Method and device for monitoring humidity of wind turbine generator set
Through multi-sensor data analysis and machine learning models, the faults and abnormalities of humidity sensors in wind turbines are identified, and the accuracy and positioning problems of humidity monitoring in the existing technology are solved, and the monitoring effect and operation and maintenance efficiency are improved.
Patent Information
- Application Number
- CN202210187150.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-02-28
- Publication Date
- 2024-12-20
- Estimated Expiration
- 2042-02-28
AI Technical Summary
When monitoring the humidity of wind turbines, the humidity sensor failure leads to monitoring failure, insufficient accuracy, and inability to accurately locate the cause of humidity abnormality.
By obtaining multiple humidity sensor data of the wind turbine, determine whether the sensor is faulty, and based on the data of the unfailed sensor, extract the feature vector and input it into the pre-trained model to identify humidity abnormalities and causes, including average, maximum, minimum, standard Statistics such as difference, variance, temperature bin statistics and abnormal data share.
It effectively avoids monitoring failure caused by humidity sensor failure, improves the accuracy of humidity monitoring, and accurately locates the causes of humidity abnormalities, such as dehumidifier failure or the hatch door not closed, reducing operation and maintenance costs and power generation losses.
Smart Images
Figure CN114562431B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure generally relates to the field of electric power technology, and more specifically, to a method and device for monitoring humidity of a wind turbine generator set. Background Art
[0002] There are abundant wind resources at sea and in coastal areas, but the environmental humidity is high, which is easy to corrode wind turbine equipment. In addition, offshore operation and maintenance are difficult and costly, so the humidity control requirements for offshore units are high. The unit mainly adjusts the operating environment of the equipment through the environmental control system. In addition to heating and heat dissipation equipment, the environmental control system of the offshore unit will also be equipped with dehumidification equipment. When the humidity in the unit reaches a certain value, the dehumidification equipment is started to reduce the humidity in the unit. If the dehumidifier fails or the hatch and cabinet doors of the unit are not closed in time, the internal components of the unit will be exposed to a high humidity environment for a long time, and condensation is prone to occur on the surface of the component, causing corrosion, resulting in short circuit or damage to the component, and also affecting its insulation effect and service life, reducing the performance of the component, and even causing the entire unit to shut down. If the humidity is too low, static electricity and other problems are prone to occur.
[0003] Monitoring the humidity of the unit allows on-site maintenance personnel to promptly discover and handle problems, prevent components from working in abnormal humidity environments for a long time, reduce component damage, ensure the smooth operation of the unit, and reduce the costs and power generation losses caused by component damage, replacement, and downtime. At present, the existing technology mainly monitors the humidity in the unit through humidity sensors. When the humidity reaches a fixed limit, the dehumidification equipment is started to dehumidify; when the humidity in the unit is too high for a long time, an abnormal humidity fault is reported.
[0004] However, the accuracy of existing humidity monitoring methods needs to be further improved. Summary of the invention
[0005] An exemplary embodiment of the present disclosure provides a method and device for monitoring humidity of a wind turbine generator set, which can accurately monitor the humidity condition of the wind turbine generator set.
[0006] According to a first aspect of an embodiment of the present disclosure, there is provided a method for monitoring the humidity of a wind turbine generator set, comprising: acquiring operating data of the wind turbine generator set in a current time period, wherein the operating data comprises: humidity data of at least one humidity sensor, different types of the at least one humidity sensor being used to detect the humidity of different parts of the wind turbine generator set; determining whether each humidity sensor fails based on the humidity data of each humidity sensor of the wind turbine generator set; and determining whether the humidity of the part detected by the humidity sensor that does not fail is abnormal based on the humidity data of the humidity sensors of the wind turbine generator set that does not fail.
[0007] Optionally, based on the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set, the step of determining whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal includes: based on the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set, determining whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal and the cause of the abnormality.
[0008] Optionally, the step of determining whether the humidity of the part detected by the non-faulty humidity sensor of the wind turbine generator set is abnormal based on the humidity data of the non-faulty humidity sensor of the wind turbine generator set includes: determining whether the humidity of the part detected by the non-faulty humidity sensor is abnormal based on the humidity data of the non-faulty humidity sensor of the wind turbine generator set by referring to the humidity data of all wind turbine generator sets in the current time period.
[0009] Optionally, based on the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set, the step of determining whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal includes: extracting the feature vector of the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set by referring to the humidity data of all the wind turbine generator sets in the current time period; inputting the extracted feature vector into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal, or a prediction result of whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal and the cause of the abnormality.
[0010] Optionally, the operating data also includes: ambient temperature data; wherein the characteristic vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed includes: at least one of the following statistical quantities of the humidity data of each humidity sensor that has not failed: mean value; maximum value; minimum value; median; standard deviation; variance; the average humidity value in each temperature interval obtained by temperature bin statistics; the proportion of abnormal data exceeding the full-field humidity threshold range corresponding to this humidity sensor; wherein the full-field humidity threshold range corresponding to this humidity sensor is based on: the average value of the humidity data of this humidity sensor of all wind turbine generator sets that have not failed within the current time period.
