Dissertations / Theses on the topic 'Electrical machine design and optimization'

The interest on rotating electrical machines adopting permanent magnets (PMs) has increased during the past few decades, representing now a fashionable design option in a number of fields as industrial processing, transportation, actuators, household appliances and power plants. The issues related to an increasing electrical energy demand and consumption, have generated a tendency to research electrical drives with high efficiency, pushing electrical machines technology to further improvements. The introduction of permanent magnets based on rare earth, experienced since the 1960's, gave a great input in the development of innovative machine topologies. On the other hand, the increase and the instability of rare earth PMs price, especially between 2010 and 2013, have directed the research of rare earth free alternatives, or machines using a smaller amount of PMs. Actually, the current trend in the industrial and academic research focused on developing high performance electric motors and generators, among different electrical machines, highlights the Synchronous Reluctance (SynRel) and the Permanent Magnet Synchronous Reluctance (PMASynRel) motors as best candidates to satisfy the future energy and efficiency requirements. This thesis is comprehensively dedicated to theoretical and experimental analysis and design of the Synchronous Reluctance (SynRel) and the Permanent Magnet Assisted Synchronous Reluctance (PMASynRel) machines. In particular, it will be focused on electrical machines which power ratings are ranging from fraction of Watts to some hundred kWatts, for vehicular traction and house-hold appliances. SynRel and PMASynRel motors exhibit many technical advantages, like simple and robust structure, high torque density, high efficiency, small space required for PMs, high degrees of freedom in the design, high operating speed range, high overload capability, low back EMF (null in case of SynRel motors), leading to a safe behavior in case of inverter failure. Furthermore, thanks to the appropriate vector control algorithm, the performance in terms of efficiency and torque become highly competitive. The subject matter covered in the thesis is organized into three Parts, each including a certain number of chapters. % At the beginning of each Part, a brief summary is proposed aiming to describe the main content of that Part, the goals and the anticipation of the main results. Part I includes seven chapters summarizing the research activities carried out during the Ph.D. period. The first six chapters are dedicated to electrical machines for vehicular traction, while chapter 7 investigate on motors for house-hold appliances. Chapter 1 is mainly introductory and meant to provide the basics information to understand which are the pros and cons and the features of the machines under study. Chapter 2 summarize the design criteria and the electrical requirements of ferrite PMASynRel machine for traction application. The electromechanical performance have been evaluated and compared, in terms of torque and power. A brief description of the electric supply system in order to accurately and efficiently manage the motor for achieving the requested performance is presented. Chapter 3 highlights the influence and benefits of using ferrite magnets on the machine performance, emphasizing the importance of a careful evaluation of the magnet volume in order to increase the performance while reducing the used quantity. Chapter 4 is devoted to the sensitivity analysis of the machine performance in terms of torque ripple with respect to the geometrical design. An optimization algorithm has been performed in order to investigate and determine a rotor geometry which maximize the torque and reduces the torque ripple. The impact of the geometrical parameters is taken into account and the sensitivity of the optimal solution to the geometry variation is pointed out. This chapter highlights the difficulty to get a robust geometry as far as the torque ripple reduction is concerned. Finally, a few experimental results on a Synchronous Reluctance motor prototype will be presented, compared with Finite Element Analysis simulations for validation. Chapter 5 deals with the design and optimization of a high speed PMASynRel motor considering the driving cycles of an electric vehicle. A procedure is employed to evaluate the most effective design area, which has to be considered for the global optimization. Both results and advantages of the adopted methodology are highlighted. Further analysis on traction machines are going to be presented in Chapter 6. A comparison between ferrite and sintered NdFeB PMASynRel, SynRel and a Surface mounted PM (SPM) machines performance is deeply investigated. Chapter 7, the last of this first part, will highlight the advantages in using SynRel and PMASynRel motors for house-hold appliances. The main purpose of this chapter is to discuss the features of these motors as a valid substitute to commercial motors actually used for washing machines application. Part II is dedicated to the analytical modeling of SynRel machines with the challenge of predicting accurately the air-gap field of the machine taking into account the effect of the rotor flux barriers. This Part is divided into two chapters. Chapter 8 explains the hypothesis on which the analytical model is based, the calculation of stator Magneto Motive Force through winding function and describes the reluctance network equivalent circuit for a SynRel motor with one and two flux barriers per pole. The computation of the parameters of the model, the air-gap flux density and finally some comparison with Finite Element Analysis are presented. In Chapter 9 SynRel motors with split-phase stator winding sets supplied by multiple inverters have been investigated as an increasingly attractive solution for fault-tolerant, rugged, magnet-free vehicle traction drives. As an extension to the previous chapter, an analytical procedure to model and simulate a SynRel motor, with a split-phase stator winding, through a magnetic equivalent circuit (MEC) technique, has been introduced. As an output, the air-gap flux density of the SynRel motor can be computed at any operating point. Part III, finally, presents some experimental measurements carried out for two prototypes of SynRel and PMASynRel machines, with the purpose of comparing the results achieved in the motor optimization presented in Chapter 7.
In questi ultimi anni l’interesse per le macchine elettriche rotanti facenti uso di magneti permanenti ha riscontrato uno sviluppo sempre pi `u crescente. Tali macchine rappresentano un mondo alternativo alle tradizionali macchine sincrone e ad induzione, e vengono considerate ad oggi soluzioni promettenti in svariati settori, come quello industriale, per il trasporto, come attuatori, elettro domestici e per l’impiego in impianti di potenza. I problemi legati all aumento della domanda di energia elettrica ed al suo consumo, hanno generato una tendenza alla ricerca di azionamenti ad alta efficienza, spingendo la tecnologia delle macchine elettriche classiche ad ulteriori miglioramenti. L’introduzione dei magneti permanenti che utilizzano terre rare, gia dagli anni 60, hanno incentivato e permesso lo sviluppo di diversi tipi di macchina innovativi. Tuttavia, l’aumento e l’instabilita del prezzo delle terre rare, tra il 2010 ed il 2013, ha diretto la ricerca verso soluzioni di macchine alternative senza magneti permanenti, o con una quantita ridotta di tali materiali, pur soddisfando le specifiche di progetto. Al momento, la ricerca industriale e quella accademica sono entrambe focalizzate allo sviluppo di motori e generatori elettrici con elevate prestazioni, tra i diversi tipi di macchine elettriche esistenti, i motori sincroni a riluttanza (SynRel) ed a riluttanza assistita da magneti permanenti (PMASynRel) risultano essere degli ottimi candidati per il soddisfacimento delle specifiche energetiche e di efficienza, sempre piu stringenti, che verranno richieste ai motori nel prossimo futuro. Questo lavoro di tesi e interamente dedicato all’analisi teorica e sperimentale ed alla progettazione di motori sincroni a riluttanza (SynRel) e motori sincroni a riluttanza assistita da magneti permanenti (PMASynRel). In particolare, l’attenzione sar`a p osta su macchine elettriche in un campo di potenza che varia dalle centinaia di Watt alle decinedi kiloWatt, principalmente per applicazioni come veicoli elettrici ed elettro domestici. Tali macchine presentano una serie di vantaggi tecnologici che le portano ad avere prestazioni, soprattutto nel campo degli azionamenti a velo cit`a variabile (VSD), competitive rispetto ad esempio alle macchine ad induzione tradizionali o quelle a magneti permanenti. La struttura semplice e robusta, l’utilizzo ridotto di magneti permanenti, i gradi di liberta nella progettazione combinate ad un’elevata densita di coppia, alta efficienza elevate caratteristiche di sovraccarico ed un ampio campo di velocita, sono tutte caratteristiche che hanno permesso di collo care le macchine SynRel e PMASynRel in una posizione di rilievo. Inoltre, grazie all’aumento dei convertitori moderni a frequenza variabile e sistemi di controllo digitale, le prestazioni di questo tip o di motori, in termini di coppia ed efficienza, sono diventate altamente competitive rispetto ai tradizionali azionamenti con motori ad induzione.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2016.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 283-285).
Improvement of performance of robots has necessitated technological advances in control algorithms, mechanical structures, and electric machines. Running, legged robots have presented challenges in the area of electric machinery in particular. In addition to the low-speed, high-torque, low-mass requirements on the machines, the act of running results in an unconventional drive cycle that consists of brief periods of high torque followed by long stretches of minimal torque requirement, a performance envelope that is not matched by commercially-available machines. An optimized motor would dissipate the minimum possible power over the given drive cycle, lowering temperatures and potentially reducing required battery mass or extending range. These performance requirements have motivated faster modeling techniques to enable optimization of designs for these unconventional applications. This thesis presents a novel, fast modeling method for permanent magnet synchronous machines consisting of a hybrid model comprising an explicit Maxwell solution and a Flux Tube solution. The Maxwell solution is performed for the rotor and airgap of the machine, where geometries are simple and materials are homogeneous. The stator, with its geometric complexities and non-linear materials, is modeled with a lumped-parameter model based on ux tubes. The two models are then stitched together, forced to be self-consistent with boundary conditions, and allowed to converge. This captures effects such as cogging torque as well as saturation of the core materials. The method is approximately four orders of magnitude faster than a reference finite element program (0.01 s versus 100 s) for the same accuracy. The modeling method is implemented for two topologies of surface-mount permanent-magnet machines, an internal-rotor machine and an external-rotor machine. It is then used to optimize machine design to a given drive cycle, including effects of core loss. A machine is built to demonstrate the validity of the model and optimization method and test results match predictions of instantaneous torque to within 5% at the worst point. Cogging torque is another aspect of performance that is important to machines for robotics and other applications. These pulsations in torque caused by magnet alignment with geometric features in the stator result in undesired vibrations and issues with control. One method, based on skew, for reducing or eliminating cogging torque is explored, and a simple analytical technique to predict the eect of skew is presented. Based on the machine optimized for the Cheetah, two additional machines were built to explore the effects of cogging: a skewed-rotor machine, and a skewed- stator machine. Each demonstrated reduction of a particular cogging harmonic or all of the cogging. The skewed machines reduced cogging by approximately 85%. Novel magnet shapes which further reduce cogging are presented and finite element modeling suggests that they can further reduce cogging by 60% over a straight skew. The design and optimization tools developed herein and described above were used to optimize a motor for the MIT Cheetah Robot. The resulting motor showed nearly an order of magnitude increase in torque density when compared to commercial, off-the-shelf machines (1.3 kg vs 820 g and 10 Nm vs 28 Nm) with simultaneous improvements to efficiency.
by Matthew G. Angle.
Ph. D.

