Дисертації з теми "Advanced machine controls"

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Mechanical Engineering, May, 2020
Cataloged from the official PDF of thesis.
Includes bibliographical references (pages 144-150).
Many conventional problems in ocean engineering remain challenging due to the stochastic nature of ocean waves, viscous effects of the flow, nonlinear resonance, etc., and the combination of these factors. Data-driven techniques is an prospective approach complementary to traditional methods to model physical problems since data from experiments, field tests or high-fidelity simulations are mostly informative about actual physical systems. Machine learning algorithms, especially kernel based methods have very good generalization capability as well as statistical inference. This thesis targets to establish a framework that how we can use data from real-time measurements or data gathered from experiments and field tests and simulations to provide an alternative approach for physical modeling or practical engineering solutions.
In this thesis, we mainly target two different types of problems-mapping highly nonlinear physical relations and predicting time series, to prove the feasibility of such a framework. More specifically, one problem is the short-term wave prediction based on realtime measurements and its application to the advanced controls of renewable energy. The other one is the modeling of nonlinear viscous hydrodynamic loads of ships and offshore platforms. The Support Vector Machines (SVM) is used in solving both the problems. In the thesis, the SVM regression model are developed for the realtime short-term forecast of wave elevations and wave excitation forces. Optimal controllers aiming to reduce the structural loads or optimize energy capture with the knowledge of the forecasted wave force are established for offshore floating wind turbines and wave energy converters.
A series of CFD simulations of a rectangular barge with bilge keels are conducted and validated, along with the experiment data of a fixed offshore cylindrical platform, to serve as the baseline data set to model the nonlinear viscous hydrodynamic loads. Using the wave elevations and ship roll kinematics as features, the SVM regression models are trained and tested to predict the nonlinear hydrodynamic loads. The influence of the stochastic effect and different feature selections and kernel selections are discussed in the thesis as well. Key words: Machine learning, SVM regression, short-term forecast, model predictive control, nonlinear viscous hydrodynamic loads
by Yu Ma.
Ph. D.
Ph.D. Massachusetts Institute of Technology, Department of Mechanical Engineering

Le projet de cette thèse consiste en la définition d'une génératrice multiphasée tolérante aux défauts permettant d'accroitre la fiabilité d'une chaîne de production électrique. En effet, quelle que soit l'énergie primaire fossile (pétrole, gaz, charbon) ou renouvelable (hydraulique, éolien, biomasse, solaire) utilisée en amont, il faut pour la transformer en énergie électrique une chaîne de conversion basée sur une génératrice électrique et ce quel que soit le niveau de puissance (du kW au MW). Cet élément, basiquement de structure triphasée, est sujet à pannes lors de la perte d'une phase. Aussi, en multipliant le nombre de phases, la production peut perdurer pour peu qu'il reste trois phases actives. Néanmoins, même si la génération est toujours possible, il faut définir une commande intelligente s'adaptant au défaut et pilotant le convertisseur de puissance de manière à maintenir une production d'énergie électrique propre respectant les règles imposées par le réseau électrique national
The project of this thesis consists in the definition of a fault tolerant multiphase generator allowing to increase the reliability of a production line. In fact, whatever the fossil fuel (oil, gas, coal) or renewable (hydraulic, wind, biomass, solar) primary energy used upstream, it is necessary to transform it into electric energy a conversion chain based on an electric generator. whatever the power level (from kW to MW). This element, basically of three-phase structure, is subject to failures during the loss of a phase. Also, by multiplying the number of phases, the production can continue if there remain three active phases. Nevertheless, even if the generation is still possible, it is necessary to define an intelligent control that adapts to the fault and drives the power converter so as to maintain a clean power generation complying with the rules imposed by the national power grid

Synchronous machines are extensively used for home appliances and industrial applications thanks to their fast dynamic response, good overload capability and high energy density. A precise knowledge of the rotor position is required to control efficiently this kind of motors. In most of the applications resolvers or absolute encoders are installed on the rotor shaft. The employment of position sensors leads to significant drawbacks such as the increased size and cost of the system and a lower reliability of the drive, caused by additional hardware and cabling. In sensorless drives motor position is estimated and employed in the machine control. Thus, no position sensor is required by the drive and all the drawbacks entailed by the sensor are eliminated. Moreover, the position estimation could be useful for redundancy in case of system failures. Therefore, position estimation techniques are object of great interest in the electric drives field. Position estimation techniques can be divided into two main categories: methods that are suitable for medium or high speed and techniques suitable for low speed or standstill operations. In the former group the motor position is estimated through a reconstruction of the permanent magnet flux or back electromotive force (back-EMF). In case of synchronous reluctance machines it is possible to reconstruct the extended active flux or back-EMF. Stator voltages and currents measurements are needed for these reconstruction methods. Since these signals amplitude is proportional to the rotor speed, position estimation can be successfully performed only for medium and high speed machine operations. In the low speed range, sensorless schemes exploit the rotor magnetic anisotropy. Thus, position can be estimated only for anisotropic motors, i.e. synchronous reluctance motors (SynRM), permanent magnet assisted synchronous reluctance motors (PMA-SynRM) and interior permanent magnet synchronous motors (IPMSM). The rotor anisotropy is recognized thanks to an high frequency voltage injection in the stator windings. Several injection techniques have been proposed, differing from the signal typology. In particular, high frequency sinusoidal or square-wave carriers are often applied. The position information is usually extracted from the current response through a heterodyning demodulation that entails the use of low pass filters in the position estimator, limiting its dynamic. The aim of the research was proposing a new algorithm to estimate the rotor position from the HF current response, getting rid of the demodulation and its weaknesses. Thus, the ellipse fitting technique has been proposed. Robustness against signal processing delay effects and a reduced number of required filters are the main advantages of this novel approach. The inverse problem related to the ellipse fitting is solved implementing a recursive least squares algorithm. The proposed ellipse fitting technique is not affected by signal processing delay effects, and it requires the tuning of only one parameter, called forgetting factor, making the studied method suitable for industrial application thanks to its minimal setup effort. Besides the ellipse fitting technique for rotor position estimation, two other topics have been studied: - Computation of self-sensing capabilities of synchronous machines. - Online incremental inductances identification for SynRM.
Synchronous machines are extensively used for home appliances and industrial applications thanks to their fast dynamic response, good overload capability and high energy density. A precise knowledge of the rotor position is required to control efficiently this kind of motors. In most of the applications resolvers or absolute encoders are installed on the rotor shaft. The employment of position sensors leads to significant drawbacks such as the increased size and cost of the system and a lower reliability of the drive, caused by additional hardware and cabling. In sensorless drives motor position is estimated and employed in the machine control. Thus, no position sensor is required by the drive and all the drawbacks entailed by the sensor are eliminated. Moreover, the position estimation could be useful for redundancy in case of system failures. Therefore, position estimation techniques are object of great interest in the electric drives field. Position estimation techniques can be divided into two main categories: methods that are suitable for medium or high speed and techniques suitable for low speed or standstill operations. In the former group the motor position is estimated through a reconstruction of the permanent magnet flux or back electromotive force (back-EMF). In case of synchronous reluctance machines it is possible to reconstruct the extended active flux or back-EMF. Stator voltages and currents measurements are needed for these reconstruction methods. Since these signals amplitude is proportional to the rotor speed, position estimation can be successfully performed only for medium and high speed machine operations. In the low speed range, sensorless schemes exploit the rotor magnetic anisotropy. Thus, position can be estimated only for anisotropic motors, i.e. synchronous reluctance motors (SynRM), permanent magnet assisted synchronous reluctance motors (PMA-SynRM) and interior permanent magnet synchronous motors (IPMSM). The rotor anisotropy is recognized thanks to an high frequency voltage injection in the stator windings. Several injection techniques have been proposed, differing from the signal typology. In particular, high frequency sinusoidal or square-wave carriers are often applied. The position information is usually extracted from the current response through a heterodyning demodulation that entails the use of low pass filters in the position estimator, limiting its dynamic. The aim of the research was proposing a new algorithm to estimate the rotor position from the HF current response, getting rid of the demodulation and its weaknesses. Thus, the ellipse fitting technique has been proposed. Robustness against signal processing delay effects and a reduced number of required filters are the main advantages of this novel approach. The inverse problem related to the ellipse fitting is solved implementing a recursive least squares algorithm. The proposed ellipse fitting technique is not affected by signal processing delay effects, and it requires the tuning of only one parameter, called forgetting factor, making the studied method suitable for industrial application thanks to its minimal setup effort. Besides the ellipse fitting technique for rotor position estimation, two other topics have been studied: - Computation of self-sensing capabilities of synchronous machines. - Online incremental inductances identification for SynRM.

