Automated nonlinear system modelling with multiple neural networks

This article discusses the identification of nonlinear dynamic systems using multi-layer perceptrons (MLPs). It focuses on both structure uncertainty and parameter uncertainty, which have been widely explored in the literature of nonlinear system identification. The main contribution is that an integrated analytic framework is proposed for automated neural network structure selection, parameter identification and hysteresis network switching with guaranteed neural identification performance. First, an automated network structure selection procedure is proposed within a fixed time interval for a given network construction criterion. Then, the network parameter updating algorithm is proposed with guaranteed bounded identification error. To cope with structure uncertainty, a hysteresis strategy is proposed to enable neural identifier switching with guaranteed network performance along the switching process. Both theoretic analysis and a simulation example show the efficacy of the proposed method.

Thermal fatigue analysis of solar panel structure for micro-satellite applications

Thermal fatigue analysis based on 2D finite difference and 3D finite element methods is carried out to study the performance of solar panel structure during micro-satellite life time. Solar panel primary structure consists of honeycomb structure and composite laminates. The 2D finite difference (I-DEAS) model yields predictions of the temperature profile during one orbit. Then, 3D finite element analysis (ANSYS) is applied to predict thermal fatigue damage of solar panel structure. Meshing the whole structure with 2D multi-layer shell elements with sandwich option is not efficient, as it misses thermal response of the honeycomb structure. So we applied a mixed approach between 3D solid and 2D shell elements to model the solar panel structure without the sandwich option.

Development of a Polymer-carbon Nanotubes based Economic Solar Collector

A low cost solar collector was developed by using polymeric components as opposed to metal and glass components of traditional solar collectors. In order to utilize polymers for the absorber of the solar collector, Carbon Nanotubes (CNT) has been added as a filler to improve the thermal conductivity and the solar absorptivity of polymers. The solar collector was designed as a multi-layer construction with considering the economic manufacturing. Through the mathematical heat transfer analysis, the performance and characteristics of the designed solar collector have been estimated. Furthermore, the prototypes of the proposed system were built and tested at a state-of-the-art solar simulator facility to evaluate the actual performance of the developed solar collector. The cost-effective polymer-CNT solar collector, which achieved efficiency as much as that of a conventional glazed flat plate solar panel, has been successfully developed.

Deep Salience: Visual Salience Modeling via Deep Belief Propagation

Visual salience is an intriguing phenomenon observed in biological neural systems. Numerous attempts have been made to model visual salience mathematically using various feature contrasts, either locally or globally. However, these algorithmic models tend to ignore the problem’s biological solutions, in which visual salience appears to arise during the propagation of visual stimuli along the visual cortex. In this paper, inspired by the conjecture that salience arises from deep propagation along the visual cortex, we present a Deep Salience model where a multi-layer model based on successive Markov random fields (sMRF) is proposed to analyze the input image successively through its deep belief propagation. As a result, the foreground object can be automatically separated from the background in a fully unsupervised way. Experimental evaluation on the benchmark dataset validated that our Deep Salience model can consistently outperform eleven state-of-the-art salience models, yielding the higher rates in the precision-recall tests and attaining the best F-measure and mean-square error in the experiments.

Design, analysis and performance of a polymer-carbon nanotubes based economic solar collector

A low cost flat plate solar collector was developed by using polymeric components as opposed to metal and glass components of traditional flat plate solar collectors. In order to improve the thermal and optical properties of the polymer absorber of the solar collector, Carbon Nanotubes (CNT) were added as a filler. The solar collector was designed as a multi-layer construction with an emphasis on low manufacturing costs. Through the mathematical heat transfer analysis, the thermal performance of the collector and the characteristics of the design parameters were analyzed. Furthermore, the prototypes of the proposed collector were built and tested at a state-of-the-art solar simulator facility to evaluate its actual performance. The inclusion of CNT improved significantly the properties of the polymer absorber. The key design parameters and their effects on the thermal performance were identified via the heat transfer analysis. Based on the experimental and analytical results, the cost-effective polymer-CNT solar collector, which achieved a high thermal efficiency similar to that of a conventional glazed flat plate solar panel, was successfully developed.

