A method for monitoring abnormal state of a variable pitch bearing of a wind turbine

Abstract

The application relates to the technical field of variable pitch bearing monitoring, in particular to a variable pitch bearing abnormal state monitoring method for a wind turbine. The method comprises the following steps: multiple monitoring points are set according to variable pitch bearing equipment parameters, and monitoring submodules are arranged at the monitoring points; a first host obtains acoustic emission signals collected by each monitoring submodule according to a preset monitoring time node, and generates a first diagnosis result and a feedback data packet; and a central control host generates processing parameters of the feedback data packet according to the first diagnosis result. By setting multiple monitoring points, acoustic emission sensors are used to collect stress wave signals emitted by the variable pitch bearing when internal deformation, cracks, corrosion and external friction occur, and a first host is used to perform edge feature extraction and data analysis on the collected data on site, so that real-time health state monitoring, abnormal early warning and fault diagnosis of the wind turbine are realized, maintenance cost is reduced, and equipment availability is improved.

Description

Technical Field

[0001] This application relates to the field of pitch bearing monitoring technology, and in particular to a method for monitoring abnormal conditions of pitch bearings in wind turbine generators. Background Technology

[0002] Pitch bearings are a crucial component of wind turbines, and common failure modes include cracking of the inner and outer rings, rolling element breakage, and cage damage. When these bearings malfunction, they can cause jamming of the pitch control, preventing the pitch from opening or closing, leading to rotor imbalance and even runaway. If cracks in the inner and outer rings of the pitch bearing are not detected in time, they can result in blade detachment, and prolonged operation can further damage the hub and expand the damage area.

[0003] The pitch bearing is a slewing bearing, which is a typical low-speed, heavy-load, incompletely rotating bearing. Its periodic signals cannot be collected, so it is impossible to use vibration methods to monitor and analyze its abnormal conditions online. In addition, the impact generated by early defects inside the pitch bearing at low speed is insufficient to cause the tested equipment to generate a sufficiently detectable impact signal or a propagable audio signal. Furthermore, due to the low speed and the extremely close frequency of its fault characteristics, it is not easy to distinguish its fault characteristics using impact pulse methods or acoustic recognition methods. Summary of the Invention

[0004] The purpose of this application is to provide a method for monitoring the abnormal state of pitch bearings in wind turbine generators in order to solve the above-mentioned technical problems, thereby achieving real-time health status monitoring and early warning of abnormal faults of pitch bearings and reducing operation and maintenance costs.

[0005] In some embodiments of this application, multiple monitoring points are set up, and acoustic emission sensors are used to collect stress wave signals emitted by the pitch bearing due to internal deformation, cracks, corrosion, and external friction. The collected data is then processed by a primary host on-site to extract edge features and perform data analysis, thereby enabling real-time monitoring and fault warning of the wind turbine's health status.

[0006] In some embodiments of this application, the primary host performs an initial judgment on the monitoring status of the pitch bearing, and the central control host pulls and analyzes the fault data based on the primary diagnostic results and network conditions, thereby constructing a two-level fault analysis model to realize real-time health status monitoring, anomaly warning and fault diagnosis of wind turbine units, reduce maintenance costs and improve equipment availability.

[0007] In some embodiments of this application, a method for monitoring abnormal conditions of wind turbine pitch bearings is provided, including:

[0008] Multiple monitoring points are set according to the parameters of the pitch bearing equipment, and a monitoring sub-module is set at each monitoring point;

[0009] The primary host acquires the acoustic emission signals collected by each monitoring submodule according to the preset monitoring time nodes, and generates primary diagnostic results and feedback data packets;

[0010] The central control unit generates processing parameters for feedback data packets based on the primary diagnostic results.

[0011] In some embodiments of this application, generating a primary diagnostic result based on the processing result includes:

[0012] Establish a sequence of monitoring points A, A = (a1, a2, ..., an), where ai is the i-th monitoring point and n is the number of monitoring points;

[0013] Acquire the acoustic emission signals of each monitoring point sequentially according to the monitoring point sequence A;

[0014] Preprocess the acoustic emission signals from the monitoring points to generate acoustic emission characteristic data for the monitoring points;

[0015] Generate sub-feedback data packets for monitoring points based on acoustic emission characteristic data:

[0016] Establish a sequence of sub-feedback data packets B, B = (b1, b2, ..., bn), where bi is the sub-feedback data packet of the i-th monitoring point at the current time node;

[0017] Establish a primary diagnostic model based on historical operating parameters;

[0018] Based on the primary diagnostic model, the fault risk value of each monitoring point is generated, and a fault risk value sequence C is established, C = (c1, c2, ..., cn), where ci is the fault risk value of the i-th monitoring point at the current time node;

[0019] A first-level diagnostic result is generated based on the fault risk value sequence C.