As civil engineering projects become increasingly ambitious, the earthworks that form a significant part of these projects become increasingly critical. Geotechnical engineering is concerned with the measurement , modelling and simulation of soil under various conditions for the application within civil engineering. In triaxial testing, an actuator applies a force or displacement to a soil sample. Dynamic triaxial testing is concerned with applying forces conforming to a rapidly changing pattern. The simplest case is concerned with applying a force in a sinusoidal pattern. The accuracy of models produced during triaxial testing depends largely on the accuracy of the equipment used for testing. As such, the control of the applied force is of great importance. The response of the soil sample directly influences the closed loop behaviour of the system. The soil sample response is unknown at the start of the test , can vary throughout a test and depends on various parameters, many of which are unobservable. This research investigates the issues faced in developing accurate control of dynamic triaxial testing equipment. Models of the load frame and various nonlinear soil sample responses are built for the purpose of simulation and controller design. Various control strategies are developed, implemented and tested, including the use of PID control, adaptive repetitive control and sliding mode control techniques. The recursive least squares algorithm is investigated and used for online identification in order to compensate for unknown and time varying plant parameters. Simulation and testing methods are investigated and several control methods are tested in simulation, on the target triaxial load frame using and finally using embedded processing hardware. Development and implementation techniques are also used including rapid prototyping and hardware-in-the-loop testing.

Effective control of vibration in rotating machinery is a major concern in many industries and research institutions. With the ever-increasing drive for higher operating speeds, the need for developing vibration mitigation methods that cater for the arduous operating conditions that consequently arise is paramount. Phenomena that may have been insignificant in relatively low-speed rotating machines begin to gain importance with increasing speed, an example of which is the gyroscopic effect. This thesis is aimed at enriching the knowledge base on rotordynamic vibration control. Independent Modal Control (IMC) is addressed, within the context of rotating machinery. A study is performed on various actuation technologies that may be used to implement active vibration control. The well-known problem of balancing rotating machinery is also considered. The first-order modal filters based on Structure Preserving Transformations (SPTs) are capable of decoupling a rotor dynamic system into individual modes of vibration, such that IMC may be performed. Unlike traditional control schemes, the method based on first-order modal filters does not require the imposition of highly restrictive conditions on the system (classical damping). As a result, gyroscopic effects - which are substantial in high-speed rotating machinery - and non-classical damping may be fully accounted for in the modal domain. The main problem pertaining to this method arises from the fact that the response of the controlled system is linked with the stability of the modal filters. As such, if the filters are unstable, the controlled response is eventually overcome by noise. This thesis explores the spectral properties of the modal filter with a view to understanding the factors that affect its stability; some interesting findings on the filter eigenvalues are presented. Furthermore, the question as to whether filter stability is an essential requirement is addressed. The relationship between the rotordynamic system and the modal filter is also investigated. An illustration of the techniques developed in relation to IMC using first-order modal filters is presented in the form of a FE simulation on a realistic aero-engine model. The implementation of active vibration control in a dynamic system is realised through the application of control forces by actuators. In the case of rotating machines, these would normally be located at the bearings. Actuation may be achieved from a variety of technologies such as electromagnetic, piezoelectric, magnetostrictive, ultrasonic etc. This thesis conducts a study on some popular actuation technologies, with the aim of finding an effective alternative to the ubiquitous squeeze film damper. The merits and drawbacks of the various technologies are compared. Also, some novel design concepts are proposed, and (in some cases) their viability demonstrated through calculations. It is well-known that rotor unbalance is usually the main source of vibration in rotating machines. Thus, improvements in procedures for balancing such machines are continuously being sought. With increasing in-service operating speeds and ever more stringent standards, traditional balancing methods progressively become inadequate. One of the reasons for this is the inability of balancing tests to capture the contribution of patterns of unbalance that excite higher modes of vibration, as the tests speeds are usually lower than in-service speeds. This thesis proposes a robust balancing approach that utilises additional information on rotor unbalance, in the form of a covariance matrix, to improve the balancing procedure. The method is illustrated in a FE model of a rotating machine, and is shown to be superior to the traditional method.

Cette thèse porte sur des méthodes innovantes de contrôle de l'état de santé des machines tournantes par l’analyse des signaux vibratoires. En effet, la surveillance de l’état de santé des machines contribue à des améliorations substantielles des points de vue économique et de sureté. Afin d’y aboutir, l’une des manières les plus populaires est de recueillir les vibrations de la machine. La plupart de ces vibrations sont directement liées au comportement périodique des sous-systèmes de la machine tels que les arbres de rotation, engrenages, champs électriques rotationnels, etc. Cette connaissance peut être exploitée afin de concevoir une méthodologie adaptée à chaque type de défaut. Cette thèse s’intéresse aux étapes de la mise en œuvre de cette méthodologie. En règle générale, la première condition préalable à l’analyse avancée de l’information récoltée est la disponibilité de la vitesse instantanée de rotation. Cette vitesse doit être connue car la plupart des techniques du traitement du signal sont adaptées aux conditions de fonctionnement stationnaires. Ainsi, la connaissance de la vitesse permettra de compenser les fluctuations de vitesse, par exemple par le ré-échantillonnage angulaire du signal de vibration. Malgré l’existence d’outils de mesure permettant l’estimation de la vitesse tels que les codeurs et les tachymètres, cette thèse étudie le potentiel d’estimer la vitesse instantanée de rotation à partir des signaux vibratoires. Après l'estimation de la vitesse et le ré-échantillonnage angulaire, une étape suivante courante consiste à séparer le signal en composantes déterministes et stochastiques. Dans ce sens, l’efficacité et l’applicabilité de la procédure d'édition du cepstre sont analysées. Ensuite, différentes méthodes de filtrage sont appliquées au signal résiduel afin d’améliorer le rapport signal sur bruit. Pour cette fin, les méthodes existantes utilisant des critères conventionnels sont étudiées en parallèles avec une nouvelle méthodologie aveugle de filtrage. La dernière étape du processus de traitement consiste à diagnostiquer le défaut potentiel. Ainsi, des indicateurs statistiques sont calculés sur le signal obtenu après traitement et suivis dans le temps pour vérifier leurs variations. Dans de nombreux cas, la signature du défaut présente un comportement cyclostationaire. Par conséquent, cette thèse examine également différentes techniques d'analyse de la cyclostationarité. Enfin, les performances des différentes méthodes de traitement sont validées sur deux ensembles de données expérimentales de vibrations issues de boîtes de vitesses d’éoliennes
This Ph.D. dissertation targets innovative methods for vibration-based condition monitoring of rotating machinery. Substantial benefits can be achieved from an economical and a safety point of view using condition monitoring. One of the most popular methods to gather information about the state of machine parts is through the analysis of machine vibrations. Most of these vibrations are directly linked to periodical behavior of subsystems within the machine like e.g. rotating shafts, gears, rotating electrical fields, etc. This knowledge can be exploited to enable faultdependent processing schemes. This dissertation investigates how to implement and utilize these processing schemes and details the steps in such a procedure. Typically, the first prerequisite for advanced analysis is the availability of the instantaneous rotation speed. This speed needs to be known since most frequency-based analysis techniques assume stationary behavior. Knowledge of the speed thus allows for compensating speed fluctuations, for example through angular resampling of the vibration signal. While there are hardware-based solutions for speed estimation using angle encoders or tachometers, this thesis investigates the potential in vibration signals for speed estimation. After speed estimation and angular resampling, a common next step is to separate the signal into deterministic and stochastic components. The cepstrum editing procedure is examined for its efficacy and applicability. Afterwards, different filtering methods are inspected as to improve the signal-to-noise ratio of the signal content of interest. Existing methods using conventional criteria are investigated together with a novel blind filtering methodology. The final step in the multi-step processing scheme is to search for the potential fault. Statistical indicators can be calculated on the processed time domain signal and tracked over time to check for increases. In many cases, the fault signature exhibits cyclostationary behavior. Therefore this dissertation also examines different cyclostationary analysis techniques. Lastly, the performance of the different processing methods is validated on two experimental vibration data sets of wind turbine gearboxes