[0011] Optionally, based on the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set, the step of determining whether the humidity of the part detected by the unmalfunctioning humidity sensor is abnormal also includes: filtering out humidity data under a specific operating state from the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set; resampling the filtered humidity data; wherein, the step of extracting the feature vector of the humidity data of the unmalfunctioning humidity sensor of the wind turbine generator set by referring to the humidity data of the entire wind turbine generator sets in the current time period includes: extracting the feature vector of the resampled humidity data by referring to the humidity data of the entire wind turbine generator sets in the current time period.
[0012] Optionally, based on the humidity data of each humidity sensor of the wind turbine generator set, the step of determining whether each humidity sensor fails includes: based on the humidity data of each humidity sensor of the wind turbine generator set, determining at least one of the range value of the humidity data, the abnormal humidity data amount of the humidity data, and the abnormal jump data amount of the humidity data; when at least one of the range value is less than the dead value threshold corresponding to the humidity sensor, the abnormal humidity data amount exceeds the abnormal humidity data amount threshold corresponding to the humidity sensor, and the abnormal jump data amount exceeds the abnormal jump data amount threshold corresponding to the humidity sensor, it is determined that the humidity sensor of the wind turbine generator set has failed.
[0013] Optionally, the different parts include at least one of the following items: a nacelle, a hub, a variable flow layer, and a control layer.
[0014] Optionally, it also includes: a humidity sensor that prompts the user that the wind turbine generator set has a fault; and prompts the user that the part of the wind turbine generator set with abnormal humidity and the cause of the abnormality.
[0015] Optionally, the abnormal cause includes at least one of the following items: dehumidifier failure, cabinet door or cabin door not closed in time.
[0016] According to a second aspect of an embodiment of the present disclosure, there is provided a device for monitoring the humidity of a wind turbine generator set, comprising: a data acquisition unit, configured to acquire operating data of the wind turbine generator set in a current time period, wherein the operating data comprises: humidity data of at least one humidity sensor, different types of humidity sensors in the at least one humidity sensor are used to detect the humidity of different parts of the wind turbine generator set; a fault identification unit, configured to determine whether a fault occurs in each humidity sensor of the wind turbine generator set based on the humidity data of each humidity sensor of the wind turbine generator set; and a humidity abnormality identification unit, configured to determine whether the humidity of the part detected by the humidity sensor of the wind turbine generator set that has not failed is abnormal based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed.
[0017] Optionally, the humidity anomaly identification unit is configured to: determine whether the humidity of a location detected by a humidity sensor that has not failed is abnormal and the cause of the abnormality based on humidity data of the humidity sensor that has not failed of the wind turbine generator set.
[0018] Optionally, the humidity anomaly identification unit is configured to determine whether the humidity of the part detected by the non-faulty humidity sensor of the wind turbine generator set is abnormal based on the humidity data of the non-faulty humidity sensor of the wind turbine generator set by referring to the humidity data of the whole field wind turbine generator set in the current time period.
[0019] Optionally, the humidity anomaly identification unit is configured to: extract the feature vector of the humidity data of the non-faulty humidity sensor of the wind turbine by referring to the humidity data of the entire wind turbine generator set in the current time period; input the extracted feature vector into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal, or a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal and the cause of the abnormality.
[0020] Optionally, the operating data also includes: ambient temperature data; wherein the characteristic vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed includes: at least one of the following statistical quantities of the humidity data of each humidity sensor that has not failed: mean value; maximum value; minimum value; median; standard deviation; variance; the average humidity value in each temperature interval obtained by temperature bin statistics; the proportion of abnormal data exceeding the full-field humidity threshold range corresponding to this humidity sensor; wherein the full-field humidity threshold range corresponding to this humidity sensor is based on: the average value of the humidity data of this humidity sensor of all wind turbine generator sets that have not failed within the current time period.
[0021] Optionally, the humidity anomaly identification unit is further configured to: filter out humidity data under a specific operating state from the humidity data of the humidity sensors of the wind turbine generator set that are not faulty; and resample the filtered humidity data; wherein the humidity anomaly identification unit is configured to: extract a feature vector of the resampled humidity data by referring to the humidity data of all wind turbine generator sets in the current time period.
[0022] Optionally, the fault identification unit is configured to: determine at least one of the range value of the humidity data, the abnormal humidity data amount of the humidity data, and the abnormal jump data amount of the humidity data based on the humidity data of each humidity sensor of the wind turbine generator set; when at least one of the range value is less than the dead value threshold corresponding to the humidity sensor, the abnormal humidity data amount exceeds the abnormal humidity data amount threshold corresponding to the humidity sensor, and the abnormal jump data amount exceeds the abnormal jump data amount threshold corresponding to the humidity sensor, it is determined that the humidity sensor of the wind turbine generator set has a fault.
[0023] Optionally, the different parts include at least one of the following items: a nacelle, a hub, a variable flow layer, and a control layer.
[0024] Optionally, it also includes: an output unit configured to prompt a user that a humidity sensor of the wind turbine generator set has failed; and prompt a user that a portion of the wind turbine generator set with abnormal humidity and a cause of the abnormality.
[0025] Optionally, the abnormal cause includes at least one of the following items: dehumidifier failure, cabinet door or cabin door not closed in time.
[0026] According to a third aspect of an embodiment of the present disclosure, a computer-readable storage medium storing a computer program is provided. When the computer program is executed by a processor, the processor is prompted to execute the method for monitoring humidity of a wind turbine generator set as described above.
