CN-122014540-A - State evaluation and fault early warning method and system for wind generating set

CN122014540ACN 122014540 ACN122014540 ACN 122014540ACN-122014540-A

Abstract

The invention discloses a method and a system for evaluating the state of a wind generating set and early warning faults, and relates to the technical field of monitoring the state of the wind generating set, wherein the method comprises the steps of obtaining operation monitoring data and control feedback data of the wind generating set; the method comprises the steps of constructing an equivalent load characterization quantity based on wind condition parameters and unit operation state parameters, determining response deviation characteristics of unit components relative to load response references based on the equivalent load characterization quantity to obtain component state evaluation results, constructing pitch control residual errors and yaw control residual errors based on control feedback data to obtain control chain abnormality evaluation results, constructing a multi-source evidence set and carrying out Bayesian network probabilistic reasoning to obtain abnormality risk confidence and abnormality source judgment results, matching early warning grades and outputting the results. The invention realizes the distinction of wind condition disturbance, control chain abnormality and component degradation abnormality, improves the accuracy, reliability and engineering practicability of state evaluation and fault early warning, and reduces the false alarm rate and the false alarm rate.

Inventors

He tengfei
OU MINGWEN
Kang Xunhong

Assignees

湖南电气职业技术学院

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (10)

1. The wind generating set state evaluation and fault early warning method is characterized by comprising the following steps of: S1, acquiring operation monitoring data and control feedback data of a wind generating set, wherein the operation monitoring data comprise wind condition parameters and set operation state parameters, and the control feedback data comprise pitch control data, yaw control data and corresponding execution feedback data; s2, constructing an equivalent load representation quantity representing the load transfer relation of wind condition disturbance to the unit based on the wind condition parameters and the unit operation state parameters; S3, based on the equivalent load characterization quantity and the unit operation state parameter, determining a response deviation characteristic of the unit component relative to a corresponding load response reference, and obtaining a component state evaluation result; s4, based on the control feedback data, constructing a pitch control residual error and a yaw control residual error, and obtaining a control chain abnormality evaluation result; S5, constructing a multi-source evidence set based on the component state evaluation result, the control chain abnormality evaluation result and the equivalent load characterization quantity, and carrying out probabilistic reasoning on the multi-source evidence set based on a Bayesian network to obtain an abnormality risk confidence degree and an abnormality source judgment result; And S6, matching the corresponding early warning grade according to the abnormal risk confidence and the abnormal source judgment result, and outputting a wind generating set state evaluation result and a fault early warning result.
2. The method for evaluating the state of a wind turbine generator system and warning the fault according to claim 1, wherein step S1 specifically comprises: Acquiring operation monitoring data and control feedback data of a wind generating set, wherein the operation monitoring data comprise wind condition parameters and set operation state parameters, the wind condition parameters comprise real-time wind speed, wind direction deviation and turbulence intensity, and the set operation state parameters comprise generator rotating speed, output power and actual values of blade pitch angles; The control feedback data comprises pitch control data, yaw control data and corresponding execution feedback data, the pitch control data comprises a target pitch angle instruction sequence output by a pitch controller, the yaw control data comprises a target yaw angle instruction sequence output by the yaw controller, and the execution feedback data comprises an actual pitch angle feedback sequence corresponding to the target pitch angle instruction sequence and an actual cabin azimuth angle feedback sequence corresponding to the target yaw angle instruction sequence; and performing time alignment, missing value compensation, abnormal value elimination and data format unified processing on the operation monitoring data and the control feedback data, and performing aggregation according to a preset analysis window to form a data set to be analyzed.
3. The method for evaluating the state of a wind turbine generator system and warning the fault according to claim 2, wherein step S2 specifically comprises: Extracting characteristics of the wind condition parameters to obtain wind condition disturbance description results; Based on the wind condition disturbance description result and the unit operation state parameter, a load mapping model is established to represent the corresponding relation among the wind condition parameter, the unit operation state parameter and the unit load level; And inputting the wind condition parameters and the unit running state parameters in the current analysis window into the load mapping model to obtain a load characterization result corresponding to the current analysis window.
4. The method for evaluating the state of a wind generating set and pre-warning faults according to claim 3, wherein in the step S2, the load characterization result comprises a thrust direction load characterization result and a torque direction load characterization result, the thrust direction load characterization result is used for characterizing a load level transmitted along a main shaft direction after the wind wheel is winded, and the torque direction load characterization result is used for characterizing a load level transmitted by a rotation driving force of the wind wheel to an input side of the generator through the main shaft and a transmission chain; and uniformly sorting the thrust direction load characterization result and the torque direction load characterization result to form the equivalent load characterization quantity.
