CN-121993361-A - Wind generating set fault diagnosis method, device and storage medium based on intermediate observer

CN121993361ACN 121993361 ACN121993361 ACN 121993361ACN-121993361-A

Abstract

The invention provides a wind generating set fault diagnosis method and device based on an intermediate observer and a storage medium, and relates to the technical field of wind generating system fault diagnosis. The method comprises the steps of collecting and preprocessing a plurality of key parameters of a wind generating set, screening core time-varying parameters to form a set, determining convex polyhedron vertex matrix parameters based on the set, establishing a convex polyhedron linear variable parameter model, deriving intermediate variables, combining the model with the intermediate variables to construct an intermediate observer, constructing a fault reconstruction model according to the combination of proportion and integral of output errors of the observer, deriving stability conditions of an error system, converting the error system into a linear matrix inequality, solving to obtain an observer gain matrix, substituting the gain matrix into the intermediate observer, and outputting fault estimated values through the observer and the fault reconstruction model. The invention integrates the middle observer and the convex polyhedron linear variable parameter model, breaks the dependence of the traditional method on the strict matching of the state and the fault, and realizes the whole-flow standardization of fault estimation.

Inventors

LI SHANGLIN
OuYang Panyi
CHEN QIMING
LI WANZE
CHEN MINGXIA
Xue Enchi

Assignees

桂林理工大学
桂林信息科技学院

Dates

Publication Date: 20260508
Application Date: 20251212

Claims (10)