Thesis: S.M., Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2017.
S.M. !c Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science 2017
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 91-94).
For an insurance company, assessing risk exposure for Property Damage (PD), and Business Interruption (BI) for large commercial clients is difficult because of the heterogeneity of that exposure, within a single client (account), and between different divisions, and regions, where the client is active. Traditional risk assessment models attempt to scale up the single location approach used in personal lines: A large amount of data is collected to profile a sample of the locations and based on this information the risk is then inferred and somewhat subjectively assessed for the whole account. The assumption is that the risk characteristics at the largest locations are representative of all locations, and moreover, that risk is proportional to the size of the location. This approach is both ineffective and inefficient. Thus our first goal is to build a better risk assessment model through machine learning based on clients' data from internal sources. Further, we define a new problem, to predict whether a specific contract would be profitable or unprofitable for the insurance company. This problem turns out to be an imbalance classification, which attracts the second half of our research efforts in this thesis. In Chapter 2, we first review related literature on state-of-the-art risk assessment models in the field of insurance. Later in the chapter we move to the imbalance classification problems and review some popular and effective solutions researchers have proposed. In Chapter 3, we describe the data structure, provide some preliminary analysis over certain attributes and discuss the preprocessing techniques used for feature construction. In Chapter 4, we propose a new model with the objective to develop a new risk index which represents clients' potential future risk level. We then compare the performance of our new index with the original risk index used by the insurance company and computational results show that our new index successfully captures clients' financial loss pattern, while the original risk score used by the insurance company fails to do so. In Chapter 5, we propose a multi-layer algorithm to predict whether a specific contract would be profitable or unprofitable for the insurance company. Simulation shows that we can accurately label more than 83 percent of the contracts on record and that our proposed algorithm outperforms traditional classifiers such as Support Vector Machines and Random Forests. Later in the chapter, we define a new imbalance classification problem and propose a hybrid method to improve the recall percentage and prediction accuracy of Support Vector Machines. The method incorporates unsupervised learning techniques into the classical Support Vector Machines algorithm and achieves satisfying results. In Chapter 6, we conclude the thesis and provide future research guidance. This thesis builds models and trains algorithms based on real world business data from a global leading insurance and reinsurance company.
by Ziyu Wang.
S.M.
S.M. !c Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science

Advances in electrical machinery with high efficiencies could significantly reduce the cost of industrial and residential energy systems, thereby reducing fossil fuel needs and emissions. Electrical machine design is a comprehensive process based on several factors, including economic factors, material limitations, specifications and special application-dependent factors. At the same time, machine design is a multi-physics task comprising of electric design, magnetic design, insulation design, thermal design and mechanical design. However, the out-of-date conventional machine design can neither reflect the advances in the past 30 years, nor exploit the trade-offs between design factors from the multi-physics nature of the electrical machine. This work focus on the development a fast and efficient method for the design and optimization of Surface Mount Permanent Magnet (SMPM) machines and induction machines, as influenced by the energy source, mechanical loads, thermal effects, and the up-to-date developments in materials and manufacturing capabilities. A new analytical design method is developed for the electromagnetic design of SMPM machines. Both distributed and concentrated winding types of SMPM machines are considered and compared. Based on the proposed electromagnetic analytical design method and a generic thermo-mechanical machine design model [1], an innovative and computationally efficient electromagnetic-thermo-mechanical integrated design method is developed for SMPM machines. Particle Swarm Optimization (PSO) is applied in a novel way based on this integrated design method for the multi-objective design optimization of SMPM machines. With the proposed method, the thermal and mechanical design is no longer treated separately and heuristically as in the traditional design, but has been systemically integrated with the electromagnetic design; the effect of power source, cooling capability, thermal limits, and up-to-date material capabilities are also reflected in the design and optimization. Superior designs compared to traditional designs can be achieved with PSO based multi-objective optimization. The proposed integrated design approach also has the merit of good computational efficiency and provides a significant time reduction of the design cycle compared to finite element analysis. A novel electromagnetic analytical design method of induction machines has been developed, which needs only six prime design variables but is able to design induction machines in fine details. The advantage over the traditional and other existing design method is that this proposed method does not have the heuristic selection of the design variables and does not need manual design iterations. The computing time is almost negligible and the design cycle is significantly reduced compared to the tradition machine design.

Modern electric machine drives, particularly three phase permanent magnet machine drive systems represent an indispensable part of high power density products. Such products include; hybrid electric vehicles, large propulsion systems, and automation products. Reliability and cost of these products are directly related to the reliability and cost of these systems. The compatibility of the electric machine and its drive system for optimal cost and operation has been a large challenge in industrial applications. The main objective of this dissertation is to find a design and control scheme for the best compromise between the reliability and optimality of the electric machine-drive system. The effort presented here is motivated by the need to find new techniques to connect the design and control of electric machines and drive systems. A highly accurate and computationally efficient modeling process was developed to monitor the magnetic, thermal, and electrical aspects of the electric machine in its operational environments. The modeling process was also utilized in the design process in form finite element based optimization process. It was also used in hardware in the loop finite element based optimization process. The modeling process was later employed in the design of a very accurate and highly efficient physics-based customized observers that are required for the fault diagnosis as well the sensorless rotor position estimation. Two test setups with different ratings and topologies were numerically and experimentally tested to verify the effectiveness of the proposed techniques. The modeling process was also employed in the real-time demagnetization control of the machine. Various real-time scenarios were successfully verified. It was shown that this process gives the potential to optimally redefine the assumptions in sizing the permanent magnets of the machine and DC bus voltage of the drive for the worst operating conditions. The mathematical development and stability criteria of the physics-based modeling of the machine, design optimization, and the physics-based fault diagnosis and the physics-based sensorless technique are described in detail. To investigate the performance of the developed design test-bed, software and hardware setups were constructed first. Several topologies of the permanent magnet machine were optimized inside the optimization test-bed. To investigate the performance of the developed sensorless control, a test-bed including a 0.25 (kW) surface mounted permanent magnet synchronous machine example was created. The verification of the proposed technique in a range from medium to very low speed, effectively show the intelligent design capability of the proposed system. Additionally, to investigate the performance of the developed fault diagnosis system, a test-bed including a 0.8 (kW) surface mounted permanent magnet synchronous machine example with trapezoidal back electromotive force was created. The results verify the use of the proposed technique under dynamic eccentricity, DC bus voltage variations, and harmonic loading condition make the system an ideal case for propulsion systems.

Interior Permanent-magnet Machines (IPMs) have seen wide usage in industry due to their robustness, high efficiency, and low manufacturing cost. Among various IPM topologies, V-shape IPMs have been claimed to exhibit higher power densities. However, designing an IPM has been a challenging task due to the complexity of the rotor structure and magnetic saturation. The objective of this work is to set forth a rigorous design paradigm for V-shape IPMs based on multi-objective optimization. The proposed approach aims to address the multi-disciplinary nature of the design process by incorporating electromagnetic, structural, and thermal analyses. The electromagnetic analysis combines the advantage of an analytical field solution and a magnetic-equivalent-circuit (MEC) approach so it is both computationally efficient and is able to address magnetic nonlinearity in the rotor. The structural analysis uses static stress/strain methods to compute the stress distribution within critical regions in the machine. A thermal analysis based on thermal-equivalent-circuit (TEC) is proposed and the impact of including the thermal analysis on the machine metrics is discussed. Using the proposed approaches, the computational cost is significantly reduced compared to traditional FEA methods. A multi-objective design paradigm incorporating relevant design constraints is set forth. A performance comparison between IPMs and SPMs is conducted for a 2.5-hp motor application.

This work covers many research aspects of anisotropic synchronous motors, which are synchronous reluctance (SyR), permanent magnet assisted synchronous reluctance (PMaSyR) and interior permanent magnet (IPM) machines. In fact, all these kinds of machines exhibit quite a strong reluctance torque component, hence the name anisotropic. From the early 2000s, the design of electric machines started to deeply rely on finite element analysis (FEA) coupled to automatic optimization algorithms. This workflow enabled the machine designer to make fewer initial sizing hypotheses and to explore a wider design space. The drawbacks of this approach are that the time required is long and that the computational resources needed are quite large. However, the computing performances have always been improving over the years, especially when multi-processor architectures became widespread. Therefore nowadays it is common to employ tens or even hundreds of cores on cluster PCs to perform FEA during optimization runs. The thesis is structured as follows. The first part gives the background knowledge needed to develop the topics covered in the following. This comprehends an introduction to the machines studied, some general knowledge about magnetic materials, some basic concepts about the differential evolution (DE) algorithm, and the drawing of fluid rotor flux-barriers. The second part deals with the analytical modeling of SyR and PMaSyR machines. The complete model is nonlinear and may become convoluted to develop especially in an industrial environment. Therefore, using simplifying assumptions, a handful of simple equations can be derived. This simple model is also extended and applied to asymmetric rotor structures, which try to compensate torque harmonics. The third part focuses on applied multi-objective optimizations coupled to FEA for many different case studies. In particular, a SyR motor (SyRM) for pumping applications is optimized, prototyped and tested. Then, a feasibility study on a very low speed PMaSyR motor is carried out through multi-objective optimization. After that, high speed SyRMs are studied and optimized to understand the power limits of this kind of machine. Finally, the DE multi-objective optimization algorithm is also applied to improve the sensorless-control capabilities of anisotropic machines by design.
Questo lavoro analizza molti aspetti di ricerca dei motori sincroni anisotropi, che includono le macchine sincrone a riluttanza pura (SyR), a riluttanza assistita da magneti (PMaSyR) e le macchine a magneti permanenti interni (IPM). Infatti, tutte queste macchine esibiscono una forte componente di riluttanza, da cui il nome anisotrope. Dai primi anni 2000, la progettazione di macchine elettriche ha cominciato a basarsi in modo consistente sull’analisi agli elementi finiti (FEA) accoppiata ad algoritmi di ottimizzazione automatici. Questo flusso di lavoro permette al progettista di fare un minor numero di ipotesi preliminari e di esplorare uno spazio di progetto più ampio. Gli svantaggi di questo approccio sono che il tempo richiesto è lungo e che le risorse computazionali richieste possono essere elevate. Tuttavia, le prestazioni dei computer migliorano di anno in anno, e in particolar modo con la diffusione delle architetture a multi-processore. Pertanto oggigiorno è comune impiegare decine o persino centinaia di core su cluster di PC per effettuare analisi agli elementi finiti durante un’ottimizzazione. La tesi è strutturata nel seguente modo. La prima parte copre le conoscenze di base necessarie a sviluppare gli argomenti trattati nel seguito. C’è quindi un’introduzione alle macchine studiate, delle conoscenze generali sui materiali magnetici e ferromagnetici, alcuni concetti di base sull’algoritmo di ottimizzazione differential evolution (DE) utilizzato, e il disegno delle barriere fluide dei rotori di macchine a riluttanza. Nella seconda parte si sono sviluppati modelli analitici di macchine SyR e PMaSyR. Il modello completo è non lineare e può diventare abbastanza complesso da sviluppare, specialmente in un contesto industriale. Pertanto, usando alcune ipotesi semplificative, si possono derivare alcune semplici equazioni di progetto. Questo modello semplice è anche esteso e applicato a strutture di rotore asimmetriche, che tentano di compensare alcune armoniche di coppia. La terza parte si concentra sull’applicazioni di ottimizzazioni multiobiettivo accoppiate a FEA per alcuni casi di studio. In particolare, si è ottimizzato, prototipato e testato un motore SyRper pompe centrifughe. Poi, è stato condotto uno studio di fattibilità per un motore PMaSyR attraverso ottimizzazioni multi-obiettivo. Dopodiché si sono studiati motori SyRper alte velocità e si sono dedotti i limiti di potenza di questa macchina. Infine l’ottimizzazione DE multi-obiettivo è stata anche applicata per migliorare le capacità di controllo sensorless delle macchine anisotrope già in fase di progetto.

