This article provides the scope of the semi-plenary lecture. A fault diagnosis system for a wind turbine typically has a modular structure, each module being dedicated to specific components or to performance monitori...
详细信息
ISBN:
(纸本)9781509006595
This article provides the scope of the semi-plenary lecture. A fault diagnosis system for a wind turbine typically has a modular structure, each module being dedicated to specific components or to performance monitoring. Important issues arising in the application of model-based fault diagnosis in an industrial context are discussed and illustrated on a performance monitoring module based on the wind turbine power curve. They include automated model identification (accounting for outliers and less frequent operating conditions), fault effect analysis, tuning and validation of fault detection algorithms. More case studies are considered in the presentation, and research perspectives are pointed out.
This paper presents an incremental way to design the decision module of a diagnostic system by resorting to dynamic weighting ensembles of classifiers. The method is applied for sensor fault detection and isolation in...
详细信息
This paper presents an incremental way to design the decision module of a diagnostic system by resorting to dynamic weighting ensembles of classifiers. The method is applied for sensor fault detection and isolation in a doubly fed induction generator for a wind turbine application. A bank of observers generates a set of residuals. These signals are progressively fed into a dynamic weighting ensembles algorithm, called Learn ++ NC, for fault classification. The proposed algorithm incrementally learns the residuals-faults relationships and classifies the faults including multiple new classes, based on a dynamically weighted consult and vote mechanism that combines the outputs of the base-classifiers of the ensemble.
In this paper, a fully stochastic setting is proposed to design indicators aimed at detecting and isolating faults in a dynamical system. Modelling uncertainties, as well as process and measurement noise are accounted...
详细信息
In his paper, a fully stochastic setting is proposed to design indicators aimed at detecting and isolating faults in a dynamical system. Modeling uncertainties, as well as process and measurement noise are accounted f...
详细信息
In his paper, a fully stochastic setting is proposed to design indicators aimed at detecting and isolating faults in a dynamical system. Modeling uncertainties, as well as process and measurement noise are accounted for. Two distinct sets of fault indicators are determined, one for fault detection, he other one for fault isolation. Each indicator is obtained by solving an optimization problem aimed at maximizing a statistical distance (namely he Kullback divergence) between he distribution of the indicator under different faulty and fault free working modes. The solution of he problem relies on statistical experiment design for he stochastic characterization of he system.
The diagnosis systems considered in this paper rely on the inconsistency between the actual process behaviour and its expected behaviour as described by an analytical model. The inconsistency is exhibited in signals c...
详细信息
The diagnosis systems considered in this paper rely on the inconsistency between the actual process behaviour and its expected behaviour as described by an analytical model. The inconsistency is exhibited in signals c...
详细信息
The diagnosis systems considered in this paper rely on the inconsistency between the actual process behaviour and its expected behaviour as described by an analytical model. The inconsistency is exhibited in signals called residuals. Two methods for residual generation are presented in a tutorial way: the parity space and the observer based approaches. Linear and nonlinear models are successively considered as a basis for the design of the residual generators.
暂无评论