1. The wind generating set fault diagnosis method based on the intermediate observer is characterized by comprising the following steps of: Collecting a plurality of key parameters of a wind generating set, preprocessing the key parameters, and selecting a plurality of core time-varying parameters from the preprocessed key parameters to obtain a core time-varying parameter set; Determining convex polyhedron vertex matrix parameters based on the core time-varying parameter set, establishing a convex polyhedron linear variable parameter model based on the convex polyhedron vertex matrix parameters, and deriving an intermediate variable for releasing the relation between the state and the fault based on the convex polyhedron linear variable parameter model; Constructing an intermediate observer based on the convex polyhedron linear variable parameter model and the intermediate variable after derivation, and constructing a fault reconstruction model based on the combination of the proportion and the integral of the error output by the intermediate observer; Deriving an error system stability condition based on a Lyapunov theory, converting the error system stability condition into a linear matrix inequality, and obtaining an observer gain matrix by solving the linear matrix inequality; and substituting the observer gain matrix into the intermediate observer, and outputting a fault estimated value of the wind generating set through the intermediate observer and the fault reconstruction model.
2. The method for diagnosing a failure of a wind turbine generator set according to claim 1, wherein collecting a plurality of key parameters of the wind turbine generator set and preprocessing the plurality of key parameters, selecting a core time-varying parameter from the preprocessed key parameters, comprises: Collecting design parameters, operation data and environment interference data of a wind generating set, and preprocessing the design parameters, the operation data and the environment interference data, wherein the preprocessing is to sequentially perform outlier rejection processing, data smoothing processing and data standardization processing; Selecting core time-varying parameters from the preprocessed design parameters, operation data and environmental disturbance data Aggregation , S is the number of time-varying parameters and is used for representing the dynamic change of the operation condition of the unit.
3. The wind turbine generator system fault diagnosis method according to claim 2, wherein determining convex polyhedron vertex matrix parameters based on the core time-varying parameter set, establishing a convex polyhedron linear variable parameter model based on the convex polyhedron vertex matrix parameters, and deriving intermediate variables for releasing the state and fault relation based on the convex polyhedron linear variable parameter model, comprises: Based on core time-varying parameter set Determining convex polyhedron vertex matrix parameters by adopting an extremum sampling method, wherein the convex polyhedron vertex matrix parameters comprise a state matrix, a control input matrix, an interference distribution matrix, a fault distribution matrix and an output matrix, wherein the state matrix is used for representing the change characteristics of a system state under each vertex, the control input matrix is used for representing the influence of a pitch angle control instruction on the system state, the fault distribution matrix is used for representing the influence path of faults on the state, the interference distribution matrix is used for representing the effect of external interference on the state, and the output matrix is used for representing the mapping relation between state quantity and measurement output; constructing a convex polyhedron linear variable parameter model of the wind generating set based on the convex polyhedron vertex matrix parameters, wherein the convex polyhedron linear variable parameter model is expressed as: , Wherein, the The derivative of the system state vector of the wind generating set; is a state matrix, is obtained by weighted summation of the state matrices of the vertexes, , As a weighting function, and For characterizing the contribution degree of each vertex to the current system dynamics, The system state vector is a system state vector of the wind generating set and comprises wind wheel rotating speed, pitch angle, generator rotating speed, gearbox input torque, stator current, terminal voltage, main shaft vibration displacement and converter output current; for the control input matrix, the control input matrix of each vertex is weighted and summed to obtain, , The control input vector comprises a pitch angle control command and a generator excitation current control command; In the form of a fault distribution matrix, The fault vector comprises a blade crack fault, a gearbox abrasion fault, a generator turn-to-turn short circuit fault and a sensor drift fault; in order to provide a matrix of interference distribution, Is an external interference vector, and comprises wind speed interference, grid voltage interference and mechanical loss interference; Is an output matrix; the output vector comprises a measurable wind wheel rotating speed, a measurable pitch angle, a measurable generator terminal voltage, a measurable stator current, a measurable gearbox oil temperature and a measurable main shaft vibration acceleration; Deriving an intermediate variable for releasing the state-to-fault relationship based on the convex hull linear variable parametric model, the intermediate variable being expressed as: , Wherein, the The design matrix is an adjustable design matrix, and the adjustable design matrix is obtained according to fault type setting.
4. A wind generating set fault diagnosis method according to claim 3, wherein constructing an intermediate observer based on the convex polyhedron linear variable parameter model and the derived intermediate variable, and constructing a fault reconstruction model based on a combination of a proportion and an integral of an output estimation error of the intermediate observer, comprises: For intermediate variables Obtaining a dynamic relation between the convex polyhedron linear variable parameter model and the intermediate variable by derivation, wherein the dynamic relation is expressed as follows: , Wherein, the As an intermediate variable Is used for the purpose of determining the derivative of (c), As the rate of change of the fault vector, for characterizing the rate of development of the fault, The derivative of the system state vector of the wind generating set; based on dynamic relations Constructing an intermediate observer, wherein the intermediate observer is used for synchronously estimating a system state vector, an intermediate variable and an output vector of the wind generating set, and the expression of the intermediate observer is as follows: , Wherein, the Is a system state vector Is used for the estimation of the (c), As an intermediate variable Is used for the estimation of the (c), For outputting vectors Is used for the estimation of the (c), Is a fault vector Is a preliminary estimate of (1); Is a parameter dependent observer gain matrix for adjusting the response sensitivity of the observer to output errors, an , Is the first An observer gain matrix of convex polyhedron vertices, An intermediate variable gain matrix for parameter dependence for adapting system time-varying conditions to optimize intermediate variable estimation performance, an , Is the first An intermediate variable gain matrix of convex polyhedron vertices, For outputting an estimated error, characterizing a deviation of the observer output from the actual measured value; Constructing a fault reconstruction model based on a proportional and integral combination of output estimation errors of the intermediate observer, wherein the fault reconstruction model is expressed as: , Wherein, the In the form of a proportional gain matrix, In order to integrate the gain matrix, To output an estimated error from 0 to The integral term of the moment of time, 。