In this thesis, we discussed various possibilities of integrating scheduling decisions with information and constraints from Advanced Process Control (APC) systems in semiconductor Manufacturing. In this context, important questions were opened regarding the benefits of integrating scheduling and APC. An overview on processes, scheduling and Advanced Process Control in semiconductor manufacturing was done, where a description of semiconductor manufacturing processes is given. Two of the proposed problems that result from integrating bith systems were studied and analyzed, they are :Problem of Scheduling with Time Constraints (PTC) and Problem of Scheduling with Equipement health Factor (PEHF). PTC and PEHF have multicriteria objective functions.PTC aims at scheduling job in families on non-identical parallel machines with setup times and time constraints.Non-identical machines mean that not all miachines can (are qualified to) process all types of job families. Time constraints are inspired from APC needs, for which APC control loops must be regularly fed with information from metrology operations (inspection) within a time interval (threshold). The objective is to schedule job families on machines while minimizing the sum of completion times and the losses in machine qualifications.Moreover, PEHF was defined which is an extension of PTC where scheduling takes into account the equipement Health Factors (EHF). EHF is an indicator on the state of a machine. Scheduling is now done by considering a yield resulting from an assignment of a job to a machine and this yield is defined as a function of machine state and job state.

The pace of technological change is increasing and sophisticated customer driven markets are forcing rapid machine evolution, increasing complexity and quality, and faster response. To survive and thrive in these markets, machine builders/suppliers require absolute customer and market orientation, focusing on rapid provision of solutions rather than products. Their production systems will need to accommodate unpredictable changes while maintaining financial and operational efficiency with short lead and delivery times. Real-Virtuality (R-V) systems are an innovative environment to address these requirements by facilitating enhanced support in machine system design utilising integrated real-virtual environments centred on concurrent machine system development and realization. This environment supports not only machine system design but also the development of the control system at the same time. Utilising the Real-Virtual Mapping Environment (RVME), 3-D simulation machine models can perform actual machine operations in real-time when coupled with the real machine controller. This provides a more understandable, reliable and transparent machine function and performance. The research study explores different types of controller verification methods and proposes a new method which employs the use of a control signal emulator. The research study has fomulated a novel technique for emulating quadrature encoder signals to provide virtual closed loop control of servomotors. The deployment of a control signal emulator technique makes the system unique and removes its dependency on specific hardware. Enabling the real-time data from the signal emulation environment eases the task of realising a real-time machine simulator. To evaluate the proposed architecture, three case studies were performed. The results have shown that it is possible to create verified and validated machine control programs with no modification needed when applied to the real machine. The migration from the virtual to the real world is totally seamless. The result from the study show that the virtual machine is able to operate and respond as a real machine in real-time. This opens up the unexplored potential of integrated 3-D virtual technology. The real-time 3-D simulation virtual machine will enable commissioning and training to be conducted at an earlier stage in the design process (without having to wait for the real machine to be built). Furthermore, various test scenarios can also be developed and tested on the system which helps to provide a better lofriderstanding of the machine behaviours and responses. This research study has made an original contribution in the field of machine system development. It has contributed a novel approach of using emulated control signals to provide machine control programmers with a platform to test their application programs at machine level which involves both discrete digital signals and continuous signals. The real-time virtual environment extends the application domain for the use of simulation. The architecture proposed is generic; to be exact it is not constrained to a specific industrial control system or to a specific simulation vendor.

This thesis presents the design and analysis of specialised control systems for switched reluctance (SR) and permanent magnet (PM) machines in vehicular electric applications. Control systems for operation in motoring and power generation are considered for both the types of machines. The SR machine operation considered in this thesis is mainly focused towards the application of aero-engine starter/generators. The control designs for PM machines are formulated considering general fault-tolerant and isolated multiphase PM machines which can be applied in the majority of safety-critical vehicular power and propulsion applications. The SR motoring mode presented in this thesis considers the control design for operation from zero speed to a high speed range, while SR generation mode is confined to the high speed range, such as for the requirements of aero-engine starter/generator operation. This thesis investigates applied control methods for both single-pulse and chopping modes of operation. Classical excitation control versus peak current control and the introduction of a zero-voltage interval are compared for SR motor operation. Optimized excitation control versus two classical forms of excitation control are developed and compared for SR generator operation. Studies include simulation of a 12/8 250kW machine and experimental work on a 6/4 300W machine. The PM motoring and power generation considered in this thesis focuses on a special class of PM machines and drives which are specifically designed for fault-tolerant operation. Optimized control strategies for the operation of PM machines with the parallel H-bridge per-phase converter architecture are investigated. Mathematical modelling of the machine and drive with a consideration of harmonics is presented. The developed control methods are then evaluated by means of finite-element model based simulations of a 125kW five phase surface PM rotor machine and an interior PM rotor machine.