[0027] According to a fourth aspect of an embodiment of the present disclosure, a controller of a wind turbine is provided, the controller comprising: a processor; and a memory storing a computer program, wherein when the computer program is executed by the processor, the processor is prompted to execute the method for monitoring the humidity of the wind turbine as described above.
[0028] According to a fifth aspect of an embodiment of the present disclosure, a controller of a wind farm is provided, the controller comprising: a processor; and a memory storing a computer program, which, when executed by the processor, prompts the processor to execute the method for monitoring humidity of a wind turbine generator set as described above.
[0029] According to the method and device for monitoring humidity of a wind turbine generator set according to the exemplary embodiment of the present disclosure, it is possible to effectively avoid humidity monitoring failure caused by humidity sensor failure, improve the accuracy of humidity monitoring of the wind turbine generator set, and accurately locate the cause of humidity abnormality.
[0030] Additional aspects and / or advantages of the present general inventive concept will be set forth in part in the following description and in part will be apparent from the description or may be learned through practice of the present general inventive concept. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] The above and other objects and features of the exemplary embodiments of the present disclosure will become more apparent through the following description in conjunction with the accompanying drawings which exemplarily illustrate the embodiments, in which:
[0032] Figure 1 A flow chart showing a method for monitoring humidity of a wind turbine generator system according to an exemplary embodiment of the present disclosure;
[0033] Figure 2 A flow chart showing a method for monitoring humidity of a wind turbine generator system according to another exemplary embodiment of the present disclosure;
[0034] Figure 3 A flowchart showing a method of determining whether a humidity sensor fails according to an exemplary embodiment of the present disclosure;
[0035] Figure 4 A flowchart showing a method of determining whether the humidity of a portion detected by a humidity sensor that has not failed is abnormal according to an exemplary embodiment of the present disclosure;
[0036] Figure 5 A histogram showing cabin humidity distribution according to an exemplary embodiment of the present disclosure;
[0037] Figure 6 An example of obtaining a humidity warning boundary according to an exemplary embodiment of the present disclosure is shown;
[0038] Figure 7 An example of a cabin humidity sensor failure is shown according to an exemplary embodiment of the present disclosure;
[0039] Figure 8 An example of abnormal cabin humidity according to an exemplary embodiment of the present disclosure is shown;
[0040] Fig. 9 A structural block diagram of a device for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0041] Reference will now be made in detail to the embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like parts throughout. The embodiments will be described below with reference to the drawings in order to explain the present disclosure.
[0042] Figure 1 A flow chart showing a method for monitoring humidity of a wind turbine according to an exemplary embodiment of the present disclosure.
[0043] As an example, the method for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure may be executed by a controller of a wind farm (ie, a wind farm controller), or may be executed by a controller of a wind turbine generator set.
[0044] Reference Figure 1 , in step S10, the operation data of the wind turbine generator set in the current time period is obtained.
[0045] The operation data includes: humidity data of at least one humidity sensor, wherein different types of humidity sensors in the at least one humidity sensor are used to detect humidity at different parts of the wind turbine generator set. The humidity data of the humidity sensor is humidity data detected by the humidity sensor.
[0046] As an example, the different parts may include at least one of the following items: a nacelle, a hub, a converter layer (for example, may include a converter), and a control layer (for example, may include a controller). It should be understood that other parts of the wind turbine generator set that need to monitor humidity may also be included, and the present disclosure will not repeat them here.
[0047] It should be understood that different types of humidity sensors are divided according to the parts they detect. For example, different types of humidity sensors may include at least one of the following items: a humidity sensor for detecting cabin humidity (i.e., cabin humidity sensor), a humidity sensor for detecting hub humidity (i.e., hub humidity sensor), a humidity sensor for detecting variable layer humidity (i.e., variable layer humidity sensor), and a humidity sensor for detecting control layer humidity (i.e., control layer humidity sensor).
[0048] As an example, the current time period may be the last N days or the last M hours, where N and M are numbers greater than 0.
[0049] In step S20, based on the humidity data of each humidity sensor of the wind turbine generator set, it is determined whether each humidity sensor fails.
[0050] As an example, based on the humidity data of each humidity sensor of the wind turbine generator set, at least one of the range value of the humidity data, the abnormal humidity data amount of the humidity data, and the abnormal jump data amount of the humidity data can be determined; when at least one of the range value is less than the dead value threshold corresponding to the humidity sensor, the abnormal humidity data amount exceeds the abnormal humidity data amount threshold corresponding to the humidity sensor, and the abnormal jump data amount exceeds the abnormal jump data amount threshold corresponding to the humidity sensor, it is determined that the humidity sensor of the wind turbine generator set has failed.
[0051] As an example, the range value of the humidity data of the humidity sensor is the difference between the maximum value and the minimum value in the humidity data.
[0052] As an example, the amount of abnormal humidity data of the humidity data of the humidity sensor is the number of humidity values in the humidity data that exceed the humidity threshold range corresponding to the humidity sensor.
[0053] As an example, the amount of abnormal jump data of the humidity data of a humidity sensor is the number of differential values among all differential values of the humidity data that exceed the differential threshold range corresponding to the humidity sensor; wherein, all differential values of the humidity data include: the difference between the humidity values of every two adjacent detection time points in the humidity data, that is, the difference between the humidity value detected at each detection time point in the humidity data and the humidity value detected at the previous detection time point.