Analysis, modeling, design and optimization of future communications systems: from theory to practice

Esta tese descreve uma framework de trabalho assente no paradigma multi-camada para analisar, modelar, projectar e optimizar sistemas de comunicação. Nela se explora uma nova perspectiva acerca da camada física que nasce das relações entre a teoria de informação, estimação, métodos probabilísticos, teoria da comunicação e codificação. Esta framework conduz a métodos de projecto para a próxima geração de sistemas de comunicação de alto débito. Além disso, a tese explora várias técnicas de camada de acesso com base na relação entre atraso e débito para o projeto de redes sem fio tolerantes a atrasos. Alguns resultados fundamentais sobre a interação entre a teoria da informação e teoria da estimação conduzem a propostas de um paradigma alternativo para a análise, projecto e optimização de sistemas de comunicação. Com base em estudos sobre a relação entre a informação recíproca e MMSE, a abordagem descrita na tese permite ultrapassar, de forma inovadora, as dificuldades inerentes à optimização das taxas de transmissão de informação confiáveis em sistemas de comunicação, e permite a exploração da atribuição óptima de potência e estruturas óptimas de pre-codificação para diferentes modelos de canal: com fios, sem fios e ópticos. A tese aborda também o problema do atraso, numa tentativa de responder a questões levantadas pela enorme procura de débitos elevados em sistemas de comunicação. Isso é feito através da proposta de novos modelos para sistemas com codificação de rede (network coding) em camadas acima da sua camada física. Em particular, aborda-se a utilização de sistemas de codificação em rede para canais que variam no tempo e são sensíveis a atrasos. Isso foi demonstrado através da proposta de um novo modelo e esquema adaptativo, cujos algoritmos foram aplicados a sistemas sem fios com desvanecimento (fading) complexo, de que são exemplos os sistemas de comunicação via satélite. A tese aborda ainda o uso de sistemas de codificação de rede em cenários de transferência (handover) exigentes. Isso é feito através da proposta de novos modelos de transmissão WiFi IEEE 801.11 MAC, que são comparados com codificação de rede, e que se demonstram possibilitar transferência sem descontinuidades. Pode assim dizer-se que esta tese, através de trabalho de análise e de propostas suportadas por simulações, defende que na concepção de sistemas de comunicação se devem considerar estratégias de transmissão e codificação que sejam não só próximas da capacidade dos canais, mas também tolerantes a atrasos, e que tais estratégias têm de ser concebidas tendo em vista características do canal e a camada física.

A neural network seismic detector

This experimental study focuses on a detection system at the seismic station level that should have a similar role to the detection algorithms based on the ratio STA/LTA. We tested two types of neural network: Multi-Layer Perceptrons and Support Vector Machines, trained in supervised mode. The universe of data consisted of 2903 patterns extracted from records of the PVAQ station, of the seismography network of the Institute of Meteorology of Portugal. The spectral characteristics of the records and its variation in time were reflected in the input patterns, consisting in a set of values of power spectral density in selected frequencies, extracted from a spectro gram calculated over a segment of record of pre-determined duration. The universe of data was divided, with about 60% for the training and the remainder reserved for testing and validation. To ensure that all patterns in the universe of data were within the range of variation of the training set, we used an algorithm to separate the universe of data by hyper-convex polyhedrons, determining in this manner a set of patterns that have a mandatory part of the training set. Additionally, an active learning strategy was conducted, by iteratively incorporating poorly classified cases in the training set. The best results, in terms of sensitivity and selectivity in the whole data ranged between 98% and 100%. These results compare very favorably with the ones obtained by the existing detection system, 50%.