Everyone is noting that Permanent Magnet Synchronous Machines (PMSMs) are more and more invading home, commercial and industrial applications thanks to their superior performance and efficiency, high torque and power density. Due to these features, this kind of machines is widely employed in several fields of technology. In the transportation branch, electric traction based on PMSM is becoming a strong competitor with conventional propulsion systems based on either combustion engines or induction machines. Clear examples are pure and hybrid electric vehicles, trains etc... Due to the always more pressing energy label constraints, PMSM are conquering the market of large home appliances, e.g. washing machines and dryers, refrigerators and air conditioner. Finally, robotics and high precision systems, machine tool and industrial applications are always more employing PMSM for high performance and efficiency devices. In recent years there has been a growing interest in Synchronous Reluctance Machines (SyRMs) since they feature superior reliability and are cost effective due to the absence of permanent magnets in the rotor. Due to these relevant characteristics, the possibility of replacing PMSM with SyRMs in existing applications is very attractive for people working in all the aforementioned fields. The motivations of this thesis are driven by industrial needs. In particular, the topics of this research have been developed according to specific requirements provided by E.E.I. S.p.a. (Equipaggiamenti Elettronici Industriali - Vicenza, Italy), the industrial partner that has been supporting the scholarship of my PhD. Two main research themes have been determined regarding the control of synchronous machines, in particular IPMSMs and SyRMs. First, the possibility to develop a control algorithm able to exploit the wide speed range operation capability of the machine. This is a crucial aspect in the design of an electric drive that allows avoiding the oversizing of both the power converter and the motor, which in turn brings to space and money savings. The second topic consists in the analysis of open research topics regarding the sensorless control of PMSM. Position estimation algorithms allow avoiding the use of position sensors leading to advantages in term of cost and reliability of the drive. For these reasons sensorless applications are very attractive from an industrial point of view. However, the performances of conventional algorithms for the position estimation at zero or low speed degrade in case of heavy saturation condition and they could be prone to instability.
Le macchine sincrone a magneti permanenti (PMSM) sono sempre più utilizzate in applicazioni domestiche, commerciali e industriali per via della loro superiore efficienza, coppia elevata e alta densità di potenza. Grazie a queste caratteristiche, questo tipo di macchine viene ampiamente utilizzato in diversi campi. Nel settore dei trasporti, la trazione elettrica basata su PMSM sta diventando un forte concorrente dei sistemi di propulsione convenzionali basati su motori a combustione o macchine a induzione. Chiari esempi sono i veicoli elettrici puri e ibridi, i treni, ecc. A causa dei vincoli sempre più stringenti delle classi energetiche, le macchine PMSM stanno conquistando il mercato dei grandi elettrodomestici, ad es. lavatrici e asciugatrici, frigoriferi e condizionatori. Infine, la robotica e i sistemi di alta precisione, le macchine utensili e le applicazioni industriali impiegano sempre più macchine a magneti permanenti per dispositivi ad alte prestazioni ed efficienza. Negli ultimi anni c'è stato un crescente interesse per le macchine a riluttanza sincrona (SyRMs) poiché presentano una superiore affidabilità e sono economicamente vantaggiose a causa dell'assenza di magneti permanenti nel rotore. Per questo la possibilità di sostituire i motori a magneti permanenti presenti nelle esistenti applicazioni con macchine sincrone a riluttanza è di grande interesse per chi lavora nei campi precedentemente citati. Esigenze industriali hanno determinato le attività descritta in questa tesi. In particolare, gli argomenti di ricerca sono stati scelti e sviluppati in base alle specifiche esigenze di E.E.I. S.p.a. (Equipaggiamenti Elettronici Industriali - Vicenza, Italia), il partner industriale che ha finanziato la mia borsa di studio. Due principali temi di ricerca riguardanti il controllo delle macchine sincrone, in particolare IPMSM e SyRM, sono stati individuati. Innanzitutto, la possibilità di sviluppare un algoritmo di controllo in grado di sfruttare la capacità di questi motori di lavorare in un ampio intervallo di velocità. Questo è un aspetto cruciale nella progettazione di un azionamento elettrico poichè consente di evitare il sovradimensionamento del convertitore di potenza e del motore, che a sua volta porta a risparmi di spazio e denaro. Il secondo argomento riguarda l'analisi del controllo sensorless dei motori sincroni. Gli algoritmi di stima della posizione consentono di evitare l'uso di sensori che portando vantaggi in termini di costi e affidabilità dell'azionamento. Per queste ragioni le applicazioni sensorless sono molto interessanti da un punto di vista industriale. Tuttavia, le prestazioni degli algoritmi convenzionali per la stima della posizione a velocità bassa o nulla peggiorano in caso di condizioni di forte saturazione e potrebbero portare perfino all’instabilità dell’algoritmo.

Pickable is a prototype, in the early stages of development, of an industrial CDPR developed by Tecnalia Research and Innovation within the "PicknPack" project, launched by the European Union to promote the development of flexible automated systems for the packaging of fresh or packaged food products. Current industrial solutions for pick and place applications are either delta parallel robots or serial arms SCARA robots but, in recent years, PKM based pick and place robots have begun to become more and more important in the pick and place market. The idea behind Pickable is to combine the principles of CDPR and PKM to build an alternative to the usual industrial manipulators that has workspace agility, a better footprint, and lower material cost. Pickable has some special elements compared to market standards; the CDPR itself with an industrial PC, cables, winches, and a platform integrating a pneumatic feeding system and two small motors to move the end-effector; and a vision system, consisting of a camera, a conveyor, an additional PC, and a graphical interface. Developed in collaboration with the LIRMM research laboratory, the aim of this Master Thesis was to carry on with the development of the machine in order to have all the components working and integrated with each other so it will then be possible to focus on the optimization and analysis of more specialized topics of interest. As a first step, the vision system was integrated with the manipulator, through the development of TCP/IP socket connections to allow the exchange of crucial information for operations and security. The second step involved the creation of a state machine to allow the management of the various functions, which was then also integrated with the GUI. The last step was the execution of tests to evaluate the preliminary dynamics performance of the machine and highlight the parameters that have the greatest influence.

Le principal objectif de cette thèse vise à estimer de manière automatique la qualité de la traduction de langue parlée (Spoken Language Translation ou SLT), appelée estimation de confiance (Confidence Estimation ou CE). Le système de SLT génère les hypothèses représentées par les séquences de mots pour l'audio qui contient parfois des erreurs. En raison de multiples facteurs, la sortie de SLT, ayant une qualité insatisfaisante, pourrait causer différents problèmes pour les utilisateurs finaux. Par conséquent, il est utile de savoir combien de confiance les tokens corrects pourraient être trouvés au sein de l'hypothèse. L'objectif de l'estimation de confiance consistait à obtenir des scores qui quantifient le niveau de confiance ou à annoter les tokens cibles en appliquant le seuil de décision (par exemple, seuil par défaut = 0,5). Dans le cadre de cette thèse, nous avons proposé un boîte à outils, qui consiste en un framework personnalisable, flexible et en une plate-forme portative, pour l'estimation de confiance au niveau de mots (Word-level Confidence Estimation ou WCE) de SLT.En premier lieu, les erreurs dans le SLT ont tendance à se produire sur les hypothèses de la reconnaissance automatique de la parole (Automatic Speech Recognition ou ASR) et sur celles de la traduction automatique (Machine Translation ou MT), qui sont représentées par des séquences de mots. Ce phénomène est étudié par l'estimation de confiance (CE) au niveau des mots en utilisant les modèles de champs aléatoires conditionnels (Conditional Random Fields ou CRF). Cette tâche, relativement nouvelle, est définie et formalisée comme un problème d'étiquetage séquentiel dans lequel chaque mot, dans l'hypothèse de SLT, est annoté comme bon ou mauvais selon un ensemble des traits importants. Nous proposons plusieurs outils servant d’estimer la confiance des mots (WCE) en fonction de notre évaluation automatique de la qualité de la transcription (ASR), de la qualité de la traduction (MT), ou des deux (combiner ASR et MT). Ce travail de recherche est réalisable parce que nous avons construit un corpus spécifique, qui contient 6.7k des énoncés pour lesquels un quintuplet est normalisé comme suit : (1) sortie d’ASR, (2) transcription en verbatim, (3) traduction textuelle, (4) traduction vocale et (5) post-édition de la traduction. La conclusion de nos multiples expérimentations, utilisant les traits conjoints entre ASR et MT pour WCE, est que les traits de MT demeurent les plus influents, tandis que les traits de ASR peuvent apporter des informations intéressantes complémentaires.En deuxième lieu, nous proposons deux méthodes pour distinguer des erreurs susceptibles d’ASR et de celles de MT, dans lesquelles chaque mot, dans l'hypothèse de SLT, est annoté comme good (bon), asr_error (concernant les erreurs d’ASR) ou mt_error (concernant les erreurs de MT). Nous contribuons donc à l’estimation de confiance au niveau de mots (WCE) pour SLT par trouver la source des erreurs au sein des systèmes de SLT.En troisième lieu, nous proposons une nouvelle métrique, intitulée Word Error Rate with Embeddings (WER-E), qui est exploitée afin de rendre cette tâche possible. Cette approche génère de meilleures hypothèses de SLT lors de l'optimisation de l'hypothèse de N-meilleure hypothèses avec WER-E.En somme, nos stratégies proposées pour l'estimation de la confiance se révèlent un impact positif sur plusieurs applications pour SLT. Les outils robustes d’estimation de la qualité pour SLT peuvent être utilisés dans le but de re-calculer des graphes de la traduction de parole ou dans le but de fournir des retours d’information aux utilisateurs dans la traduction vocale interactive ou des scénarios de parole aux textes assistés par ordinateur.Mots-clés: Estimation de la qualité, Estimation de confiance au niveau de mots (WCE), Traduction de langue parlée (SLT), traits joints, Sélection des traits
The main aim of this thesis is to investigate the automatic quality assessment of spoken language translation (SLT), called Confidence Estimation (CE) for SLT. Due to several factors, SLT output having unsatisfactory quality might cause various issues for the target users. Therefore, it is useful to know how we are confident in the tokens of the hypothesis. Our first contribution of this thesis is a toolkit LIG-WCE which is a customizable, flexible framework and portable platform for Word-level Confidence Estimation (WCE) of SLT.WCE for SLT is a relatively new task defined and formalized as a sequence labelling problem where each word in the SLT hypothesis is tagged as good or bad accordingto a large feature set. We propose several word confidence estimators (WCE) based on our automatic evaluation of transcription (ASR) quality, translation (MT) quality,or both (combined/joint ASR+MT). This research work is possible because we built a specific corpus, which contains 6.7k utterances for which a quintuplet containing: ASRoutput, verbatim transcript, text translation, speech translation and post-edition of the translation is built. The conclusion of our multiple experiments using joint ASR and MT features for WCE is that MT features remain the most influent while ASR features can bring interesting complementary information.As another contribution, we propose two methods to disentangle ASR errors and MT errors, where each word in the SLT hypothesis is tagged as good, asr_error or mt_error.We thus explore the contributions of WCE for SLT in finding out the source of SLT errors.Furthermore, we propose a simple extension of WER metric in order to penalize differently substitution errors according to their context using word embeddings. For instance, the proposed metric should catch near matches (mainly morphological variants) and penalize less this kind of error which has a more limited impact on translation performance. Our experiments show that the correlation of the new proposed metric with SLT performance is better than the one of WER. Oracle experiments are also conducted and show the ability of our metric to find better hypotheses (to be translated) in the ASR N-best. Finally, a preliminary experiment where ASR tuning is based on our new metric shows encouraging results.To conclude, we have proposed several prominent strategies for CE of SLT that could have a positive impact on several applications for SLT. Robust quality estimators for SLT can be used for re-scoring speech translation graphs or for providing feedback to the user in interactive speech translation or computer-assisted speech-to-text scenarios.Keywords: Quality estimation, Word confidence estimation (WCE), Spoken Language Translation (SLT), Joint Features, Feature Selection