Cette thèse présente le dimensionnement d’une machine électrique à rotor disque pour la traction de véhicules hybrides. Un état de l’art complet sur les machines électriques à rotor disque permet de montrer que la machine à flux axial à rotor central est la plus adaptée. Différentes géométries du circuit magnétique sont successivement étudiées et comparées par simulations par éléments finis. Pour maximiser le rendement de la machine, les pertes d’origine électromagnétique sont étudiées. Un modèle analytique des pertes par effet de peau dans les conducteurs de cuivre est proposé et validé par éléments finis. Un second modèle analytique estime les pertes par courant induits dans les aimants permanents pour tous les points de travail de la machine se basant sur uniquement trois simulations par éléments finis permettant ainsi un gain de temps important lors du dimensionnement. Deux méthodes de dimensionnement sont comparées : le dimensionnement manuel par essais/erreurs et l’optimisation multi-objectifs. Cette dernière méthode automatise le dimensionnement et permet une optimisation plus pointue et une forte amélioration des performances. Dans ce cas, la densité de couple a par exemple été augmentée de 29%. La machine à flux axial dimensionnée est comparée avec une machine à flux radial pour véhicules hybrides. Cette comparaison révèle le potentiel de la machine à flux axial avec notamment une densité de couple augmentée de plus de 20%. Enfin, deux prototypes ont été construits et mesurés pour valider les simulations
This PhD presents the design of a disc rotor electric machine for the traction of hybrid vehicles. A complete state of the art enables the selection of the internal rotor axial flux machine which is the most suited to this application. Different geometries of the magnetic circuit are successively studied and compared with finite elements simulations. To maximize the efficiency of the machine losses generated in the magnetic circuit are studied. An analytic model on the skin effect in the copper conductors is proposed and validated with finite elements simulations. A second analytic model estimates the eddy current losses in the permanent magnets for every operating point of the machine based on only three finite element simulations enabling an important time-saving. Two dimensioning methods are compared: the manual dimensioning based on a tries/errors method and the multi-objectives optimization. This last method automates the dimensioning and enables a more refined optimization and a strong improvement of the performances. For example, the torque density has been improved by 29% in that case. The designed axial flux machine is compared to a state of the art radial flux machine for hybrid vehicle. This comparison shows the potential of the axial flux machine with an improvement of the torque density by more than 20%. Finally two prototypes have been built and measured to validate the simulations

This thesis deals with the analysis, design and test of three-phase high efficiency electrical motors, with particular reference to motors with a rotor winding. At first, the background and the motivations of this work are described. The bibliography on the subjects is deeply examined and a selection of the most relevant papers can be found in the reference. In this scenario, the main objective of this thesis are illustrated. The Line-Start (LS) Synchronous Machine (SyM) design is a subject under investigation since the beginning of the last century, when solid state power converters was not available to drive SyMs. The LS SyM diffusion was limited by the intrinsic difficulties in its design and by the availability of the cheaper and more robust Induction Machine (IM). The working principle of IM and LS SyM are briefly described, as well as the state of the art of the techniques of analysis. Recently, there is a renewed interest on LS SyMs due to the new efficiency requirements and fast analysis techniques are required for the LS SyM design. A Finite-Element (FE) aided analytical model is developed to simulate the LS SyM dynamic. The aim is to develop a model that gives reliable solutions with limited computational efforts compared with other analysis techniques. With this procedure, the LS SyM rotor parameters can be quickly calibrated to fulfill the dynamic load requirements. An innovative analysis technique of LS SyM steady state condition is described. Such an analysis is carried out in the same reference frame used for classical SyMs. It is shown that the analysis can be used to optimize some machine parameters. The issues in LS SyM manufacturing are introduced, with particular reference to the die casting process. The possibility to apply the recent improvements in the SRM design to LS SyM is discussed from the manufacturing point of view. Stochastic optimization has been adopted for the design of electrical motors to reduce the torque ripple, increase the average torque and reduce the losses. The LS SyM torque ripple reduction, achieving at the same time a high average torque, is an important issue even though this topic is not treated extensively in the literature for LS SyM. For this reason, a stochastic optimization is considered in this thesis for the design of a new LS SyM lamination. The analysis is applied on a small size, 2-pole, three-phase LS SyM as this category is still not found in the motor market. The optimization is carried out considering the necessity to achieve a robust design, suitable for the industrial production, as such a LS SyM must be competitive with the workhorse of electrical motors, the IM. One of the most promising design is prototyped. Its performance are compared with the corresponding IM. To demonstrate the feasibility in adopting LS SyM in the large-scale production, an innovative LS SyM design is proposed. The main aim is to use the same lamination for motors of different number of poles so as to reduce the manufacturing cost. A tradeoff between contrasting aspects is necessary in the design step. The performance achievable by these rotor structures are quantified. An analytical model that describes the mutual interaction between coupled electrical circuits in machines with complex rotor structure is developed. Such a model is useful to analyze the parasitic torques in the torque characteristic of motors with rotor cage such as IM and LS SyM. The literature reveals that this topic has been discussed extensively for IM. As regards LS SyM, there is a lack of theoretical studies regarding harmonic phenomena due to the complex machine structure. This part of the thesis aims to fill this gap. The high and unstable cost of rare-earth PMs, together with the advances in solid-state control technology, leads designers to reconsider IM for variable speed drive (VSD) applications. To the aim of making the IM suitable for the full-speed sensorless control, a particular cage design is considered. An intentionally created saliency is introduced in the rotor so as to allow the rotor position to be estimated by means of a high frequency (HF) injected signal in the stator winding also at zero-speed. Different experimental tests are carried out on IMs with asymmetrical rotor cage to validate the analysis techniques and quantify the achievable performance. As far as the HF signal injection sensorless technique is concerned, the cross-saturation differential inductance of SyMs represents an issue. It causes a rotor position estimation error, reducing the region in which such technique is effective. The proper-ties of the cross-saturation inductance are deeply discussed. It is originally shown that the cross-saturation inductance depends from certain machine parameters. With such an analysis, a designer can consider the effect of the cross-saturation inductance in any model-based control algorithm. A rotor winding is added in Surface-mounted permanent-magnet machine (SPM) to create a HF anisotropy that is useful to detect the rotor position by means of a HF signal injection. Such a configuration is called ”ringed-pole”. In literature, this technique has been used on small-size machines. In certain configuration, the presence of the additional rotor winding causes significant rotor losses. This part of the thesis studies the rotor losses in ringed pole machines by means of FE analysis and analytical models. The aim is to investigate if the ringed-pole technique can be adopted also for large machines from the point of view of additional losses. With few exceptions, the work described in this thesis is always supported by means of experimental measurements. Dedicated experiments has been designed. Their results are compared with those achieved with analytical models or FE analysis.
Questo lavoro di tesi è incentrato sull’analisi, la progettazione e la prototipazione di macchine elettriche trifase ad alto rendimento, con particolare riferimento a motori dotati di avvolgimenti rotorici. Inizialmente si descrivono le motivazioni di questo lavori di tesi e il contesto in cui essa si inserisce, illustrandone i principali obiettivi. Una dettagliata analisi bibliografica è alla base del lavoro svolto. Una selezione di questi lavori si trova nelle referenze. I motori sincroni autoavvianti (LS SyM) sono stati introdotti nella prima metà del novecento e la loro progettazione è soggetto di ricerca sin da allora. Essi non si sono mai affermati a causa della loro difficile progettazione e per la disponibilità del più robusto ed economico motore ad induzione (IM). Dopo aver descritto il principio di funzionamento di IM e LS SyM, se ne illustrano le tecniche di analisi sviluppate fino al giorno d’oggi. Negli ultimi anni vi è un rinnovato interesse verso i LS SyM grazie agli stringenti requisiti di rendimento. Vi è quindi la necessità di tecniche di progettazione veloci ed affidabili per LS SyM. I risultati di simulazioni agli elementi finiti sono stati combinati a modelli analitici per descrivere la complessa dinamica di LS SyM. L’obiettivo è quello di ottenere una risposta sufficientemente precisa in tempi molto più brevi rispetto ad altre tecniche di analisi. In questo modo si rende possibile una rapida e precisa calibrazione dei parametri rotorici necessari per soddisfare determinati requisiti di carico dinamico. Parte di questa tesi è dedicata allo sviluppo di una tecnica di analisi per LS SyM in condizioni di regime. Tale analisi `e condotta nello stesso sistema di riferimento usato nei classici modelli per macchine sincrone non autoavvianti. Si mostra che l’analisi proposta permette anche di ottimizzare alcuni parametri di macchina. Negli ultimi anni vi sono stati numerosi sviluppi nella progettazione di macchine sincrone a riluttanza, con o senza l’assistenza di magneti permanenti. In questa tesi si è voluto investigare sulla possibilità di applicare tali sviluppi ai LS SyM, tenendo in considerazione i vincoli costruttivi legati alla presenza della gabbia rotorica. Lo scopo è quello di ridurre il volume di magneti permanenti utilizzati per contenere i costi di produzione. Si è affrontato il problema dell’industrializzazione dei LS SyM, con particolare riferimento al processo di pressofusione del rotore. Nell’intento di ridurre il ripple di coppia, incrementare la coppia media e ridurre le perdite dei motori elettrici, recenti lavori propongono l’utilizzo di algoritmi di ottimizzazione stocastica nella fase di progettazione. I suddetti obiettivi sono basilari anche per LS SyM, anche se per questo tipo di motori la letteratura è meno fornita. Per questo motivo si è voluto utilizzare un algoritmo di ottimizzazione nella fase di progettazione della lamiera di un LS SyM. L’analisi è applicata ad un LS SyM trifase a 2 poli di piccola taglia, dato che ancora non si trovano nei cataloghi dei principali costruttori. L’ottimizzazione è sviluppata considerando la necessità di ottenere un progetto robusto e comunque adatto alla produzione industriale, dato che tale LS SyM deve essere competitivo con l’ormai consolidato IM. Una promettente struttura rotorica è stata prototipata. Le prestazioni ottenute sono confrontate con quelle del corrispondente IM. Si è proposta un innovativa configurazione di LS SyM per dimostrare la fattibilità del loro utilizzo su scala industriale. Lo scopo è quello di utilizzare la stessa lamiera per motori con un diverso numero di poli, riducendo di conseguenza il costo di produzione. Per fare ciò è necessario un compromesso tra aspetti contrastanti nel progetto. In questa parte di tesi, si è voluto quantificare le prestazioni ottenibili da tali geometrie nelle diverse configurazioni. In questa tesi si è sviluppato un modello analitico per caratterizzare l’interazione di circuiti elettrici accoppiati in strutture complesse quali quelle dei LS SyM. Questa analisi mira ad essere uno strumento per la determinazione analitica delle coppie parassite in motori dotati di gabbia rotorica come LS SyM e IM. La letteratura riporta un gran numero di lavori riguardanti la descrizione di coppie parassite nella caratteristica di coppia di motori IM. In LS SyM, l’analisi delle coppie parassite è molto più complessa a causa della struttura di macchina. In letteratura, gli studi analitici riguardanti gli effetti di armoniche di MMF in motori LS SyM sono pochi ed incompleti. L’elevato ed instabile prezzo dei magneti permanenti, assieme allo straordinario sviluppo dell’elettronica allo stato solido, ha spinto a riconsiderare il motore ad induzione per applicazioni a velocità variabile. In questo scenario, si è considerato un avvolgimento rotorico a gabbia di scoiattolo in cui i conduttori sono asimmetrici. Tale asimmetria permette il riconoscimento sensorless della posizione rotorica tramite iniezione di segnali ad alta frequenza negli avvolgimenti di statore anche a velocità molto basse. Sono stati condotti test sperimentali su prototipi di IM con gabbia asimmetrica allo scopo di verificare le tecniche di analisi e di quantificare le prestazioni ottenibili da tali geometrie. Proseguendo l’analisi delle problematiche riscontrate in controlli di tipo sensorless con iniezione di segnale, si sono approfondite le proprietà della mutua induttanza differenziale causata dal fenomeno della saturazione incrociata tra asse d e q in macchine sincrone. Essa causa un errore nella stima della posizione rotorica, riducendo di fatto l’applicabilità del controllo sensorless con iniezione di segnale. Dopo aver discusso in dettaglio le propriet`a di tale induttanza, si `e dimostrato che essa dipende da alcuni parametri di macchina. Con i risultati ottenuti, può essere intrapresa una serie di accorgimenti nel controllo della macchina volta a mitigare l’effetto negativo dell’induttanza mutua dovuta alla saturazione incrociata. Uno o più avvolgimenti rotorici possono essere introdotti anche in motori sincroni a magneti permanenti superficiali, allo scopo di estendere l’applicabilità del controllo sensorless con iniezione di segnale anche a questo tipo di motori. In questo tipo di macchine, denominate ”ringed-pole”, tali avvolgimenti rotorici possono essere sede di perdite importanti nel funzionamento a regime. In letteratura, questa tecnologia è stata applicata a motori di piccola taglia. In questo contesto, si sono studiate le perdite rotoriche di macchine ”ringed-pole” tramite analisi agli elementi finiti e modelli analitici. Lo scopo è quello di verificare se l’uso di tale tecnologia può essere esteso a macchine di taglia superiore dal punto di vista delle perdite rotoriche. Con poche eccezioni, gli argomenti di questa tesi sono validati tramite misure sperimentali. I risultati delle prove sperimentali sono confrontati con quelli provenienti da modelli analitici o da analisi agli elementi finiti.

