade monitoring points, bearing monitoring points, generator monitoring points, gearbox monitoring points, and the like.
Specifically, according to parameters of the monitoring points and the operation influence relation among the monitoring points, corresponding association degrees are generated, and the higher the association degree between the two monitoring points is, the greater the possibility of simultaneous faults between the two monitoring points is. And establishing an association relation network among the monitoring points according to the association degree among the monitoring points, and improving the efficiency of extracting the space characteristic parameters of the dynamic load data of each monitoring point, thereby improving the prediction accuracy of the space evaluation model.
In a preferred embodiment of the present application, when generating the feedback time node, the method includes:
acquiring historical operation parameters of the wind turbine generator, and generating historical operation time and fault frequency according to the historical operation parameters;
generating a first historical evaluation value D1 according to the historical operation time length;
generating a second historical evaluation value D2 according to the fault frequency;
generating a monitoring evaluation value f according to the first historical evaluation value D1 and the second historical evaluation value D2;
f= e1+e2×d2, wherein e1 is a preset first weight coefficient, e2 is a preset second weight coefficient, and e1+e2=1;
and setting a time interval t between adjacent feedback time nodes according to the monitoring evaluation value f.
Specifically, the value ranges of the first historical evaluation value and the second historical evaluation value are the same, the longer the historical operation time is, the higher the corresponding first historical evaluation value is, the higher the corresponding second historical evaluation value is, the higher the fault frequency is, and the higher the monitoring evaluation value is, namely the higher the probability of the fault of the wind turbine generator is.
Specifically, when the time interval t is set, it includes:
presetting a first monitoring evaluation value interval (F1, F2), a second monitoring evaluation value interval (F2, F3) and a third monitoring evaluation value interval (F3, F4);
if the monitoring evaluation value f is in the preset first monitoring evaluation value interval, setting the time interval as a preset first time interval t1, namely t=t1; if the monitoring evaluation value f is in the preset second monitoring evaluation value interval, setting the time interval as a preset second time interval t2, namely t=t2; if the monitoring evaluation value f is in the preset third monitoring evaluation value interval, the set time interval is a preset third time interval t3, namely t=t3; and t1 > t2 > t3.
Specifically, in the above embodiment, the monitoring evaluation value is generated according to the historical operation parameters of the wind turbine generator, the time interval between the feedback time nodes is dynamically adjusted according to the monitoring evaluation value, when the operation time of the wind turbine generator is longer, the probability of fault risk is greater, and the time interval between the feedback time nodes is shortened, so that the fault risk is early warned in time, the fault is removed, and the safe operation of the wind turbine generator is ensured.
In a preferred embodiment of the present application, the generating of the spatial evaluation value includes;
acquiring monitoring data packets of all monitoring points according to the feedback time node, and preprocessing the monitoring data packets to generate feedback data packets;
generating initial evaluation values of all monitoring points according to the feedback data packet and the space evaluation model;
sequentially selecting target monitoring points, and correcting initial evaluation values of the target monitoring points according to feedback data packets of primary associated monitoring points and secondary associated monitoring points of the target monitoring points;
generating a space evaluation value g of the target monitoring point according to the correction result;
a spatial evaluation value array G, g= (G1, G2 … gn) is established, where gi is the spatial evaluation value of the i-th monitoring point.
Specifically, feedback data packages of the secondary associated monitoring points and the primary associated monitoring points are obtained through a space evaluation model, and space characteristic data in the feedback data packages are extracted, so that an initial evaluation value is corrected, the mutual influence among all sub-devices is fully considered, and the space evaluation value is more accurate.
Specifically, the time series evaluation value generation includes:
establishing a time sequence evaluation period of the monitoring point according to the current feedback time node;
acquiring corresponding feedback data packets of each feedback time node in a time sequence evaluation period;
generating a characteristic value fluctuation curve according to all the feedback data packets;
generating a time sequence evaluation value h of the current feedback time node of the monitoring point according to the characteristic value fluctuation curve and the time sequence evaluation model;
a time series evaluation value sequence H, h= (H1, H2 … hn) is established, wherein hi is the time series evaluation value of the i-th monitoring point.
