Adaptive data-driven fault-tolerant control strategy for optimal power extraction in presence of broken rotor bars in wind turbine.

Broken rotor bars of a Squirrel Cage Induction Generator (SCIG) impact significantly the quality and quantity of produced electrical energy from Wind Energy Conversion Systems (WECS) because: (1) they change the characteristics of SCIG entailing the invalidity of the designed Maximum Power Point Trucking (MPPT) control and pitch angle control, (2) they increase mechanical and thermal stresses, as well as the harmonic content in stator currents. Therefore, this paper proposes an efficient fault tolerant control strategy in order to mitigate the aforementioned consequences. This strategy is based on a hybrid (mechanic-electric) fault tolerant control (FTC) allowing obtaining optimal power extraction in presence of broken rotor bars. A failure (one or two broken bars) is detected in early stage based on the use of a data-driven (machine learning) approach. The latter uses a discriminative feature space (frequency and magnitude of stator harmonic currents) in order to represent and separate the normal operation mode from faulty modes (broken bars). Simulations using MATLAB demonstrated a nominal current in healthy rotor bars in wind turbine operation zone-3 and new optimal power extraction despite one or two broken bars.

Self adaptive learning scheme for early diagnosis of simple and multiple switch faults in multicellular power converters.

This paper proposes a scheme based on the use of unsupervised machine learning approach and a drift detection mechanism in order to perform an early fault diagnosis of simple and multiple stuck-opened/stuck-closed switches in multicellular converters. Only the data samples representing the normal operation conditions are used in order to be adapted to the case where no data is available about faulty behaviors. A health indicator measuring the dissimilarity between normal and current operation conditions is built in order to detect a drift (degradations) in early stage. When a degradation (fault) is detected, the isolation is achieved by taking into account the discrete dynamics of switches. The features related to the latter are extracted in order to build a feature space allowing to separate the faulty behavior (zone or class) of the different switches. The proposed scheme is evaluated using real data samples representing different normal/simple/multiple switch fault scenarios issued from a test rig.

A Bi-Criteria Active Learning Algorithm for Dynamic Data Streams.

Active learning (AL) is a promising way to efficiently build up training sets with minimal supervision. A learner deliberately queries specific instances to tune the classifier's model using as few labels as possible. The challenge for streaming is that the data distribution may evolve over time, and therefore the model must adapt. Another challenge is the sampling bias where the sampled training set does not reflect the underlying data distribution. In the presence of concept drift, sampling bias is more likely to occur as the training set needs to represent the whole evolving data. To tackle these challenges, we propose a novel bi-criteria AL (BAL) approach that relies on two selection criteria, namely, label uncertainty criterion and density-based criterion. While the first criterion selects instances that are the most uncertain in terms of class membership, the latter dynamically curbs the sampling bias by weighting the samples to reflect on the true underlying distribution. To design and implement these two criteria for learning from streams, BAL adopts a Bayesian online learning approach and combines online classification and online clustering through the use of online logistic regression and online growing Gaussian mixture models, respectively. Empirical results obtained on standard synthetic and real-world benchmarks show the high performance of the proposed BAL method compared with the state-of-the-art AL methods.

Active learning for classifying data streams with unknown number of classes.

The classification of data streams is an interesting but also a challenging problem. A data stream may grow infinitely making it impractical for storage prior to processing and classification. Due to its dynamic nature, the underlying distribution of the data stream may change over time resulting in the so-called concept drift or the possible emergence and fading of classes, known as concept evolution. In addition, acquiring labels of data samples in a stream is admittedly expensive if not infeasible at all. In this paper, we propose a novel stream-based active learning algorithm (SAL) which is capable of coping with both concept drift and concept evolution by adapting the classification model to the dynamic changes in the stream. SAL is the first AL algorithm in the literature to explicitly take account of these concepts. Moreover, using SAL, only labels of samples that are expected to reduce the expected future error are queried. This process is done while tackling the problem of sampling bias so that samples that induce the change (i.e., drifting samples or samples coming from new classes) are queried. To efficiently implement SAL, the paper proposes the application of non-parametric Bayesian models allowing to cope with the lack of prior knowledge about the data stream. In particular, Dirichlet mixture models and the stick breaking process are adopted and adapted to meet the requirements of online learning. The empirical results obtained on real-world benchmarks demonstrate the superiority of SAL in terms of classification performance over the state-of-the-art methods using average and average class accuracy.