The topic of this thesis is the development and the implementation of advanced vibration processing techniques for machine condition monitoring and diagnostics with two fields of applications: the quality control of I.C. engines by means of cold tests, and the monitoring and control of harvesting processes. The cold test, i.e. the final test after the assembly line and before shipping the engine to the customer, consists of the final quality control of the engine in a non-combustion state. Techniques for engine condition monitoring based on the analysis of vibration signals are widely used. However, these techniques are often applied to engine tests in firing conditions. This thesis addresses the use of several signal processing tools as a means for the monitoring and the diagnosis of assembly faults through the cold test technology. Firstly, an approach based on the use of Symmetrized Dot Patterns for the visual characterization of vibration signatures is proposed in order to obtain reliable thresholds for the pass/fail decision after the cold test. Secondly, the fault identification is discussed on the basis of the cyclostationary modelling of the signals. The first-order cyclostationarity is exploited through the analysis of the Time Synchronous Average (TSA). Subsequently, secondorder cyclostationarity is analysed by means of the Wigner-Ville Distribution (WVD), Wigner-Ville Spectrum (WVS) and Mean Instantaneous Power (MIP). Moreover, Continuous Wavelet Transform (CWT) is presented and compared with the WVD and WVS. The choice of different wavelet functions and some methods for the CWT map optimization (i.e. purification method and the average across the scale vi method (TDAS)) are also considered. Moreover, the capabilities of the Instantaneous Angular Speed (IAS) in detecting assembly faults have been tested. It is worth noting that the cyclostationary and time-frequency technique capabilities have been verified for both simulated and real signals. The experimental results indicate that the image correlation of Symmetrised Dot Patterns is a good solution that can be used in the cold test technology in order to increase its efficiency and fault detection capability. Moreover, it will be proved that the first order cyclostationary analysis is able to identify the presence of assembly faults but it is not appropriate to localise the faults. The second order analysis overcomes this problem indicating the angular position of the mechanical part affected by the fault. This is achieved by means of a correlation between the results obtained from the cyclostationarity analysis and the angular position of the mechanical events. Concerning the time-frequency analysis, the WVS as well as the CWT, using both Morlet mother wavelet and TDAS method can be considered good tools to characterise the transients due to the fault events in the timefrequency domain. Thanks to this research study it is possible to understand which of the above-mentioned techniques is effective for an easy and fast quality control and for the diagnosis of the considered assembly faults. Moreover, the limits and drawbacks of both monitoring and diagnostic procedures are shown. The originality of the first part of the research mainly concerns the use of vibration measurements for the quality control of engines at the end of the assembly line while the greater part of methods used for cold test applications focuses on pressure and torque measurements. The second part of this thesis concerns the analysis of relationships between the harvesting process parameters relative to a nonconventional harvesting machine and its vibration response. Common and uncommon features extracted from a segmentation analysis have been correlated with the harvesting process efficiency in order to define the optimal monitoring feature subset. Moreover, the Discrete Wavelet Transform method is performed in order to find the vii frequency range mostly characterised by impulsive components. In addition, some outlines obtained through the vibro-acoustic analysis performed in the angular domain are also given. Two different indoor and outdoor test rigs have been built to test the machine under different setting conditions in order to evaluate their influence over the vibration response of the threshing unit. The test results are used to identify how the vibration generation is linked to the crop distribution during the threshing process. Good correlations have been obtained by analysing the concave middle radial signal and by calculating the relationships that exist between some time domain features and the efficiency parameters. These features can be assumed as good indexes in explaining the crop distribution between the rotor and the concave and, consequently, the efficiency of the process. Moreover, it will be shown that the vibroacoustic features selected are well-connected to the different sources of the concave excitation. The main original contribution of this second part concerns the use of the vibration signal as an effective way to monitor the harvesting process. It can also be considered as a proper quality control indicator for the user during field operations.

Permanent magnet motors are known to provide higher torque and better efficiency than induction motors. They have found applications such as propulsion, motion control, traction, etc in industry. To reduce production costs by eliminating the use of position sensor, many sensorless drive schemes have emerged to fulfil this aim. Most of these sensorless algorithms utilise the back-EMF and magnetic saliency of PM motors to predict the rotor position, which is necessary for any closed loop vector controlled drive implementation. At zero or low speed operation, most sensorless schemes failed to perform well because of the inaccuracy in determining the small induced back-EMF. Contrary, at high-speed, the sampling rate of the rotor position must be sufficiently high. Moreover, the stator resistance varies after prolonged operation due to dissipated heat from the motor, which is significant when the motor is loaded and stalled. The author has proposed several sensorless algorithms to tackle these problems. Compensations are made to take account of the non-linearity of the switching devices in the inverter and the effect of dead time phenomenon. Small signal analyses of PM machine models and systems that include the speed and current regulators are carried out to ensure that they are stable for all operating conditions. Simulations and experimental results are presented to demonstrate that the sensorless drives work in practical implementations. Stability analyses are conducted to verify that the proposed sensorless algorithms are stable in both motoring and regenerating regions.

