Search

This paper presents an improved analysis of a novel Programmable Power-factor-corrected-Based Hybrid Muitipulse Power Rectifier (PFC-HMPR) for utility interface of power electronic converters. The proposed hybrid multipuise rectifier is composed of an ordinary three-phase six-pulse diode-bridge rectifier (Graetz bridge) with a parallel connection of single-phase switched converters in each three-phase rectifier leg. In this paper, the authors present a complete discussion about the controlled rectifiers' power contribution and also a complete analysis concerning the total harmonic distortion of current that can be achieved when the proposed converter operates as a conventional 12-pulse rectifier. The mathematical analysis presented in this paper corroborate, with detailed equations, the experimental results of two 6-kW prototypes implemented in a laboratory. Index Terms--AC motor drives, high power drives for trolley-bus systems, high power factor three-phase rectifiers, multipulse rectifiers, tractions applications, 12-pulse rectifiers.

This paper demonstrates an application of the parametric cerebellar model articulation controller (P-CMAC) network--a neural structure derived from Albus' CMAC algorithm and Takagi-Sugeno-Kang parametric fuzzy inference systems. It resembles the original CMAC proposed by Albus in the sense that it is a local network, i.e., for a given input vector, only a few of the networks neurons will be active and will effectively contribute to the corresponding network output. The internal mapping structure is built in such a way that it implements, for each CMAC memory location, one linear parametric equation of the network input strengths. First, a new approach to design neural optimal control (NOC) systems is proposed. Gradient-descent techniques are still used here to adjust network weights, but this approach has many differences when compared to classical error backpropagation algorithm. Then, P-CMAC is used to control the output voltage of a proton exchange membrane fuel cell (PEM-FC), by means of NOC. The proposed control system allows the definition of an arbitrary performance/cost criterion to be maximized/minimized, resulting in an approximated optimal control strategy. Practical results of PEM-FC voltage behavior at different load conditions are shown, to demonstrate the effectiveness of the NOC algorithm. Index Terms--Control systems, fuel cells (FCs), neural networks, optimal control.

This paper presents a dynamic electrochemical model for representation, simulation, and evaluation of performance of small size generation systems emphasizing particularly proton exchange membrane fuel-cell (PEMFC) stacks. The results of the model are used to predict the output voltage, efficiency, and power of FCs as a function of the actual load current and of the constructive and operational parameters of the cells. Partial and total load insertion and rejection tests were accomplished to evaluate the dynamic response of the studied models. The results guarantee a better analytical performance of these models with respect to former ones with a consequent reduction in time and costs of projects using FCs as the primary source of energy. Additionally, this electrochemical model was tested for the SR-12 Modular PEM Generator, a stack rated at 500 W, manufactured by Avista Laboratories, for the Ballard Mark V FC and for the BCS 500-W stack. Index Terms--Automation, control, fuel cells (FCs), modeling and simulation.

This paper shows fundamentals and applications of the parametric cerebellar model arithmetic computer (P-CMAC) network: a neural structure derived from the Albus CMAC algorithm and Takagi-Sugeno--Kang parametric fuzzy inference systems. It resembles the original CMAC proposed by Albus in the sense that it is a local network, (i.e., for a given input vector, only a few of the networks nodes--or neurons--will be active and will effectively contribute to the corresponding network output). The internal mapping structure is built in such a way that it implements, for each CMAC memory location, one linear parametric equation of the network input strengths. This mapping ran be corresponded to a hidden layer in a multilayer perceptron (MLP) structure. The output of the active equations are then weighted and averaged to generate the actual outputs to the network. A practical comparison between the proposed network and other structures is, thus, accomplished. P-CMAC, MLP, and CMAC networks are applied to approximate a nonlinear function. Results show advantages of the proposed algorithm based on the computational efforts needed by each network to perform nonlinear function approximation. Also, P-CMAC is used to solve a practical problem at mobile telephony, approximating an RF mapping at a given region to help operational people while maintaining service quality. Index Terms--Cerebellar model arithmetic computers (CMACs), communication systems, mobile communication, neural networks.