[0020] In some embodiments of this application, the process of establishing the fault risk value series includes:

[0021] Establish a sequence of standard values ​​D for feature evaluation indicators based on historical parameters, where D = (Δd1, Δd2, ..., Δdm), where Δd1 is the standard value of the i-th feature evaluation indicator and m is the number of feature evaluation indicators.

[0022] Based on the sequence of sub-feedback data packets B, bi is sequentially set as the target sub-reflection data packet;

[0023] Generate a sequence of reference values ​​for feature evaluation indicators in the target sub-feedback data packet, D1, where D1 = (d1, d2, ..., dm), and di is the reference value of the i-th feature evaluation indicator in the target sub-feedback data packet.

[0024] Generate the fault risk value c of the monitoring point corresponding to the target sub-feedback data packet;

[0025]

[0026] Where βi is the influence factor of the i-th feature evaluation index, and Q is a fixed coefficient.

[0027] The fault risk value for each monitoring point is generated sequentially.

[0028] In some embodiments of this application, when generating a first-level diagnostic result based on the fault risk value sequence C, the following steps are included:

[0029] Preset a first fault risk threshold C1 and a second fault risk threshold C2;

[0030] If C1 < ci ≤ C2, the i-th monitoring point is designated as a first-level anomaly monitoring point;

[0031] If ci > C2, the i-th monitoring point is designated as a secondary anomaly monitoring point;

[0032] A warning evaluation value k is generated based on all abnormal monitoring points;

[0033] The first early warning evaluation value range (K1, K2) and the second early warning evaluation value range (K2, K3) are preset;

[0034] If the early warning evaluation value K is within the preset first early warning evaluation value range, a first-level early warning instruction is generated;

[0035] If the early warning evaluation value K is within the preset second early warning evaluation value range, a second-level early warning instruction is generated.

[0036] In some embodiments of this application, when generating an early warning evaluation value k based on all abnormal monitoring points, the following are included:

[0037] The first reference evaluation value H1 is generated based on the number of first-level abnormal monitoring points at the current monitoring time point;

[0038] A second reference evaluation value H2 is generated based on the number of secondary abnormal monitoring points at the current monitoring time point;

[0039] The early warning evaluation value k for each monitoring time node is generated based on the first reference evaluation value H1 and the second reference evaluation value H2.

[0040] k = e1 * H1 + e2 * H2, where e1 is the preset first weight coefficient and e2 is the preset second weight coefficient.

[0041] In some embodiments of this application, when setting the processing parameters of the feedback data packet based on the primary diagnostic results, the following are included:

[0042] The central control unit obtains warning commands based on the primary diagnostic results;

[0043] If no warning instruction is given, the central control unit generates a compression instruction and sends it to the primary control unit;

[0044] When the warning instruction is a Level 1 warning instruction, the central control host obtains the feedback data packet of the current monitoring time node;

[0045] A historical correction model was constructed based on historical monitoring data;

[0046] The fault risk values ​​of each monitoring point are adjusted based on the feedback data packets at the current time point and the historical correction model.

[0047] A Level 1 maintenance plan is generated based on the revised results;

[0048] When the warning instruction is a level 2 warning instruction, the central control host obtains the feedback data packet of the current monitoring time node and generates a level 2 maintenance plan.

[0049] In some embodiments of this application, the construction of the history correction model includes:

[0050] The time interval between the current monitoring time node and the previous maintenance time node is set as the historical correction cycle;

[0051] Generate a sequence of average fault risk values ​​P, P = (p1, p2, ..., pn). Where pi is the average fault risk value of the i-th monitoring point within the historical correction period;

[0052] Based on the average fault risk series P, the correction coefficient α for the fault risk value of each monitoring point is set sequentially.

[0053] In some embodiments of this application, setting the correction coefficient α includes:

[0054] The first average fault risk interval (P1, P2), the first average fault risk interval (P2, P3), and the third average fault risk interval (P3, P4) are preset.

[0055] If pi is within the preset first fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset first correction coefficient α1, that is, α=α1;

[0056] If pi is within the preset second fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset second correction coefficient α1, that is, α=α2;

[0057] If pi is within the preset third fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset third correction coefficient α3, that is, α=α3; and 1<α1<α2<α3.