High-Speed Solid-Rotor Induction Machine – Electromagnetic Calculation and Design

Within the latest decade high-speed motor technology has been increasingly commonly applied within the range of medium and large power. More particularly, applications like such involved with gas movement and compression seem to be the most important area in which high-speed machines are used. In manufacturing the induction motor rotor core of one single piece of steel it is possible to achieve an extremely rigid rotor construction for the high-speed motor. In a mechanical sense, the solid rotor may be the best possible rotor construction. Unfortunately, the electromagnetic properties of a solid rotor are poorer than the properties of the traditional laminated rotor of an induction motor. This thesis analyses methods for improving the electromagnetic properties of a solid-rotor induction machine. The slip of the solid rotor is reduced notably if the solid rotor is axially slitted. The slitting patterns of the solid rotor are examined. It is shown how the slitting parameters affect the produced torque. Methods for decreasing the harmonic eddy currents on the surface of the rotor are also examined. The motivation for this is to improve the efficiency of the motor to reach the efficiency standard of a laminated rotor induction motor. To carry out these research tasks the finite element analysis is used. An analytical calculation of solid rotors based on the multi-layer transfer-matrix method is developed especially for the calculation of axially slitted solid rotors equipped with wellconducting end rings. The calculation results are verified by using the finite element analysis and laboratory measurements. The prototype motors of 250 – 300 kW and 140 Hz were tested to verify the results. Utilization factor data are given for several other prototypes the largest of which delivers 1000 kW at 12000 min-1.

Understanding deep architectures and the effect of unsupervised pre-training

Cette thèse porte sur une classe d'algorithmes d'apprentissage appelés architectures profondes. Il existe des résultats qui indiquent que les représentations peu profondes et locales ne sont pas suffisantes pour la modélisation des fonctions comportant plusieurs facteurs de variation. Nous sommes particulièrement intéressés par ce genre de données car nous espérons qu'un agent intelligent sera en mesure d'apprendre à les modéliser automatiquement; l'hypothèse est que les architectures profondes sont mieux adaptées pour les modéliser. Les travaux de Hinton (2006) furent une véritable percée, car l'idée d'utiliser un algorithme d'apprentissage non-supervisé, les machines de Boltzmann restreintes, pour l'initialisation des poids d'un réseau de neurones supervisé a été cruciale pour entraîner l'architecture profonde la plus populaire, soit les réseaux de neurones artificiels avec des poids totalement connectés. Cette idée a été reprise et reproduite avec succès dans plusieurs contextes et avec une variété de modèles. Dans le cadre de cette thèse, nous considérons les architectures profondes comme des biais inductifs. Ces biais sont représentés non seulement par les modèles eux-mêmes, mais aussi par les méthodes d'entraînement qui sont souvent utilisés en conjonction avec ceux-ci. Nous désirons définir les raisons pour lesquelles cette classe de fonctions généralise bien, les situations auxquelles ces fonctions pourront être appliquées, ainsi que les descriptions qualitatives de telles fonctions. L'objectif de cette thèse est d'obtenir une meilleure compréhension du succès des architectures profondes. Dans le premier article, nous testons la concordance entre nos intuitions---que les réseaux profonds sont nécessaires pour mieux apprendre avec des données comportant plusieurs facteurs de variation---et les résultats empiriques. Le second article est une étude approfondie de la question: pourquoi l'apprentissage non-supervisé aide à mieux généraliser dans un réseau profond? Nous explorons et évaluons plusieurs hypothèses tentant d'élucider le fonctionnement de ces modèles. Finalement, le troisième article cherche à définir de façon qualitative les fonctions modélisées par un réseau profond. Ces visualisations facilitent l'interprétation des représentations et invariances modélisées par une architecture profonde.

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements.

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements

A sensitive fibre optic pH sensor using multiple sol–gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements

A sensitive fibre optic pH sensor using multiple sol-gel coatings

The fabrication and characterization of a fibre optic pH sensor based on evanescent wave absorption is presented. The unclad portion of a multi-mode optical fibre is coated with a pH sensitive dye, which is immobilized by the sol–gel route. The sensitivity of the device has been found to increase when multiple sol–gel coatings are used as the sensing region. The dynamic range and the temporal response of the sensor are investigated for two different dyes, namely bromocresol purple and bromocresol green. The performance of the device is evaluated in terms of the results obtained during actual measurements.