Recent technological developments have implemented the use of proportional control in prosthetic hands, giving rise to the importance of appropriate myoelectric control. EMG models in the past have assumed a linear proportionality to simplify the EMG-force relationships. However, it has been shown that a non-linear EMG-force relationship may be a more effective model. This study focused on evaluating three different control algorithms for a novel myoelectric training device, consisting of a toy car controlled by EMG signals from the distal muscles in the arm. Sixteen healthy adult subjects (5 male and 11 female) with an average age of 23.6 years (SD = 2.7) were asked to drive the car through a slalom course. Completion times as well as number of errors (wall hits, cone hits, and reversals) were recorded to evaluate performance. The NASA TLX was administered to evaluate psychometrics such as mental demand, physical demand, frustration, and overall workload. The average total errors per trial on the final day of testing using the linear proportional algorithm was found to be statistically significantly (p < 0.05) lower than digital and non-linear proportional. The average course completion time per trial and overall workload using the non-linear proportional algorithm was found to be statistically significantly (p < 0.05) lower than digital and linear proportional. These results suggest that a non-linear algorithm would be most appropriate for myoelectric control in prosthetic hands.

La sophistication grandissante des dispositifs intelligents ultra-portatifs, tels que les smartphones ou les tablettes,crée un besoin d'amélioration des performances des organes de conversion de puissance.La tendance des technologies d'acheminement de puissance évolue progressivement vers une fréquence plus élevée, une meilleure densité d'intégration et une plus grande flexibilité dans les schémas d'asservissement. La modulation dynamique de tension est utilisée dans les circuits intégrés de gestion de puissances(DVS PMICs)des transmetteurs RF alors que la modulation DVFS est utilisée dans les PMICs dédiées au CPUs et GPUs. Des DCDC flexibles et fonctionnant à haute fréquence constituent aujourd'hui la solution principale en conjonction avec des régulateurs à faible marge de tension (LDO).L'évolution vers des solutions à base de HFDCDC de faibles dimensions pose un défi sérieux en matière de 1)stabilité des boucles d'asservissement,2)de complexité des architectures de contrôle imbriquant des machines d'état asynchrones pour gérer une large dynamique de puissance de sortie et 3)de portabilité de la solutions d'une technologie à une autre.Les solutions les plus courantes atteignent aujourd'hui une gamme de 2 à 6 Mhz de fréquence de découpage grâce à l'usage de contrôleurs à hystérésis qui souffrent de la difficulté à contenir la fréquence de découpage lors des variations de la tension ou du courant en charge.Nous avons voulu dans ce travail étendre l'usage des méthodes de conception et de modélisation conventionnelles comme le modèle petit signal moyen, dans une perspective de simplification et de création de modèles paramétriques. L'objectif étant de rendre la technique de compensation flexible et robuste aux variations de procédés de fabrication ou bien aux signaux parasités inhérents à la commutation de puissance.Certes, le modèle moyen petit signal, au demeurant bien traité dans la littérature, réponds amplement à la problématique de compensation des DCDCs notamment quand la stabilité s'appuie sur le zéro naturel à haute fréquence inhérent à la résistance série ESR de la capacité de sortie, mais les HFDCDC actuels utilisent des capacités MLCC ayant une très faible ESR et font appel à des techniques de compensation paramétriques imbriquant le schéma de compensation dans la génération même du rapport cyclique. La littérature existante sur le fonctionnement de la machine d'état, se contente d'une description simpliste de convertisseurs PWM/PFM mais ne donne que très peu d'éléments sur la gestion des opérations synchrones/asynchrones alternant PWM,PFM,écrêtage de courant, démarrage ou détection de défaillance. Dans ce travail, notre études est axée sur les deux aspects suivants:1)La modélisation paramétrique et la compensation de la boucle d'asservissement de HFDCDC et 2)la portabilité de la conception de la machine d'états du contrôleur notamment lorsqu'elle intègre des transitions complexes entre les modes.Dans la première section, nous avons développé un modèle petit signal moyen d'un convertisseur Buck asservi en mode courant-tension et nous l'avons analysé pour faire apparaitre les contributions proportionnelle, intégrale et dérivé dans la boucle. Nous avons démontré la possibilité d'utiliser le retour en courant pour assurer l'amortissement nécessaire et la stabilité de la boucle pour une large dynamique de variations des conditions de charge.Dans la seconde section, nous avons développé une architecture de machine d'états sophistiquée basé sur la méthode d'Huffman avec un effort substantiel d'abstraction que nous a permis de la concevoir en description RTL pour une gestion fiable du fonctionnement asynchrone et temps réel.Notre contribution théorique a fait l'objet d'une réalisation d'un PMIC de test comportant deux convertisseurs Buck cadencés à 12MHz en technologie BiCMOS 0.5um/0.18um. Les performances clefs obtenues sont:une surtension de 50mV pendant 2us suite à l'application d'un échelon de courant de 300mA
The continuous sophistication of smart handheld devices such as smartphones and tablets creates an incremental need for improving the performances of the power conversion devices. The trend in power delivery migrates progressively to higher frequency, higher density of integration and flexibility of the control scheme. Dynamic Voltage Scaling Power Management ICs (DVS PMIC) are now systematically used for powering RF Transmitters and DVFS PMICS using Voltage and Frequency scaling are used for CPUs and GPUs. Flexible High frequency (HF) DC/DC converters in conjunction with low dropout LDOs constitute the main solution largely employed for such purposes. The migration toward high frequency/small size DCDC solutions creates serious challenges which are: 1) the stability of the feedback loop across a wide range of loading voltage and current conditions 2) The complexity of the control and often-non-synchronous state machine managing ultra large dynamics and bridging low power and high power operating modes, 3) The portability of the proposed solution across technology processes.The main stream solutions have so far reached the range of 2 to 6 MHz operation by employing systematically sliding mode or hysteretic converters that suffer from their variable operating frequency which creates EMI interferences and lead to integration problems relative to on-chip cross-talk between converters.In this work we aim at extend the use of traditional design and modeling techniques of power converters especially the average modeling technique by putting a particular care on the simplification of the theory and adjunction of flexible compensation techniques that don't require external components and that are less sensitive to process spread, or to high frequency substrate and supply noise conditions.The Small Signal Average Models, widely treated in the existing literature, might address most needs for system modeling and external compensation snubber design, especially when aiming on the high frequency natural zero of the output capacitor. However, HFDCDC converters today use small size MLCC capacitors with a very low ESR which require using alternative techniques mixing the compensation scheme with the duty cycle generation itself. The literature often provides a simplistic state machine description such as PWM/PFM operations but doesn't cover combined architectures of synchronous / non synchronous mode operations such as PWM, PFM, Current Limit, Boundary Clamp, Start, Transitional and finally Fault or Protection modes.In our work, we have focused our study on two main axes: 1) The parametric modeling and the loop compensation of HFDCDC and 2) the scalability of the control state machine and mode inter-operation. In the first part, we provided a detailed small signal averaged model of the “voltage and current mode buck converter” and we depicted it to emphasize and optimize the contributions of the Proportional, Integral and Derivative feedback loops. We demonstrated the ability to use the current feedback to damp and stabilize the converter with a wide variety of loading conditions (resistive or capacitive). In the second part, we provided architecture of the mode control state machine with different modes like the PWM, PFM, soft-start, current limit,… .The technique we have used is inspired by Huffman machine with a significant effort to make it abstract and scalable. The state machine is implemented using RTL coding based on a generic and scalable approach.The theoretical effort has been implemented inside a real PMIC test-chip carrying two 12MHz buck converters, each employing a voltage and current mode feedback loop. The chip has been realized in a 0.5um / 0.18um BiCMOS technology and tested through a dedicate Silicon validation platform able to test the analog, digital and power sections. The key performance obtained is a 50mV load transient undershoot / overshoot during 2us following a load step of 300mA (slope 0.3A/ns)

In this study, universal motion command generator systems utilizing a Field Programmable Gate Array (FPGA) and an interface board for Robotics and Computer Numerical Control (CNC) applications have been developed. These command generation systems can be classified into two main groups as polynomial approximation and data compression based methods. In the former type of command generation methods, the command trajectory is firstly divided into segments according to the inflection points. Then, the segments are approximated using various polynomial techniques. The sequence originating from modeling error can be further included to the generated series. In the second type, higher-order differences of a given trajectory (i.e. position) are computed and the resulting data are compressed via lossless data compression techniques. Besides conventional approaches, a novel compression algorithm is also introduced in the study. This group of methods is capable of generating trajectory data at variable rates in forward and reverse directions. The generation of the commands is carried out according to the feed-rate (i.e. the speed along the trajectory) set by the external logic dynamically. These command generation techniques are implemented in MATLAB and then the best ones from each group are realized using FPGAs and their performances are assessed according to the resources used in the FPGA chip, the speed of command generation, and the memory size in Static Random Access Memory (SRAM) chip located on the development board.