5. The method for diagnosing a wind turbine generator system fault according to claim 4, wherein deriving an error system stability condition based on lyapunov theory and converting the error system stability condition into a linear matrix inequality, obtaining an observer gain matrix by solving the linear matrix inequality, comprises: Constructing an error system comprising a state estimation error And intermediate variable estimation error And deriving the state estimation error to obtain a dynamic equation of the state estimation error: , And deriving the intermediate variable estimation error to obtain a dynamic equation of the state estimation error: , Wherein, the The matrix is input for the error system, , The vector is input for the error system, I is an identity matrix; Constructing a lyapunov function based on the state estimation error and the intermediate variable estimation error, the lyapunov function expressed as: , Wherein, the Estimating errors for states Is to be used in the present invention, Estimating error for intermediate variables Is to be used in the present invention, For and state estimation error The corresponding positive lyapunov matrix, Estimating errors for and intermediate variables The corresponding positive lyapunov matrix, And deriving Lyapunov function based on the dynamic equation of state estimation error and the dynamic equation of intermediate variable estimation error, and combining The Lyapunov function satisfies the stability condition after the performance index is required to be derived: , Wherein, the For the lyapunov function after derivation, the method is used for representing the energy change trend of the wind generating set system, 、 Is that A positive index scalar of the performance is provided, As the external interference vector(s), Is the uncertainty vector of the fault.
6. The method for diagnosing a wind turbine generator system according to claim 5, wherein converting the error system stability condition into a linear matrix inequality, obtaining an observer gain matrix by solving the linear matrix inequality, comprises: Converting the stabilization condition into a linear matrix inequality LMI, the linear matrix inequality LMI being a matrix positive definite matrix Matrix The linear matrix inequality condition of (2) is expressed as: , Wherein, the , , , , , For stability, fault term and interference term sub-term constraints, Is a positive Lyapunov matrix used for analyzing the stability of the wind generating set, An equivalent gain matrix for the intermediate observer related to wind turbine and generator state estimates; the vertex number of the convex polyhedron linear variable parameter model is represented and corresponds to the full working condition of the wind generating set; 、 Is that A positive index scalar of the performance is provided, Vertex of linear variable parameter model for convex polyhedron Is a state matrix of the (c) in the (c), Is a fault distribution matrix of system states associated with pitch actuator and sensor faults, For a disturbance distribution matrix in which system states are associated with wind speed turbulence and grid load fluctuations, In order to output the matrix of the matrix, 、 And In order to adjust the design parameters of the device, 、 Is that Performance constraint terms; Obtaining a positive Lyapunov matrix P and an equivalent gain matrix corresponding to the vertex by solving the inequality of the linear matrix And based on inversion formula Pair-equivalent gain matrix Inversion is carried out to obtain an observer gain matrix of each vertex And weighting and combining the observer gain matrix through a convex polyhedron weighting and combining formula Calculation is performed to obtain a real-time parameter dependent observer gain matrix The convex polyhedron weighted combination formula is expressed as: , Wherein, the As a function of the time-varying weighting, And is also provided with , The total number of vertexes of the convex polyhedron linear variable parameter model.
7. The method for diagnosing a failure of a wind turbine generator set according to any one of claims 1 to 6, further comprising, after obtaining a failure estimated value of the wind turbine generator set, the steps of: presetting a fault threshold set , For the number of fault types, the fault estimation value Compared with the corresponding fault threshold value, if , Diagnosing the corresponding type of faults and simultaneously outputting the fault amplitude If (1) For all of If true, diagnosing that the wind generating set has no fault.
8. A wind turbine generator system fault diagnosis device based on an intermediate observer, comprising: The data preprocessing module is used for collecting a plurality of key parameters of the wind generating set, preprocessing the key parameters, selecting a plurality of core time-varying parameters from the preprocessed key parameters and obtaining a core time-varying parameter set; The intermediate variable derivation module is used for determining convex polyhedron vertex matrix parameters based on the core time-varying parameter set, establishing a convex polyhedron linear variable parameter model based on the convex polyhedron vertex matrix parameters, and deriving intermediate variables for relieving the relation between the state and the fault based on the convex polyhedron linear variable parameter model; the fault reconstruction model construction module is used for constructing an intermediate observer based on the convex polyhedron linear variable parameter model and the intermediate variable after derivation, and constructing a fault reconstruction model based on the combination of the proportion and the integral of the error output by the intermediate observer; The observer gain solving module is used for deducing an error system stability condition based on the Lyapunov theory, converting the error system stability condition into a linear matrix inequality, and obtaining an observer gain matrix by solving the linear matrix inequality; the fault estimation output module is used for substituting the gain matrix of the observer into the intermediate observer and outputting the fault estimation value of the wind generating set through the intermediate observer and the fault reconstruction model.
9. An intermediate observer-based wind turbine fault diagnosis apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the intermediate observer-based wind turbine fault diagnosis method according to any one of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium storing a computer program, characterized in that the method for diagnosing a failure of a wind power generator set based on an intermediate observer according to any one of claims 1 to 7 is implemented when the computer program is executed by a processor.