Axial flux permanent magnet (AFPM) machines have recently attracted significant attention due to several reasons, such as their specific form factor, potentially higher torque density and lower losses, feasibility of increasing the number of poles, and facilitating innovative machine structures for emerging applications. One such machine design, which has promising, high efficiency particularly at higher speeds, is of the coreless AFPM type and has been studied in the dissertation together with more conventional AFPM topologies that employ a ferromagnetic core. A challenge in designing coreless AFPM machines is estimating the eddy current losses. This work proposes a new hybrid analytical and numerical finite element (FE) based method for calculating ac eddy current losses in windings and demonstrates its applicability for axial flux electric machines. The method takes into account 3D field effects in order to achieve accurate results and yet greatly reduce computational efforts. It is also shown that hybrid methods based on 2D FE models, which require semi-empirical correction factors, may over-estimate the eddy current losses. The new 3D FE-based method is advantageous as it employs minimum simplifications and considers the end turns in the eddy current path, the magnetic flux density variation along the effective length of coils, and the field fringing and leakage, which ultimately increases the accuracy of simulations. After exemplifying the practice and benefits of employing a combined design of experiments and response surface methodology for the comparative design of coreless and conventional AFPM machines with cores, an innovative approach is proposed for integrated design, prototyping, and testing efforts. It is shown that extensive sensitivity analysis can be utilized to systematically study the manufacturing tolerances and identify whether the causes for out of specification performance are detectable. The electromagnetic flux path in AFPM machines is substantially 3D and cannot be satisfactorily analyzed through simplified 2D simulations, requiring laborious 3D models for performance prediction. The use of computationally expensive 3D models becomes even more challenging for optimal design studies, in which case, thousands of candidate design evaluations are required, making the conventional approaches impractical. In this dissertation a new two-level surrogate assisted differential evolution multi-objective optimization algorithm (SAMODE) is developed in order to optimally and accurately design the electric machine with a minimum number of expensive 3D design evaluations. The developed surrogate assisted optimization algorithm is used to comparatively and systematically design several AFPM machines. The studies include exploring the effects of pole count on the machine performance and cost limits, and the systematic comparison of optimally designed single-sided and double-sided AFPM machines. For the case studies, the new optimization algorithm reduced the required number of FEA design evaluations from thousands to less than two hundred. The new methods, developed and presented in the dissertation, maybe directly applicable or extended to a wide class of electrical machines and in particular to those of the PM-excited synchronous type. The benefits of the new eddy current loss calculation and of the optimization method are mostly relevant and significant for electrical machines with a rather complicated magnetic flux path, such is the case of axial flux and of transvers flux topologies, which are a main subject of current research in the field worldwide.

La machine à induction est considérée comme l'une des machines électriques les plus utilisées dans les applications industrielles en raison de sa structure ferme, de sa robustesse, de sa simplicité, de son coût raisonnable, de sa longue durée de vie et de sa grande fiabilité. La machine à induction multiphasée, en particulier une machine à six phases, est une solution judicieuse pour augmenter la fiabilité par rapport à une machine à induction triphasée pour des applications telles que les véhicules électriques. L'augmentation du nombre de phases sur la machine peut entraîner une baisse de capacité en cas de perte d'une ou plusieurs phases. En outre, la configuration multiphasée offre d'autres avantages, tels qu'une capacité de gestion de puissance élevée en divisant la puissance requise entre les phases, des pulsations de couple réduites, des pertes de cuivre de stator réduites et des courants harmoniques de rotor réduits. Une autre solution pour améliorer les performances de la machine à induction consiste à utiliser un enroulement concentré car il présente de nombreux avantages tels que l'extrémité des spires plus courte sans chevauchement, une capacité de tolérance de panne plus élevée et une procédure de fabrication plus facile. Cependant, le principal inconvénient du bobinage concentré est la densité de flux d’entrefer de faible qualité. Compte tenu de l'inconvénient mentionné ci-dessus, l'application d'un enroulement concentré sur des moteurs à induction détériore considérablement leurs performances, ce qui rend inefficaces les avantages du bobinage concentré. Dans ce travail de recherche, l'objectif principal est d'améliorer les performances des moteurs asynchrones à six phases à enroulements concentrés. À cette fin, un nouveau schéma d'enroulement concentré réalisable, à savoir un enroulement pseudo-concentré, est proposé en utilisant une analyse de la fonction d'enroulement, qui a un harmonique fondamental plus élevé de la densité de flux d'entrefer par rapport aux enroulements concentrés conventionnels. L'utilisation de l'enroulement pseudo-concentré au lieu de l'enroulement conventionnel améliore considérablement les paramètres de performance du moteur à induction à six phases tels que la puissance de sortie, le rendement et le facteur de puissance. Pour améliorer encore les performances du moteur, un nouveau moteur à induction à six phases optimal équipé de l'enroulement pseudo-concentré est conçu et fabriqué en tenant compte des caractéristiques d'enroulement telles que l'extrémité d'enroulement plus courte et l'harmonique fondamental inférieur de la densité de flux d'entrefer par rapport à la distribution de l’enroulement conventionnel. Les résultats de performance du moteur montrent que le moteur à induction à six phases équipé de l'enroulement pseudo-concentré peut avoir des paramètres de performance élevés même par rapport au moteur équipé de l'enroulement distribué. Compte tenu des faits précités, il semble que l'utilisation d'un moteur à induction équipé du bobinage pseudo-concentré puisse être une solution appropriée pour une application de véhicule électrique. Pour en savoir plus, un moteur à induction à six phases à rotor extérieur équipé du bobinage pseudo-concentré est conçu et optimisé pour les véhicules électriques légers de faible puissance. Comme prévu, le moteur à induction à six phases à rotor externe optimal fabriqué fournit des performances acceptables pour une application dans la roue
Electrical Machine is an important research area in electrical engineering with numerous industrial applications. Induction machine is considered as one of the most employed electrical machines in industrial applications because of its firm structure, robustness, simplicity, reasonable cost, long lifetime, and high reliability. Multiphase induction machine, especially the six-phase one, is a sensible solution to increasing reliability compared to a three-phase induction machine for applications such as an electric vehicle. Increasing the number of phases on the machine can result in the ride through capability in case of loss of one or more phases. Furthermore, the multiphase configuration provides other advantages, such as high power-handling capability by dividing the required power between phases, reduced torque pulsations, reduced stator copper losses, and reduced rotor harmonic currents. Another solution in the way of improving induction machine performance is using concentrated winding because it has many advantages such as shorter non-overlapping end turns, higher fault-tolerant capability, and easier manufacturing procedure. The main drawback of concentrated winding, however, is its low-quality air gap flux density, which usually limits the winding application to permanent magnet synchronous machines. Considering the aforementioned drawback, applying concentrated winding to induction motors deteriorates their performances significantly, which renders the concentrated winding advantages ineffective. In this research work, the main goal is to improve the performance of the six-phase induction motors with concentrated windings. For this purpose, a new feasible concentrated winding layout, namely pseudo-concentrated winding, is proposed using winding function analysis, which has a higher fundamental harmonic of the air gap flux density compared to the conventional concentrated windings. Using the pseudo-concentrated winding instead of the conventional one enhances the performance parameters of the six-phase induction motor such as output power, efficiency, and power factor significantly. To enhance the motor performance, even more, a new optimum six-phase induction motor equipped with the pseudo-concentrated winding is designed and fabricated considering the winding features such as shorter end winding and lower fundamental harmonic of air gap flux density compared to the conventional distributed winding. The performance results of the motor show that the six-phase induction motor equipped with the pseudo-concentrated winding could have high-performance parameters even compared to the motor equipped with the distributed winding. Considering the aforementioned facts, it seems that using an induction motor equipped with the pseudo-concentrated winding can be an appropriate solution for electric vehicle application. To investigate more, an outer rotor six-phase induction motor equipped with the pseudo-concentrated winding is designed and optimized for light-duty low-power electric vehicles. As expected, the fabricated optimum outer rotor six-phase induction motor provides acceptable performance for the in-wheel application. In this research work, a novel pseudo-concentrated winding is introduced with better magnetic characteristics compared to the conventional concentrated windings, which allows applying it to the induction motors without having a noticeable performance drop compared to the distributed winding. The pseudo-concentrated winding has a lower air gap flux density amplitude compared to the distributed winding; therefore, for using this winding layout in induction motor application, a new stator lamination should be designed considering the magnetic characteristics of the pseudo-concentrated winding to achieve acceptable performance parameters