Specifically, a time sequence evaluation model is generated, the time sequence evaluation period of each piece of sub-equipment is established, and accumulated risks caused by running abrasion are fully considered by extracting time sequence characteristic parameters in dynamic load data, so that the identification capacity of fault hidden danger is improved, and the fault risks are early warned in time. And ensuring the safe operation of the wind turbine generator.
Specifically, when the time sequence evaluation period of the monitoring point is established, the method comprises the following steps:
presetting a first space evaluation value interval (G1, G2), a second space evaluation value interval (G2, G3) and a third space evaluation value interval (G3, G4);
acquiring a space evaluation value gi of an ith monitoring point of a current feedback time node, and setting a time sequence evaluation period duration j of the ith monitoring point according to the space evaluation value gi;
if the space evaluation value g is in the preset first space evaluation value interval, setting the time sequence evaluation period duration j as a preset first time sequence evaluation period duration j1, namely j=j1; if the space evaluation value g is in the preset second space evaluation value interval, setting the time sequence evaluation period duration j as a preset second time sequence evaluation period duration j2, namely j=j2; if the space rating value g is in the preset third space rating value interval, setting the time sequence rating period duration j as a preset third time sequence rating period duration j3, namely j=j3, and j1 is less than j2 and less than j3.
Specifically, the current feedback time node is taken as an important point, the corresponding historical feedback time node is selected according to the time sequence evaluation period, the time length of the time sequence evaluation period is dynamically adjusted through the space evaluation value of the monitoring point, when the space evaluation value is higher, the fault risk of the current monitoring point is larger, more historical feedback time node data are acquired when the time sequence evaluation is carried out, the identification capability of fault hidden danger is improved, and the fault risk is early warned in time. And ensuring the safe operation of the wind turbine generator.
In a preferred embodiment of the present application, when judging whether to generate a fault early warning according to the spatial evaluation value and the time sequence evaluation value, the method includes:
acquiring a space evaluation value gi of the ith monitoring point according to the spaceThe evaluation value gi generates a first failure probability K of the ith monitoring point i1 ；
Acquiring a time sequence evaluation value hi of the ith monitoring point, and generating a second fault probability K of the ith monitoring point according to the space evaluation value hi i2 ；
According to the first failure probability K i1 And a second failure probability K i2 Generating a fault evaluation value Pi of an ith monitoring point;
pi＝e3＊K i1 +e4＊K i2 wherein e3 is a third weight coefficient, e4 is a fourth weight coefficient, and e3+e4=1;
and judging whether to generate fault early warning according to the fault evaluation value pi.
Specifically, the third weight coefficient and the fourth weight coefficient can be set according to the equipment parameters of the monitoring points, if the possibility of the equipment failing due to self abrasion is high, the third weight coefficient should be smaller than the fourth weight coefficient, and if the possibility of the equipment failing when other equipment fails is high, the third weight coefficient should be larger than the fourth weight coefficient.
Specifically, when judging whether to generate a fault early warning according to the fault evaluation value Pi, the method includes:
presetting a first fault evaluation value threshold value P1 and a second fault evaluation value threshold value P2, wherein P1 is smaller than P2;
if pi is in a preset first fault evaluation value threshold value P1 and a preset second fault evaluation value threshold value P2s, generating a first-level fault early warning instruction of an ith monitoring point;
if pi is larger than a preset second fault evaluation value threshold value P2, generating a second fault early warning value instruction of the ith monitoring point;
and generating an overhaul plan according to all the fault early warning instructions.