[0054] It should be understood that whether the humidity sensor fails can also be determined based on other types of statistics of the humidity data of the humidity sensor. In addition, the dead value thresholds of different types of humidity sensors can be the same or different; the abnormal humidity data volume thresholds of different types of humidity sensors can be the same or different; the abnormal jump data volume thresholds of different types of humidity sensors can be the same or different; the humidity threshold ranges corresponding to different types of humidity sensors can be the same or different; the differential threshold ranges corresponding to different types of humidity sensors can be the same or different.
[0055] The following will be combined Figure 3 An exemplary embodiment of step S20 is described below.
[0056] In step S30, based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, it is determined whether the humidity of the part detected by the humidity sensor that has not failed is abnormal.
[0057] As an example, based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, it can be determined whether the humidity of the part detected by the humidity sensor that has not failed is abnormal and the cause of the abnormality.
[0058] As an example, the abnormal reason may include at least one of the following items: dehumidifier failure, cabinet door or cabin door not closed in time. It should be understood that other types of abnormal reasons may also be included, which will not be described in detail in this disclosure.
[0059] As an example, by referring to the humidity data of all wind turbines in the field (for example, all wind turbines of the same type in the field) during the current time period, it is possible to determine whether the humidity of the part detected by the non-faulty humidity sensor of the wind turbine is abnormal based on the humidity data of the non-faulty humidity sensor of the wind turbine.
[0060] As an example, step S30 may include: extracting the feature vector of the humidity data of the non-faulty humidity sensor of the wind turbine by referring to the humidity data of the entire wind turbine in the current time period; and inputting the extracted feature vector into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal, or a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal and the cause of the abnormality.
[0061] In addition, as an example, the extracted feature vector is input into a pre-trained humidity monitoring model to predict the future humidity of the part detected by the humidity sensor that has not failed, so as to achieve early warning.
[0062] As an example, the humidity monitoring model can be constructed based on the XGBoost algorithm. It should be understood that other suitable machine learning algorithms can also be used to construct the humidity monitoring model.
[0063] As an example, the operating data may also include: ambient temperature data. As an example, the feature vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed may include: at least one of the following statistics of the humidity data of each humidity sensor that has not failed: mean value; maximum value; minimum value; standard deviation; median; variance; average humidity value in each temperature interval obtained by temperature bin statistics; the proportion of abnormal data exceeding the full-field humidity threshold range corresponding to this humidity sensor; wherein the full-field humidity threshold range corresponding to this humidity sensor is based on: the average value of the humidity data of this humidity sensor of all wind turbine generator sets that have not failed in the current time period. It should be understood that other types of statistics may also be included, and the present disclosure will not be repeated here.
[0064] As an example, step S30 may also include: filtering out humidity data under a specific operating state from the humidity data of the humidity sensors of the wind turbine generator set that have not failed; resampling the filtered humidity data; wherein the step of extracting the feature vector of the humidity data of the humidity sensors of the wind turbine generator set that have not failed by referring to the humidity data of the entire wind turbine generator set in the current time period includes: extracting the feature vector of the resampled humidity data by referring to the humidity data of the entire wind turbine generator set in the current time period. Specifically, the feature vector of the resampled humidity data of the wind turbine generator set may be extracted by referring to the humidity data of the humidity sensors of the entire wind turbine generator set that have not failed in the current time period after screening and resampling.
[0065] As an example, the operation data may also include: an active power value; wherein the specific operation state may include: an operation state in which the active power value is greater than the minimum active power of the grid-connected power generation state. It should be understood that the specific operation state may also be set according to actual conditions and specific needs.
[0066] The following will be combined Figure 4 An exemplary embodiment of step S30 is described below.
[0067] In addition, as an example, the method for monitoring the humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure may further include: prompting a user that a humidity sensor of the wind turbine generator set has failed. For example, identification information of the humidity sensor of the wind turbine generator set that has failed may be displayed through a visual interface.
[0068] In addition, as an example, the method for monitoring the humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure may further include: prompting a user of the abnormal humidity portion of the wind turbine generator set and the abnormal reason. For example, the identification information of the abnormal humidity portion of the wind turbine generator set and the abnormal reason may be displayed through a visual interface.
[0069] The existing unit humidity monitoring method is completely dependent on the humidity sensor. When the humidity sensor fails, it will cause the dehumidifier to work abnormally and the humidity monitoring to report a false alarm. In addition, the existing technology determines whether the humidity abnormality occurs based on whether the humidity monitored by the humidity sensor reaches a fixed limit, and the abnormal humidity monitoring capability is weak. In addition, the existing technology cannot accurately locate the cause of the humidity abnormality, which brings inconvenience to the operation and maintenance personnel.
[0070] The exemplary embodiment of the present disclosure extracts abnormal features by comparing the humidity conditions of a single unit and the entire field of units, and can accurately locate problems such as dehumidifier failure in a unit, failure to close cabinet doors or cabin doors in time, and humidity sensor failure. It can also display humidity abnormality information to the user and promptly notify maintenance personnel to handle it, thereby preventing unit equipment from working in an abnormal humidity environment for a long time, reducing power generation losses, and lowering operation and maintenance costs.
[0071] It should be understood that when the method for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure is executed by a field control, steps S10 to S30 may be uniformly executed on each wind turbine generator set.
[0072] Figure 2 A flow chart showing a method for monitoring humidity of a wind turbine generator system according to another exemplary embodiment of the present disclosure is shown. In this embodiment, the method is executed by a field control.
