Wind generating set and gearbox abnormity prediction method and device thereof
By analyzing the operating data of the gearbox, extracting specific data sequences, determining trends and combining historical trends, we successfully predict abnormalities in the gearbox, solving the problem of gearbox abnormal prediction in wind turbines and improving the operating reliability of the generator set.
Patent Information
- Application Number
- CN202310612013.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-26
- Publication Date
- 2024-11-26
AI Technical Summary
It is difficult to predict abnormalities of gear boxes in wind turbines. The existing technology fails to effectively predict abnormalities of gear boxes, affecting the normal operation of the generator set.
By obtaining the operating data of the bearing, lubrication system and cooling system of the gearbox, extracting the serial incremental and decreasing data sequences, determining the operating trend, and predicting the abnormal situation of the gearbox based on the predetermined abnormal trend.
Effectively predict abnormal situations in the gearbox, discover potential problems in advance, avoid generator set failures, and improve operational reliability.
Smart Images

Figure CN119026041A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure generally relates to the field of wind power generation technology, and more specifically, to a method and device for predicting abnormalities of a wind power generator set and a gearbox thereof. Background Art
[0002] The gearbox is an important component of the wind turbine. If the gearbox fails to detect an abnormality before reporting it, it will have a serious impact on the wind turbine. Therefore, it is very important to predict the abnormality of the gearbox. Summary of the invention
[0003] The embodiments of the present disclosure provide a method and device for predicting abnormality of a wind turbine generator set and a gearbox thereof, which can effectively predict abnormal conditions of the gearbox.
[0004] In a general aspect, a method for predicting an abnormality of a gearbox of a wind turbine generator set is provided, comprising: acquiring operating data of the gearbox of the wind turbine generator set within a predetermined time period, wherein the operating data comprises at least one of the following: operating data of a bearing of the gearbox, operating data of a lubrication system of the gearbox, and operating data of a cooling system of the gearbox; extracting a plurality of data sequences from the operating data, wherein the plurality of data sequences comprises a continuously increasing data sequence and a continuously decreasing data sequence in the operating data; determining an operating trend of the gearbox based on the plurality of data sequences; and predicting an abnormal condition of the gearbox based on the operating trend and a predetermined abnormal operating trend.
[0005] Optionally, predetermined statistical indicators for each data sequence are determined; multiple data sequences are classified according to the attributes of the data in the data sequence; indicator change trends of the predetermined statistical indicators in each category after classification are determined; and the operating trend of the gearbox is determined based on the indicator change trend.
[0006] Optionally, extracting multiple data sequences from the operating data includes: classifying the operating data according to attributes of the data in the operating data; clustering the operating data under each classified category to obtain at least one cluster of data; and extracting multiple data sequences from the current data for each cluster of data in the at least one cluster of data.
[0007] Optionally, before predicting the abnormal condition of the gearbox based on the operating trend and a predetermined abnormal operating trend, it also includes: acquiring historical operating data of the gearbox as first historical operating data; segmenting the first historical operating data according to a predetermined duration and time sequence to obtain multiple segments of first sub-historical operating data; for each segment of the first sub-historical operating data, extracting multiple data sequences in the current first sub-historical operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current first sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data, determining the operating trend of the gearbox before the abnormal condition occurs in the gearbox as the abnormal operating trend.
[0008] Optionally, based on multiple data sequences of each section of the first sub-historical operating data, the operating trend of the gearbox before the abnormal situation occurs in the gearbox is determined, including: determining predetermined statistical indicators for each data sequence; performing the following processing on each section of the first sub-historical operating data: classifying the multiple data sequences of the current first sub-historical operating data according to the attributes of the data in the data sequence; determining the indicator change trend of the predetermined statistical indicator under each category after classification; based on the indicator change trend under each category of each section of the first sub-historical operating data, determining the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0009] Optionally, based on the indicator change trend under each category of each section of the first sub-historical operating data, the operating trend of the gearbox before the abnormal situation occurs in the gearbox is determined, including: performing the following processing for each category: comparing the indicator change trend under the current category of each section of the first sub-historical operating data; based on the comparison result, determining the first sub-historical operating data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operating data; determining the abnormal operating data based on the estimated abnormal operating data of each category; determining the indicator change trend under each category of the abnormal operating data as the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0010] Optionally, multiple data sequences of each segment of the first sub-historical operating data determine the operating trend of the gearbox before the abnormal situation occurs in the gearbox, including: obtaining historical operating data of gearboxes operating normally in other wind turbines as second historical operating data; segmenting the second historical operating data according to a predetermined duration and time sequence to obtain multiple segments of second sub-historical operating data; for each segment of the second historical operating data, extracting multiple data sequences in the current second sub-historical operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current second sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data and the multiple data sequences of each segment of the second sub-historical operating data, determining the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0011] Optionally, multiple data sequences of each section of first sub-historical operating data and multiple data sequences of each section of second sub-historical operating data determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox, including: determining predetermined statistical indicators for each data sequence; performing the following processing on each section of first sub-historical operating data and each section of second sub-historical operating data respectively: classifying multiple data sequences of current sub-historical operating data according to attributes of the data in the data sequence; determining indicator change trends of predetermined statistical indicators under each category after classification; determining the operating trend of the gearbox before an abnormal situation occurs in the gearbox based on the indicator change trends under each category of each section of second sub-historical operating data and the indicator change trends under each category of each section of first sub-historical operating data.
[0012] Optionally, based on the indicator change trend under each category of each section of the second sub-historical operating data and the indicator change trend under each category of each section of the first sub-historical operating data, the operating trend of the gearbox before the abnormal situation occurs in the gearbox is determined, including: performing the following processing for each category: comparing the indicator change trend under the current category of each section of the first sub-historical operating data and each section of the second sub-historical operating data; based on the comparison result, determining the sub-historical operating data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operating data; determining the abnormal operating data based on the estimated abnormal operating data of each category; determining the indicator change trend under each category of the abnormal operating data as the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0013] Optionally, before determining the indicator change trend under each category of abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox, it also includes: obtaining historical fault information of the gearbox; based on the historical fault information, determining that the abnormal operation data is actual abnormal operation data.
[0014] Optionally, the attributes of the data include: bearing temperature, oil sump temperature of the lubrication system, oil sump inlet oil temperature of the lubrication system, oil pressure of the lubrication system oil entering the gearbox, oil pressure of the lubrication system oil flowing out of the gearbox, inlet valve water pressure of the cooling system and outlet valve water pressure of the cooling system.
[0015] In another general aspect, a gearbox abnormality prediction device for a wind turbine generator set is provided, comprising: an acquisition unit, configured to acquire operating data of the gearbox of the wind turbine generator set within a predetermined time period, wherein the operating data comprises at least one of the following: operating data of bearings of the gearbox, operating data of a lubrication system of the gearbox and operating data of a cooling system of the gearbox; an extraction unit, configured to extract a plurality of data sequences from the operating data, wherein the plurality of data sequences comprise a continuously increasing data sequence and a continuously decreasing data sequence in the operating data; a determination unit, configured to determine an operating trend of the gearbox based on the plurality of data sequences; and a prediction unit, configured to predict an abnormal condition of the gearbox based on the operating trend and a predetermined abnormal operating trend.
[0016] Optionally, the determination unit is further configured to determine predetermined statistical indicators for each data sequence; classify multiple data sequences according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicators under each category after classification; and determine the operating trend of the gearbox based on the indicator change trend.
[0017] Optionally, the extraction unit is further configured to extract multiple data sequences from the running data, including: classifying the running data according to the attributes of the data in the running data; clustering the running data under each classified category to obtain at least one cluster of data; and for each cluster of data in the at least one cluster of data, extracting multiple data sequences from the current data.
[0018] Optionally, the prediction unit is further configured to obtain historical operating data of the gearbox as first historical operating data before predicting the abnormal condition of the gearbox based on the operating trend and a predetermined abnormal operating trend; segment the first historical operating data according to a predetermined duration and time sequence to obtain multiple segments of first sub-historical operating data; for each segment of the first sub-historical operating data, extract multiple data sequences in the current first sub-historical operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current first sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before the abnormal condition occurs in the gearbox as the abnormal operating trend.
[0019] Optionally, the prediction unit is also configured to determine predetermined statistical indicators for each data sequence; perform the following processing on each segment of the first sub-historical operating data: classify multiple data sequences of the current first sub-historical operating data according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicators under each category after classification; based on the indicator change trend under each category of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before the gearbox has an abnormal situation.
[0020] Optionally, the prediction unit is also configured to perform the following processing for each category: compare the indicator change trend under the current category of each segment of the first sub-historical operation data; based on the comparison result, determine the first sub-historical operation data corresponding to the abnormal situation of the gearbox, and use it as the estimated abnormal operation data; determine the abnormal operation data based on the estimated abnormal operation data of each category; determine the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox.
[0021] Optionally, the prediction unit is also configured to obtain historical operating data of gearboxes that are operating normally in other wind turbines as second historical operating data; segment the second historical operating data according to a predetermined duration and time sequence to obtain multiple segments of second sub-historical operating data; for each segment of the second historical operating data, extract multiple data sequences in the current second sub-historical operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current second sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data and the multiple data sequences of each segment of the second sub-historical operating data, determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox.
[0022] Optionally, the prediction unit is also configured to determine predetermined statistical indicators for each data sequence; perform the following processing on each segment of the first sub-historical operating data and each segment of the second sub-historical operating data: classify multiple data sequences of the current sub-historical operating data according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicator under each category after classification; based on the indicator change trend under each category of each segment of the second sub-historical operating data and the indicator change trend under each category of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox.
[0023] Optionally, the prediction unit is also configured to perform the following processing for each category: compare the indicator change trends under the current category of each first sub-historical operating data and each second sub-historical operating data; based on the comparison results, determine the sub-historical operating data corresponding to the abnormal situation of the gearbox, and use it as the estimated abnormal operating data; determine the abnormal operating data based on the estimated abnormal operating data of each category; determine the indicator change trend under each category of the abnormal operating data as the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0024] Optionally, the prediction unit is also configured to obtain historical fault information of the gearbox before determining the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox; and determine that the abnormal operation data is actual abnormal operation data based on the historical fault information.