L'objectif de ce travail de thèse est de définir et de présenter une nouvelle approche de la conception sous contraintes de machines électriques. Celle-ci permet aux électrotechniciens d'utiliser l'ordinateur non pas seulement comme un outil d'analyse mais aussi comme un outil ayant de réelles facultés de dimensionnement automatique. Cette approche offre de grands avantages comme: -la gestion des contraintes d'un cahier des charges, -la gestion de l'interdépendance des phénomènes physiques intervenant dans une machine, -ou encore l'optimisation des solutions trouvées. Cependant sa grande originalité est qu'elle assure automatiquement la génération du logiciel de conception. Pour cela elle utilise comme connaissance de base, un modèle analytique de la machine à laquelle on s'intéresse, et comme moyen, des techniques de calcul symbolique et de programmation automatique. Afin de guider le processus de conception, elle emploie des algorithmes génériques d'optimisation numérique sous contraintes. Cette approche est notamment appliquée, dans ce mémoire, au dimensionnement de machines asynchrones
The aim of this work is to define and to present a new approach to the design with constraints of electrical devices. It gives to electrical engineers and searchers the opportunity to use computers not only as analysis tools but also as tools that have real capabilities to make automatic designs. This approach offers great advantages like the ability: -to take into account constraints on the specifications of the device, -to take into account the interdependence of all the physical phenomenon that occur, - to find optimal solutions. However its main originality is that it automatically generates the design software by using techniques of symbolic calculation and automatic programming. For this an analytical model of the device is used as basic knowledge. For the design process, constrained optimisation algorithms are used. In this work, the approach is particularly used for the design of induction machines

The interest for Permanent Magnet (PM) synchronous machines has increased in the recent years. This is mainly due to important developments in the field of the electrical machine design, control systems and PM technology. These factors enable the possibility of designing electrical machines with a high torque density, power density, efficiency, low cost with wide degrees of freedom. Such requirements are highly sought in the automotive field, especially following the increasingly stringent energy efficiency and air pollution emissions policies. These conditions are pushing forward the transition from the conventional internal combustion engine mobility systems to the hybrid electric and full electric vehicles for the mass market production. Moreover, further requirements, especially in high demanding traction and auxiliary machines, such as Electric Power Steering (EPS) motors, are nowadays getting more and more important. Flux weakening performance, torque quality (cogging torque and torque ripple), demagnetization, acoustic performance and robustness of the designs has to be taken into account, together with stringent dimensional, electrical and thermal constraints typical for compact integrated automotive electrical drives. Due to recent sky-rocketing price, instability, critical availability of Rare Earth (RE) metals and the related environmental issue for their sourcing, Heavy (HR) RE free, RE-free and even PM free machines are hunted for in the automotive market. The novel PM motor topologies meet such a „market pull“. The main research subject of this PhD is the development of novel analysis and design procedures of PM synchronous machines for automotive applications. The latter range from low torque / power auxiliary systems, i.e. EPS machines, to high torque / high power applications, such as traction machines. It provides a comprehensive evaluation of various synchronous machine topologies. Particular attention is devoted to the study of HRE-free solutions and the development of Finite Element Analysis (FEA) and analytical procedures for the optimization of the motor designs. As mentioned above, these methodologies take into account several electromechanical constraints and various performance targets. As regards the traction machines, the investigation is focused to the permanent magnet assisted reluctance (PMASR) topology. The interest around the PMASR motors has grown in the last years especially as consequence of the price crisis of rare earth magnets. In fact these machines represent a potential low cost replacement of conventional PM machines due to several technical advantages, such as wide speed range, competitive torque density and efficiency, high overload capability and robust structure. On the other had, the most important drawback is represented by the high torque ripple. In the area of the electric power steering motors, several topologies, both with isotropic anisotropic rotor structures and equipped with fractional slot concentrated windings, are evaluated. Some novel motor designs, which employ the reluctance and flux concentration principle, are introduced, optimized by means of FEA procedures and experimentally validated in detail, including the acoustic and PM performance. Further considerations are derived on the robustness of the solutions against the manufacturing imperfections. Comparing to the isotropic topologies, these innovative motor configurations show higher torque density, extended speed range, optimal torque quality and acoustic performance, ensuring lower active cost and structural simplicity. Therefore, they represent attractive candidates for high performance automotive applications.
Negli ultimi anni l'interesse nel settore delle macchine sincrone a magneti permanenti ha riscontrato una notevole crescita. Questo e' stato determinato principalmente da importanti sviluppi nel campo della progettazione delle macchine elettriche, nei sistemi di controllo e nelle tecnologie dei magneti permanenti. Questi fattori hanno reso possibile la progettazione di macchine elettriche con elevata densita' di coppia, densita' di potenza, rendimento, basso costo unitamente ad ampi gradi di liberta' nella geometria e struttura della macchina. Tali requisiti sono fortemente ricercati nel campo automotive, soprattutto a seguito dell'introduzione di sempre piu' stringenti normative sull'efficienza energetica e sulle emissioni atmosferiche inquinanti. Questi aspetti hanno accelerato la transizione da sistemi di mobilita' tradizionali con motori a combustione interna a veicoli ibridi-elettrici e puramente elettrici, nella produzione industriale di massa. Inoltre, ulteriori requisiti, specialmente in macchine da trazione ed ausiliarie ad alte prestazioni, quali ad es. motori per servosterzo elettrico, stanno attualmente diventando sempre piu' importanti. Prestazioni in deflussaggio, qualita' della coppia (coppia di impuntamento ed oscillazione della coppia), smagnetizzazione, prestazioni acustiche e robustezza delle configurazioni devono essere tenute in conto, insieme a stringenti vincoli dimensionali, elettrici e termici tipici di azionamenti elettrici compatti ed integrati. A causa dell'incremento significativo ed instabilita' dei prezzi delle terre rare, della criticita' dei loro approvvigionamenti ed i relativi impatti ambientali, soluzioni prive di terre rare pesanti, senza terre rare od addirittura senza magneti permanenti, sono fortemente ricercate nel settore automotive. La principale tematica di ricerca di questo dottorato riguarda lo sviluppo di innovative procedure di analisi e sintesi di macchine sincrone a magneti permanenti per applicazioni automotive. Queste ultime spaziano da sistemi ausiliari a bassa coppia / bassa potenza, nella fattispecie motori per servosterzo elettrico, fino a applicazioni at alta coppia / alta potenza, quali ad es. motori da trazione. Questo lavoro fornisce una valutazione complessiva di differenti tipologie di macchine. Particolare attenzione e' dedicata allo studio di soluzioni con magneti permanenti privi di terre rare pesanti ed allo sviluppo di procedure di ottimizzazione ad elementi finiti ed analitiche della geometria del motore. Come menzionato precedentemente, tali metodologie tengono in considerazione numerosi vincoli elettromeccanici e differenti obiettivi. Per quanto riguarda le macchine da trazione, la ricerca si e' focalizzata sulla topologia di motori a riluttanza assistiti da magneti permanenti. L‘interesse attorno a tali macchine e' cresciuto negli ultimi anni soprattutto a seguito della crisi dei prezzi dei magneti permanenti a terre rare. Infatti, tali macchine rappresentano una soluzione alternativa a basso costo rispetto a motori a magneti permanenti tradizionali a causa di importanti vantaggi tecnici quali ad es. un ampio regime di velocita', competitivi valori di densita' di coppia ed efficienza, elevata capacita' di sovraccarico ed una robustezza della struttura. D'altra parte, uno dei principali svantaggi e' rappresentato dall'elevata oscillazione della coppia. Nel campo dei motori per servosterzo elettrico, differenti topologie, sia dotate di rotore isotropo che anisotropo ed equipaggiate con avvolgimenti concentrati, sono state valutate. Alcune nuove configurazioni, che sfruttano il principio della riluttanza e della concentrazione di flusso, sono introdotte, ottimizzate per mezzo di procedure ad elementi finiti ed infine validate per mezzo di misure sperimentali, includendo valutazioni delle prestazioni acustiche e quelle dei magneti permanenti. Ulteriori considerazioni sono tratte per quanto concerne la robustezza delle soluzioni nei confronti delle imperfezioni costruttive. Rispetto alle macchine isotrope, tali configurazioni sviluppano una densita' di coppia piu' alta, esibiscono un piu' ampio intervallo di velocita', maggiore rendimento raggiungendo un qualita' della coppia e prestazioni acustiche competitive ad un costo inferiore e con una buona semplicita' costruttiva. Esse si dimostrano, pertanto, dei candidati ottimali per applicazioni automotive ad alte prestazioni.

