Pedro Pereira

The need for implementing low cost, fully integrated RF wireless transceivers has motivated the widespread use CMOS technology. However, in the particular case for voltage-controlled oscillators (VCO) where ever more stringent specifications in terms of phase-noise must be attained, the design of the on-chip LC tank is a challenging task, where fully advantage of the actual technologies characteristics must be pushed to nearly its limits. To overcome phase-noise limitations arising from the low quality factor of integrated inductors, optimization design methodologies are usually used. In this paper a model-based optimization approach is proposed. In this work the characterization of the oscillator behaviour is guaranteed by a set of analytical models describing each circuit element performance. A set of working examples for UMC130 technology, aiming the minimization of both VCO phase noise and power consumption, is addressed. The results presented, illustrate the potential of a GA optimization procedure design methodology yielding accurate and timely efficient oscillator designs. The validity of the results is checked against HSPICE/RF simulations.

In this paper an optimization-based approach for the design of RF integrated inductors is addressed. For the characterisation of the inductor behaviour the double pi-model is used. The use of this model is twofold. On one hand it enables the generation of the inductor characterisation in a few seconds. On the other hand its integration into the optimization procedure is straightforward. For the evaluation of the model element values analytical expressions based on technology parameters as well as on the device geometric characteristics are used. The use of a technology-based methodology for the evaluation of the model parameters grants the adaptability of the model to any technology. The inductor analytical characterization is integrated into an optimization-based tool for the automatic design of RF integrated inductors. This tool uses a modified genetic algorithm (MGA) optimization procedure, which has proved its validation in previous work. Due to the design parameter constraints nature as well as the topology constraints, discrete variables optimization techniques are used. The accuracy of the results is checked against a non-commercial software.

High temperature superconducting (HTS) machines are recognized to offer several advantageous features when comparing to conventional ones. Amongst these, highlights the decrease in weight and volume of the machines, due to increased current density in conductors or the absence of iron slots' teeth; or the decrease in AC losses and consequent higher efficiency of the machines, even accounting for cryogenics. These concepts have been already demonstrated and some machines have even achieved commercial stage. In this paper, several alternative approaches are applied to electrical motors employing HTS materials. The first one is an all superconducting linear motor, where copper conductors and permanent magnets are replaced by Bi-2223 windings and trapped flux magnets, taking advantage of stable levitation due to flux pinning, higher current densities and higher excitation field. The second is an induction disk motor with Bi-2223 armature, where iron, ironless and hybrid approaches are compared. Finally, an innovative command strategy, consisting of an electronically variable pole pairs' number approach, is applied to a superconducting hysteresis disk motor. All these concepts are being investigated and simulation and experimental results are presented.

This paper presents the design of a test rig for an all HTS linear synchronous motor. Although this motor showed to have several unattractive characteristics, its design raised a number of problems which must be considered in future HTS machines design. HTS electromagnetic properties led to the development of new paradigms in electrical machines and power systems, as e. g. in some cases iron removal and consequent assembly of lighter devices. This is due to superconductor's ability to carry high currents with minimum losses and consequent generation in the surrounding air of flux densities much higher than the allowed by ferromagnetic saturation. However, severe restrictions in HTS power devices design that goes further beyond cryogenic considerations must be accounted in. This is usually the case when BSCCO tapes are used as conductors. Its bending limitations and the presence of flux components perpendicular to tape surface, due to the absence of iron, have to be considered for it may turn some possible applications not so attractive or even practically unfeasible. An all HTS linear synchronous motor built by BSCCO tapes as armature conductors and two trapped-flux YBCO bulks in the mover was constructed and thrust force measurements are starting to be performed. Although the device presents severe restrictions due to the exposed and other reasons, it allowed systematising its design. A pulsed-field magnetiser to generate opposite fluxes for both YBCO bulks is also detailed. Thrust force numerical predictions were already derived and presented.

Inductive superconducting fault current limiters have already demonstrated their technical viability in electrical networks. Its architecture and robustness make them potentially adequate for distribution networks, and this type of devices is considered as an enabling technology for the advent of embedded generation with renewable energy sources. In order to promote the growth and maturity of these superconducting technologies, fast design tools must be developed, allowing simulating devices with different materials in grids with diverse characteristics. This work presents advances in the development of such tool, which, at present stage, is an effective alternative to software simulations by finite elements methods, reducing dramatically computation time. The algorithms are now compared with experimental results from a laboratory scale prototype, showing the need to refine them.

The sand pile model, in conjunction with Bean model, is often applied to describe single grain bulk superconductors. However, in several applications such as electric motors, multiseeded bulks are needed, due to the need to increase sample dimensions. In this paper, an extension of the sand pile model is presented in order to manage this type of materials. Multiseeded HTS bulk superconductors, produced, e.g., by the top-seeded melt growth process, are characterized by intra- and intergrain currents, and these are reflected in the model. However, identifying these currents from flux density measurements is not straightforward, when considering more than one grain. In fact, the number of currents increases with the number of grains, and these have to be identified from the measured field surface. A method to identify these currents based on genetic algorithms is validated with artificial data and then used in real measurements.

In this paper, an innovative approach for monitoring home electric power quality indicators is presented. Using an electric power analysis device (for this work purpose it will be a smart-meter) and a personal computer it is proved that it is possible to monitor and register electric power quality anomalies, such as long interruptions, voltage dips/swells and frequency oscillations. Through an application developed in Java, a user can view real-time electric parameters, check for electric power quality anomalies and assess load diagram of previous days. Experimental results regarding the application performance are also presented with the respective conclusions.

The key step of the analog active filter design is the optimal selection of component values due to manufactured series (E12, E24, E48, E96 and E192). In this paper, four simulation-based metaheuristics are applied to optimize four active filters using commercials available ICs as building blocks. The emphasis of this work is applying Richardson extrapolation-based sensitivity analysis in the optimization process of analog active filters. Indeed, Richardson extrapolation technique facilitates the calculation of the partial derivatives for the sensitivity using the simulation-based evaluation, without an explicit mathematical expression. Viability and benefits of the sensitivity analysis are highlighted. Monte Carlo analysis is performed in order to investigate robustness of the proposed sensitivity analysis of the active filters in case of component value variations due to specified tolerances of manufactured series.

Abstract This paper proposes a hybrid methodology for the evaluation of integrated inductors sensitivity against technological/geometrical parameters variation. The obtained results are used in an optimization-based design environment for integrated inductors, as a way of guaranteeing that obtained solutions are robust against parameter variation. For the inductor characterization, a lumped element model is used, where each element value is evaluated through physics based equations. The sensitivity of the inductor characterization to parameter variations is evaluated at two levels. At the physical level, the sensitivity of the model element values to technological/geometrical parameters variations is computed through an equation-based strategy. Then, the sensitivity of the inductor characterization to the model parameter variations is obtained through a simulation-based approach, where the Richardson extrapolation technique is used for the calculation of the partial derivatives. Several examples considering the evaluation of sensitivity of both inductance and quality factor of two inductors in \{UMC130\} technology are presented. Obtained results are compared against Monte-Carlo simulations.

In this paper we deal with analog active filter design using discrete components taking into consideration tolerance effects. Sensitivity analysis is performed to determine the most influential components in the considered circuit, thus relative higher precision is offered to those parameters. Further, an in-loop optimization technique is considered, thus actual IC models are handled. An application example is presented. HSpice simulation results, supported by Monte Carlo analysis, are given to highlight efficiency of the proposed selection technique.