[0025] Optionally, the attributes of the data include: bearing temperature, oil sump temperature of the lubrication system, oil sump inlet oil temperature of the lubrication system, oil pressure of the lubrication system oil entering the gearbox, oil pressure of the lubrication system oil flowing out of the gearbox, inlet valve water pressure of the cooling system and outlet valve water pressure of the cooling system.
[0026] In another general aspect, a computer-readable storage medium storing instructions is provided, wherein when the instructions are executed by at least one computing device, the at least one computing device is prompted to execute any of the above-mentioned methods for predicting gearbox abnormality of a wind turbine generator set.
[0027] In another general aspect, a system is provided comprising at least one computing device and at least one storage device storing instructions, wherein the instructions, when executed by the at least one computing device, cause the at least one computing device to execute any of the above-described methods for predicting gearbox abnormalities of a wind turbine generator set.
[0028] In another general aspect, a wind turbine generator set is provided, comprising the gearbox abnormality prediction device of the wind turbine generator set as described above.
[0029] According to the wind turbine generator set and the gearbox abnormality prediction method and device of the embodiment of the present disclosure, the continuously increasing data sequence and the continuously decreasing data sequence are extracted from the gearbox operation data, and the operation trend of the gearbox can be determined based on the continuously increasing data sequence and the continuously decreasing data sequence, and then combined with the predetermined abnormal operation trend, the upcoming abnormal situation of the gearbox is predicted. Therefore, through the present disclosure, the abnormal situation of the gearbox can be effectively predicted.
[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 embodiments of the present disclosure will become more apparent through the following description in conjunction with the accompanying drawings showing the embodiments, in which:
[0032] Figure 1 is a schematic diagram showing a main transmission chain of a wind turbine generator set according to an embodiment of the present disclosure;
[0033] Figure 2 is a flow chart showing a method for predicting abnormality of a gearbox of a wind turbine generator set according to an embodiment of the present disclosure;
[0034] Figure 3 is a flow chart showing the time series characteristics of part of the data of normal bearing temperature of the wind turbine generator set according to an embodiment of the present disclosure;
[0035] Figure 4 is a flow chart showing the time series characteristics of some data of the oil pool temperature and the oil pool inlet oil temperature of the wind turbine generator set according to the embodiment of the present disclosure being normal;
[0036] Figure 5 is a flow chart showing the time series characteristics of part of the data of the normal oil pressure of the wind turbine generator set according to an embodiment of the present disclosure;
[0037] Figure 6 is a flow chart showing the time series characteristics of part of the data of normal water pressure of the wind turbine generator set according to an embodiment of the present disclosure;
[0038] Figure 7 is a schematic diagram showing the relationship between the outlet oil pressure and the inlet oil pressure under low oil pressure when the oil pressure of the wind turbine generator set according to the embodiment of the present disclosure is normal;
[0039] Figure 8 is a schematic diagram showing the relationship between the outlet oil pressure and the inlet oil pressure under high oil pressure when the oil pressure of the wind turbine generator set according to an embodiment of the present disclosure is normal;
[0040] Fig. 9 is a schematic diagram showing a data sequence of a continuous monotonically increasing bearing temperature according to an embodiment of the present disclosure;
[0041] Fig.10 is a schematic diagram showing a type of statistical indicators of bearing temperature according to an embodiment of the present disclosure;
[0042] Fig.11 is a schematic diagram showing another type of statistical indicator of bearing temperature according to an embodiment of the present disclosure;
[0043] Fig.12 is a schematic diagram showing the mean values of multiple data series of daily water pressure according to an embodiment of the present disclosure;
[0044] Fig.13 is a schematic diagram showing the water pressure when the water cooling pump of the embodiment of the present disclosure is started;
[0045] Fig.14 is a schematic diagram showing an abnormal sequence of inlet oil temperature according to an embodiment of the present disclosure;
[0046] Fig.15 is a schematic diagram showing a type of statistical indicator of oil temperature in an oil pool according to an embodiment of the present disclosure;
[0047] Fig.16 is a schematic diagram showing a type of statistical indicator of oil temperature at an oil pool inlet according to an embodiment of the present disclosure;
[0048] Fig.17 is a schematic diagram showing the inlet and outlet oil pressures of an embodiment of the present disclosure;
[0049] Fig.18 is a schematic diagram showing a comparison of different clustering results of inlet oil pressures of different units according to an embodiment of the present disclosure;
[0050] Fig.19 is a schematic diagram showing a comparison of different clustering results of outlet oil pressures of different units according to an embodiment of the present disclosure;
[0051] Fig. 20 is a schematic diagram showing an example of high-speed shaft temperature anomaly according to an embodiment of the present disclosure;
[0052] Fig.21 is a schematic diagram showing a gearbox abnormality prediction process according to an embodiment of the present disclosure;
[0053] Fig. 22 is a schematic diagram showing the warning results of wind turbine generator sets 2# and 7# according to an embodiment of the present disclosure;
[0054] Fig.23 is a schematic diagram showing the early warning result of wind turbine generator set 1# according to an embodiment of the present disclosure;
[0055] Fig.24 is a block diagram showing a gearbox abnormality prediction device for a wind turbine generator set according to the present disclosure. DETAILED DESCRIPTION
[0056] The following specific embodiments are provided to help the reader obtain a comprehensive understanding of the methods, devices and / or systems described herein. However, after understanding the disclosure of the present application, various changes, modifications and equivalents of the methods, devices and / or systems described herein will be clear. For example, the order of operations described herein is only an example and is not limited to those orders set forth herein, but can be changed as will be clear after understanding the disclosure of the present application, except for operations that must occur in a specific order. In addition, for greater clarity and simplicity, the description of features known in the art may be omitted.
[0057] The features described herein can be implemented in different forms and should not be construed as being limited to the examples described herein. Rather, the examples described herein have been provided to illustrate only some of the many possible ways to implement the methods, devices, and / or systems described herein, which will be clear after understanding the disclosure of the present application.
[0058] As used herein, the term "and / or" includes any one of the associated listed items and any combination of any two or more.
[0059] Although terms such as "first", "second", and "third" may be used herein to describe various members, components, regions, layers, or portions, these members, components, regions, layers, or portions should not be limited by these terms. Instead, these terms are only used to distinguish one member, component, region, layer, or portion from another member, component, region, layer, or portion. Therefore, without departing from the teachings of the examples described herein, the first member, first component, first region, first layer, or first portion referred to in the examples may also be referred to as the second member, second component, second region, second layer, or second portion.
[0060] In the specification, when an element (such as a layer, a region, or a substrate) is described as being “on”, “connected to”, or “coupled to” another element, the element may be directly “on”, “connected to”, or “coupled to” another element, or one or more other elements may be present therebetween. Conversely, when an element is described as being “directly on”, “directly connected to”, or “directly coupled to” another element, there may be no other elements present therebetween.
[0061] The terms used herein are only used to describe various examples and are not intended to limit the disclosure. Unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. The terms "comprise", "include" and "have" indicate the presence of the described features, quantities, operations, components, elements and / or combinations thereof, but do not exclude the presence or addition of one or more other features, quantities, operations, components, elements and / or combinations thereof.
[0062] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by a person of ordinary skill in the art to which the present disclosure belongs after understanding the present disclosure. Unless explicitly defined as such herein, terms (such as those defined in a general dictionary) should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and should not be interpreted in an idealized or overly formal manner.
[0063] Furthermore, in the description of examples, when it is considered that a detailed description of a well-known related structure or function would cause vague interpretation of the present disclosure, such a detailed description will be omitted.
[0064] The present disclosure provides a method and device for predicting abnormalities of a wind turbine and its gearbox, which can solve the above problems. The method for predicting abnormalities of the gearbox of a wind turbine disclosed in the present disclosure can be applied to a server, or to a controller of a single wind turbine, or to a master controller of a wind farm, which is not limited by the present disclosure. The server, controller and wind turbine can be connected wirelessly or by wire, which is not limited here. The above server can be a single server, or a server cluster composed of several servers, or a cloud computing platform or a virtualization center. The following is an explanation using a server as an example.
[0065] The server obtains operating data of the gearbox of the wind turbine generator set within a predetermined time period, wherein the operating data includes at least one of the following: operating data of the bearings of the gearbox, operating data of the lubrication system of the gearbox, and operating data of the cooling system of the gearbox; then extracts multiple data sequences from the operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the operating data; then determines the operating trend of the gearbox based on the multiple data sequences; and further, based on the operating trend and a predetermined abnormal operating trend, can effectively predict abnormal conditions of the gearbox.
[0066] For ease of understanding, the following briefly introduces the relevant knowledge of the gearbox involved in the present disclosure.
[0067] The main transmission chain of the wind turbine converts the mechanical energy absorbed by the impeller into electrical energy, connects the nacelle base and the hub, and provides braking and locking functions for impeller maintenance. The main transmission chain generally includes a shaft system, a gearbox (cooling and lubrication), a generator, and related monitoring modules, such as Figure 1 As shown. Among them, the gearbox converts low-speed, high-torque mechanical energy into high-speed, low-torque mechanical energy required by the generator. A multi-stage planetary structure can be used. The medium-speed permanent magnet generator and the gearbox are highly integrated, which shortens the size of the generator and the gearbox, makes the transmission structure compact, occupies less space, and makes the transmission connection more reasonable.
[0068] 1. Operating mechanism of gearbox
[0069] The operating mechanism of the gearbox and its contained lubrication system and cooling system is crucial for obtaining relevant measurement point data for analyzing the gearbox. Based on the operating mechanism of the gearbox and its contained lubrication system and cooling system, the operating data of the gearbox can be analyzed to diagnose the health status of the overall operation of the gearbox.