Les véhicules électriques (VE) ou les véhicules électriques hybrides (VEH) offrent de nombreux avantages par rapport aux véhicules à moteur à combustion interne classiques. Selon les tendances récentes, la demande en VE(H) efficaces devrait augmenter considérablement. Pour une gamme haute puissance, la technologie des moteurs à aimants permanents a été le choix privilégié dans les véhicules électriques hybrides. La demande croissante de moteurs à haut rendement est en corrélation directe avec la demande d'aimants puissants (NdFeB ou SmCo) utilisant des terres rares. La disponibilité et la production des terres rares sont très critiques particulièrement pour les terres rares lourdes. L'objectif de cette thèse de doctorat est donc de concevoir un moteur Halbach à rotor extérieur pour une application VE(H) avec recyclage et réutilisation faciles des aimants. En outre, le projet vise à étudier et à proposer la fabrication d'un aimant Halbach utilisé dans les moteurs de forte puissance pour application VE.Tout d'abord, la fabrication d'un aimant Halbach utilisant un aimant NdFeB fritté avec et sans liant a été étudiée. L'étude montre que la fabrication d'une configuration de Halbach à l'aide d'un aimant collé est beaucoup plus facile et plus rentable que la fabrication d'un aimant fritté. La caractérisation d'un aimant NdFeB lié utilisé pour fabriquer un aimant Halbach a également été réalisée. Diverses voies de recyclage des aimants frittés et liés ont été analysées; on peut en déduire que les aimants collés sont beaucoup plus faciles à recycler, de manière rentable et respectueuse de l'environnement. La thèse propose également un moyen de recyclage pour l'aimant collé utilisé dans le moteur.Deuxièmement, un moteur à aimant Halbach collé a été conçu en modélisation éléments finis 2D et 3D. Pour obtenir un moteur très efficace et compact, on a utilisé un bobinage à pas fractionnaire. Les propriétés de l'aimant Halbach ont été calculées à l'aide du modèle éléments finis et comparées au modèle analytique. Les résultats obtenus par les deux approches étaient similaires. De plus, l'impact des combinaisons nombre d’encoches-pôles sur les pertes moteur et le couple a été étudié, en particulier les pertes Joule (compte tenu de toutes les contraintes de conception). Différentes stratégies pour utiliser des aimants recyclés à faible rémanence sont également présentées. L'utilisation d'un aimant recyclé avec une augmentation de la longueur axiale du moteur pourrait être le meilleur choix compte tenu de différents facteurs, notamment la fabrication de l'aimant Halbach. Sur la base de différentes études paramétriques, une conception du moteur a été proposée et un prototype a été construit. Il a été montré qu'un aimant Halbach de grande puissance pouvait être construit de manière économique avec un aimant NdFeB collé. La densité de flux d'entrefer du rotor, mesurée sur le prototype, est en étroite concordance avec les valeurs calculées.De plus, la méthodologie WIRE (Weighted Index of Recycling and Energy) a été présentée pour comparer différentes conceptions de moteurs en fonction de leur performance et de leur recyclabilité. La méthode développée produit deux indices basés sur-Facilité de recyclage du moteur en ce qui concerne le matériau, le montage et le démontage des aimants.-Impact d'un aimant recyclé sur la consommation d'énergie d'un moteur pendant sa durée de vie.En utilisant ces deux indices, on peut facilement analyser les avantages et les inconvénients des différentes conceptions sur la base de la recyclabilité et de l'efficacité énergétique. La conception proposée a été évaluée à l'aide cette méthode et on montre que le moteur est facile à monter et à démonter. De plus, l’assemblage moteur (sans colle) permet une extraction facile des aimants et une réutilisation directe. L'indice énergétique évalué du moteur montre l'impact de l'utilisation d'un aimant recyclé et sa viabilité pour les applications VE dans différents scénarios
Electric vehicles (EVs) or Hybrid electric vehicles (HEVs) offer many advantages over the conventional IC engine vehicles. According to recent trends, the demand for efficient (H)EVs is expected to grow significantly. For a high-power range, permanent magnet based motor technology has been the preferred choice for motors deployed in (H)EVs. Growing demand of highly efficient motors is in direct correlation to the demand of strong magnets (NdFeB or SmCo), which uses rare earth elements (REE). The availability and supply of REEs specially heavy REEs is very critical. Therefore, the aim of this doctoral thesis is to design an outer rotor Halbach motor for a (H)EV application with easy recycling and reuse of the magnet. Further, the project aims to investigate and propose the manufacturing of a Halbach magnet used in a high power motor EV applications.Firstly, the manufacturing of Halbach magnet using a sintered and a bonded NdFeB magnet was investigated. The study shows that the manufacturing of Halbach array using a bonded magnet is much easier and more cost effective than the sintered magnet. The characterisation of a bonded NdFeB magnet used for manufacturing a Halbach magnet was also performed. Various recycling routes for both sintered and bonded magnets were analysed and it can be inferred that bonded magnets are much easier to recycle in a cost effective and environment friendly manner. The thesis also proposes the recycling route for the bonded magnet used in the motor.Secondly, a motor with bonded Halbach magnet was designed using 2D and 3D FEM. To achieve a highly efficient and compact motor, fractional slot tooth coil winding was used. The properties of Halbach magnet was calculated using FEM model and benchmarked against the analytical model. The results obtained from the two approaches were in close agreement. Further, the impact of slot pole combinations on motor losses and the subsequent torque were investigated, specifically eddy loss (considering all the design constraints). Different strategies to use recycled magnet with lower remanence is also presented. It is shown that using a recycled magnet with increased axial length of the motor could be the best choice considering different factors, specially manufacturing of the Halbach magnet. Based on different parametric studies a design of the motor was proposed and prototype was built. It was demonstrated that a high power Halbach magnet could be built economically using a bonded NdFeB magnet. The airgap flux density of the rotor, measured on the prototype is in close agreement with the calculated values.Additionally, WIRE (Weighted Index of Recycling and Energy) methodology was presented to benchmark different motor designs on the basis of performance and recy- clability. The method developed produces two indices based on:• Ease of motor recyclability considering material, assembly and disassembly of magnets.• Impact of a recycled magnet on the energy consumption of a motor during its operational lifetime.Using both the above indices, one can easily analyse the pros and cons of different motor designs on the basis of recyclability and energy efficiency. The proposed motor design was evaluated using the developed method and it is shown that the motor is easy to assemble and disassemble. In addition, the motor assembly (glue free) enables easy magnet extraction and direct reuse. The evaluated energy index of the motor shows the impact of using a recycled magnet and its viability for EV applications in different scenarios

La quasi-totalité des études de machines synchrones à aimants permanents (MSAP) pour les applications aux véhicules hybrides concernent les performances uniquement sur quelques points particuliers d’un cycle de fonctionnement du véhicule (le point de base, le point à grande vitesse ou le point le plus sollicité). Cependant, ces machines électriques fonctionnent souvent à différents couples et à différentes vitesses. Cette thèse s’intéresse donc à l’étude des performances de MSAP sur l’ensemble d’un cycle de fonctionnement en vue de les optimiser sur cycle. Durant cette thèse, l’auteur a contribué à développer les modèles de couple, de défluxage, de pertes cuivre et de pertes magnétiques et les méthodes de calcul de ces pertes à vide et en charge pour les quatre MSAP dont trois machines à concentration de flux et une machine à aimants en surface du rotor et pour trois cycles de fonctionnement : NEDC, Artemis-Urbain et Artemis-Routier. Une validation expérimentale de ces modèles a été effectuée sur un banc d’essai moteur avec deux prototypes de MSAP. Ensuite, les MSAP ont été dimensionnées en vue d’une minimisation des pertes sur cycle et du courant efficace du point de base. Cette combinaison a pour but d’augmenter le rendement de la machine électrique et de minimiser la dimension de l’onduleur de tension associée. Ce problème d’optimisation multi-objectif a été réalisé en utilisant l'algorithme génétique, Non-Dominated Sorting Genetic Algorithm (NSGA-II). Ainsi, un Front de Pareto des solutions optimales peut être déduit. Les impacts des modèles de pertes (à vide et en charge) sur l’optimisation sur cycle des machines sont étudiés et l’intérêt de chaque modèle est présenté. Les modèles et méthodes de calcul proposés peuvent être appliqués à tous les cycles de fonctionnement, à différentes MSAP et à différentes applications
Almost all studies of permanent magnet synchronous machines (PMSM) for for hybrid vehicle applications relate to their performances on a specific point of a driving cycle of the vehicle (the base point, the point at high speed or the most used point). However, these machines often operate at different torques and at different speeds. This thesis studies therefore PMSM performances in order to optimize during an entire driving cycle. In this thesis, the author contributed to develop models of torque, field weakening, copper losses and iron losses and methods of calculating these losses at no-load and at load for four MSAP (three concentrated flux machine and a surface mounted PMSM) and for three driving cycles (New Eurepean Driving Cycle, Artemis-Urban and Artemis-Road). An experimental validation of these models was realized on a test bench with two prototypes of MSAP. Then, the MSAP were sized for a minimization of average power losses during the cycle and of the RMS current at the base point. This combination is designed to increase the efficiency of the electrical machine and minimize the size of the associated voltage inverter. This problem of multi-objective optimization was performed using the genetic algorithm, Non-Dominated Sorting Genetic Algorithm (NSGA-II). Thus, a Pareto front of optimal solutions can be derived. The impacts of loss models (at no-load and at load) on the PMSM optimization during the cycle are studied and the interest of each model is presented. Models and calculation methods proposed in this thesis can be applied to all cycles, at different MSAP and for other applications

The overarching goal of this work is to address the design of next-generation, high torque density electrical machines through numerical optimization using an integrated thermal-electromagnetic design tool that accounts for advanced cooling technology. A parametric thermal model of electric machines was constructed and implemented using a finite difference approach incorporating an automated, self segmenting mesh generation. A novel advanced cooling technology is proposed to improve thermal transport in the machine by removing heat directly from the windings via heat exchangers located between the winding bundles. Direct winding heat exchange (DWHX) requires high convective transport and low pressure loss. The heat transfer to pressure drop tradeoff was addressed by developing empirically derived Nusselt number and friction factor correlations for micro-hydrofoil enhanced meso-channels. The parametric thermal model, advanced cooling technique, Nusselt number and friction factor correlations were combined with a parametric electromagnetic model for electric machines. The integrated thermal-electromagnetic model was then used in conjunction with particle swarm optimization to determine optimal conceptual designs. The Nusselt number correlation achieves an R² value of 0.99 with 95% of the data falling within ± 2.5% similarly the friction factor correlation achieves an R² value of 0.92 with 95% of the data falling within ± 10.2%. The integrated thermal-electromagnetic design tool, incorporating DWHX, generated an optimized 20 kW permanent magnet electric machine design achieving a torque density of 23.2 N-m/L based on total system volume.