Specifically, the primary fault early warning instruction means that the current equipment has fault risk, overhauling should be carried out on the same day, and in the monthly overhauling, the secondary associated monitoring point of the current monitoring point is subjected to key overhauling. The secondary fault early warning instruction means that the current equipment has larger operation risk, faults possibly occur at any time, shutdown maintenance needs to be carried out immediately, primary association monitoring points and secondary association monitoring points of the monitoring points need to be subjected to key maintenance in monthly maintenance, fault hidden dangers are removed timely, and safe operation of the wind turbine generator is guaranteed.
According to the first conception, the space characteristic data of the wind turbine generator equipment are extracted by utilizing a convolutional neural network technology, and a space evaluation model is established. And extracting time sequence characteristics of each piece of sub-equipment of the wind turbine generator by using a cyclic neural network technology, and establishing a time sequence evaluation model. And performing fault diagnosis on each piece of sub equipment through the space evaluation model and the time sequence evaluation model, and improving the accuracy of fault diagnosis.
According to the second conception, the association relation among all the sub-equipment is established by generating the space evaluation model, and the space characteristic parameters in the dynamic load data of the wind turbine generator are dynamically captured by presetting the feedback time node, so that the fault diagnosis accuracy of all the monitoring points is improved.
According to the third conception, the time sequence evaluation period of each piece of sub equipment is established by generating the time sequence evaluation model, and the accumulated risk caused by running abrasion is fully considered by extracting the time sequence characteristic parameters in dynamic load data, so that the identification capability of fault hidden danger is improved, and the fault risk is early warned in time. And ensuring the safe operation of the wind turbine generator.
The foregoing is merely a preferred embodiment of the present application, and it should be noted that modifications and substitutions can be made by those skilled in the art without departing from the technical principles of the present application, and these modifications and substitutions should also be considered as being within the scope of the present application.
Claims (10)
1. A wind turbine generator system fault diagnosis method based on dynamic load is characterized by comprising the following steps:
establishing a space evaluation model and a time sequence evaluation model;
generating a plurality of monitoring points according to the wind turbine generator equipment parameters, and generating a feedback time node according to the wind turbine generator equipment parameters and the time sequence evaluation model;
and generating a space evaluation value and a time sequence evaluation value of each monitoring point according to a preset feedback time node, and judging whether fault early warning is generated or not according to the space evaluation value and the time sequence evaluation value.
2. The method for diagnosing a wind turbine generator fault based on dynamic load according to claim 1, wherein the generating the plurality of monitoring points comprises:
generating a monitoring point number array A, A= (a 1, a2 … an), wherein n is the number of monitoring points, and ai is the ith monitoring point;
setting characteristic data of each monitoring point;
sequentially selecting target monitoring points ai according to the monitoring point number sequence A;
generating a correlation degree array Bi, bi= (b) between each target monitoring point and the rest monitoring points according to the space evaluation model i1 ，b i2 …b in ) Wherein b ii The degree of association between the ith monitoring point and the target monitoring point ai.
3. The method for diagnosing a wind turbine generator fault based on dynamic load according to claim 2, wherein when generating the plurality of monitoring points, the method further comprises:
presetting a first association threshold C1 and a second association threshold C2, wherein C1 is smaller than C2;
if b ii When the relation between the first relation threshold C1 and the second relation threshold C2 is set, the ith monitoring point and the target monitoring point ai are set as primary relation monitoring points;
if b ii And when the relation degree is larger than the second relation degree threshold C2, setting the ith monitoring point and the target monitoring point ai as secondary relation monitoring points.
4. The method for diagnosing a wind turbine generator fault based on dynamic load as claimed in claim 3, wherein when generating the feedback time node, the method comprises:
acquiring historical operation parameters of the wind turbine generator, and generating historical operation time and fault frequency according to the historical operation parameters;
generating a first historical evaluation value D1 according to the historical operation time length;
generating a second historical evaluation value D2 according to the fault frequency;
generating a monitoring evaluation value f according to the first history evaluation value D1 and the second history evaluation value D2;
f= e1+e2×d2, wherein e1 is a preset first weight coefficient, e2 is a preset second weight coefficient, and e1+e2=1;
and setting a time interval t between adjacent feedback time nodes according to the monitoring evaluation value f.