The methods used today for the commissioning and validation of industrial machines requires theconstruction of physical prototypes. Those prototypes help the engineers to e.g. validate if theprogram code meant to control a machine works as intended. In recent years the development ofnew techniques for the commissioning and validation of industrial machines has changed rapidlythanks to the development of new software. The method used in this thesis is called simulationin the loop. Another method that can be benecial to use is hardware in the loop. Using thosemethods for the commissioning of a machine is called virtual commissioning. The simulation inthe loop method is used to simulate both the machine and the control system that operate thatmachine. This is called a digital twin, a virtual copy of the physical hardware and its control systemthat can be used without the need for a real prototype to be available.The software used in this thesis comes all from the company Siemens and those are TIA Portal,Mechatronics Concept Designer, SIMIT and PLCSim Advanced. By using those programs it waspossible to build a digital twin with rigid body dynamics and its control system of the industrialmodel that was given by the company Renholmen AB. This model contained all the necessarycomponents needed for a virtual commissioning project to be done without the need to be at thefactory oor.The results showed that it was possible to achieve a real time simulation, allowing the possibilityto trim the controller parameters without the need of a physical prototype. Design errors were alsofound thanks to the results of the simulation.This new technique has shown to be a useful tool due to most of the work could be done on a digitalmodel of the machine. Simulations can reduce the time to market for industrial machines and alsohelp engineers to validate and optimize the product at an early stage. This tool that can be usedto validate industrial machines before they are created.

Thesis (Ph. D.)--Missouri University of Science and Technology, 2009.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed May 27, 2009) Includes bibliographical references (p. 126-130).

The efficacy of the energy management systems at dealing with energy consumption in buildings has been a topic with a growing interest in recent years due to the ever-increasing global energy demand and the large percentage of energy being currently used by buildings. The scale of this sector has attracted research effort with the objective of uncovering potential improvement avenues and materializing them with the help of recent technological advances that could be exploited to lower the energetic footprint of buildings. Specifically, in the area of heating, ventilating and air conditioning installations, the availability of large amounts of historical data in building management software suites makes possible the study of how resource-efficient these systems really are when entrusted with ensuring occupant comfort. Actually, recent reports have shown that there is a gap between the ideal operating performance and the performance achieved in practice. Accordingly, this thesis considers the research of novel energy management strategies for heating, ventilating and air conditioning installations in buildings, aimed at narrowing the performance gap by employing data-driven methods to increase their context awareness, allowing management systems to steer the operation towards higher efficiency. This includes the advancement of modeling methodologies capable of extracting actionable knowledge from historical building behavior databases, through load forecasting and equipment operational performance estimation supporting the identification of a building’s context and energetic needs, and the development of a generalizable multi-objective optimization strategy aimed at meeting these needs while minimizing the consumption of energy. The experimental results obtained from the implementation of the developed methodologies show a significant potential for increasing energy efficiency of heating, ventilating and air conditioning systems while being sufficiently generic to support their usage in different installations having diverse equipment. In conclusion, a complete analysis and actuation framework was developed, implemented and validated by means of an experimental database acquired from a pilot plant during the research period of this thesis. The obtained results demonstrate the efficacy of the proposed standalone contributions, and as a whole represent a suitable solution for helping to increase the performance of heating, ventilating and air conditioning installations without affecting the comfort of their occupants.
L’eficàcia dels sistemes de gestió d’energia per afrontar el consum d’energia en edificis és un tema que ha rebut un interès en augment durant els darrers anys a causa de la creixent demanda global d’energia i del gran percentatge d’energia que n’utilitzen actualment els edificis. L’escala d’aquest sector ha atret l'atenció de nombrosa investigació amb l’objectiu de descobrir possibles vies de millora i materialitzar-les amb l’ajuda de recents avenços tecnològics que es podrien aprofitar per disminuir les necessitats energètiques dels edificis. Concretament, en l’àrea d’instal·lacions de calefacció, ventilació i climatització, la disponibilitat de grans bases de dades històriques als sistemes de gestió d’edificis fa possible l’estudi de com d'eficients són realment aquests sistemes quan s’encarreguen d'assegurar el confort dels seus ocupants. En realitat, informes recents indiquen que hi ha una diferència entre el rendiment operatiu ideal i el rendiment generalment assolit a la pràctica. En conseqüència, aquesta tesi considera la investigació de noves estratègies de gestió de l’energia per a instal·lacions de calefacció, ventilació i climatització en edificis, destinades a reduir la diferència de rendiment mitjançant l’ús de mètodes basats en dades per tal d'augmentar el seu coneixement contextual, permetent als sistemes de gestió dirigir l’operació cap a zones de treball amb un rendiment superior. Això inclou tant l’avanç de metodologies de modelat capaces d’extreure coneixement de bases de dades de comportaments històrics d’edificis a través de la previsió de càrregues de consum i l’estimació del rendiment operatiu dels equips que recolzin la identificació del context operatiu i de les necessitats energètiques d’un edifici, tant com del desenvolupament d’una estratègia d’optimització multi-objectiu generalitzable per tal de minimitzar el consum d’energia mentre es satisfan aquestes necessitats energètiques. Els resultats experimentals obtinguts a partir de la implementació de les metodologies desenvolupades mostren un potencial important per augmentar l'eficiència energètica dels sistemes de climatització, mentre que són prou genèrics com per permetre el seu ús en diferents instal·lacions i suportant equips diversos. En conclusió, durant aquesta tesi es va desenvolupar, implementar i validar un marc d’anàlisi i actuació complet mitjançant una base de dades experimental adquirida en una planta pilot durant el període d’investigació de la tesi. Els resultats obtinguts demostren l’eficàcia de les contribucions de manera individual i, en conjunt, representen una solució idònia per ajudar a augmentar el rendiment de les instal·lacions de climatització sense afectar el confort dels seus ocupants

Thesis (MScEng)--University of Stellenbosch, 2002.
ENGLISH ABSTRACT: The use of commercial-off-the-shelf (COTS) components and emerging technologies in satellite systems has become increasingly popular over the past few years. This is mainly due to advances in radiation shielding and system-level reliability improving techniques. The use of a new generation commercial processor in the design of a satellite's on-board computer (OBC) is now considered a feasible option. The aim of this thesis was to evaluate the use of a commercial grade ARM processor in a low earth orbit (LEO) microsatellite on-board computer. The process began with the selection of the most suitable ARM processor for an aBC design. A typical aBC system was developed for the chosen processor, in order to test its functionality and performance in anOBC design.
AFRIKAANSE OPSOMMING: Die afgelope paar jaar het die gebruik van gewone kommersiële komponente en die nuutste tegnologie in satelliet stelsels heelwat toegeneem. Dit kan grootliks toeskryf word aan die vordering in bestralings afskerming en stelselvlak betroubaarheid tegnieke. Die gebruik van 'n nuwe generasie kommersiële verwerker in die ontwerp van 'n satellite aanboord rekenaar (AR) is nou prakties uitvoerbaar. Die doel van hierdie tesis was om die gebruik van 'n ARM verwerker in 'n lae aardwentelbaan mikrosatelliet aanboord rekenaar te evalueer. Eerstens moes die mees geskikte kommersiële ARM verwerker vir 'n AR ontwerp gevind word. Daarna is 'n tipiese AR stelselontwikkel om die verwerker se funksionaliteit en werkverrigting te toets.