Minimizing cost and optimization of nonlinear problems are important for industries in order to be competitive. The need of optimization strategies provides significant benefits for companies when providing quotes for products. Accurate and easily attained estimates allow for less waste, tighter tolerances, and better productivity. The Nelder-Mead Simplex method with exterior penalty functions was employed to solve optimum machining parameters. Two case studies were presented for optimizing cost and time for a multiple tools scenario. In this study, the optimum machining parameters for milling operations were investigated. Cutting speed and feed rate are considered as the most impactful design variables across each operation. Single tool process and scalable multiple tool milling operations were studied. Various optimization methods were discussed. The Nelder-Mead Simplex method showed to be simple and fast.

The objective of this dissertation is to study the switched reluctance machine (SRM) electromagnetic design and optimization. The research of electric machines is mostly driven by the motivation for higher efficiency and lower cost. The demands for high-performance electric machines also come from the development of emerging industries, such as electric vehicles (EV), hybrid electric vehicles (HEV), renewable energy conversion, energy storage and precision manufacturing. The additional requirements for those applications include volume, weight, speed, torque, reliability, fault tolerance capability, etc. The focus of the research effort is on the high speed and high torque applications, where the SRM stands out compared to other types of machines. The conventional design method significantly depends on the designer’s experience, which uses equivalent magnetic circuit models, and therefore the SRM design is not well developed. A novel SRM electromagnetic design and optimization method is developed, which uses the current-fed FEA simulation as the SRM performance estimation tool. This method serves as the main innovation of this research work. First, the proposed method is applicable to any SRM topologies and dimension, and no detailed modeling of a specific SRM configuration is required in advance. Therefore, an automated SRM design and optimization approach is developed. Secondly, great accuracy of the SRM electromagnetic analysis, e.g. flux density, torque, and current calculation, is achieved by using FEA simulation instead of simplified magnetic circuit approximations. This contribution is particularly significant when considering the poor accuracy of conventional SRM analytical analysis methods, where several assumptions and approximations are used. Lastly, the proposed design method takes the typical SRM control strategy into account, where the excitation current profile is characterized as a trapezoid. This method adapts the flux linkage of the first FEA simulation result to specify the excitation current profile for the second FEA simulation, so the calculated SRM performance in FEA simulation agrees with the measurement on a practical machine. The proposed SRM design and optimization method is used for a 12/8 SRM rotor design and for a complete 4/2 SRM design. These design examples validate the applicability of the proposed method to different SRM configurations and dimensions. Detailed design steps are presented for both design cases, and the selection of the parametric design variables are also discussed. The optimization results are demonstrated using multi-dimension diagrams, where the optimal design with the highest torque can be easily identified. The FEA simulation results are compared to the experimental results of a fabricated SRM prototype, and good agreement is found. In addition, a new rotor configuration with a flux bridge is proposed for an ultra high speed SRM design. The primary motivation of this rotor topology is to reduce the windgae losses and the acoustic noise at a high speed of 50,000 rpm. However, care must be taken for the flux bridge design, and the impact of different flux bridge thicknesses to the SRM performance is studied. Meanwhile, the manufacturing difficulties and the mechanical stresses should also be considered when fabricating the flux-bridge rotor. As a result, two SRM prototypes are built, and the two rotors are one without a flux bridge and one with a flux bridge. The prototypes are tested at different speeds (10,000 rpm, 20,000 rpm and 50,000 rpm) respectively, and the experimental results show good agreement with the FEA simulation results.

This paper is a report on master thesis project conducted in cooperation with SMARTMOTOR AS and NTNU. The research for electrical downhole drilling machine has developed for decades. Permanent Magnet is a new addition in this arena. SMARTMOTOR AS built a downhole drilling machine which is needed to compare with machines from other competitors of the market. This machine has 10% cogging torque over its rated torque. The reason of this cogging torque and possible solutions has been investigated throughout this project. And a new geometry and design of the rotor (with 10 poles) is suggested for the machine in this report. Measurement shows the newer design is less sensitivity to eccentricity. at nominal load and speed. Lastly, flux weakening and magnet working point of the machine also investigated. In this report it is concluded that out of 3 experimental designs, the design with less epoxy and more magnet offer better performance regarding voltage waveform and vibrations. It is superior in terms of torque per weight. The other designs however allow easier manufacturing, better efficiency and shorter length of machine. It It is therefore concluded that the last design with less epoxy can be a suitable alternative rotor design for the built up machine which can reduce the existing cogging torque upto 83.4%. The end part of the paper describes a template-style of a generic electromagnetic modeling tool for the analysis and optimization of Electrical Machines. Further accurate virtual prototypes can then be produced to help designers provide answers on the performance of specific machine designs rapidly. And further investigations to identify the design characteristics of the perfect machine. A two and three-dimensional FEA model for a generator and motor can be created in minutes, investigations to identify the design characteristics of the perfect machine. Optimization tool assists designers to find the 'best' solution automatically. The tool is structured to allow creation and analysis of customized geometries, including special proprietary features. A new Matlab-based script is developed in order to handle the optimization process of electrical machine which will be used in the finite element analysis. The script saves the designer’s valuable time. It is concluded that ‘fmincon-constrained nonlinear minimization’ method in new optimization tool of MATLAB is more accurate and less time consuming compared to the method for optimization. The report explains a modern design procedure which uses both analytical and numerical analysis. The numerical analysis uses finite element analysis that is performed in Comsol mostly.

Machine learning techniques promise to greatly accelerate structural design and optimization. In this thesis, deep learning and active learning techniques are applied to different non-convex structural optimization problems. Finite Element Analysis (FEA) based standard optimization methods for aircraft panels with bio-inspired curvilinear stiffeners are computationally expensive. The main reason for employing many of these standard optimization methods is the ease of their integration with FEA. However, each optimization requires multiple computationally expensive FEA evaluations, making their use impractical at times. To accelerate optimization, the use of Deep Neural Networks (DNNs) is proposed to approximate the FEA buckling response. The results show that DNNs obtained an accuracy of 95% for evaluating the buckling load. The DNN accelerated the optimization by a factor of nearly 200. The presented work demonstrates the potential of DNN-based machine learning algorithms for accelerating the optimization of bio-inspired curvilinearly stiffened panels. But, the approach could have disadvantages for being only specific to similar structural design problems, and requiring large datasets for DNNs training. An adaptive machine learning technique called active learning is used in this thesis to accelerate the evolutionary optimization of complex structures. The active learner helps the Genetic Algorithms (GA) by predicting if the possible design is going to satisfy the required constraints or not. The approach does not need a trained surrogate model prior to the optimization. The active learner adaptively improve its own accuracy during the optimization for saving the required number of FEA evaluations. The results show that the approach has the potential to reduce the total required FEA evaluations by more than 50%. Lastly, the machine learning is used to make recommendations for modeling choices while analyzing a structure using FEA. The decisions about the selection of appropriate modeling techniques are usually based on an analyst's judgement based upon their knowledge and intuition from past experience. The machine learning-based approach provides recommendations within seconds, thus, saving significant computational resources for making accurate design choices.
Doctor of Philosophy
This thesis presents an innovative application of artificial intelligence (AI) techniques for designing aircraft structures. An important objective for the aerospace industry is to design robust and fuel-efficient aerospace structures. The state of the art research in the literature shows that the structure of aircraft in future could mimic organic cellular structure. However, the design of these new panels with arbitrary structures is computationally expensive. For instance, applying standard optimization methods currently being applied to aerospace structures to design an aircraft, can take anywhere from a few days to months. The presented research demonstrates the potential of AI for accelerating the optimization of an aircraft structures. This will provide an efficient way for aircraft designers to design futuristic fuel-efficient aircraft which will have positive impact on the environment and the world.

Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (page 83).
Tacit is a software tool for designing and solving 2D truss structures. The Tacit web application expands these features to allow users to design truss structures with integrated analysis and optimization tools. Tacit was intended for single user inter- action, but much of engineering and design work is team-based and collaborative. To examine the effects of collaboration on design optimization problems, it is crucial for users to be able to easily share and view the structure designs of their teammate. This thesis presents a new version of Tacit with these collaboration features along with a new method for assessing the similarity of pairs of structures created by collaborators, and the results and analysis of a user study conducted where participants collaborated on design optimization problems with these new functionalities. Results showed that collaboration and the collaborative software tools were both effective in improving performance. The new software functionalities improved the efficiency of collaboration but unintentionally reduced the amount of physical collaboration.
by Eunice Lin.
M. Eng.

This master‘s thesis deals with problems of optimization of cooling in electrical machines. This work includes introduction to the theory of mathematical optimization and brief introduction to optimization problems. This thesis deals with using of program Ansys Workbench for the thermal analysis in electrical machines and optimization of their cooling. Thesis contains thermal analyse of specific asynchronous motor with the squirrel cage and optimization of its cooling.