5. The method for diagnosing a wind turbine generator system fault based on dynamic load as claimed in claim 4, wherein the setting of the time interval t comprises:
presetting a first monitoring evaluation value interval (F1, F2), a second monitoring evaluation value interval (F2, F3) and a third monitoring evaluation value interval (F3, F4);
if the monitoring evaluation value f is in the preset first monitoring evaluation value interval, setting the time interval as a preset first time interval t1, namely t=t1; if the monitoring evaluation value f is in the preset second monitoring evaluation value interval, setting the time interval as a preset second time interval t2, namely t=t2; if the monitoring evaluation value f is in the preset third monitoring evaluation value interval, the set time interval is a preset third time interval t3, namely t=t3; and t1 > t2 > t3.
6. The method for diagnosing a wind turbine generator system fault based on dynamic load as claimed in claim 4, wherein the generating of the spatial evaluation value comprises;
acquiring monitoring data packets of all monitoring points according to the feedback time node, and preprocessing the monitoring data packets to generate feedback data packets;
generating initial evaluation values of all monitoring points according to the feedback data packet and the space evaluation model;
sequentially selecting target monitoring points, and correcting initial evaluation values of the target monitoring points according to feedback data packets of primary associated monitoring points and secondary associated monitoring points of the target monitoring points;
generating a space evaluation value g of the target monitoring point according to the correction result;
a spatial evaluation value array G, g= (G1, G2 … gn) is established, where gi is the spatial evaluation value of the i-th monitoring point.
7. The method for diagnosing a wind turbine generator system fault based on dynamic load as claimed in claim 6, wherein generating the time series evaluation value comprises:
establishing a time sequence evaluation period of the monitoring point according to the current feedback time node;
acquiring corresponding feedback data packets of each feedback time node in a time sequence evaluation period;
generating a characteristic value fluctuation curve according to all the feedback data packets;
generating a time sequence evaluation value h of a current feedback time node of the monitoring point according to the characteristic value fluctuation curve and the time sequence evaluation model;
a time series evaluation value sequence H, h= (H1, H2 … hn) is established, wherein hi is the time series evaluation value of the i-th monitoring point.
8. The method for diagnosing a wind turbine generator system fault based on dynamic load as claimed in claim 7, wherein when the time sequence evaluation period of the monitoring point is established, the method comprises:
presetting a first space evaluation value interval (G1, G2), a second space evaluation value interval (G2, G3) and a third space evaluation value interval (G3, G4);
acquiring a space evaluation value gi of an ith monitoring point of a current feedback time node, and setting a time sequence evaluation period duration j of the ith monitoring point according to the space evaluation value gi;
if the space evaluation value g is in the preset first space evaluation value interval, setting the time sequence evaluation period duration j as a preset first time sequence evaluation period duration j1, namely j=j1; if the space evaluation value g is in the preset second space evaluation value interval, setting the time sequence evaluation period duration j as a preset second time sequence evaluation period duration j2, namely j=j2; if the space rating value g is in the preset third space rating value interval, setting the time sequence rating period duration j as a preset third time sequence rating period duration j3, namely j=j3, and j1 is less than j2 and less than j3.
9. The method for diagnosing a wind turbine generator fault based on dynamic load according to claim 8, wherein the determining whether to generate a fault pre-warning according to the spatial evaluation value and the time sequence evaluation value comprises:
acquiring a space evaluation value gi of an ith monitoring point, and generating a first fault probability K of the ith monitoring point according to the space evaluation value gi i1 ；
Acquiring a time sequence evaluation value hi of an ith monitoring point, and generating a second fault probability K of the ith monitoring point according to the space evaluation value hi i2 ；
According to the first failure probability K i1 And a second failure probability K i2 Generating a fault evaluation value Pi of an ith monitoring point;
pi＝e3＊K i1 +e4＊K i2 wherein e3 is a third weight coefficient, e4 is a fourth weight coefficient, and e3+e4=1;
and judging whether to generate fault early warning according to the fault evaluation value pi.