This paper presents the guidelines for simulation of fuel-cell (FC) stacks by using a computer-controlled high-power converter, which drives actual electric loads, or injects power to the grid. The FC output static and dynamic characteristics are closely reproduced in such a way the actual loads are seamlessly driven as if they were supplied by the simulated FC. The simulator characteristics include the membrane temperature and humidity, efficiency, flow of the reactants, cooling air fan and water pumps, the actual air environmental temperature and humidity, and the regimen of operation of the actual electrical load. Any type of FC of ordinary size can be simulated without having to use hydrogen with improved safety, variety of tests, flexibility, and demo facilities. Those features allied to the low cost of this FC simulator contribute for market analysis and life-cycle studies of a site installation. Index Terms--Alternative energy, computer control, fuel cells (FCs), interconnection, modeling, rectifiers, simulation.

This paper presents a comprehensive review of the industrial applications of artificial neural networks (ANNs), in the last 12 years. Common questions that arise to practitioners and control engineers while deciding how to use NNs for specific industrial tasks are answered. Workable issues regarding implementation details, training and performance evaluation of such algorithms are also discussed, based on a judiciously chronological organization of topologies and training methods effectively used in the past years. The most popular ANN topologies and training methods are listed and briefly discussed, as a reference to the application engineer. Finally, ANN industrial applications are grouped and tabulated by their main functions and what they actually performed on the referenced papers. The authors prepared this paper bearing in mind that an organized and normalized review would be suitable to help industrial managing and operational personnel decide which kind of ANN topology and training method would be adequate for their specific problems. Index Terms--Architecture, industrial control, neural network (NN) applications, training.

This paper presents the design, analysis, simulation, and modeling of a high-torque low-speed multiphase permanent-magnet brushless machine. The machine fits an in-wheel motor arrangement to be used for electric vehicle applications. This paper presents issues regarding the high-level modeling comprised of a transient model in conjunction with their corresponding experimental evaluation. Analysis was made to combine the modeling efforts with the expected behavior concerned with mutual inductance and armature reaction effects, so as to have realistic simulation results verified by the experimental setup. Comprehensive experimental results corroborate the work. Index Terms--Brushless machines, digital signal processors, electric machines, modeling, motor drives.

This paper describes the development of a neural-network-based prediction of mooring forces of a deep-sea oil exploitation production process. The evolvement of a neural network simulator for analysis of the dynamic behavior of a system consisting of a turret-floating production storage and offloading (FPSO) system and a shuttle ship in tandem configuration is described. The turret-FPSO is a vessel with a cylindrical anchoring system fixed to the sea bed my mooring lines and a shuttle ship is connected during the oil transference. This system has quite complex dynamics owing to interactions of the forces and moments due to current, wind, and waves. In general, the mathematical model that represents the dynamics of these connected floating units involves a set of nonlinear equations requiring several parameters difficult to be obtained. In order to deal with such complexities, a neural network has been devised to simulate an FPSO tandem system. This approach opens new horizons for maintenance of mooring lines, preventing commons of the ships. Index Terms--Neural networks, offshore simulation, oil exploitation.

Artificial intelligence techniques, such as fuzzy logic, neural network, and genetic algorithm, are recently showing a lot of promise in the application of power electronic systems. The paper describes the control strategy development, design, and experimental performance evaluation of a fuzzy-logic-based variable-speed wind generation system that uses a cage-type induction generator and double-sided pulsewidth-modulated (PWM) converters. The system can feed a utility grid maintaining unity power factor at all conditions or can supply an autonomous load. The fuzzy-logic-based control of the system helps to optimize efficiency and enhance performance. A complete 3.5-kW generation system has been developed, designed, and thoroughly evaluated by laboratory tests, in order to validate the predicted performance improvements. The system gives excellent performance and can easily be translated to a larger size in the field. Index Terms - Fuzzy logic, intelligent control, power electronics, wind generation.

The optimal operation of an oil well requires the periodic measurement of temperature and pressure at the downhole. In this paper, acoustic waves are used to transmit data to the surface through the pipeline column of the well, making up a wireless transmission system. Binary data is transmitted in two frequencies, using frequency-shift keying modulation. Such transmission faces problems with noise, attenuation, and, at pipeline joints, multiple reflections and nonlinear distortion. Hence, conventional demodulation techniques do not work well in this case. The neural network presented here classifies signals received by the receiver to estimate transmitted data, using a linear-vector-quantization-based network, with the help of a preprocessing procedure that transforms time-domain incoming signals in three-dimensional images. The results have been successfully verified. The neural network estimation principles presented in this paper can be easily applied to other patterns and time-domain recognition applications. Index Terms--Acoustic data transmission, neural network, oil pipeline.