[0073] In step S101, a list of wind farm units is obtained.
[0074] In step S102, the operation data of a single unit is obtained according to the list.
[0075] As an example, the SCADA operation data of each unit for a certain period of time can be obtained according to the list, which may include real-time data of active power, cabin humidity, hub humidity, control layer humidity, variable layer humidity, and ambient temperature. The data acquisition cycle can be adjusted according to actual conditions, and the data of each unit within the data acquisition cycle shall not be less than 10,000.
[0076] In step S103, based on the humidity data of the unit, it is determined whether the humidity sensor of the unit is faulty.
[0077] In step S104, the humidity data of the humidity sensor of the unit that has failed is deleted. For example, when it is determined that the cabin humidity sensor of the unit has failed, the cabin humidity data is deleted from the acquired operation data of the unit.
[0078] In step S105, humidity data in a power generation state in which active power exceeds a power threshold is screened out.
[0079] As an example, the power threshold may be the minimum active power (kW) of the unit in the grid-connected power generation state, which may be selected according to actual conditions.
[0080] In step S106, the humidity data of each unit is resampled using a preset sampling window.
[0081] The sampling window can be selected according to the resampling effect. After resampling, the amount of data for each unit should be no less than 200.
[0082] In step S107, it is determined that the data of the wind farm units have been processed.
[0083] In step S108, the amount of effective data of the wind farm is counted.
[0084] In step S109, it is determined whether the effective data amount is greater than the minimum data amount threshold of the entire field.
[0085] As an example, if the full-field effective data volume after resampling meets the calculation requirements, that is, the full-field effective data volume is greater than the full-field minimum data volume threshold, step S110 is performed. For example, the full-field minimum data volume threshold may be 2000.
[0086] In step S110, a feature vector of the resampled humidity data of each unit is extracted.
[0087] In step S111, the extracted feature vector is input into the trained humidity monitoring model.
[0088] In step S112, it is determined whether all units in the wind farm have completed model calculation.
[0089] According to the method for monitoring the humidity of a wind turbine generator set of an exemplary embodiment of the present disclosure, by real-time monitoring of the humidity of the set, abnormal humidity problems of the set can be discovered in time, and the causes of the abnormalities can be accurately located, such as humidity sensor failure, dehumidifier failure, cabinet door or cabin door not closed in time, etc., and a visual interface is provided to display abnormal information, which is convenient for operation and maintenance personnel to troubleshoot problems, conduct inspections and maintenance, and provide sufficient maintenance time for the operation and maintenance personnel; problems such as component damage, performance degradation, and reduced service life caused by long-term operation of wind power equipment in an abnormal humidity environment can be prevented, thereby reducing the loss of power generation and reducing operation and maintenance costs.
[0090] Figure 3 A flow chart illustrating a method of determining whether a humidity sensor fails according to an exemplary embodiment of the present disclosure.
[0091] Reference Figure 3 In step S201, for each unit, after deleting the null value, the maximum and minimum values in the humidity data of each humidity sensor of the unit are counted, and the range D1 and difference D2 of the humidity data of each humidity sensor of the unit are calculated.
[0092] In step S202, for each unit, the amount of abnormal humidity data Q1 with humidity value < P1 or humidity value > P2 in the humidity data of each humidity sensor of the unit, and the amount of abnormal jump data Q2 with humidity difference D2 < P3 or D2 > P4 are counted.
[0093] P1 indicates the minimum humidity (%) that can be achieved in the corresponding unit part (for example, the cabin); P2 indicates the maximum humidity (%) that can be achieved in the corresponding unit part; P3 indicates the lower limit of the humidity difference threshold (%), and it is recommended that it should not be greater than -5%; P4 indicates the upper limit of the humidity difference threshold (%), and it is recommended that it should not be less than 5%.
[0094] In step S203, if Q1>P5 or Q2>P6 or D1<P7 of the humidity data of the humidity sensor, it is determined that the humidity sensor is faulty.
[0095] As an example, the abnormal humidity data volume threshold P5, the abnormal jump data volume threshold P6 and the data dead value threshold P7 can be determined by analyzing the humidity data statistics and mechanism. For example, the data dead value threshold P7 is recommended to be no greater than 0.1%.
[0096] Figure 4 A flowchart illustrating a method of determining whether the humidity of a portion detected by a humidity sensor that has not failed is abnormal according to an exemplary embodiment of the present disclosure.
[0097] Reference Figure 4 In step S301, humidity data under a specific operating state is filtered out from humidity data of humidity sensors of the unit that have not failed; and the filtered humidity data is resampled.
[0098] In step S302, based on the resampled humidity data of each unit, the average value, maximum value, minimum value, median value, and standard deviation of the resampled humidity data of each humidity sensor of the unit are counted;
[0099] In addition, considering the impact of ambient temperature on humidity, the average humidity of the resampled humidity data of each humidity sensor of the unit in each temperature range is counted according to the ambient temperature bin. For example, the temperature range can span about 5°C.
[0100] In addition, based on the average value N1 of the resampled humidity data of the cabin humidity sensors of all units, the average value N2 of the resampled humidity data of the hub humidity sensors of all units, the average value N3 of the resampled humidity data of the transformer flow layer sensors of all units, and the average value N4 of the resampled humidity data of the control layer humidity sensors of all units, the proportion of abnormal humidity data of each unit is counted, and the humidity feature vector of each unit is constructed based on the above statistics.