[0070] The principles of the lubrication system and cooling system are as follows: The gearbox of the wind turbine can have two types: air cooler (V12) and water cooler (V11). Specifically, the gearbox will be equipped with a lubrication system and cooling system (water cooling, air cooling) to ensure the forced lubrication and cooling of the gear meshing points and bearings of the gearbox. The lubrication system generally uses a gear pump to suck oil from the inside of the gearbox. The pumped lubricating oil first passes through the overflow valve (16bar) and enters the filter. When the pressure difference between the inlet and outlet of the filter is greater than 3.5bar, it means that the filter element of the filter is blocked. After the oil is filtered, it will enter the temperature control valve. When the oil temperature is ≧45℃, the external circulation is started, that is, it passes through the cooler all the way back to the gearbox. When the oil temperature is >60℃, the lubricating oil completely passes through the cooler to fully dissipate the heat of the oil. The cooler part has an overflow valve (bar) and a heater to ensure that the pressure can be relieved when the pressure suddenly increases and the heating can be started when the temperature is low. In this way, the lubricating oil circulates and cools continuously, continuously taking away the heat generated by the gearbox, and maintaining the oil temperature not exceeding 75℃. The gear pump of the gear box can also achieve a certain cooling and heat dissipation effect when running at low speed or high speed, but since it does not go through external circulation, the heat dissipation effect on the oil pool and gear box bearings is relatively small.
[0071] The oil pool temperature participates in the control of the following contents: low-speed operation of the gear pump, high-speed operation of the gear pump, start-up of the lubricating oil external circulation radiator temperature control valve, low-speed operation of the radiator fan, high-speed operation of the radiator fan, triggering of the high oil temperature warning, triggering of the high oil temperature timeout fault, triggering of the high oil temperature fault. The bearing temperature participates in the control of the following contents: high-speed operation of the gear pump, low-speed operation of the radiator fan, high-speed operation of the radiator fan, triggering of the high bearing temperature warning, triggering of the high bearing temperature fault.
[0072] The gearbox water cooling mainly relies on the water cooling system to achieve the heat dissipation of the gearbox. When the water cooling pump is started, the heat dissipation effect is significantly improved, the temperature of the oil when it enters the oil pool (hereinafter referred to as the inlet oil temperature) drops significantly, and the gearbox oil pool oil temperature and gearbox bearing temperature also show a slow downward trend.
[0073] The air cooling of the gearbox mainly relies on the external cooling fan of the wind turbine to achieve forced air cooling of the gearbox. When the temperature control valve and the cooling fan are started, the heat dissipation effect is significantly improved, the inlet oil temperature of the oil pool drops significantly, and the oil temperature of the gearbox oil pool and the gearbox bearing temperature also show a slow downward trend.
[0074] Through data analysis and comparison, it can be seen that the bearing temperature of the gearbox with water cooling is lower than that of the gearbox with air cooling. Without considering the model and configuration, the water cooling effect of the gearbox is better than the air cooling effect.
[0075] Based on the operating mechanism of the gearbox described above, the inventors found that the common fault phenomena and mechanisms of the gearbox are as follows:
[0076] Common abnormalities in gearboxes are mainly temperature abnormalities and pressure abnormalities. Temperature abnormalities mainly include the temperature of the gearbox bearing when running at low speed (hereinafter referred to as the low-speed shaft temperature), the temperature of the bearing when running at high speed (hereinafter referred to as the high-speed shaft temperature), and abnormal oil temperature in the oil pool. Pressure abnormalities mainly include abnormal oil pressure when the oil enters the gearbox (hereinafter referred to as the inlet oil pressure), abnormal oil pressure when the oil flows out of the gearbox (hereinafter referred to as the outlet oil pressure), and abnormal water pressure of the cooling system (water cooling) inlet and outlet valves (hereinafter referred to as the inlet water pressure and outlet water pressure).
[0077] (1) Abnormal oil pool temperature
[0078] There are two main reasons for high oil pool temperature. First, the heat generated by the gearbox exceeds the design requirements, and the rated cooling capacity of the cooling system cannot cool down the excess heat. Second, the heat generated by the gearbox meets the design requirements, but the cooling system cannot cool down the generated heat due to insufficient design, reduced cooling capacity, control strategy, and reduced oil quality. Third, an abnormality occurs in the data measurement link, resulting in excessively high oil temperature.
[0079] After discovering that the oil temperature is abnormal, you can analyze it in combination with the gearbox inlet oil pressure and outlet oil pressure. If the oil pressure is low, it is generally believed that there is a problem with the pump oil circuit of the lubrication system; if the oil pressure is normal, it is generally believed that there is a problem with the cooling function of the cooling system. Whether the oil temperature is abnormal can be analyzed by the inlet oil temperature. If the inlet oil temperature is low, it means that the heat dissipation is normal. If the inlet oil temperature is high, it means that the heat dissipation is abnormal, that is, the oil temperature is abnormal. The higher and lower thresholds can be set as needed.
[0080] The main reasons for abnormal oil temperature in the oil pool can be roughly divided into the following categories:
[0081] a. The pump oil circuit of the lubrication system is abnormal, including: problems with the gear pump, filter screen blockage, overflow and other abnormalities. The overflow abnormality is mainly manifested as: the lubricating oil does not pass through the cooler, but directly flows back to the gear box from the overflow valve, causing the system temperature to continue to rise and eventually alarm.
[0082] b. Abnormal cooling of the cooling system: After the lubricating oil flows out of the filter, it first passes through the temperature control valve, and then directly enters the oil distributor or enters the oil distributor after passing through the cooler. The cooling problem of the cooling system may be caused by the lubricating oil not entering the cooler (temperature control valve problem) or the cooling capacity problem of the cooler itself.
[0083] b.1. Temperature control valve problem: The temperature control valve can change the flow direction of the lubricating oil according to the change of oil temperature. When the oil temperature is low, the temperature control valve does not open; when the oil temperature is high, the temperature control valve controls the lubricating oil to pass through the cooler before entering the gearbox. When the temperature control valve fails, the lubricating oil temperature is high and it cannot switch from the straight oil circuit to the cooling oil circuit. The high-temperature lubricating oil cannot be cooled, and the temperature of the system will continue to rise until the temperature alarm is sounded.
[0084] b.2 Cooler problem.
[0085] The air cooler uses air to force cool the oil (V12), and the water cooler uses coolant to cool the oil (V11).
[0086] Air cooler problems are often caused by heat dissipation components such as motors and cooling fans, which lead to a decrease in cooling capacity.
[0087] Water cooler problems are often caused by: first, insufficient water coolant and pipe leakage; second, abnormalities in the water pump itself; third, blockage of the water coolant filter, resulting in reduced cooling capacity.
[0088] c. The oil in the gearbox deteriorates and impurities increase, resulting in reduced lubrication and heat dissipation effects.
[0089] d. Temperature sensor - PT100 sensor or circuit has an abnormality, resulting in inaccurate temperature measurement.
[0090] e. The control parameters, stability and adaptability of wind turbines are insufficient, resulting in inadaptability to some environmental conditions.
[0091] (2) Abnormal bearing temperature
[0092] The main reasons for abnormal bearing temperature may be PT100 abnormality, bearing wear, electrical corrosion, abnormal heat dissipation, control strategy and other abnormalities. The temperature of the bearing can be used to judge the abnormal bearing temperature. Specifically, a measurement point is set when the bearing is running at low speed, called the low-speed shaft temperature, and two measurement points are set when the bearing is running at high speed, called the high-speed shaft 1 temperature and the high-speed shaft 2 temperature respectively; when the low-speed shaft temperature, high-speed shaft 1 temperature and high-speed shaft 2 temperature are all high, it is generally because the oil temperature is high, resulting in a decrease in the heat dissipation of the bearing, and the oil pool oil temperature is diagnosed abnormally; when one of the three temperatures far exceeds the other two temperature values, the possibility of abnormal lubrication system and cooling system is ruled out, and the possible reasons for the high temperature can be diagnosed based on the characteristics of the bearing.
[0093] a. Bearing lubrication problem
[0094] If the lubrication oil circuit of the bearing is abnormal or the lubrication oil is insufficient, the bearing temperature will be high.
[0095] b. Problems with the bearing itself
[0096] The bearing roller raceway has faults such as peeling and cracks, which causes the heat to increase during operation and causes the bearing to alarm for high temperature.
[0097] (3) Abnormal pressure in the gearbox
[0098] Gearbox pressure abnormalities generally include abnormal oil pressure and abnormal water (i.e., coolant in the cooling system) pressure, which are mainly divided into: inlet oil pressure (involving V11 and V12), outlet oil pressure (involving V11 and V12), inlet water pressure (involving V11), and outlet water pressure (involving V11).
[0099] Abnormal oil pressure is mainly caused by pump oil, oil circulation pipeline, filter, PT100, etc. Abnormal water pressure is mainly caused by water cooling pump, water cooling circulation pipeline, filter, PT100, etc.
[0100] The main characteristics of oil pressure are as follows: Oil pressure shows periodic high and low changes, mainly because of changes in oil temperature. High oil pressure is caused by low oil pool temperature and high viscosity of lubricating oil, while low oil pressure is caused by high oil temperature and low viscosity of lubricating oil and low resistance. Therefore, there is a negative correlation between oil temperature and oil pressure.
[0101] The main water pressure characteristics are as follows: when the oil pool temperature or bearing temperature reaches the set value, the water cooling circulation system needs to be started to cool down, the water cooling pump starts, the outlet water pressure increases, the inlet water pressure decreases, the oil pool temperature and the bearing temperature will show a downward trend. After the temperature drops to the set value, the water cooling pump stops.
[0102] The wind turbine generator set and the gearbox abnormality prediction method and device disclosed in the present invention are described in detail below with reference to the accompanying drawings.
[0103] The present invention proposes a method for predicting abnormality of a gearbox of a wind turbine generator set. Figure 2 1 is a flow chart showing a method for predicting abnormality of a gearbox of a wind turbine generator set according to an embodiment of the present disclosure. Figure 2 The method for predicting abnormality of a gearbox of a wind turbine generator set comprises the following steps:
[0104] In step S201, operating data of a gearbox of a wind turbine generator set within a predetermined time period is obtained, wherein the operating data includes at least one of the following: operating data of a bearing of the gearbox, operating data of a lubrication system of the gearbox, and operating data of a cooling system of the gearbox.