[0058] In some embodiments of this application, when setting the processing parameters of the feedback data packet based on the primary diagnostic results, the method further includes:

[0059] When the central control host obtains the feedback data packet of the current feedback time node, it sets the first-level storage duration t of each sub-feedback data packet in the central control host according to the fault risk value sequence C;

[0060] When the storage duration of the sub-feedback data packet exceeds the preset first-level storage duration t, a compression instruction for the sub-feedback data packet is generated.

[0061] Compared with the prior art, the beneficial effects of the method for monitoring abnormal conditions of pitch bearings in wind turbine generators according to the embodiments of this application are as follows:

[0062] By setting up multiple monitoring points, acoustic emission sensors are used to collect stress wave signals emitted by the pitch bearing due to internal deformation, cracks, corrosion, and external friction. The collected data is then processed on-site by a primary host unit to extract edge features and perform data analysis, enabling real-time monitoring and fault warning of the wind turbine's health status.

[0063] The primary control unit makes an initial judgment on the monitoring status of the pitch bearing. The central control unit pulls and analyzes the fault data based on the primary diagnostic results and network conditions, thereby constructing a two-level fault analysis model. This enables real-time health monitoring, anomaly warning, and fault diagnosis of wind turbine units, reducing maintenance costs and improving equipment availability. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating a preferred embodiment of a method for monitoring abnormal conditions of a wind turbine pitch bearing. Detailed Implementation

[0065] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0066] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0067] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0068] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0069] like Figure 1 As shown, a preferred embodiment of this application provides a method for monitoring abnormal conditions of a wind turbine pitch bearing, comprising:

[0070] S101: Multiple monitoring points are set according to the parameters of the pitch bearing equipment, and a monitoring sub-module is set at each monitoring point;

[0071] S102: The primary host acquires the acoustic emission signals collected by each monitoring submodule according to the preset monitoring time nodes, and generates primary diagnostic results and feedback data packets;

[0072] S103: The central control host generates processing parameters for feedback data packets based on the first-level diagnostic results.

[0073] Specifically, a primary host is installed on a single wind turbine, and multiple monitoring points are set on each pitch bearing. Each monitoring point is equipped with a monitoring submodule. The primary host is preferably an acoustic emission host, which can extract edge features from the collected acoustic emission signals on site and save the corresponding acoustic emission feature data locally or transmit it to the central control host.

[0074] Specifically, when generating a primary diagnostic result based on the processing results, the following are included:

[0075] Establish a sequence of monitoring points A, A = (a1, a2, ..., an), where ai is the i-th monitoring point and n is the number of monitoring points;

[0076] Acquire the acoustic emission signals of each monitoring point sequentially according to the monitoring point sequence A;

[0077] Preprocess the acoustic emission signals from the monitoring points to generate acoustic emission characteristic data for the monitoring points;

[0078] Generate sub-feedback data packets for monitoring points based on acoustic emission characteristic data:

[0079] Establish a sequence of sub-feedback data packets B, B = (b1, b2, ..., bn), where bi is the sub-feedback data packet of the i-th monitoring point at the current time node;

[0080] Establish a primary diagnostic model based on historical operating parameters;

[0081] Based on the primary diagnostic model, the fault risk value of each monitoring point is generated, and a fault risk value sequence C is established, C = (c1, c2, ..., cn), where ci is the fault risk value of the i-th monitoring point at the current time node;

[0082] A first-level diagnostic result is generated based on the fault risk value sequence C.

[0083] Specifically, acoustic emission feature data is formed by filtering, denoising, amplifying, frequency domain transforming, and extracting acoustic emission features from the collected acoustic emission signals.

[0084] Specifically, acoustic emission characteristic signals refer to the stress wave parameters emitted when materials undergo internal deformation, cracking, corrosion, or external friction.

[0085] Specifically, multiple monitoring and evaluation indicators are set based on the extracted acoustic emission characteristic data, and corresponding standard values ​​are established.

[0086] Specifically, when establishing the fault risk value series, the following is included:

[0087] Establish a sequence of standard values ​​D for feature evaluation indicators based on historical parameters, where D = (Δd1, Δd2, ..., Δdm), where Δd1 is the standard value of the i-th feature evaluation indicator and m is the number of feature evaluation indicators.

