Validation of wind turbine LPV model

Summary This study is concerned with the robust nonlinear filtering problem for nonlinear discrete‐time stochastic system with multiplicative noise uncertainties, unknown external disturbances, and packet dropouts. The focus of this paper is to design a filter with predictor–corrector structure such that the upper bound on the state estimation error variance is minimized in the presence of multiplicative noise, unknown external disturbances, and packet dropouts. Thus, a robust nonlinear filter based on the method to obtain the upper bound on variances of multiplicative noises, unknown disturbances, and packet dropouts is designed. Further stability analysis shows that the proposed filter has robustness against multiplicative noises, unknown external disturbances, and packet dropouts. Simulation results show that the proposed filter is more effective than extended Kalman filter and other robust extended Kalman filter. Copyright © 2017 John Wiley & Sons, Ltd.

show abstract

“…k 1 , k 2 , k 3 , k 4 , k 5 , and k 6 are parameters depending on the considered drive. The output equations are given by

h (x_{k}) = [\begin{matrix} x_{5, k} \\ x_{3, k}^{2} + x_{4 . k}^{2} \end{matrix}] .

For the simulation, the model parameters and inputs are set to the values given in Table .…”

Section: Numerical Simulation Examplementioning

confidence: 99%

Robust nonlinear filter for nonlinear systems with multiplicative noise uncertainties, unknown external disturbances, and packet dropouts

Zhang

Geng

et al. 2017

Intl J Robust & Nonlinear

View full text Add to dashboard Cite

show abstract

“…The VAF value obtained with this identification method applied to angle pitch estimation is 97% with noise and 100% without. The difference between the reference and the estimated value of the same variable at each time instant (the sample time is 0.01 second) is representing in the figure (3), which is called residue which is used to decide about the faults. In the case of this figure, the residue signal is non null between time t = 15000s and t = 25000s, it shows that a fault occurs in this interval time and the system is in nominal case outside.…”

Section: B Lpv Pitch System Modelmentioning

confidence: 99%

“…This procedure of fault detection is based either on the knowledge about the system or on the model of the system [3]. Model based fault detection is often necessary to obtain a good diagnosis of faults [4], the methods used in model based diagnosis are state observer based scheme, parity equations based scheme and parameter estimation.…”

Section: Introductionmentioning

confidence: 99%

An LPV modelling and fault diagnosis in wind turbine benchmark system

Aubrun

Puig²,

Abdelkrim³

et al. 2017

IJMIC

Self Cite

View full text Add to dashboard Cite

In this paper, we present a fault detection approach of a linear parameter varying (LPV) system using residuals generator based on the use of the predictor based LPV subspace identification approach (LPV PBSID). To address this problem of fault detection in the pitch subsystem considered in the wind turbine benchmark introduced in IFAC SAFEPROCESS 2009, the failure events are caused by jumps in the damping ratio and natural frequency values of the model. The damping ratio is increased and the natural frequency is decreased with the hydraulic pressure variation. Satisfactory results have been obtained of the proposed approach using the proposed fault scenarios.

show abstract

Validation of wind turbine LPV model

Cited by 2 publications

References 16 publications

Robust nonlinear filter for nonlinear systems with multiplicative noise uncertainties, unknown external disturbances, and packet dropouts

Robust nonlinear filter for nonlinear systems with multiplicative noise uncertainties, unknown external disturbances, and packet dropouts

An LPV modelling and fault diagnosis in wind turbine benchmark system

Contact Info

Product

Resources

About