[0105] Specifically, the wind turbine generator set disclosed in the present invention may be a medium-speed wind turbine generator set, a doubly-fed wind turbine generator, etc., which is not limited in the present disclosure. The above-mentioned predetermined time length can be set as needed, which is not limited in the present disclosure. In step S201, the operating data of the bearings, lubrication system and cooling system in the gearbox within the predetermined time length can be obtained, and the operating data of one or both of the bearings, lubrication system and cooling system within the predetermined time length can also be obtained, which is not limited in the present disclosure. It should be noted that the gearbox disclosed in the present invention not only includes bearings, lubrication systems and cooling systems, but also may include other components. Since the operating data used in the present disclosure only needs to involve the operating data of the bearings, the operating data of the lubrication system and the operating data of the cooling system, only the bearings, lubrication systems and cooling systems in the gearbox are reflected.
[0106] As an example, taking the operation data of the bearing, lubrication system and cooling system in the gearbox within a predetermined time as an example, the SCADA data of the wind turbine generator set one week before the current time can be obtained, and then the wind speed, rotation speed, power, ambient temperature, gearbox related sensor measurement points and other data can be extracted from the data, mainly involving temperature sensor, wind sensor, speed sensor, pressure sensor data, and all are structured data, and the time interval period can be 3-7s. The SCADA data extraction measurement point details can be shown in Table 1 below,
[0107] Table 1 SCADA data measurement points
[0108]
[0109] The above gearbox main sensor measurement point data can include:
[0110] 1. Floating point type (REAL) data, such as the oil pool temperature and bearing temperature used in the gearbox control logic in Table 2;
[0111] Table 2 Gearbox control logic
[0112]
[0113]
[0114] 2. Boolean type (BOOL) data. Such as the gear pump low speed running state (such as low speed running or not low speed running), gear pump high speed running state, filter pressure difference, gear box cooling fan low speed running state, gear box cooling fan high speed running state, gear heater action state.
[0115] It should be noted that the relevant measuring point data of the above-mentioned gearbox may include but are not limited to: the low-speed shaft temperature of the gearbox, the high-speed shaft 1 temperature, the high-speed shaft 2 temperature, the oil pool temperature, the oil pool inlet oil temperature, the gearbox outlet oil pressure, the gearbox inlet oil pressure, the outlet water pressure of the cooling system inlet and outlet valves, the inlet water pressure, etc., among which the low-speed shaft indicates that the bearing is in a low-speed running state, and the high-speed shaft indicates that the bearing is in a high-speed running state.
[0116] As an example, still taking the operation data of the bearing, lubrication system and cooling system in the gearbox within a predetermined time as an example, after obtaining the operation data of the bearing, lubrication system and cooling system in the gearbox within a predetermined time, the obtained operation data can be preprocessed, and the data preprocessing may include but is not limited to: data cleaning, format standardization, vacancy value processing, feature variable extraction or reconstruction, elimination of noise and data that does not conform to business logic, unit maintenance process data, etc. After a series of preprocessing, clean data is obtained, and then analysis and modeling are performed. It should be noted that, in general, the amount of the original data obtained must reach 60% or more of the theoretical data amount before entering the preprocessing link, and the amount of the preprocessed data must be no less than 1.5 hours. For example, the above data preprocessing may include: 1) converting the time column data into a date format; 2) eliminating abnormal data formats at each measuring point; 3) removing duplicates, empty spaces, and NA; 4) eliminating noise values and abnormal values of related variables, and screening data that conform to business logic and specific working conditions; 5) reconstructing feature variables.
[0117] In step S202, a plurality of data sequences are extracted from the operating data, wherein the plurality of data sequences include a continuously increasing data sequence and a continuously decreasing data sequence in the operating data.
[0118] Specifically, after the data preprocessing is completed, multiple data sequences can be extracted from the processed operating data. The multiple data sequences can include: continuously increasing data sequences and continuously decreasing data sequences in the operating data. Through such data sequences, some useless discrete data can be avoided and the accuracy of the prediction can be improved. The multiple data sequences can also include: specific sequences identified by the relevant control logic of the gearbox, such as the data sequence of the temperature control valve being half open, the data sequence of the temperature control valve being fully open and completely circulating externally, etc. Such data sequences are generally targeted at specific faults and can achieve better prediction results under characteristic fault scenarios.
[0119] As an example, the above-mentioned continuously increasing data sequence can be a continuous monotonically increasing data sequence, so as to avoid the appearance of partially decreasing fragments in the continuously increasing data sequence, making it impossible to accurately predict the extracted data sequence; the above-mentioned continuously decreasing data sequence can be a continuous monotonically decreasing data sequence, so as to avoid the appearance of partially increasing fragments in the continuously decreasing data sequence, making it impossible to accurately predict the extracted data sequence.
[0120] According to an embodiment of the present disclosure, extracting multiple data sequences from the operating data may include: classifying the operating data according to the attributes of the data in the operating data; clustering the classified operating data under each category to obtain at least one cluster of data; and extracting multiple data sequences from the current data for each cluster of data in the at least one cluster of data. According to this embodiment, in order to ensure the accuracy of the prediction, the operating data may be clustered before extraction, so that the data sequences extracted from each cluster of data after clustering have greater similarity, and the differences between the data sequences of different clusters are as large as possible, so that a relatively accurate operating trend of the gearbox can be determined.
[0121] Specifically, the above clustering process can adopt the method of machine learning, that is, the operation data is classified according to the attributes of the data in the operation data, such as the operation data of one attribute is classified into one category, and the operation data under each category after classification can be input into the pre-trained clustering model to obtain the corresponding clustering results, such as obtaining at least one cluster of data, the similarity of the data objects in the same cluster after clustering is as large as possible, and the difference of the data objects not in the same cluster after clustering is also as large as possible, so that the data sequence extracted from each cluster of data after clustering can relatively accurately determine the operation trend of the gearbox. It should be noted that since the amount of operation data disclosed in the present invention is large and relatively dense, clustering processing using machine learning is more suitable for this scenario and can also achieve better clustering effects.
[0122] According to an embodiment of the present disclosure, the attributes of the above data may include but are not limited to: bearing temperature, oil sump temperature of the lubrication system, oil sump inlet oil temperature of the lubrication system, oil pressure of the lubrication system oil entering the gearbox, oil pressure of the lubrication system oil flowing out of the gearbox, inlet valve water pressure of the cooling system and outlet valve water pressure of the cooling system.
[0123] As an example, taking the data attributes including bearing temperature and oil pool temperature as an example, multiple operating data are divided into two categories, one is bearing temperature and the other is oil pool temperature. The operating data under the two categories are clustered respectively. If they are input into a pre-trained clustering model, the corresponding clustering results are obtained, and then the data sequence extracted from each cluster of data after clustering can determine a relatively accurate operating trend of the gearbox.
[0124] In step S203, the operation trend of the gearbox is determined based on multiple data sequences. The above operation trend can be reflected by but not limited to the following features: data time series feature, data trend feature, data distribution feature.
[0125] Specifically, a non-parametric verification method (such as Mann-Kendall) combined with multi-order lag cross validation can be used to process multiple data sequences to obtain the operating trend of the gearbox. Since the amount of operating data disclosed in the present invention is large and relatively dense, a non-parametric verification method (such as Mann-Kendall) combined with multi-order lag cross validation is more suitable for this scenario. Therefore, a non-parametric verification method (such as Mann-Kendall) combined with multi-order lag cross validation is used to determine the operating trend of the gearbox, and a relatively accurate prediction result can be obtained.
[0126] According to an embodiment of the present disclosure, determining the operating trend of a gearbox based on multiple data sequences may include: determining a predetermined statistical indicator for each data sequence; classifying multiple data sequences according to the attributes of the data in the data sequence; determining the indicator change trend of the predetermined statistical indicator under each category after classification; and determining the operating trend of the gearbox based on the indicator change trend. According to this embodiment, after extracting a continuously increasing data sequence and a continuously decreasing data sequence from the operating data of the gearbox, the predetermined statistical indicator of each data sequence may be calculated, and the data sequences may be classified according to the data attributes, so that the change trend of the gearbox may be conveniently and quickly determined according to the change trend of the predetermined statistical indicator under each category.
[0127] The above-mentioned predetermined statistical indicators include but are not limited to: mean value, maximum value, upper quartile value, range, coefficient of variation, variance, standard deviation, etc. In general, different statistical indicators can be used for data with different attributes. For example, the maximum value and the average value can be used for the export oil pressure. These two indicators can better highlight the abnormal state of the export oil pressure and make it easy to observe abnormal conditions. The maximum value and the average value can also be used for the import oil pressure. Similarly, these two indicators can better highlight the abnormal state of the import oil pressure and make it easy to observe abnormal conditions. For data with other attributes, appropriate statistical indicators can also be determined through experiments or experience, which is not limited by the present disclosure.
[0128] As an example, after obtaining multiple data sequences, the average value of each data sequence can be calculated, and the multiple data sequences can be classified according to the attributes of the data in the data sequence. For example, the data sequence of one attribute is classified into one category. Assuming that one of the attributes is the outlet oil pressure, the change trend of the average value corresponding to the classified inlet oil pressure can be determined. For other attributes, the adapted statistical indicators can be calculated, so that the change trend of the statistical indicators under each category can be used as the operating trend of the gearbox. To put it simply, the changing trend of the statistical indicators is used to characterize the operating trend of the gearbox, and the changing trend of the statistical indicators under each category can be used to characterize the operating trend of the gearbox, that is, the operating trend of the gearbox may include the changing trends of the statistical indicators under multiple categories.
[0129] As an example, the statistical indicators can be processed by non-parametric verification methods (such as Mann-Kendall) combined with multi-order lagged cross-validation to obtain the operating trend of the gearbox.
[0130] As an example, taking the data attributes including bearing temperature and oil pool temperature, and the predetermined statistical indicator of the two attributes being the average value, multiple data series are divided into two categories, one is the data series corresponding to the bearing temperature, and the other is the data series corresponding to the oil pool temperature, and the changing trends of the average values under the two categories are determined. The changing trends of the average values under each category can be used as the operating trend of the gearbox.
[0131] In step S204, an abnormal condition of the gearbox is predicted based on the operating trend and the predetermined abnormal operating trend.