[0088] Based on the sequence of sub-feedback data packets B, bi is sequentially set as the target sub-reflection data packet;

[0089] Generate a sequence of reference values ​​for feature evaluation indicators in the target sub-feedback data packet, D1, where D1 = (d1, d2, ..., dm), and di is the reference value of the i-th feature evaluation indicator in the target sub-feedback data packet.

[0090] Generate the fault risk value c of the monitoring point corresponding to the target sub-feedback data packet;

[0091]

[0092] Where βi is the influence factor of the i-th feature evaluation index, and Q is a fixed coefficient.

[0093] The fault risk value for each monitoring point is generated sequentially.

[0094] It is understood that in the above embodiments, by establishing multiple monitoring points, collecting real-time acoustic emission signals, and generating fault risk values ​​for each monitoring point based on preprocessed data, timely warnings are given for potential defects and operational risks at each monitoring point, thereby enabling safe prediction of the wind turbine's operating status, timely elimination of fault risks, and ensuring the safe operation of the wind turbine.

[0095] In a preferred embodiment of this application, generating a first-level diagnostic result based on the fault risk value sequence C includes:

[0096] Preset a first fault risk threshold C1 and a second fault risk threshold C2;

[0097] If C1 < ci ≤ C2, the i-th monitoring point is designated as a first-level anomaly monitoring point;

[0098] If ci > C2, the i-th monitoring point is designated as a secondary anomaly monitoring point;

[0099] A warning evaluation value k is generated based on all abnormal monitoring points;

[0100] The first early warning evaluation value range (K1, K2) and the second early warning evaluation value range (K2, K3) are preset;

[0101] If the early warning evaluation value K is within the preset first early warning evaluation value range, a first-level early warning instruction is generated;

[0102] If the early warning evaluation value K is within the preset second early warning evaluation value range, a second-level early warning instruction is generated.

[0103] Specifically, a Level 1 warning indicates that there is a high probability of a fault in the current wind turbine. The central control unit needs to retrieve historical fault data to conduct a comprehensive analysis of the current health status of the wind turbine. A Level 2 warning indicates that there is a risk of fault in the current wind turbine. The central control unit needs to immediately generate a maintenance command to repair the internal defects of the wind turbine in a timely manner.

[0104] Specifically, when generating an early warning evaluation value k based on all abnormal monitoring points, it includes:

[0105] The first reference evaluation value H1 is generated based on the number of first-level abnormal monitoring points at the current monitoring time point;

[0106] A second reference evaluation value H2 is generated based on the number of secondary abnormal monitoring points at the current monitoring time point;

[0107] The early warning evaluation value k for each monitoring time node is generated based on the first reference evaluation value H1 and the second reference evaluation value H2.

[0108] k = e1 * H1 + e2 * H2, where e1 is the preset first weight coefficient and e2 is the preset second weight coefficient;

[0109] Specifically, the first and second reference evaluation values ​​have the same range. The more first-level anomaly monitoring points there are, the larger the corresponding first reference evaluation value will be. The more second-level anomaly monitoring points there are, the larger the corresponding second reference evaluation value will be. The larger the warning evaluation value, the greater the possibility that there is a defect in the pitch bearing of the current wind turbine.

[0110] In a preferred embodiment of this application, when setting the processing parameters of the feedback data packet based on the primary diagnostic results, the following is included:

[0111] The central control unit obtains warning commands based on the primary diagnostic results;

[0112] If no warning instruction is given, the central control unit generates a compression instruction and sends it to the primary control unit;

[0113] When the warning instruction is a Level 1 warning instruction, the central control host obtains the feedback data packet of the current monitoring time node;

[0114] A historical correction model was constructed based on historical monitoring data;

[0115] The fault risk values ​​of each monitoring point are adjusted based on the feedback data packets at the current time point and the historical correction model.

[0116] A Level 1 maintenance plan is generated based on the revised results;

[0117] When the warning instruction is a level 2 warning instruction, the central control host obtains the feedback data packet of the current monitoring time node and generates a level 2 maintenance plan.

[0118] Specifically, the first-level compression instruction means compressing the current feedback data packet and temporarily storing it locally. When the compressed file reaches a preset threshold, it is transmitted to the central control host according to the network parameters, thereby improving the data transmission efficiency.

[0119] Specifically, its Level 1 and Level 2 maintenance plans refer to the maintenance personnel carrying out maintenance on monitoring points with abnormalities in sequence according to the fault risk values ​​of each monitoring point, so as to eliminate fault risks.