5. The method for evaluating the state of a wind turbine generator system and warning against faults as claimed in claim 4, wherein in step S3, the load response reference is established by: Taking the equivalent load characterization quantity as input and taking a component response parameter corresponding to the component to be evaluated as output, and establishing a regression model describing the mapping relation between the load condition and the component response; And inputting the equivalent load characterization quantity corresponding to the current analysis window into the regression model to obtain a reference response result of the part to be evaluated under the current load condition.
6. The method for evaluating the state of a wind turbine generator system and warning against faults as claimed in claim 5, wherein in step S3, the response offset characteristic is a normalized offset; The standardized offset is obtained by comparing the difference value between the actual response value of the part to be evaluated in the current analysis window and the reference response result with the response fluctuation scale under the same load condition in the historical health sample; And classifying the current state of the part to be evaluated according to whether the standardized offset exceeds the allowable fluctuation range, the exceeding degree and the continuous occurrence time of the reference response, so as to form a part state evaluation result.
7. The method for evaluating the state of a wind turbine generator system and warning the fault according to claim 6, wherein step S4 specifically comprises: Extracting a target pitch angle instruction sequence of a pitch controller, an actual pitch angle feedback sequence of a pitch driver, a target yaw angle instruction sequence of a yaw controller and an actual cabin azimuth angle feedback sequence fed back by a yaw encoder from the control feedback data; calculating a pitch transient residual sequence, a pitch response time delay and a pitch accumulated deviation based on the target pitch angle instruction sequence and the actual pitch angle feedback sequence; calculating a yaw transient residual sequence and a yaw start delay based on the target yaw angle command sequence and the actual nacelle azimuth feedback sequence; and comparing the variable pitch instantaneous residual sequence, the variable pitch response time delay, the variable pitch accumulated deviation, the yaw instantaneous residual sequence and the yaw starting delay with control residual reference standards corresponding to the historical healthy operation stage to obtain a control chain abnormality evaluation result.
8. The method for evaluating the state of a wind turbine generator system and warning the fault according to claim 7, wherein step S5 specifically comprises: Constructing a component side evidence based on the component state evaluation result, constructing a control side evidence based on the control chain abnormality evaluation result, constructing a load side evidence based on an equivalent load characterization quantity, and forming a multi-source evidence set by the component side evidence, the control side evidence and the load side evidence; Inputting the multi-source evidence set into a pre-established Bayesian network, wherein the Bayesian network comprises abnormal source nodes and evidence nodes corresponding to the multi-source evidence set, and the abnormal source nodes comprise wind condition excitation abnormality, control chain execution abnormality and component body degradation abnormality; calculating posterior probability of each abnormal source category under the current evidence condition according to the corresponding evidence states of various evidences in the multi-source evidence set in the current analysis window; And taking the abnormal source category with the maximum posterior probability as an abnormal source judging result, and outputting the maximum posterior probability value as an abnormal risk confidence coefficient.
9. The method for evaluating the state of a wind turbine generator system and warning the fault according to claim 8, wherein step S6 specifically comprises: According to the abnormal risk confidence coefficient obtained in the step S5, matching the abnormal risk confidence coefficient with a preset abnormal risk confidence coefficient interval, and determining an early warning level corresponding to the current analysis window, wherein the early warning level comprises a normal state, a concerned state, an early warning state and an alarm state; when the early warning level is matched, carrying out auxiliary adjustment by combining the abnormal source judging result obtained in the step S5, wherein when the abnormal source judging result is wind condition excitation abnormality, constraint processing is carried out on the early warning level; and generating a wind generating set state evaluation result and a fault early warning result corresponding to the current analysis window based on the abnormal source judgment result, the abnormal risk confidence coefficient and the early warning level.
10. A wind turbine generator system status evaluation and fault pre-warning system for implementing the wind turbine generator system status evaluation and fault pre-warning method according to any one of claims 1 to 9, comprising: The data acquisition module is used for acquiring operation monitoring data and control feedback data of the wind generating set, wherein the operation monitoring data comprise wind condition parameters and set operation state parameters, and the control feedback data comprise pitch control data, yaw control data and corresponding execution feedback data; The load characterization module is used for constructing an equivalent load characterization quantity for characterizing the load transfer relationship of wind condition disturbance to the unit based on the wind condition parameters and the unit operation state parameters; the state evaluation module is used for determining response deviation characteristics of the unit components relative to corresponding load response references based on the equivalent load characterization quantity and the unit operation state parameters to obtain component state evaluation results; The control evaluation module is used for constructing a pitch control residual error and a yaw control residual error based on the control feedback data to obtain a control chain abnormality evaluation result; The probability reasoning module is used for constructing a multi-source evidence set based on the component state evaluation result, the control chain abnormality evaluation result and the equivalent load characterization quantity, and carrying out probability reasoning on the multi-source evidence set based on a Bayesian network to obtain an abnormality risk confidence degree and an abnormality source judgment result; And the early warning output module is used for matching the corresponding early warning grade according to the abnormal risk confidence and the abnormal source judgment result and outputting a wind generating set state evaluation result and a fault early warning result.