[0132] The above-mentioned predicted abnormal conditions of the gearbox may include an upcoming abnormality, and may also include the location of the cause of the abnormality, provide reference and guidance for problem solving, and may also include other contents, which are not limited by the present disclosure. The above-mentioned abnormal operation trend may be based on the historical operation data of the gearbox, design simulation data, horizontal comparison of multiple wind turbines and multiple wind farms, vertical comparison of a single wind turbine, and fault data, etc., to compare and analyze and identify the operation trend of the gearbox. The specific process will be described in detail later, and will not be discussed here. After determining the operation trend of the gearbox, it can be compared with the abnormal operation trend. If the determined operation trend is consistent with the abnormal operation trend, it can be determined that the gearbox is expected to have an abnormal situation and what causes the abnormality. In addition, it is also possible to determine that the gearbox is expected to have an abnormal situation by determining that the operation trend is consistent with the abnormal operation trend in a certain proportion. For example, the determined operation trend is 90% or 80% consistent with the abnormal operation trend. It can also be determined that the gearbox is expected to have an abnormal situation, and the abnormal situation expected to occur is the same as the abnormal situation corresponding to the abnormal operation trend.
[0133] According to an embodiment of the present disclosure, before predicting the abnormal condition of the gearbox based on the operation trend and the predetermined abnormal operation trend, the historical operation data of the gearbox can be obtained as the first historical operation data; the first historical operation data can be segmented according to the predetermined time length and time sequence to obtain multiple segments of first sub-historical operation data; for each segment of the first sub-historical operation data, multiple data sequences in the current first sub-historical operation data are extracted, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current first sub-historical operation data; based on the multiple data sequences of each segment of the first sub-historical operation data, the operation trend of the gearbox before the abnormal condition of the gearbox occurs is determined as the abnormal operation trend. According to this embodiment, after a large amount of historical operation data of the gearbox is segmented according to the predetermined time length, the multiple data sequences of each segment of the first sub-historical operation data are extracted, so that the relatively accurate and reliable abnormal operation trend of the gearbox can be determined through the longitudinal comparison of the data of the gearbox of the wind turbine generator set.
[0134] As an example, still taking the operation data of the bearings, lubrication system and cooling system in the gearbox within a predetermined time as an example, the SCADA data of the wind turbine generator set in the first half of the year before the current time can be collected. In order to match the predetermined time of the actual prediction, the SCADA data of the first half of the year can be divided into multiple sections according to the predetermined time and time sequence, and then multiple data sequences in each section of operation data are extracted to determine the operation trend of the gearbox before the abnormal situation occurs in the gearbox. This can not only be conveniently applied to actual predictions, but also can realize the longitudinal comparison of the data of the gearbox of the wind turbine generator set through the comparison between multiple sections of operation data, and then determine a relatively accurate and reliable abnormal operation trend.
[0135] As an example, suppose that this wind turbine generator set triggered bearing temperature fault, abnormal inlet water pressure, abnormal outlet water pressure, abnormal oil pressure and other faults during the six months before the current time. At this time, data such as wind speed, rotation speed, power, ambient temperature, gearbox related sensor measurement points, etc. can be obtained. It mainly involves temperature sensor, wind sensor, speed sensor, pressure sensor data and all are structured data, and the time interval period can be 3-7s. For each of the above data of the wind turbine generator set, it can be segmented according to the predetermined duration and time sequence, and then multiple data sequences in each segment of data can be extracted to analyze whether there are data abnormality features before the wind turbine generator set is abnormal. If there are data abnormality features, the corresponding features can be used as abnormal operation trends.
[0136] It should be noted that each type of wind turbine has its own corresponding abnormal situation, but no matter what type of wind turbine, the corresponding abnormal operation trend can be determined by the method disclosed in the present invention. The following is an introduction to the abnormal situation of only one type of wind turbine. Generally, the data at different measuring points will have their own time series characteristics when the gearbox is operating normally, and the data at different measuring points before the gearbox is abnormal will also have their own time series characteristics. The following describes the time series characteristics of the data at different measuring points when the gearbox is operating normally and before the abnormality occurs.
[0137] When the gearbox is operating normally:
[0138] The timing characteristics of the bearing temperature: the temperature of high-speed shaft 2 is higher than that of high-speed shaft 1, and the temperature of high-speed shaft 1 is higher than that of low-speed shaft. The bearing temperature shows a trend of periodic increase and decrease. As the heat of the gearbox accumulates, the bearing temperature gradually rises. When the temperature rises to the critical point of heat dissipation, the wind turbine starts to dissipate heat (external circulation is turned on, external circulation is fully turned on, low-speed / high-speed cooling fans, and water cooling pumps are started), and the bearing temperature gradually decreases. When the gearbox and related components are abnormal, the timing characteristics will not conform to the timing characteristics of the wind turbine design. The following shows the timing characteristics of some data with normal bearing temperatures of wind turbines, such as Figure 3 As shown, the red curve is the temperature of high-speed shaft 1, the green curve is the temperature of high-speed shaft 2, and the blue curve is the temperature of the low-speed shaft.
[0139] Time series characteristics of oil pool temperature and oil pool inlet oil temperature: When the gearbox lubrication system and cooling system are normal, after the oil pool temperature and oil pool inlet oil temperature reach the set value, the wind turbine will implement different heat dissipation strategies to control the oil pool temperature and oil pool inlet oil temperature within a reasonable range. The following shows the time series characteristics of some data of normal oil pool temperature and oil pool inlet oil temperature of the wind turbine, such as Figure 4 As shown, the red curve is the oil pool oil temperature and the blue curve is the oil pool inlet temperature.
[0140] The timing characteristics of the gearbox inlet and outlet oil pressure, as well as the gearbox coolant inlet and outlet water pressure, where the oil pressure characteristics are mainly based on the pressure changes caused by temperature rise and fall, and then generate two types of data: high oil pressure and low oil pressure. The following shows the timing characteristics of some data with normal oil pressure and water pressure of wind turbines, such as Figure 5 and Figure 6 As shown, Figure 5 The red curve is the inlet oil pressure, and the blue curve is the outlet oil pressure. Figure 6 The red curve is the inlet water pressure, the blue curve is the outlet water pressure, and the disclosure also shows two types of oil pressure data, such as Figure 7 and Figure 8 As shown, Figure 7 It is the relationship between the inlet oil pressure and the outlet oil pressure of the low oil pressure type. Figure 8 It is the relationship between the inlet oil pressure and the outlet oil pressure under the high oil pressure type. The two relationships are that the oil pressure changes with the gearbox oil temperature, resulting in changes in oil viscosity and resistance, and the impact on the oil pressure.
[0141] When the gearbox lubrication system and cooling system are abnormal, the bearing temperature and oil pool temperature will not drop or drop very little after reaching the heat dissipation critical point, resulting in a shorter time for the bearing temperature to reach its peak value, a higher peak value, and a longer duration. At the same time, when the heat dissipation effect decreases, the oil pool inlet oil temperature will fluctuate slightly, and the oil pool inlet temperature full range, variance, standard deviation and other indicators will drop significantly until the data at each measuring point of the gearbox reaches the fault value, causing the wind turbine to shut down.
[0142] When the gearbox was running before the abnormality occurred:
[0143] The analysis is mainly focused on abnormal bearing temperature, abnormal oil pool temperature, abnormal oil pool inlet temperature, abnormal inlet and outlet oil pressure, and abnormal inlet and outlet water pressure. Specifically, based on statistical indicators, trend analysis, machine learning analysis, and the operating mechanism of the gearbox, when the oil pool temperature, bearing temperature, oil pressure, and water pressure reach the set values, the wind turbine will execute corresponding lubrication, cooling, and other control strategies, and then detect whether the data of the relevant measuring points of the gearbox have returned to a reasonable range that meets the design requirements of the wind turbine gearbox, identify and quantify the characteristics of the data, and diagnose the current health status of the gearbox. In general, the abnormal data of each measuring point of the gearbox is visualized. This disclosure lists a schematic diagram of abnormal conditions that may occur in a wind turbine, which can be seen in detail. Figure 9-Figure 20 It should be noted that other types of wind turbines may have their own corresponding abnormal conditions, which are not limited in this disclosure. Fig. 9 is a schematic diagram showing a data sequence of a bearing temperature that increases monotonically continuously according to an embodiment of the present disclosure, Fig.10 is a schematic diagram showing a type of statistical index of the bearing temperature of an embodiment of the present disclosure, wherein RANGE represents the range, VAR represents the mean square error, Fig.11 is a schematic diagram showing another type of statistical index of the bearing temperature of an embodiment of the present disclosure, wherein SD represents standard deviation, CV represents coefficient of variation, KURTOSIS represents kurtosis of continuous subsequence temperature of a high-speed shaft, Fig.12 is a schematic diagram showing the mean values of multiple data series of daily water pressure according to an embodiment of the present disclosure, Fig.13 is a schematic diagram showing the water pressure when the water cooling pump of the embodiment of the present disclosure is started, wherein red represents the inlet water pressure, blue represents the outlet water pressure, Fig.14is a schematic diagram showing an abnormal sequence of inlet oil temperature according to an embodiment of the present disclosure, Fig.15 is a schematic diagram showing a type of statistical indicators of the oil pool oil temperature of an embodiment of the present disclosure, wherein RANGE represents range, VAR represents mean square error, SD represents standard deviation, CV represents coefficient of variation, KURTOSIS represents kurtosis of high-speed shaft continuous subsequence temperature, Fig.16 is a schematic diagram showing a type of statistical index of the oil pool inlet oil temperature of an embodiment of the present disclosure, wherein RANGE represents range, VAR represents mean square error, SD represents standard deviation, CV represents coefficient of variation, KURTOSIS represents kurtosis of high-speed shaft continuous subsequence temperature, Fig.17 is a schematic diagram showing the inlet and outlet oil pressures of an embodiment of the present disclosure, Fig.17 Different colors represent the inlet and outlet oil pressures of different units. The upper violin graph represents the gearbox outlet oil pressure, and the lower violin graph represents the inlet oil pressure. The unit with a high pressure value is a normal unit, and the unit with a low pressure value is an abnormal unit. Fig.18 is a schematic diagram showing the comparison of different clustering results of inlet oil pressure of different units according to an embodiment of the present disclosure, wherein: Fig.18 1 in the middle represents low-pressure data, 2 represents high-pressure data, and different lines represent different units. Fig.19 is a schematic diagram showing the comparison of different clustering results of outlet oil pressures of different units according to an embodiment of the present disclosure, wherein: Fig.19 1 in the middle represents low-pressure data, 2 represents high-pressure data, and different lines represent different units. Fig. 20 Schematic diagram showing an example of high-speed shaft temperature anomaly according to an embodiment of the present disclosure.