[0101] As an example, based on the statistical analysis of the full-field data, the humidity of the cabin, hub, converter layer, and control layer can be set with an upper and lower floating margin based on the full-field average value, and the corresponding full-field humidity threshold range can be set to (N1-lower margin 1, N1+upper margin 1) / (N2-lower margin 2, N2+upper margin 2) / (N3-lower margin 3, N3+upper margin 3) / (N4-lower margin 4, N4+upper margin 4). The abnormal data proportions L1, L2, L3, and L4 of the resampled humidity data of each humidity sensor of each unit that exceed the corresponding full-field threshold range are counted separately.
[0102] In step S303, the extracted feature vector is input into the trained humidity monitoring model.
[0103] As an example, the training data containing abnormal humidity in the cabin, hub, converter layer, and control layer caused by problems such as dehumidifier failure and cabin door or cabinet door not closed can be used to extract feature vectors through the above process, input into the XGBoost algorithm model, and complete model training. Model training requires data from at least 2,000 units.
[0104] The above parameters need to be adjusted according to the actual situation. For example, P1 and P2 can be determined by analyzing the histogram. Taking cabin humidity as an example, Figure 5 The figure shows the distribution histogram of cabin humidity of about 300 units. It can be seen from the figure that the cabin humidity is basically between 20% and 80%. When setting the threshold, the range can be appropriately expanded considering the actual situation.
[0105] Figure 6Taking the cabin as an example, the method of obtaining the lower margin 1 and the upper margin 1 is explained. First, the cabin humidity of each unit is obtained according to the wind field, and the scattered points represent the cabin humidity at different temperatures of different units. Based on the cabin humidity of the entire field, the upper and lower margins of the cabin humidity are analyzed based on the average value of the entire field. The threshold range of the cabin humidity of the entire field is (average cabin humidity of the entire field - lower margin 1, average cabin humidity of the entire field + upper margin 1), which is represented by a dotted line. When the average value ± margin of the cabin humidity of the entire field exceeds the range of (0%, 100%), the margin at this temperature should be adjusted according to the actual situation to ensure that the threshold range of the cabin humidity of the entire field is within the range of (0%, 100%). Figure 6 The threshold range of the full-field cabin humidity is (full-field average cabin humidity - 35%, full-field average cabin humidity + 18%), thereby determining that the lower margin 1 can be 35% and the upper margin 1 can be 18%.
[0106] According to the method for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure, data of 5,305 units were tested, and a total of 49 units were reported to have sensor failures or humidity anomalies, with an accuracy rate of 100%.
[0107] Figure 7 The picture shows that the cabin humidity sensor of a certain unit is faulty and the cabin humidity data is dead. Figure 8 The data of a unit with abnormal cabin humidity is shown. The cabin dehumidifier of this unit is faulty. When the humidity in the cabin is too high, the dehumidifier does not work properly. The cabin humidity fluctuates around 77%, with a maximum of over 80%, while the average cabin humidity of the entire field is only around 55%, which is a large difference.
[0108] The present disclosure can timely discover humidity sensor failure problems and prevent dehumidifiers from working abnormally; accurately locate the causes of abnormal humidity, such as dehumidifier failure, cabinet doors or cabin doors not being closed, etc., and can promptly notify on-site staff to provide support for maintenance personnel to troubleshoot problems and select maintenance equipment.
[0109] The present invention fully considers the influence of temperature on humidity. On the basis of extracting the humidity characteristics of a single unit, it adopts a method of comparing the humidity with that of all units in the field, screens abnormal humidity data, constructs abnormal characteristics, reduces the influence of environmental factors and unit equipment models, effectively realizes abnormal unit humidity monitoring, and prevents power generation loss and increased operation and maintenance costs caused by unit components working in an abnormal humidity environment.
[0110] The present invention adopts an effective data cleaning and processing method to eliminate abnormal data, filter out valid data under power generation status, and reduce the interference of abnormal data by resampling to ensure the reliability of data.
[0111] The present invention constructs an effective unit humidity anomaly monitoring model based on the XGBoost algorithm, which can locate the specific cause of the fault and clarify the problems of the dehumidifier and cabinet door.
[0112] Fig. 9 A structural block diagram of a device for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure is shown.
[0113] like Fig. 9 As shown, the device for monitoring humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure includes: a data acquisition unit 10 , a fault identification unit 20 , and a humidity anomaly identification unit 30 .
[0114] Specifically, the data acquisition unit 10 is configured to acquire operating data of the wind turbine generator set in a current time period, wherein the operating data includes: humidity data of at least one humidity sensor, and different types of humidity sensors in the at least one humidity sensor are used to detect the humidity of different parts of the wind turbine generator set.
[0115] The fault identification unit 20 is configured to determine whether each humidity sensor of the wind turbine generator system fails based on humidity data of each humidity sensor of the wind turbine generator system.
[0116] The humidity abnormality identification unit 30 is configured to determine whether the humidity of a portion detected by a humidity sensor that has not failed of the wind turbine generator set is abnormal based on the humidity data of the humidity sensor that has not failed of the wind turbine generator set.