10. The method for diagnosing a wind turbine generator system fault based on dynamic load according to claim 9, wherein the determining whether to generate the fault early warning according to the fault evaluation value Pi comprises:
presetting a first fault evaluation value threshold value P1 and a second fault evaluation value threshold value P2, wherein P1 is smaller than P2;
if pi is in a preset first fault evaluation value threshold value P1 and a preset second fault evaluation value threshold value P2s, generating a first-level fault early warning instruction of an ith monitoring point;
if pi is larger than a preset second fault evaluation value threshold value P2, generating a second fault early warning value instruction of the ith monitoring point;
and generating an overhaul plan according to all the fault early warning instructions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311525723.9A CN117738851A (en) | 2023-11-15 | 2023-11-15 | Wind turbine generator system fault diagnosis method based on dynamic load |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311525723.9A CN117738851A (en) | 2023-11-15 | 2023-11-15 | Wind turbine generator system fault diagnosis method based on dynamic load |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117738851A true CN117738851A (en) | 2024-03-22 |
Family
ID=90278474
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311525723.9A Pending CN117738851A (en) | 2023-11-15 | 2023-11-15 | Wind turbine generator system fault diagnosis method based on dynamic load |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117738851A (en) |
-
2023
- 2023-11-15 CN CN202311525723.9A patent/CN117738851A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Besnard et al. | A model for the optimization of the maintenance support organization for offshore wind farms | |
| CN108072524B (en) | Wind turbine generator gearbox bearing fault early warning method | |
| Besnard et al. | Reliability-Centred Asset Maintenance—A step towards enhanced reliability, availability, and profitability of wind power plants | |
| CN108335021B (en) | Method for evaluating state reliability of wind turbine and maintenance decision optimization | |
| CN116187593B (en) | Power distribution network fault prediction processing method, device, equipment and storage medium | |
| CN103925155B (en) | The self-adapting detecting method that a kind of Wind turbines output is abnormal | |
| Hu et al. | Using domain knowledge features for wind turbine diagnostics | |
| CN109188227A (en) | A kind of double feed wind power generator Condition assessment of insulation method and system | |
| CN107654342A (en) | A kind of abnormal detection method of Wind turbines power for considering turbulent flow | |
| El-Metwally et al. | Reliability assessment of wind turbine operating concepts using reliability block diagrams (RBDs) | |
| CN114412726A (en) | Wind turbine generator system variable pitch bearing detection method and device and wind turbine generator system | |
| CN108074045B (en) | Wind turbine generator complete machine vulnerability analysis and fault sequencing method and electronic terminal | |
| Botsaris et al. | Systemic assessment and analysis of factors affect the reliability of a wind turbine | |
| CN114004991A (en) | Fault identification method and device for wind turbine generator | |
| May et al. | Condition based maintenance for offshore wind turbines: the effects of false alarms from condition monitoring systems | |
| CN117149498A (en) | Power plant fault diagnosis method and system | |
| CN117738851A (en) | Wind turbine generator system fault diagnosis method based on dynamic load | |
| CN114047442B (en) | Fault ride-through detection method, device and system for new energy station | |
| CN115345370A (en) | Vine Copula model-based wind turbine generator operation linkage analysis method | |
| CN113848347A (en) | Health state detection method for wind meter of wind driven generator | |
| Ye et al. | Using SCADA data fusion by swarm intelligence for wind turbine condition monitoring | |
| CN112699598A (en) | Intelligent diagnosis method and device for abnormal oil temperature of gear box | |
| Fazli et al. | Wind turbine fault detection and isolation robust against data imbalance using KNN | |
| CN114046228B (en) | Wind turbine generator abnormality diagnosis method and system | |
| CN116502543B (en) | Wind turbine generator set wind deviation recognition method and device, electronic equipment and storage medium |
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 |