[0144] The following is only for the above Figure 9-Figure 20 The middle part of the figure (such as Figure 9-11 , Figure 12-13 ) are described in detail.
[0145] Characteristics of abnormal gearbox bearing temperature: After the bearing temperature reaches the set value, the gear pump runs at high speed and the cooling system is started (air cooling, water cooling). The bearing temperature gradually cools down with the circulation of the lubricating oil, and there will be a slow downward trend. Generally, the duration, number and maximum value of the continuous increase of the bearing temperature are relatively stable and fixed. However, when the bearing itself is abnormal or the heat dissipation is abnormal, the sequence of the continuous increase of the bearing temperature and its statistical indicators will show data characteristics that are biased towards the design of the wind turbine generator set, such as Figure 9-11 .
[0146] When the water pressure of the gearbox cooling system is abnormal, the following characteristics are reflected: the daily water cooling inlet and outlet valve pressure has the characteristics of multiple continuous increases and decreases. Due to the large changes in the data time series, the water pressure data can be moved averaged for 1 minute to smooth the water pressure data. The data of continuous decrease and increase of water pressure for more than 5 minutes and above can be identified and extracted. When the oil pool temperature and bearing temperature reach the conditions for water cooling to start heat dissipation, the statistical characteristics of the subsequence data of the water cooling pressure when the water cooling pump is started (cluster = 1) and the water cooling pump is not started (cluster = 2) are calculated. It can be found that the water pressure has an obvious downward trend and abnormality, such as Fig.12 and Fig.13 ,in, Fig.12 1 represents the continuous sequence of water pressure when water cooling is started, 2 represents the continuous sequence of water pressure when water cooling is not started, and the dotted line and solid line represent the comparison of percentile and mean, respectively. The oil pressure characteristics are similar to the water pressure characteristics and are not shown here.
[0147] For the abnormal features shown in the above figure, the method disclosed in the present invention can be used to divide the data of each attribute in the operating data (such as bearing temperature data, water pressure data) into multiple segments according to a predetermined duration and time order, and then extract multiple data sequences in each segment of data, such as a continuously increasing data sequence and a continuously decreasing data sequence. By performing trend analysis on these data sequences, the data form of the features reflected in the above figure can be extracted and used as the abnormal operation trend. When predicting abnormal situations subsequently, it is only necessary to use the determined abnormal operation trend without the help of graphics, thus avoiding the step of generating graphics.
[0148] According to an embodiment of the present disclosure, based on multiple data sequences of each section of the first sub-historical operating data, determining the operating trend of the gearbox before an abnormal situation occurs in the gearbox may include: determining a predetermined statistical indicator for each data sequence; performing the following processing on each section of the first sub-historical operating data: classifying multiple data sequences of the current first sub-historical operating data according to the attributes of the data in the data sequence; determining the indicator change trend of the predetermined statistical indicator under each category after classification; determining the operating trend of the gearbox before an abnormal situation occurs based on the indicator change trend under each category of each section of the first sub-historical operating data. According to this embodiment, after extracting continuously increasing data sequences and continuously decreasing data sequences from the operating data of the gearbox, the predetermined statistical indicator of each data sequence can be calculated, and the data sequences can be classified according to the data attributes, so that the change trend of the gearbox can be conveniently and quickly determined according to the change trend of the predetermined statistical indicator under each category.
[0149] The above-mentioned predetermined statistical indicators include but are not limited to: mean value, maximum value, upper quartile value, range, coefficient of variation, variance, standard deviation, etc. In general, different statistical indicators can be used for data with different attributes. For example, the maximum value and the average value can be used for the export oil pressure. These two indicators can better highlight the abnormal state and make it easy to observe the abnormal situation. The maximum value and the average value can also be used for the import oil pressure. Similarly, these two indicators can better highlight the abnormal state and make it easy to observe the abnormal situation. For data with other attributes, appropriate statistical indicators can also be determined through experiments or experience, which is not limited by the present disclosure.
[0150] According to an embodiment of the present disclosure, based on the indicator change trend under each category of each section of the first sub-historical operating data, determining the operating trend of the gearbox before the abnormal situation occurs in the gearbox can include: performing the following processing for each category: comparing the indicator change trend under the current category of each section of the first sub-historical operating data; based on the comparison result, determining the first sub-historical operating data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operating data; determining the abnormal operating data based on the estimated abnormal operating data of each category; determining the indicator change trend under each category of the abnormal operating data as the operating trend of the gearbox before the abnormal situation occurs in the gearbox. According to this embodiment, by comparing the indicator change trend under each section of the operating data, the longitudinal comparison of the data of the gearbox of the wind turbine generator set can be conveniently and quickly realized, thereby determining a relatively accurate and reliable abnormal operating trend of the gearbox.
[0151] As an example, taking the data attributes including bearing temperature and oil pool temperature, and the predetermined statistical index of the two attributes being the average value as an example, multiple data sequences are divided into two categories, one is the data sequence corresponding to the bearing temperature, and the other is the data sequence corresponding to the oil pool temperature. For the data sequence corresponding to the bearing temperature, the index change trend under the current category of each segment of the first sub-historical operation data can be compared. According to the comparison result, the first sub-historical operation data corresponding to the abnormal situation of the gearbox can be determined. If the comparison result shows that the index change trend of one segment of the first sub-historical operation data is very different from the index change trend of other segments of the first sub-historical operation data, it can be determined that the segment of the first sub-historical operation data is the estimated abnormal operation data. Similarly, for the data sequence corresponding to the oil pool temperature, a segment of estimated abnormal operation data can also be determined. After obtaining two segments of estimated abnormal operation data, it can be determined whether the two segments of estimated abnormal operation data are the same segment of sub-historical operation data. If so, it can be determined that the estimated abnormal operation data is the abnormal operation data. Otherwise, the estimated abnormal operation data can be re-determined, or new factors can be added to determine the abnormal operation data, which is not limited by the present disclosure. After the abnormal operation data is determined, the indicator change trend under each category of the abnormal operation data can be used as the operation trend of the gearbox before the abnormal situation occurs in the gearbox.
[0152] According to an embodiment of the present disclosure, based on the first historical operation data, determining the operation trend of the gearbox before the abnormal situation occurs in the gearbox may include: obtaining the historical operation data of the gearbox in normal operation in other wind turbines as the second historical operation data; segmenting the second historical operation data according to a predetermined time length and time sequence to obtain multiple segments of second sub-historical operation data; for each segment of the second historical operation data, extracting multiple data sequences in the current second sub-historical operation data, wherein the multiple data sequences include a continuously increasing data sequence and a continuously decreasing data sequence in the current second sub-historical operation data; based on the multiple data sequences of each segment of the first sub-historical operation data and the multiple data sequences of each segment of the second sub-historical operation data, determining the operation trend of the gearbox before the abnormal situation occurs in the gearbox. According to this embodiment, further combined with the operation data of the gearbox in normal operation of other units, a more accurate and reliable abnormal operation trend of the gearbox can be determined.
[0153] As an example, still taking the operation data of the bearing, lubrication system and cooling system in the gearbox within a predetermined time as an example, the SCADA data of the wind turbine generator set in the first half of the current time and the SCADA data of the gearbox of other wind turbine generator sets in normal operation in the first half of the current time can be collected. Assuming that the wind turbine generator set triggers bearing temperature fault, abnormal inlet water pressure, abnormal outlet water pressure, abnormal oil pressure and other faults in the first half of the current time, the wind speed, rotation speed, power, ambient temperature, gearbox related sensor measurement points and other data in the operation data of all units can be extracted at this time, mainly involving temperature sensor, wind sensor, speed sensor, pressure sensor data and all are structured data, and the time interval period can be 3-7s. Through the longitudinal comparison of the operation data of the wind turbine generator set and the horizontal comparison of the operation data of other wind turbine generator sets, it is analyzed whether the wind turbine generator set has data abnormality characteristics before the abnormality occurs. If there are data abnormality characteristics, the corresponding characteristics can be used as abnormal operation trends.
[0154] According to an embodiment of the present disclosure, multiple data sequences of each section of the first sub-historical operating data and multiple data sequences of each section of the second sub-historical operating data determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox, which may include: determining a predetermined statistical indicator for each data sequence; performing the following processing on each section of the first sub-historical operating data and each section of the second sub-historical operating data: classifying the multiple data sequences of the current sub-historical operating data according to the attributes of the data in the data sequence; determining the indicator change trend of the predetermined statistical indicator under each category after classification; based on the indicator change trend under each category of each section of the second sub-historical operating data and the indicator change trend under each category of each section of the first sub-historical operating data, determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox. According to this embodiment, after extracting a continuously increasing data sequence and a continuously decreasing data sequence from the operating data of the gearbox, the predetermined statistical indicator of each data sequence can be calculated, and the data sequences can be classified according to the data attributes, so that the change trend of the gearbox can be conveniently and quickly determined according to the change trend of the predetermined statistical indicator under each category.
[0155] According to an embodiment of the present disclosure, based on the indicator change trend under each category of each section of the second sub-historical operation data and the indicator change trend under each category of each section of the first sub-historical operation data, determining the operation trend of the gearbox before the abnormal situation occurs in the gearbox can include: performing the following processing for each category: comparing the indicator change trend under the current category of each section of the first sub-historical operation data and each section of the second sub-historical operation data; based on the comparison result, determining the sub-historical operation data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operation data; based on the estimated abnormal operation data of each category, determining the abnormal operation data; determining the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox. According to this embodiment, by comparing the indicator change trend under each section of the operation data, the longitudinal comparison of the data of the gearbox of the wind turbine generator set and the lateral comparison with the operation data of other wind turbine generator sets can be conveniently and quickly realized, thereby determining a relatively accurate and reliable abnormal operation trend of the gearbox.
[0156] According to an embodiment of the present disclosure, before determining the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation of the gearbox occurs, the historical fault information of the gearbox can be obtained; based on the historical fault information, the abnormal operation data is determined to be the actual abnormal operation data. According to this embodiment, further combined with the fault information of the gearbox, a more accurate abnormal operation trend can be determined.