[0117] As an example, the humidity anomaly identification unit 30 may be configured to determine whether the humidity of a portion detected by a humidity sensor that is not faulty of the wind turbine generator set is abnormal and the cause of the abnormality based on humidity data of the humidity sensor that is not faulty of the wind turbine generator set.
[0118] As an example, the humidity anomaly identification unit 30 can be configured to determine whether the humidity of the part detected by the non-faulty humidity sensor of the wind turbine generator set is abnormal based on the humidity data of the non-faulty humidity sensor of the wind turbine generator set by referring to the humidity data of the whole field wind turbine generator set in the current time period.
[0119] As an example, the humidity anomaly identification unit 30 can be configured to: extract the feature vector of the humidity data of the non-faulty humidity sensor of the wind turbine by referring to the humidity data of the entire wind turbine generator set in the current time period; input the extracted feature vector into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal, or a prediction result of whether the humidity of the part detected by the non-faulty humidity sensor is abnormal and the cause of the abnormality.
[0120] As an example, the operating data may also include: ambient temperature data; wherein the feature vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed may include: at least one of the following statistics of the humidity data of each humidity sensor that has not failed: mean value; maximum value; minimum value; median; standard deviation; variance; the average humidity value in each temperature interval obtained by temperature bin statistics; the proportion of abnormal data exceeding the full-field humidity threshold range corresponding to this humidity sensor; wherein the full-field humidity threshold range corresponding to this humidity sensor is based on: the average value of the humidity data of this humidity sensor of all wind turbine generator sets that have not failed within the current time period.
[0121] As an example, the humidity anomaly identification unit 30 can also be configured to: filter out humidity data under a specific operating state from the humidity data of the humidity sensors of the wind turbine generator set that are not faulty; resample the filtered humidity data; wherein the humidity anomaly identification unit 30 can be configured to: extract a feature vector of the resampled humidity data by referring to the humidity data of all wind turbine generator sets in the current time period.
[0122] As an example, the fault identification unit 20 can be configured to: determine at least one of the extreme value of the humidity data, the abnormal humidity data amount of the humidity data, and the abnormal jump data amount of the humidity data based on the humidity data of each humidity sensor of the wind turbine generator set; when at least one of the extreme value is less than the dead value threshold corresponding to the humidity sensor, the abnormal humidity data amount exceeds the abnormal humidity data amount threshold corresponding to the humidity sensor, and the abnormal jump data amount exceeds the abnormal jump data amount threshold corresponding to the humidity sensor, it is determined that the humidity sensor of the wind turbine generator set has a fault.
[0123] As an example, the different parts may include at least one of the following items: a nacelle, a hub, a flow conversion layer, and a control layer.
[0124] As an example, the device for monitoring the humidity of a wind turbine generator set according to an exemplary embodiment of the present disclosure may also include: an output unit (not shown), the output unit being configured to prompt a user that a humidity sensor of the wind turbine generator set has failed; and prompt a user that the part of the wind turbine generator set where the humidity is abnormal and the cause of the abnormality.
[0125] As an example, the abnormal cause may include at least one of the following items: a dehumidifier failure, a cabinet door or a cabin door not being closed in time.
[0126] It should be understood that the specific processing performed by the device for monitoring humidity of a wind turbine generator set according to the exemplary embodiment of the present disclosure has been referred to in detail. Figures 1 to 8The details have been described in detail and will not be repeated here.
[0127] It should be understood that the various units in the device for monitoring humidity of a wind turbine generator set according to the exemplary embodiment of the present disclosure may be implemented as hardware components and / or software components. Those skilled in the art may implement the various units, for example, using a field programmable gate array (FPGA) or an application specific integrated circuit (ASIC) according to the processing performed by the various defined units.
[0128] An exemplary embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the method for monitoring the humidity of a wind turbine generator set as described in the above exemplary embodiment. The computer-readable storage medium is any data storage device that can store data read by a computer system. Examples of computer-readable storage media include: read-only memory, random access memory, read-only optical disk, magnetic tape, floppy disk, optical data storage device, and carrier wave (such as data transmission through the Internet via a wired or wireless transmission path).
[0129] According to an exemplary embodiment of the present disclosure, a controller of a wind turbine generator set includes: a processor (not shown) and a memory (not shown), wherein the memory stores a computer program, and when the computer program is executed by the processor, it prompts the processor to execute the method for monitoring the humidity of the wind turbine generator set as described in the above exemplary embodiment.
[0130] According to an exemplary embodiment of the present disclosure, a controller of a wind farm includes: a processor (not shown) and a memory (not shown), wherein the memory stores a computer program, and when the computer program is executed by the processor, it prompts the processor to execute the method for monitoring the humidity of a wind turbine generator set as described in the above exemplary embodiment.
[0131] Although some exemplary embodiments of the present disclosure have been shown and described, it will be appreciated by those skilled in the art that modifications may be made to these embodiments without departing from the principles and spirit of the present disclosure, the scope of which is defined by the claims and their equivalents.