Description

Wind generating set fault diagnosis method, device and storage medium based on intermediate observer Technical Field The invention mainly relates to the technical field of wind power generation system fault diagnosis, in particular to a wind power generation unit fault diagnosis method and device based on an intermediate observer and a storage medium. Background As core equipment in the field of new energy, the wind generating set is often operated in a field complex environment, is easily influenced by wind speed turbulence, power grid fluctuation, mechanical fatigue and other factors, and causes various faults such as blade cracks, gear box abrasion, turn-to-turn short circuit of a generator, sensor drift and the like. If the faults are not diagnosed in time, unplanned shutdown of a unit and reduction of power generation efficiency are caused by light faults, and serious safety accidents such as gear box bursting, main shaft breakage and the like are caused by heavy faults, so that huge economic losses are caused, and the fault diagnosis technology is crucial to guaranteeing safe and stable operation of a wind power system. In the existing wind generating set fault diagnosis method, the technology based on the traditional state observer (such as a Lunberg observer) is most widely applied, and the core logic of the method is to reversely infer fault information by establishing a system model and comparing errors of model output and actual measurement output. However, the method has three key bottlenecks in practical application, namely, firstly, strict matching condition limitation exists, the traditional observer design needs to meet the constraint of strict matching of a fault distribution matrix and a system output matrix, but the wind generating set is a complex system with strong nonlinearity and multivariable coupling, the influence path of faults on states (fault distribution matrix) is dynamically changed along with working conditions such as wind speed, load and the like, the fixed matching condition is difficult to meet, the applicability of the observer is limited, secondly, the fault estimation precision and response speed are insufficient, the traditional observer structure is fixed, adjustable parameters are few, the tracking hysteresis of time-varying faults (such as gradual aggravation of the abrasion degree of a gearbox) is obvious, the influence of external disturbance such as fluctuation of the rotation speed of a wind wheel and the voltage of a power grid is easy to cause, the estimation error is large, the accurate capture of early tiny faults is difficult to realize, thirdly, the robustness is weak, the wind generating set continuously faces uncertain factors such as wind speed shock, turbulent interference, mechanical vibration and the like, the traditional method is not aiming at a special-term suppression mechanism for the interference design, the fault is easy to mask the fault or the true fault signal, the diagnosis result is easy to report, the decision rate is high, the failure rate is easy, and the operation is reliable, and the support dimension is provided. Disclosure of Invention The invention aims to solve the technical problem of providing a wind generating set fault diagnosis method based on an intermediate observer, which comprises the following steps of: S1, collecting a plurality of key parameters of a wind generating set, preprocessing the key parameters, and selecting a plurality of core time-varying parameters from the preprocessed key parameters to obtain a core time-varying parameter set; S2, determining convex polyhedron vertex matrix parameters based on the core time-varying parameter set, establishing a convex polyhedron linear variable parameter model based on the convex polyhedron vertex matrix parameters, and deriving an intermediate variable for relieving the relation between the state and the fault based on the convex polyhedron linear variable parameter model; S3, constructing an intermediate observer based on the convex polyhedron linear variable parameter model and the intermediate variable after derivation, and constructing a fault reconstruction model based on the combination of the proportion and the integral of the error output by the intermediate observer; s4, deducing an error system stability condition based on a Lyapunov theory, converting the error system stability condition into a linear matrix inequality, and obtaining an observer gain matrix by solving the linear matrix inequality; s5, substituting the observer gain matrix into the intermediate observer, and outputting a fault estimated value of the wind generating set through the intermediate observer and the fault reconstruction model. The technical scheme for solving the technical problems is as follows, the wind generating set fault diagnosis device based on the intermediate observer comprises: The data preprocessing module is used for collecting a plurality of key parameters of the win