[0157] Normally, when an abnormality occurs in the bearing, lubrication system or cooling system of the gearbox, data from one or more measurement points of the wind turbine will trigger the corresponding warning and fault protection mechanism of the wind turbine. By analyzing and mining these real fault data, data support can be provided for early warning of abnormalities, location of fault causes and improvements.
[0158] As an example, since the above vertical comparison and horizontal comparison only determine the abnormal operation trend based on different movement trends, it is very likely that the following situation exists: although the operation trend at a certain time is quite special and different from the operation trends of other units at that time, the operation trend is not necessarily an abnormal operation trend. If combined with the above real fault information, it is more accurate to determine whether the abnormal operation trend is accurate. It should be noted that the above fault information includes but is not limited to warning information, fault protection mechanism information, and may also include fault data involved in simulation, etc.
[0159] According to an embodiment of the present disclosure, the above-mentioned historical operation data can be classified according to the attributes of the data; the historical operation data under each category after classification can be clustered to obtain at least one cluster of data; each cluster of data in the at least one cluster of data can be segmented and processed, and the corresponding data sequence can be extracted, which is not limited by the present disclosure.
[0160] In order to facilitate the understanding of the present disclosure, Fig.21 Describe the system. Fig.21 The present invention is a schematic diagram showing the gearbox abnormality prediction process of the embodiment of the present invention. Through the operation mechanism of the gearbox, after the oil pool temperature and the bearing temperature reach the set value, the action sequence and action result of the wind turbine generator set are identified through the characteristics of the gearbox operation data. The characteristics of the operation data of the wind turbine generator set that does not act according to the set requirements and the effect after the action does not reach the expected level are identified and extracted, so as to give an early warning of the abnormal characteristics before the gearbox operation is abnormal. Specifically, as Fig.21 As shown in the figure, the process of gearbox abnormality prediction mainly includes:
[0161] Step 1: Obtain SCADA data of several units within the Xt period (predetermined duration), judge the data volume of the obtained SCADA data, and if the judgment result meets the requirements, pre-process the obtained SCADA data, and judge the data volume after pre-processing again. If the judgment result meets the requirements, proceed to the next step.
[0162] Step 2: According to the relevant measurement point data of the gearbox and the operating mechanism of the gearbox (such as control strategy parameters), the pre-processed data is identified and marked, and continuous monotonically increasing subsequences and continuous monotonically decreasing subsequences of different time periods and different lengths in the pre-processed data are extracted, and the statistical indicators of the relevant characteristics of the gearbox are calculated based on the extracted subsequences. Trend analysis is performed through statistical indicators, and multi-order lagged cross-validation judgment can also be performed. In this process, clustering (DBSCAN and other methods) can also be introduced to cluster the data. The time point for introducing clustering can be set as needed, such as clustering after acquiring the data, or clustering after data preprocessing, which is not limited in this disclosure. After the above processing, feature quantization is achieved.
[0163] Step 3: Comprehensively judge the calculation results in step 2 to determine the abnormal situation of the gearbox. This may include determining the abnormal operation trend of the gearbox, and comparing the current operation trend of the gearbox with the abnormal operation trend, and other comprehensive judgments, which are not limited by the present disclosure.
[0164] Step 4: When it is determined that the gearbox is about to have an abnormal situation, the warning result is output. For example, 5 warning results Alarm1-5 can be set, where Alarm1 indicates that the temperature of the gearbox bearing (low-speed shaft, high-speed shaft 1, high-speed shaft 2) is abnormal, Alarm2 indicates that the oil temperature of the gearbox oil pool is abnormal, Alarm3 indicates that the inlet oil temperature of the gearbox oil pool is abnormal, Alarm4 indicates that the inlet and outlet oil pressure of the gearbox is abnormal, and Alarm5 indicates that the cooling pressure of the water-cooled inlet and outlet valves of the gearbox is abnormal. Of course, the sum can include other warning results, and the present disclosure is not limited to them.
[0165] In summary, the method disclosed in the present invention mainly uses statistical knowledge, machine learning, big data and other technologies, integrates information such as the operating mechanism of the gear image, and based on a large amount of online SCADA multi-sensor measurement point data of the wind turbine generator set, realizes accurate and reliable identification and early warning of the health status and abnormal characteristics of the gearbox of the wind turbine generator set, for example, including identification of common abnormal conditions such as the low-speed shaft, high-speed shaft, oil pool oil, heat dissipation, wear, and oil / water filter blockage of the gearbox.
[0166] As an example, by testing the data of a faulty wind turbine generator set, the present disclosure can accurately and reliably detect abnormalities related to the gearbox of the wind turbine generator set about one week in advance of the wind turbine generator set fault reporting time.
[0167] For all wind turbines in multiple wind farms, corresponding data are collected and predicted through the present disclosure, and finally abnormal bearing temperatures of wind turbines 1#, 2#, and 7#, abnormal gearbox inlet and outlet oil pressures, and abnormal gearbox heat dissipation water cooling inlet and outlet water pressures are identified.
[0168] Assuming that wind turbine generator sets 2# and 7# in a certain wind farm report a high-speed shaft temperature overtime fault on January 15, 2023 and January 23, 2023, the method disclosed in the present invention is used to run the data of the corresponding wind turbine generator set one month before the failure, and the high-speed shaft temperature high warning can be triggered on January 14, 15, 17, and 23. On the basis of the warning of high bearing temperature, the reason for the high bearing temperature can be identified through the data as the abnormal heat dissipation of the gearbox. The warning result can be seen Fig. 22 , Fig. 22 Only part of the result data between the start time and the end time of the warning is given, where alarm is 1 for abnormal warning, alarm is 0 for no abnormal warning, wtid represents the number of the wind turbine, and rectime_day represents the date, that is, the date of the operation data. One day can correspond to multiple subsequences. The display form of the warning results is not limited to Fig. 22 , and may also be in other forms, which are not limited by the present disclosure.
[0169] Assume that the 1# unit of another wind farm reports a high-speed shaft temperature overtime fault on December 7, 2022. The method disclosed in the present invention is used to run the data of the unit one month before the fault. The model can trigger high bearing temperature, oil pool temperature, and oil pool inlet temperature warnings on November 21, 24-26, more than one week before the fault. Based on the warning abnormality, the data can identify that the cause of the abnormality is caused by abnormal gearbox heat dissipation, rather than abnormal gearbox pump oil. The warning results are shown in Fig.23 , Fig.23 Only part of the result data between the start time and the end time of the warning is given, where alarm is 1 for abnormal warning, alarm is 0 for no abnormal warning, and rectime_day represents the date, that is, the date of the running data. One day can correspond to multiple subsequences. The display form of the warning results is not limited to Fig.23 , and may also be in other forms, which are not limited by the present disclosure.
[0170] The method disclosed in the present invention has been tested on 70 million SCADA data of more than 30 wind turbines. It can accurately and early identify and warn different measurement point data and different abnormal types corresponding to gearbox abnormalities.
[0171] Fig.24 is a block diagram showing a gearbox abnormality prediction device for a wind turbine generator set disclosed in the present invention, such as Fig.24 As shown, the device includes an acquisition unit 240, an extraction unit 242, a determination unit 244 and a prediction unit 246.
[0172] The acquisition unit 240 is configured to acquire the operating data of the gearbox of the wind turbine generator set within a predetermined time period, wherein the operating data includes at least one of the following: operating data of the bearings of the gearbox, operating data of the lubrication system of the gearbox and operating data of the cooling system of the gearbox; the extraction unit 242 is configured to extract multiple data sequences from the operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the operating data; the determination unit 244 is configured to determine the operating trend of the gearbox based on the multiple data sequences; the prediction unit 246 is configured to predict abnormal conditions of the gearbox based on the operating trend and a predetermined abnormal operating trend.
[0173] According to an embodiment of the present disclosure, the determination unit 244 is further configured to determine predetermined statistical indicators for each data sequence; classify multiple data sequences according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicators under each category after classification; and determine the operating trend of the gearbox based on the indicator change trend.
[0174] According to an embodiment of the present disclosure, the extraction unit 242 is also configured to extract multiple data sequences from the operating data, including: classifying the operating data according to the attributes of the data in the operating data; clustering the operating data under each classified category to obtain at least one cluster of data; and for each cluster of data in the at least one cluster of data, extracting multiple data sequences from the current data.
[0175] According to an embodiment of the present disclosure, the prediction unit 246 is also configured to obtain historical operating data of the gearbox as first historical operating data before predicting the abnormal condition of the gearbox based on the operating trend and a predetermined abnormal operating trend; segment the first historical operating data according to a predetermined duration and time sequence to obtain multiple segments of first sub-historical operating data; for each segment of the first sub-historical operating data, extract multiple data sequences in the current first sub-historical operating data, wherein the multiple data sequences include a continuously increasing data sequence and a continuously decreasing data sequence in the current first sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before the abnormal condition occurs in the gearbox as the abnormal operating trend.
[0176] According to an embodiment of the present disclosure, the prediction unit 246 is also configured to determine predetermined statistical indicators for each data sequence; perform the following processing on each segment of the first sub-historical operating data: classify multiple data sequences of the current first sub-historical operating data according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicators under each category after classification; based on the indicator change trend under each category of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0177] According to an embodiment of the present disclosure, the prediction unit 246 is also configured to perform the following processing for each category: compare the indicator change trend under the current category of each section of the first sub-historical operation data; based on the comparison result, determine the first sub-historical operation data corresponding to the abnormal situation of the gearbox, and use it as the estimated abnormal operation data; determine the abnormal operation data based on the estimated abnormal operation data of each category; determine the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox.
[0178] According to an embodiment of the present disclosure, the prediction unit 246 is further configured to obtain historical operating data of gearboxes that are operating normally in other wind turbines as second historical operating data; segment the second historical operating data according to a predetermined duration and time sequence to obtain multiple segments of second sub-historical operating data; for each segment of the second historical operating data, extract multiple data sequences in the current second sub-historical operating data, wherein the multiple data sequences include continuously increasing data sequences and continuously decreasing data sequences in the current second sub-historical operating data; based on the multiple data sequences of each segment of the first sub-historical operating data and the multiple data sequences of each segment of the second sub-historical operating data, determine the operating trend of the gearbox before an abnormal situation occurs in the gearbox.