Recent research and development has resulted in a working prototype of an automated process for manufacturing IsoTruss® and other innovative open lattice composite structures which yields faster, and more predictable and consistent parts, while automatically consolidating individual members. This machine is sufficiently versatile to manufacture any type of open lattice structure fabricated from filamentary composite materials. The objectives of the research in this thesis were two-fold: (1) to validate this new process for making IsoTruss structures; and (2) to measure the compression strength and stiffness of specimens produced on the machine. In order to accomplish the first purpose, various parts were manufactured on this prototype machine, including: a six-node IsoTruss structure with single outer longitudinal members, a three-longitudinal member section of an inner longitudinal IsoTruss structure with consolidated members, and a two-bay IsoTruss panel structure. By creating and running patterns to make these parts, the hypothesis that the machine will make any geometry of IsoTruss structure was validated. The second objective of this research was accomplished by testing the compression strength and stiffness of specimens manufactured with this automated process. Buckling versus compression failure of members was examined by varying member aspect ratios. The effect of intersecting helical members was also explored, as was the effect of changing the number of braiding bobbins used to consolidate members. Testing showed that increasing the number of braiders increases consistency of the braided sleeves and reduces scatter in the results. The ratio of helical to longitudinal tows at a joint is directly related to the percent decrease in member strength at the joint. Compression failure of individual members is the preferred method of failure, because this type of failure absorbs significantly more energy. This research proves that the manufacturing process will produce even the most complex IsoTruss geometries, with the necessary consolidation of individual members. Findings also indicate that a few modifications -- such as improved bobbins, more reliable switches, more accurate pulling system, etc. -- will enable this automated process to produce composite lattice structures with superior mechanical properties.

The vehicles of tomorrow will be more sophisticated, intelligent and safe than the vehicles of today. The future is leaning towards fully autonomous vehicles. This degree project provides a data driven solution for a speed adaptation system that can be used to compute a vehicle speed for curves, suitable for the underlying driving style of the driver, road properties and weather conditions. A speed adaptation system for curves aims to compute a vehicle speed suitable for curves that can be used in Advanced Driver Assistance Systems (ADAS) or in Autonomous Driving (AD) applications. This degree project was carried out at Volvo Car Corporation. Literature in the field of speed adaptation systems and factors affecting the vehicle speed in curves was reviewed. Naturalistic driving data was both collected by driving and extracted from Volvo's data base and further processed. A novel speed adaptation system for curves was invented, implemented and evaluated. This speed adaptation system is able to compute a vehicle speed suitable for the underlying driving style of the driver, road properties and weather conditions. Two different artificial neural networks and two mathematical models were used to compute the desired vehicle speed in curves. These methods were compared and evaluated.
Morgondagens fordon kommer att vara mer sofistikerade, intelligenta och säkra än dagens fordon. Framtiden lutar mot fullständigt autonoma fordon. Detta examensarbete tillhandahåller en datadriven lösning för ett hastighetsanpassningssystem som kan beräkna ett fordons hastighet i kurvor som är lämpligt för förarens körstil, vägens egenskaper och rådande väder. Ett hastighetsanpassningssystem för kurvor har som mål att beräkna en fordonshastighet för kurvor som kan användas i Advanced Driver Assistance Systems (ADAS) eller Autonomous Driving (AD) applikationer. Detta examensarbete utfördes på Volvo Car Corporation. Litteratur kring hastighetsanpassningssystem samt faktorer som påverkar ett fordons hastighet i kurvor studerades. Naturalistisk bilkörningsdata samlades genom att köra bil samt extraherades från Volvos databas och bearbetades. Ett nytt hastighetsanpassningssystem uppfanns, implementerades samt utvärderades. Hastighetsanpassningssystemet visade sig vara kapabelt till att beräkna en lämplig fordonshastighet för förarens körstil under rådande väderförhållanden och vägens egenskaper. Två olika artificiella neuronnätverk samt två matematiska modeller användes för att beräkna fordonets hastighet. Dessa metoder jämfördes och utvärderades.

With the increasing popularity of phone application development, a number of features have surfaced that enable users to utilize extended functionalities with their phone. With a single device, a user is now able to have access to the latest trends. As a result, functionalities from various products such as personal computers or GPS devices can now be accessible in one place. While this technology is evolving at a fast rate, it is the embedded technology and hardware that drive it which facilitate the innovative designs and solutions. At the other end of the spectrum of embedded systems, companies continue to maintain legacy products that use embedded chips programmed through their flash memory. With this perceived commitment to stay loyal to their older product designs, it has become rather cumbersome for such companies to stay current with the latest hardware and software trends. The result is higher costs to the customers for their high-end products. While the average user invests in newer and faster Google phones or iPhones, it is unlikely that the same can be said for commercial products such as a refrigerator or a ball machine. What if a development framework existed that allowed customers to have access to better user interfaces and functionalities over the lifetime of their product? In this Masters report, an innovative approach is discussed which demonstrates the latest mobile phone technology combined with an existing embedded device being applied to the conversion of a low-end tennis ball machine into an affordable high-end one.
text

碩士
國立交通大學
電控工程研究所
103
Ethernet Automation Technology EtherCAT (Ethernet for control automation technology) that is established for real-time control applications now is widely adopted in Industrial automation. In this thesis, the packet congestion and data dropout in EtherCAT master ET9200 were analyzed by the packet monitor of Beckoff ET2000. Both the interleaving transmission/receive communication and the missing message estimator are proposed to improve the data loss. On the other hand, ITRI has developed an CNC control interface with the real-time EtherCAT system INtime for master of EtherCAT Internet communication, experimental results that Beckoff ET2000 measure packet transmission time indicate that the sampling time can be set to 0.25ms without data dropout. By implementing the EtherCAT CNC interface on the Tongtai CNC XZ table, all the proportional-derivative (PD) controller, the digital disturbance observer (DDOB), and the cross-coupled controller (CCC) can be realized. Furthermore, both the static nonlinear friction compensator (NFC) and the dynamic nonlinear friction compensator LuGre model were realized on the real-time EtherCAT CNC controller to significantly improve motion precision. Compared with the P Controller, the roundness was improved from 17.8 to 3.8 based on the encoder measurements by applying the proposed integrated controller. Furthermore, its roundness was also significantly reduced from 26.2 to 12.0 from the measurements of double-ball bar.

Research Doctorate - Doctor of Philosophy (PhD)
Model Predictive Control (MPC) is a popular control strategy studied in many research publications. Moreover, its acceptance by industry is slow. This thesis identifies a class of AC application, which can significantly benefit from MPC paradigm. Thus, is high performance AC drives operating in industrial environments where common mode voltage (CMV) is a critical aspect. After critical analysis of the existing MPC-based approaches, the thesis proposes a new and advanced MPC scheme called Feedback Quantization Model Predictive Control (FBQ-MPC). The proposed scheme has a number of important improvements, including integral action, advanced disturbance rejection, improved modulation performance and control over the harmonic spectrum, as well as CMV minimization. Application of the proposed FBQ-MPC method is demonstrated with two selected power converter options found as most appropriate in CMV sensitive environment. Based on the above, full models of industrial AC drive have been developed and studied by simulation and experiment. The studies have shown that AC drives based on FBQ-MPC overcome the common MPC drawback and offer prominent advantages in CMV sensitive, as well as more general, AC drive applications.