Optimization of Processing Technology for Commercial Drying of Bananas (Matooke)

Summary - Cooking banana is one of the most important crops in Uganda; it is a staple food and source of household income in rural areas. The most common cooking banana is locally called matooke, a Musa sp triploid acuminate genome group (AAA-EAHB). It is perishable and traded in fresh form leading to very high postharvest losses (22-45%). This is attributed to: non-uniform level of harvest maturity, poor handling, bulk transportation and lack of value addition/processing technologies, which are currently the main challenges for trade and export, and diversified utilization of matooke. Drying is one of the oldest technologies employed in processing of agricultural produce. A lot of research has been carried out on drying of fruits and vegetables, but little information is available on matooke. Drying of matooke and milling it to flour extends its shelf-life is an important means to overcome the above challenges. Raw matooke flour is a generic flour developed to improve shelf stability of the fruit and to find alternative uses. It is rich in starch (80 - 85%db) and subsequently has a high potential as a calorie resource base. It possesses good properties for both food and non-food industrial use. Some effort has been done to commercialize the processing of matooke but there is still limited information on its processing into flour. It was imperative to carry out an in-depth study to bridge the following gaps: lack of accurate information on the maturity window within which matooke for processing into flour can be harvested leading to non-uniform quality of matooke flour; there is no information on moisture sorption isotherm for matooke from which the minimum equilibrium moisture content in relation to temperature and relative humidity is obtainable, below which the dry matooke would be microbiologically shelf-stable; and lack of information on drying behavior of matooke and standardized processing parameters for matooke in relation to physicochemical properties of the flour. The main objective of the study was to establish the optimum harvest maturity window and optimize the processing parameters for obtaining standardized microbiologically shelf-stable matooke flour with good starch quality attributes. This research was designed to: i) establish the optimum maturity harvest window within which matooke can be harvested to produce a consistent quality of matooke flour, ii) establish the sorption isotherms for matooke, iii) establish the effect of process parameters on drying characteristics of matooke, iv) optimize the drying process parameters for matooke, v) validate the models of maturity and optimum process parameters and vi) standardize process parameters for commercial processing of matooke. Samples were obtained from a banana plantation at Presidential Initiative on Banana Industrial Development (PIBID), Technology Business Incubation Center (TBI) at Nyaruzunga – Bushenyi in Western Uganda. A completely randomized design (CRD) was employed in selecting the banana stools from which samples for the experiments were picked. The cultivar Mbwazirume which is soft cooking and commonly grown in Bushenyi was selected for the study. The static gravitation method recommended by COST 90 Project (Wolf et al., 1985), was used for determination of moisture sorption isotherms. A research dryer developed for this research. All experiments were carried out in laboratories at TBI. The physiological maturity of matooke cv. mbwazirume at Bushenyi is 21 weeks. The optimum harvest maturity window for commercial processing of matooke flour (Raw Tooke Flour - RTF) at Bushenyi is between 15-21 weeks. The finger weight model is recommended for farmers to estimate harvest maturity for matooke and the combined model of finger weight and pulp peel ratio is recommended for commercial processors. Matooke isotherms exhibited type II curve behavior which is characteristic of foodstuffs. The GAB model best described all the adsorption and desorption moisture isotherms. For commercial processing of matooke, in order to obtain a microbiologically shelf-stable dry product. It is recommended to dry it to moisture content below or equal to 10% (wb). The hysteresis phenomenon was exhibited by the moisture sorption isotherms for matooke. The isoteric heat of sorption for both adsorptions and desorption isotherms increased with decreased moisture content. The total isosteric heat of sorption for matooke: adsorption isotherm ranged from 4,586 – 2,386 kJ/kg and desorption isotherm from 18,194– 2,391 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (db) respectively. The minimum energy required for drying matooke from 80 – 10% (wb) is 8,124 kJ/kg of water removed. Implying that the minimum energy required for drying of 1 kg of fresh matooke from 80 - 10% (wb) is 5,793 kJ. The drying of matooke takes place in three steps: the warm-up and the two falling rate periods. The drying rate constant for all processing parameters ranged from 5,793 kJ and effective diffusivity ranged from 1.5E-10 - 8.27E-10 m2/s. The activation energy (Ea) for matooke was 16.3kJ/mol (1,605 kJ/kg). Comparing the activation energy (Ea) with the net isosteric heat of sorption for desorption isotherm (qst) (1,297.62) at 0.1 (kg water/kg dry matter), indicated that Ea was higher than qst suggesting that moisture molecules travel in liquid form in matooke slices. The total color difference (ΔE*) between the fresh and dry samples, was lowest for effect of thickness of 7 mm, followed by air velocity of 6 m/s, and then drying air temperature at 70˚C. The drying system controlled by set surface product temperature, reduced the drying time by 50% compared to that of a drying system controlled by set air drying temperature. The processing parameters did not have a significant effect on physicochemical and quality attributes, suggesting that any drying air temperature can be used in the initial stages of drying as long as the product temperature does not exceed gelatinization temperature of matooke (72˚C). The optimum processing parameters for single-layer drying of matooke are: thickness = 3 mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6 m/s overflow mode. From practical point of view it is recommended that for commercial processing of matooke, to employ multi-layer drying of loading capacity equal or less than 7 kg/m², thickness 3 mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6 m/s overflow mode.