Description

State evaluation and fault early warning method and system for wind generating set Technical Field The invention relates to the technical field of wind generating set state monitoring, in particular to a wind generating set state evaluation and fault early warning method and system. Background With the continuous improvement of the single-machine capacity of the wind generating set, the continuous increase of the impeller diameter and the increasingly complex running environment of a wind farm, the state monitoring and fault early warning technology of the wind generating set is gradually developed into a comprehensive evaluation technology based on multi-source data fusion from single parameter monitoring. The related art generally relies on a unit monitoring system, a state monitoring system and a control system interface to collect operation data such as wind speed, power, rotating speed, vibration, temperature, current, voltage, pitch, yaw and the like, and combines statistical analysis, a fault diagnosis model, health index calculation or predictive maintenance strategies to identify and early warn the operation states of key components of the unit. Meanwhile, as the application of the data driving method, the mechanism modeling method and the probabilistic reasoning method in the field of industrial equipment health management is deepened continuously, the state evaluation of the wind turbine generator set is developing towards the directions of load sensing, component level identification, abnormal source reason and hierarchical early warning output. However, the state evaluation and fault early warning technology of the existing wind generating set still has obvious limitations. Firstly, many existing methods directly judge abnormality based on absolute values or fixed thresholds of monitoring quantities such as vibration, temperature, power and the like, and cannot fully represent the middle process of wind condition disturbance to unit load transmission, so that influence relations among external wind environment change, control action change and component body degradation are mutually coupled, and whether component response truly deviates from a health reference is difficult to accurately identify under the same or similar load conditions. Secondly, in the prior art, attention to the execution states of a pitch control link and a yaw control link is insufficient, and anomalies such as control system execution deviation, response delay, action mismatch and the like are often mixed with mechanical component degradation anomalies, so that it is difficult to effectively distinguish between the control link anomalies and component body anomalies, and the accuracy of anomaly source identification is reduced. Thirdly, the prior art stays at the aspects of simple weighting, threshold superposition or single model classification in the aspect of multi-source information fusion, and lacks a unified probability reasoning mechanism capable of simultaneously fusing load side evidence, component side evidence and control side evidence, so that the early warning result with reliability quantification and abnormal source pointing is difficult to output, and the hierarchical early warning output adapting to the actual operation and maintenance requirements is also difficult to realize. Therefore, the prior art has difficulty in simultaneously realizing wind condition disturbance influence stripping, control chain abnormal independent identification, component state accurate evaluation and abnormal source credibility. Disclosure of Invention The present invention has been made in view of the above-described problems. Therefore, the method solves the technical problems that the existing wind generating set state evaluation and fault early warning method cannot effectively characterize the load transmission process of wind condition disturbance to the set, is difficult to distinguish between control chain abnormality and component body abnormality, has insufficient multi-source information fusion capability and lacks an abnormality source credible attribution mechanism, and how to realize accurate evaluation of the wind generating set state, abnormality source judgment and grading fault early warning output. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, an embodiment of the present invention provides a method for evaluating a state of a wind turbine generator set and early warning a fault, including the following steps: S1, acquiring operation monitoring data and control feedback data of a wind generating set, wherein the operation monitoring data comprise wind condition parameters and set operation state parameters, and the control feedback data comprise pitch control data, yaw control data and corresponding execution feedback data; s2, constructing an equivalent load representation quantity representing the load transfer relation of wind conditi