[0120] Specifically, constructing a historical correction model includes:

[0121] The time interval between the current monitoring time node and the previous maintenance time node is set as the historical correction cycle;

[0122] Generate a sequence of average fault risk values ​​P, P = (p1, p2, ..., pn). Where pi is the average fault risk value of the i-th monitoring point within the historical correction period;

[0123] Based on the average fault risk series P, the correction coefficient α for the fault risk value of each monitoring point is set sequentially.

[0124] Specifically, when setting the correction coefficient α, the following are included:

[0125] The first average fault risk interval (P1, P2), the first average fault risk interval (P2, P3), and the third average fault risk interval (P3, P4) are preset.

[0126] If pi is within the preset first fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset first correction coefficient α1, that is, α=α1;

[0127] If pi is within the preset second fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset second correction coefficient α1, that is, α=α2;

[0128] If pi is within the preset third fault risk average range, the correction coefficient α of the i-th monitoring point is set to the preset third correction coefficient α3, that is, α=α3; and 1<α1<α2<α3.

[0129] Specifically, when the primary control unit determines that a wind turbine may be at risk of failure, the central control unit retrieves and re-analyzes fault data based on early warning instructions and network conditions. By constructing a historical correction model, the fault risk value of each monitoring point is adjusted to more accurately reflect the probability of failure at each monitoring point. This provides more accurate data support for on-site maintenance personnel to diagnose faults, enabling real-time health monitoring, anomaly early warning, and fault diagnosis of wind turbines, reducing maintenance costs and improving equipment availability.

[0130] Specifically, when setting the processing parameters for the feedback data packet based on the primary diagnostic results, it also includes:

[0131] When the central control host obtains the feedback data packet of the current feedback time node, it sets the first-level storage duration t of each sub-feedback data packet in the central control host according to the fault risk value sequence C;

[0132] When the storage duration of the sub-feedback data packet exceeds the preset first-level storage duration t, a compression instruction for the sub-feedback data packet is generated.

[0133] Specifically, by setting a primary storage duration, abnormal data is stored normally for easy retrieval and analysis by the central control host. When the preset storage duration is reached, the data is compressed to reduce the memory usage of the monitoring data.

[0134] According to the first concept of this application, by setting up multiple monitoring points, acoustic emission sensors are used to collect stress wave signals emitted by the pitch bearing due to internal deformation, cracks, corrosion, and external friction. The collected data is then processed by a primary host on-site to extract edge features and perform data analysis, thereby enabling real-time monitoring and fault early warning of the wind turbine's health status.

[0135] According to the second concept of this application, the first-level host makes an initial judgment on the monitoring status of the pitch bearing, and the central control host pulls and analyzes the fault data based on the first-level diagnostic results and network conditions, thereby constructing a two-level fault analysis model to realize real-time health status monitoring, abnormal early warning and fault diagnosis of wind turbine, reduce maintenance costs and improve equipment availability.

[0136] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A method for monitoring abnormal condition of a pitch bearing of a wind turbine generator, characterized in that, The application relates to a wind turbine blade pitch bearing device fault diagnosis method and device. According to the pitch bearing device parameter, a plurality of monitoring points are set, and monitoring sub-modules are arranged at the monitoring points; A primary host obtains acoustic emission signals collected by each monitoring sub-module according to a preset monitoring time node, and generates a primary diagnosis result and a feedback data packet; A central control host generates processing parameters of the feedback data packet according to the primary diagnosis result; When the primary diagnosis result is generated according to the processing result, the method comprises the following steps: A monitoring point sequence A is established, A=(a1, a2…an), wherein ai is the i-th monitoring point, and n is the number of monitoring points; Acoustic emission signals of each monitoring point are obtained according to the monitoring point sequence A; The acoustic emission signals of the monitoring points are preprocessed to generate acoustic emission characteristic data of the monitoring points; Sub-feedback data packets of the monitoring points are generated according to the acoustic emission characteristic data: A sub-feedback data packet sequence B is established, B=(b1, b2…bn), wherein bi is the sub-feedback data packet of the i-th monitoring point at the current time node; A primary diagnosis model is established according to historical operation parameters; Fault risk values of each monitoring point are generated according to the primary diagnosis model, and a fault risk value sequence C is established, C=(c1, c2…cn), wherein ci is the fault risk value of the i-th monitoring point at the current time node; A primary diagnosis result is generated according to the fault risk value sequence C; When the fault risk value sequence is established, the method comprises the following steps: Characteristic evaluation index standard value sequence D is established according to historical parameters, D=(Delta d1, Delta d2,…Delta dm), wherein Delta di is the i-th characteristic evaluation index standard value, and m is the number of characteristic evaluation indexes; Bi is set as a target sub-reflection data packet according to the sub-feedback data packet sequence B; Characteristic evaluation index reference value sequence D1 of the target sub-feedback data packet is generated, D1=(d1, d2…dm), wherein di is the reference value of the i-th characteristic evaluation index in the target sub-feedback data packet; The fault risk value c of the monitoring point corresponding to the target sub-feedback data packet is generated; c=[ βi* (di - Δdi) 2 ]*Q; Wherein, beta i is the influence factor of the i-th characteristic evaluation index, and Q is a fixed coefficient; The fault risk values of each monitoring point are generated in sequence.

2. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 1, wherein, When the primary diagnosis result is generated according to the fault risk value sequence C, the method comprises the following steps: A first fault risk value threshold C1 and a second fault risk value threshold C2 are preset; If C1<ci<=C2, the i-th monitoring point is set as a primary abnormal monitoring point; If ci>C2, the i-th monitoring point is set as a secondary abnormal monitoring point; An early warning evaluation value k is generated according to all abnormal monitoring points; A first early warning evaluation value interval (K1, K2) and a second early warning evaluation value interval (K2, K3) are preset; If the early warning evaluation value K is in the preset first early warning evaluation value interval, a primary early warning instruction is generated; If the early warning evaluation value K is in the preset second early warning evaluation value interval, a secondary early warning instruction is generated.

3. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 2, wherein, When the early warning evaluation value k is generated according to all abnormal monitoring points, the method comprises the following steps: A first reference evaluation value H1 is generated according to the number of primary abnormal monitoring points at the current monitoring time node; A second reference evaluation value H2 is generated according to the number of secondary abnormal monitoring points at the current monitoring time node; The early warning evaluation value k of the monitoring time node is generated according to the first reference evaluation value H1 and the second reference evaluation value H2. k = e1*H1 + e2*H2, wherein e1 is a preset first weight coefficient, and e2 is a preset second weight coefficient.

4. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 3, wherein, The processing parameter of the feedback data packet is set according to the first diagnosis result, and the processing parameter includes: The central control host obtains a warning instruction according to the first diagnosis result; If there is no warning instruction, the central control host generates a compression instruction and sends it to the first host; When the warning instruction is a first-level warning instruction, the central control host obtains the feedback data packet of the current monitoring time node; A historical correction model is constructed according to historical monitoring data; The fault risk value of each monitoring point is corrected according to the feedback data packet of the current time node and the historical correction model; A first-level maintenance plan is generated according to the correction result; When the warning instruction is a second-level warning instruction, the central control host obtains the feedback data packet of the current monitoring time node and generates a second-level maintenance plan.

5. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 4, wherein, The historical correction model is constructed, and the construction includes: The time interval between the current monitoring time node and the last maintenance time node is set as a historical correction period; A fault risk average value sequence P is generated, P = (p1, p2…pn), wherein pi is the fault risk average value of the i th monitoring point in the historical correction period; The correction coefficient a of the fault risk value of each monitoring point is set according to the fault risk average value sequence P.

6. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 5, wherein, The correction coefficient a is set, and the setting includes: A first fault risk average value interval (P1, P2), a second fault risk average value interval (P2, P3) and a third fault risk average value interval (P3, P4) are preset; If pi is in the preset first fault risk average value interval, the correction coefficient a of the i th monitoring point is set as a preset first correction coefficient a1, that is, a = a1; If pi is in the preset second fault risk average value interval, the correction coefficient a of the i th monitoring point is set as a preset second correction coefficient a1, that is, a = a2; If pi is in the preset third fault risk average value interval, the correction coefficient a of the i th monitoring point is set as a preset third correction coefficient a3, that is, a = a3; and 1 < a1 < a2 < a3.

7. The wind turbine generator variable pitch bearing abnormal state monitoring method of claim 5, wherein, The processing parameter of the feedback data packet is set according to the first diagnosis result, and the processing parameter includes: When the central control host obtains the feedback data packet of the current feedback time node, the storage time t of each sub-feedback data packet in the central control host is set according to the fault risk value sequence C; When the storage time of the sub-feedback data packet is greater than the preset first storage time t, a compression instruction of the sub-feedback data packet is generated.

Images