Claims
1. A method for monitoring humidity of a wind turbine generator set, characterized in that: include: Acquiring operation data of the wind turbine generator set in a current time period, wherein the operation data includes: humidity data of at least one humidity sensor, wherein different types of humidity sensors in the at least one humidity sensor are used to detect humidity at different parts of the wind turbine generator set; Based on humidity data of each humidity sensor of the wind turbine generator set, determining whether each humidity sensor fails; By referring to the humidity data of all wind turbines in the current time period, based on the humidity data of the humidity sensors of the wind turbines that have not failed, it is determined whether the humidity of the part detected by the humidity sensors that have not failed is abnormal. Wherein, based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, the step of determining whether the humidity of the part detected by the humidity sensor that has not failed is abnormal includes: Filter out humidity data under a specific operating state from humidity data of humidity sensors of the wind turbine generator set that have not failed; Resample the filtered humidity data; Extracting a feature vector of humidity data of humidity sensors of the wind turbine generator sets that have not failed by referring to humidity data of all wind turbine generator sets in the current time period; The extracted feature vector is input into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the humidity sensor that has not failed is abnormal, or a prediction result of whether the humidity of the part detected by the humidity sensor that has not failed is abnormal and the cause of the abnormality. Among them, the step of extracting the characteristic vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed by referring to the humidity data of all the wind turbine generator sets in the current time period includes: extracting the characteristic vector of the resampled humidity data by referring to the humidity data of all the wind turbine generator sets in the current time period.
2. The method according to claim 1, characterized in that Based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, the step of determining whether the humidity of the part detected by the humidity sensor that has not failed is abnormal includes: Based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, it is determined whether the humidity of the part detected by the humidity sensor that has not failed is abnormal and the cause of the abnormality.
3. The method according to claim 1, characterized in that The operation data also includes: ambient temperature data; The characteristic vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed includes at least one of the following statistics of the humidity data of each humidity sensor that has not failed: Average value; maximum value; minimum value; median value; standard deviation; variance; average humidity value in each temperature range obtained by temperature bin statistics; the proportion of abnormal data that exceeds the overall humidity threshold range corresponding to this type of humidity sensor; The humidity threshold range for the entire field corresponding to the humidity sensor is based on the average value of humidity data of the humidity sensor of the entire field that has not failed in the current time period.
4. The method according to claim 1, characterized in that: Based on the humidity data of each humidity sensor of the wind turbine generator set, the step of determining whether each humidity sensor fails comprises: Based on the humidity data of each humidity sensor of the wind turbine generator set, determining at least one of the range value of the humidity data, the abnormal humidity data amount of the humidity data, and the abnormal jump data amount of the humidity data; When at least one of the following conditions is met: the extreme difference value is less than the dead value threshold corresponding to the humidity sensor, the abnormal humidity data volume exceeds the abnormal humidity data volume threshold corresponding to the humidity sensor, and the abnormal jump data volume exceeds the abnormal jump data volume threshold corresponding to the humidity sensor, it is determined that the humidity sensor of the wind turbine generator set has failed.
5. The method according to claim 1, characterized in that The different parts include at least one of the following items: a nacelle, a hub, a variable flow layer, and a control layer.
6. The method according to claim 2, characterized in that Also includes: A humidity sensor that alerts the user that the wind turbine generator set has failed; The user is prompted with the abnormal humidity part of the wind turbine generator set and the abnormal reason.
7. The method according to claim 2, characterized in that: The abnormal reason includes at least one of the following items: dehumidifier failure, cabinet door or cabin door not closed in time.
8. A device for monitoring humidity of a wind turbine generator set, characterized in that: include: A data acquisition unit is configured to acquire operation data of the wind turbine generator set in a current time period, wherein the operation data includes: humidity data of at least one humidity sensor, different types of humidity sensors in the at least one humidity sensor are used to detect humidity at different parts of the wind turbine generator set; a fault identification unit configured to determine whether each humidity sensor fails based on humidity data of each humidity sensor of the wind turbine generator set; The humidity abnormality identification unit is configured to determine whether the humidity of the part detected by the humidity sensor that has not failed is abnormal based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed by referring to the humidity data of the wind turbine generator sets in the current time period, Wherein, based on the humidity data of the humidity sensor of the wind turbine generator set that has not failed, the operation of determining whether the humidity of the part detected by the humidity sensor that has not failed is abnormal includes: Filter out humidity data under a specific operating state from humidity data of humidity sensors of the wind turbine generator set that have not failed; Resample the filtered humidity data; Extracting a feature vector of humidity data of humidity sensors of the wind turbine generator sets that have not failed by referring to humidity data of all wind turbine generator sets in the current time period; The extracted feature vector is input into a pre-trained humidity monitoring model to obtain a prediction result of whether the humidity of the part detected by the humidity sensor that has not failed is abnormal, or a prediction result of whether the humidity of the part detected by the humidity sensor that has not failed is abnormal and the cause of the abnormality. Among them, the operation of extracting the characteristic vector of the humidity data of the humidity sensor of the wind turbine generator set that has not failed by referring to the humidity data of all the wind turbine generator sets in the current time period includes: extracting the characteristic vector of the resampled humidity data by referring to the humidity data of all the wind turbine generator sets in the current time period.
9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is prompted to perform the method for monitoring humidity of a wind turbine generator system according to any one of claims 1 to 7.
10. A controller for a wind turbine generator set, characterized in that: The controller comprises: processor; A memory storing a computer program, which, when executed by a processor, prompts the processor to execute the method for monitoring humidity of a wind turbine generator set according to any one of claims 1 to 7.
11. A controller for a wind farm, characterized in that: The controller comprises: processor; A memory storing a computer program, which, when executed by a processor, prompts the processor to execute the method for monitoring humidity of a wind turbine generator set according to any one of claims 1 to 7.