Optimization of the Mechanical and Optical Properties of Tunable Optical Sensor Arrays (TOSA) for a Nanospectrometer in the Visible and Near Infrared Spectral Range

Tunable Optical Sensor Arrays (TOSA) based on Fabry-Pérot (FP) filters, for high quality spectroscopic applications in the visible and near infrared spectral range are investigated within this work. The optical performance of the FP filters is improved by using ion beam sputtered niobium pentoxide (Nb2O5) and silicon dioxide (SiO2) Distributed Bragg Reflectors (DBRs) as mirrors. Due to their high refractive index contrast, only a few alternating pairs of Nb2O5 and SiO2 films can achieve DBRs with high reflectivity in a wide spectral range, while ion beam sputter deposition (IBSD) is utilized due to its ability to produce films with high optical purity. However, IBSD films are highly stressed; resulting in stress induced mirror curvature and suspension bending in the free standing filter suspensions of the MEMS (Micro-Electro-Mechanical Systems) FP filters. Stress induced mirror curvature results in filter transmission line degradation, while suspension bending results in high required filter tuning voltages. Moreover, stress induced suspension bending results in higher order mode filter operation which in turn degrades the optical resolution of the filter. Therefore, the deposition process is optimized to achieve both near zero absorption and low residual stress. High energy ion bombardment during film deposition is utilized to reduce the film density, and hence the film compressive stress. Utilizing this technique, the compressive stress of Nb2O5 is reduced by ~43%, while that for SiO2 is reduced by ~40%. Filters fabricated with stress reduced films show curvatures as low as 100 nm for 70 μm mirrors. To reduce the stress induced bending in the free standing filter suspensions, a stress optimized multi-layer suspension design is presented; with a tensile stressed metal sandwiched between two compressively stressed films. The stress in Physical Vapor Deposited (PVD) metals is therefore characterized for use as filter top-electrode and stress compensating layer. Surface micromachining is used to fabricate tunable FP filters in the visible spectral range using the above mentioned design. The upward bending of the suspensions is reduced from several micrometers to less than 100 nm and 250 nm for two different suspension layer combinations. Mechanical tuning of up to 188 nm is obtained by applying 40 V of actuation voltage. Alternatively, a filter line with transmission of 65.5%, Full Width at Half Maximum (FWHM) of 10.5 nm and a stopband of 170 nm (at an output wavelength of 594 nm) is achieved. Numerical model simulations are also performed to study the validity of the stress optimized suspension design for the near infrared spectral range, wherein membrane displacement and suspension deformation due to material residual stress is studied. Two bandpass filter designs based on quarter-wave and non-quarter-wave layers are presented as integral components of the TOSA. With a filter passband of 135 nm and a broad stopband of over 650 nm, high average filter transmission of 88% is achieved inside the passband, while maximum filter transmission of less than 1.6% outside the passband is achieved.