[0179] According to an embodiment of the present disclosure, the prediction unit 246 is also configured to determine predetermined statistical indicators for each data sequence; perform the following processing on each segment of the first sub-historical operating data and each segment of the second sub-historical operating data: classify multiple data sequences of the current sub-historical operating data according to the attributes of the data in the data sequence; determine the indicator change trend of the predetermined statistical indicator under each category after classification; based on the indicator change trend under each category of each segment of the second sub-historical operating data and the indicator change trend under each category of each segment of the first sub-historical operating data, determine the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0180] According to an embodiment of the present disclosure, the prediction unit 246 is also configured to perform the following processing for each category: compare the indicator change trend under the current category of each section of the first sub-historical operating data and each section of the second sub-historical operating data; based on the comparison result, determine the sub-historical operating data corresponding to the abnormal situation of the gearbox, and use it as the estimated abnormal operating data; determine the abnormal operating data based on the estimated abnormal operating data of each category; determine the indicator change trend under each category of the abnormal operating data as the operating trend of the gearbox before the abnormal situation occurs in the gearbox.
[0181] According to an embodiment of the present disclosure, the extraction unit 242 is also configured to obtain historical fault information of the gearbox before determining the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox; based on the historical fault information, determine that the abnormal operation data is actual abnormal operation data.
[0182] According to an embodiment of the present disclosure, the attributes of the data include: bearing temperature, oil sump temperature of the lubrication system, oil sump inlet oil temperature of the lubrication system, oil pressure of the lubrication system oil entering the gearbox, oil pressure of the lubrication system oil flowing out of the gearbox, inlet valve water pressure of the cooling system and outlet valve water pressure of the cooling system.
[0183] According to an embodiment of the present disclosure, a computer-readable storage medium storing instructions is provided, wherein when the instructions are executed by at least one computing device, the at least one computing device is prompted to execute a gearbox abnormality prediction method for a wind turbine generator set as described in any of the above embodiments.
[0184] According to an embodiment of the present disclosure, a system is provided comprising at least one computing device and at least one storage device storing instructions, wherein the instructions, when executed by the at least one computing device, prompt the at least one computing device to execute a gearbox abnormality prediction method for a wind turbine generator set as described in any of the above embodiments.
[0185] According to an embodiment of the present disclosure, a wind turbine generator set is provided, comprising the gearbox abnormality prediction device of the wind turbine generator set as described above.
[0186] Although some 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 the 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 predicting abnormality of a gearbox of a wind turbine generator set. It is characterized in that include: Acquiring operating data of a gearbox of a wind turbine generator set within a predetermined time period, wherein the operating data includes at least one of the following: operating data of a bearing of the gearbox, operating data of a lubrication system of the gearbox, and operating data of a cooling system of the gearbox; Extracting a plurality of data sequences from the operating data, wherein the plurality of data sequences include a continuously increasing data sequence and a continuously decreasing data sequence from the operating data; determining an operating trend of the gearbox based on the plurality of data sequences; Based on the operating trend and a predetermined abnormal operating trend, an abnormal condition of the gearbox is predicted.
2. The gearbox abnormality prediction method according to claim 1, It is characterized in that Determining the operating trend of the gearbox based on the multiple data sequences includes: Determine predetermined statistical indicators for each data series; Classifying the multiple data sequences according to attributes of the data in the data sequences; Determine the indicator change trend of the predetermined statistical indicators under each category after classification; Based on the indicator change trend, the operation trend of the gearbox is determined.
3. The gearbox abnormality prediction method according to claim 1, It is characterized in that The extracting a plurality of data sequences from the operating data comprises: Classifying the operation data according to attributes of the data in the operation data; Clustering the classified operating data in each category to obtain at least one cluster of data; For each cluster of data in the at least one cluster of data, a plurality of data sequences in current data are extracted.
4. The gearbox abnormality prediction method according to claim 1, It is characterized in that Before predicting the abnormal condition of the gearbox based on the operation trend and the predetermined abnormal operation trend, the method further includes: Acquiring historical operation data of the gearbox as first historical operation data; Segmenting the first historical operation data according to the predetermined duration and time sequence to obtain a plurality of first sub-historical operation data segments; For each section of the first sub-historical operation data, extract multiple data sequences in the current first sub-historical operation data, wherein the multiple data sequences include a continuously increasing data sequence and a continuously decreasing data sequence in the current first sub-historical operation data; Based on a plurality of data sequences of each section of the first sub-historical operation data, an operation trend of the gearbox before an abnormal situation occurs in the gearbox is determined as the abnormal operation trend.
5. The gearbox abnormality prediction method according to claim 4, It is characterized in that The determining, based on the plurality of data sequences of each segment of the first sub-historical operation data, the operation trend of the gearbox before the abnormal situation of the gearbox occurs, comprises: Determine predetermined statistical indicators for each data series; The following processing is performed on each section of the first sub-historical operation data: multiple data sequences of the current first sub-historical operation data are classified according to the attributes of the data in the data sequence; and the indicator change trend of the predetermined statistical indicator in each category after classification is determined; Based on the indicator change trend in each category of each section of the first sub-historical operation data, the operation trend of the gearbox before the abnormal situation occurs in the gearbox is determined.
6. The gearbox abnormality prediction method according to claim 5, It is characterized in that The determining the operating trend of the gearbox before the abnormal situation occurs in the gearbox based on the indicator change trend in each category of each section of the first sub-historical operating data includes: The following processing is performed for each category: comparing the indicator change trend of each first sub-historical operation data under the current category; based on the comparison result, determining the first sub-historical operation data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operation data; determining abnormal operating data based on the estimated abnormal operating data for each category; The indicator change trend under each category of the abnormal operation data is determined as the operation trend of the gearbox before the abnormal situation occurs in the gearbox.
7. The gearbox abnormality prediction method according to claim 4, It is characterized in that The plurality of data sequences of each first sub-historical operation data segment, determining the operation trend of the gearbox before the abnormal situation of the gearbox occurs, comprises: Acquire historical operation data of gearboxes in other wind turbine generator sets that are operating normally as second historical operation data; Segmenting the second historical operation data according to the predetermined duration and time sequence to obtain a plurality of second sub-historical operation data segments; For each section of the second historical operation data, extract multiple data sequences in the current second sub-historical operation data, wherein the multiple data sequences include a continuously increasing data sequence and a continuously decreasing data sequence in the current second sub-historical operation data; Based on the multiple data sequences of each section of the first sub-historical operation data and the multiple data sequences of each section of the second sub-historical operation data, the operation trend of the gearbox before the abnormal situation of the gearbox occurs is determined.
8. The gearbox abnormality prediction method according to claim 7, It is characterized in that The multiple data sequences of each section of the first sub-historical operation data and the multiple data sequences of each section of the second sub-historical operation data, determining the operation trend of the gearbox before the abnormal situation of the gearbox occurs, comprises: Determine predetermined statistical indicators for each data series; Each section of the first sub-historical operation data and each section of the second sub-historical operation data are processed as follows: multiple data sequences of the current sub-historical operation data are classified according to the attributes of the data in the data sequence; and the indicator change trend of the predetermined statistical indicator in each category after classification is determined; Based on the indicator change trend under each category of each section of the second sub-historical operating data and the indicator change trend under each category of each section of the first sub-historical operating data, the operating trend of the gearbox before the abnormal situation of the gearbox occurs is determined.
9. The gearbox abnormality prediction method according to claim 8, It is characterized in that The determining the operation trend of the gearbox before the abnormal situation occurs in the gearbox based on the indicator change trend in each category of each section of the second sub-historical operation data and the indicator change trend in each category of each section of the first sub-historical operation data includes: The following processing is performed for each category: comparing the indicator change trend of each first sub-historical operation data and each second sub-historical operation data under the current category; based on the comparison result, determining the sub-historical operation data corresponding to the abnormal situation of the gearbox, and using it as the estimated abnormal operation data; determining abnormal operating data based on the estimated abnormal operating data for each category; The indicator change trend under each category of the abnormal operation data is determined as the operation trend of the gearbox before the abnormal situation occurs in the gearbox.
10. The gearbox abnormality prediction method according to claim 6 or 9, It is characterized in that Before determining the indicator change trend under each category of the abnormal operation data as the operation trend of the gearbox before the abnormal situation occurs in the gearbox, the method further includes: Obtaining historical fault information of the gearbox; Based on the historical fault information, it is determined that the abnormal operation data is actual abnormal operation data.
11. The gearbox abnormality prediction method according to claim 2, 3, 5 or 8, It is characterized in that The attributes of the data include: bearing temperature, oil pool temperature of the lubrication system, oil pool inlet oil temperature of the lubrication system, oil pressure of the lubrication system oil entering the gearbox, oil pressure of the lubrication system oil flowing out of the gearbox, inlet valve water pressure of the cooling system and outlet valve water pressure of the cooling system.
12. A gearbox abnormality prediction device for a wind turbine generator set, It is characterized in that include: An acquisition unit is configured to acquire operating data of a gearbox of a wind turbine generator set within a predetermined time period, wherein the operating data includes at least one of the following: operating data of a bearing of the gearbox, operating data of a lubrication system of the gearbox, and operating data of a cooling system of the gearbox; an extraction unit, configured to extract a plurality of data sequences from the operating data, wherein the plurality of data sequences include a continuously increasing data sequence and a continuously decreasing data sequence from the operating data; a determination unit configured to determine an operation trend of the gearbox based on the plurality of data sequences; The prediction unit is configured to predict an abnormal condition of the gearbox based on the operation trend and a predetermined abnormal operation trend.
13. A computer-readable storage medium storing instructions, It is characterized in that When the instructions are executed by at least one computing device, the at least one computing device is prompted to execute the gearbox abnormality prediction method for a wind turbine generator set according to any one of claims 1 to 11.
Citation Information
Patent Citations
Method for monitoring state of gear case of wind generating set
CN101566523A
Draught fan gear box state on-line evaluation model building method and on-line evaluation method
CN105205569A
Wind turbine generator gearbox fault diagnosis and analysis method based on machine learning algorithm
CN114444382A
Cited By
Gearbox ventilation quantity monitoring and early warning system and method
CN121431884A
