Decentralized fault detection and diagnosis via sparse PCA based decomposition and Maximum Entropy decision fusion

This paper proposes an approach for decentralized fault detection and diagnosis in process monitoring sensor networks. The sensor network is decomposed into multiple, potentially overlapping, blocks using the Sparse Principal Component Analysis algorithm. Local predictions are generated at each block using Support Vector Machine classifiers. The local predictions are then fused via a Maximum Entropy algorithm. Empirical studies on the benchmark Tennessee Eastman Process data demonstrated that the proposed decentralized approach achieves accuracy comparable to that of the fully centralized approach, while offering benefits in terms of fault tolerance, reusability, and scalability.     

Supervised clustering of label ranking data using label preference information

This paper studies supervised clustering in the context of label ranking data. The goal is to partition the feature space into K clusters, such that they are compact in both the feature and label ranking space. This type of clustering has many potential applications. For example, in target marketing we might want to come up with K different offers or marketing strategies for our target audience. Thus, we aim at clustering the customers’ feature space into K clusters by leveraging the revealed or stated, potentially incomplete customer preferences over products, such that the preferences of customers within one cluster are more similar to each other than to those of customers in other clusters. We establish several baseline algorithms and propose two principled algorithms for supervised clustering. In the first baseline, the clusters are created in an unsupervised manner, followed by assigning a representative label ranking to each cluster. In the second baseline, the label ranking space is clustered first, followed by partitioning the feature space based on the central rankings. In the third baseline, clustering is applied on a new feature space consisting of both features and label rankings, followed by mapping back to the original feature and ranking space. The RankTree principled approach is based on a Ranking Tree algorithm previously proposed for label ranking prediction. Our modification starts with K random label rankings and iteratively splits the feature space to minimize the ranking loss, followed by re-calculation of the K rankings based on cluster assignments. The MM-PL approach is a multi-prototype supervised clustering algorithm based on the Plackett-Luce (PL) probabilistic ranking model. It represents each cluster with a union of Voronoi cells that are defined by a set of prototypes, and assign each cluster with a set of PL label scores that determine the cluster central ranking. Cluster membership and ranking prediction for a new instance are determined by cluster membership of its nearest prototype. The unknown cluster PL parameters and prototype positions are learned by minimizing the ranking loss, based on two variants of the expectation-maximization algorithm. Evaluation of the proposed algorithms was conducted on synthetic and real-life label ranking data by considering several measures of cluster goodness: (1) cluster compactness in feature space, (2) cluster compactness in label ranking space and (3) label ranking prediction loss. Experimental results demonstrate that the proposed MM-PL and RankTree models are superior to the baseline models. Further, MM-PL is has shown to be much better than other algorithms at handling situations with significant fraction of missing label preferences.     

An IGBT Gate Driver for Feed-Forward Control of Turn-on Losses and Reverse Recovery Current

This paper addresses the problem of turn on performances of an insulated gate bipolar transistor (IGBT) that works in hard switching conditions. The IGBT turn on dynamics with an inductive load is described, and corresponding IGBT turn on losses and reverse recovery current of the associated freewheeling diode are analysed. A new IGBT gate driver based on feed-forward control of the gate emitter voltage is presented in the paper. In contrast to the widely used conventional gate drivers, which have no capability for switching dynamics optimisation, the proposed gate driver provides robust and simple control and optimization of the reverse recovery current and turn on losses. The collector current slope and reverse recovery current are controlled by means of the gate emitter voltage control in feed-forward manner. In addition the collector emitter voltage slope is controlled during the voltage falling phase by means of inherent increase of the gate current. Therefore, the collector emitter voltage tail and the total turn on losses are significantly reduced. The proposed gate driver was experimentally verified and compared to a conventional gate driver, and the results are presented and discussed in the paper.     

High-Voltage Auxiliary Power Supply Using Series-Connected MOSFETs and Floating Self-Driving Technique

This paper deals with high-voltage auxiliary switching-mode power supplies (SMPSs). An overview of the state of the art is given, and a novel solution is proposed. The proposed solution is based on a single-ended flyback or forward topology with the main switch arranged as a series connection of two metal-oxide-semiconductor field-effect transistors (MOSFETs). The bottom MOSFET is driven directly by an ordinary control circuit and gate driver, while the top MOSFET is driven by a floating self-supplied gate driver. The floating gate driver is connected to the input filter capacitors' midpoint. This gate driver plays two roles: driving of the top MOSFET and control of distribution of the blocking voltage among the series-connected MOSFETs, in steady state as well as during commutation. The series connection of lower voltage MOSFETs has two important advantages compared to that of a single high-voltage MOSFET: lower conduction losses and lower cost. When several switches are series connected, each switch supports a fraction of the total blocking voltage, and therefore, each switch can be rated for lower voltage. The total on-state resistance and the cost of such a switch arrangement are lower compared to that of a single switch that supports the full blocking voltage. The proposed SMPS is theoretically analyzed and experimentally verified. The experimental results are presented and discussed.