Cette thèse traite de la modélisation, conception, optimisation ainsi que de la validation expérimentale d'une machine à reluctance variable à double alimentation (BDFRM-Brushless Doubly-Fed Reluctance Machine) destinée aux éoliennes. La BDFRM est notamment considérée comme une alternative viable à la Machine Asynchrone à Double Alimentation (MASDA) dans les systèmes éoliens à vitesse variable. Elle maintient les avantages de coût de la MASDA tout en permettant l'utilisation d'un convertisseur de puissance réduit ainsi que la diminution des coûts d'entretien en raison de son fonctionnement sans balais. Une revue de la littérature fait apparaitre un manque de recherches concernant la définition de procédures de conception pour rendre cette machine plus populaire en général, et dans l'éolien en particulier. L'objectif principal de cette thèse est de contribuer à la maîtrise du processus de conception optimale de la BDFRM en proposant une approche méthodologique basée sur différents niveaux de modélisation et sur l'optimisation. Elle examine comment l'optimisation pourrait être appliquée à toutes les étapes de développement avec des objectifs distincts à évaluer. Plus précisément, elle se focalise sur la définition du problème d'optimisation sous contraintes et sur sa solution itérative en utilisant un algorithme déterministe couplé à des modèles semi-analytiques de différents niveaux. Les activités effectuées au cours de cette thèse peuvent être divisées en cinq parties principales. La première se réfère à l'étude de la BDFRM et de son principe de fonctionnement dans le contexte de l'énergie éolienne. La seconde partie examine les aspects de modélisation électromagnétique de la BDFRM en utilisant différentes approches. Deux modèles orientés pour l'optimisation ont été développés: le Modèle Semi-Analytique (SAM-Semi Analytical Model ) et le modèle multistatique de réseau de reluctances (MSRN-Multi-Static Reluctance Network). La mise en œuvre des modèles axés sur l'optimisation déterministe et leurs vérifications par des simulations utilisant la méthode des éléments finis (MEF) constituent la troisième partie. Il est possible de conclure qu'à partir des résultats de simulation que le SAM a un niveau de précision limité et qu'il est alors recommandé de l'utiliser dans les étapes de prédimensionnement, où le concepteur est plus intéressé par l'acquisition de résultats avec des temps de calcul rapides que par l'obtention d'une plus grande précision. Le MSRN, au contraire, présente des résultats d'une précision remarquable par rapport à la MEF, ce qui donne un compromis très intéressant entre précision et temps de calcul. Cette thèse a permis aussi de spécifier et réaliser un prototype de BDFRM en utilisant une approche d'optimisation présenté en quatrième partie. Ensuite, les données expérimentales obtenues à partir du prototype ont été confrontées aux résultats de la simulation pour valider les modèles, mettant l'accent sur le processus de modulation de flux par le rotor à réluctance, en particulier l'inductance mutuelle entre les enroulements. Bien que les résultats soient dans un sens satisfaisant pour la validation des modèles, il y a des différences qui ont exigés un examen plus approfondie. Une discussion sur les hypothèses les plus probables a donc été effectuée, celle-ci a souligné le rôle important du processus de fabrication de la machine sur ses performances. La cinquième partie explore à travers une étude de cas l'utilisation de la procédure de conception de la BDFRM proposée dans cette thèse pour les éoliennes. En conclusion générale, on peut affirmer que la BDFRM est potentiellement une bonne candidate pour être utilisée dans les systèmes éoliens. Toutefois, les aspects techniques et économiques sur ce choix doivent être encore évalués, en analysant et en comparant la solution globale du système dans le même cadre de recherche avec d'autres solutions alternatives
This thesis addresses the modeling, design and optimization with experimental validation of the Brushless Doubly-Fed Reluctance Machine (BDFRM) for wind power systems. The BDFRM is being considered as a viable alternative to the Doubly Fed Induction Machine (DFIG) in variable speed wind energy conversion systems. It keeps the cost advantages of the DFIG by allowing the use of a fractionally rated power converter and it has the advantage of reduced maintenance costs due to its brushless operation. A literature review shows that there is still a lack of researches to define a design procedure to make this machine widely used in general and in such application in particular. The main goal of this thesis is to contribute on mastering the BDFRM optimized design by proposing a methodological approach based on different modeling levels and on optimization. It discusses how optimization could be applied in all development stages with distinct objectives to be assessed. More precisely, it draws its attention on setting the optimization problem and on the iterative solution of a constrained inputs/outputs problem by using a deterministic algorithm coupled to analytical-based modeling levels. The activities performed during this thesis can be divided in five main topics. The first refers to the study of the BDFRM and its operating principles in the context of wind power. The second discusses the BDFRM electromagnetic modeling aspects using different approaches. Two optimization-oriented models have been developed: the Semi-Analytical Model (SAM) and the Multi-Static Reluctance Network model (MSRN). The implementation of the models focusing on deterministic optimization and their verification through simulations using Finite Element Analysis (FEA) are considered the third topic. It can be concluded from the simulation results that the SAM has a limited accuracy level and it is recommended to be used in early design stages, where the designer is most interested in fast computation times to test many design variation than in obtaining the results with the highest possible accuracy. The MSRN, on the contrary, presents remarkably precise results when compared to FEA, yielding a very interesting trade-off among accuracy and computation time. This thesis has also allowed to specify and realize a BDFRM prototype using an optimization approach, presented in the fourth part. Then, the experimental data obtained from the prototype has been confronted to the simulation results to validate the models, focusing on the investigation of the flux modulation process by the reluctance rotor, especially the mutual inductance among the windings. Although the results were in a sense satisfactory to validate the models, there have been differences that demanded further investigation. A discussion on the most likely hypothesis for that has been performed, indicating the significant role of the manufacturing process on machine performance. The fifth topic explores through a case study the use of the proposed BDFRM design procedure for wind power applications. As a general conclusion, it can be stated that the BDFRM is potentially a good candidate to be used in wind power systems. However, the technical and economic aspects on this choice must be still assessed, analyzing and comparing the overall system solution of distinct topologies within the same framework
Esta tese aborda a modelagem, o projeto e a otimização, com validaçãoexperimental, de máquinas de relutância duplamente alimentadas sem escovas (BDFRM)para sistemas de geração de energia eólica.A BDFRM é considerada como uma alternativa viável para o gerador de indução duplamentealimentado (DFIG) em sistemas de geração de energia eólica com variação develocidade. Ela mantém as vantagens de custo da solução com o DFIG, permitindo autilização de um conversor de frequência de potência nominal reduzida, e tem a vantagemadicional de custos de manutenção mais baixos devido a sua operação sem escovas. Umarevisão da literatura evidencia que ainda há uma necessidade de pesquisas na área parade_nir um procedimento de projeto desta máquina para torná-la amplamente utilizada emaplicações em geral e, em particular, para geração eólica.O objetivo principal desta tese é de contribuir para o domínio de técnicas de projetootimizado para a BDFRM através da proposição de uma metodologia baseada em diferentesníveis de modelagem e em otimização. Discute-se como técnicas de otimização podem seraplicadas em todas as fases de desenvolvimento com objetivos distintos. Especi_camente,a metodologia proposta se concentra na de_nição e na solução iterativa de problemas deotimização com restrições nas saídas utilizando um algoritmo determinístico acoplado amodelos semi-analíticos de diferentes níveis.As atividades realizadas durante esta tese podem ser divididas em cinco tópicos principais.O primeiro refere-se ao estudo da BDFRM e seu princípio de funcionamento no contextode geração de energia eólica. O segundo trata dos aspectos de modelagem eletromagnética da BDFRM utilizando diferentes abordagens. Dois modelos orientados à otimizaçãoforam desenvolvidos: o modelo semi-analítico (SAM) e o modelo multi-estático de redes derelutâncias (MSRN). A implementação dos modelos com foco na otimização e a veri_caçãodeles através de simulações com o método de elementos _nitos (FEA) são consideradas aterceira parte. Pode-se concluir, a partir dos resultados de simulação, que o SAM tem umaprecisão limitada e é recomendado para ser utilizado em estágios iniciais de projeto, emque o projetista está mais interessado em cálculos rápidos para testar diversas variações deprojeto do que na obtenção de resultados com a maior precisão possível. O MSRN, ao contrário, apresenta resultados precisos quando comparado aos obtidos com o FEA, resultandonum interessante custo-benefício entre precisão e tempo de cálculo. Nesta tese, fabricou-setambém um protótipo da BDFRM, o qual foi especi_cado utilizando-se otimização e osdetalhes sobre ele são introduzidos na quarta parte. Os dados experimentais obtidos com oprotótipo foram confrontados com os resultados de simulação para validação dos modelos,focando-se na investigação do processo de modulação de _uxo pelo rotor relutância, especialmentea indutância mútua entre os enrolamentos. Embora os resultados obtidos sejamsatisfatórios para validar os modelos, encontraram-se diferenças que exigiram uma investigação mais detalhada. As hipóteses mais prováveis foram investigadas e as conclusõesindicam o papel determinante do processo de fabricação no desempenho da máquina. Oquinto tópico explora através de um estudo de caso a utilização do procedimento de projetoproposto da BDFRM para aplicações de geração de energia eólica.Como conclusão geral, pode-se a_rmar que a BDFRM é potencialmente uma boa candidatapara ser utilizada em sistemas de geração de energia eólica. Contudo, aspectostécnicos e econômicos sobre essa escolha devem ainda ser avaliados, comparando-se asdiferentes topologias existentes sob o mesmo enfoque metodológico

Les travaux de recherches menés dans cette thèse s'inscrivent dans un contexte d'électrification du secteur automobile en réponse aux préoccupations environnementales. Cette thèse concerne la modélisation et l'optimisation de machines à commutateur mécanique pour la motorisation électrique de petits véhicules mono- ou bi- place. Les machines à collecteur représentent une alternative crédible par leur compétitivité, leur robustesse et leur fiabilité principalement dues à l'absence d'électronique de puissance. Néanmoins, elles nécessitent des dispositifs de compensation et d'aide à la commutation dont l'absence peuvent dégrader leurs performances. L'alimentation par commutateur mécanique rend nécessaire la prise en compte de la dynamique électrique dans la modélisation numérique par éléments finis. C'est pourquoi, il est crucial de développer un modèle permettant un couplage fort au sens de la formulation variationnelle du problème magnétique, du circuit électrique externe et enfin du problème électrocinétique de circulation des courants à l'interface du commutateur. Dans ce contexte, l'utilisation d'une plateforme de modélisation ouverte ONELAB a abouti au développement d'un modèle original de projection des propriétés physiques et des sources sur un maillage fixe. Cette méthode permet de s'affranchir du remaillage lors de processus itératifs tels que l'optimisation géométrique ou encore la prise en compte du mouvement. Cette méthode a abouti au couplage dynamique du problème magnétique et du circuit électrique externe à la formulation électrocinétique de l'ensemble balais-collecteur ainsi qu'à des perspectives de simulations multi-physiques à cette interface. Enfin, le choix d'un algorithme d'optimisation adapté aux modèles numériques (boîte noire à fort coût d'évaluation) a permis de développer un outil de dimensionnement des machines à commutateur mécanique adapté à une grande variété de structures en réponse à un cahier des charges industriel
The research work carried out in this thesis is part of a context of vehicle electrification in response to environmental concerns. This thesis focuses on the modeling and optimization of commutator machines used in powertrains of small electric vehicles. This kind of machines represents an alternate solution due to their competitiveness, robustness and reliability mainly due to the absence of power electronics. Nevertheless, they require compensation windings and commutation poles to improve their performances. Commutator power supply makes it necessary to take into account electrical dynamics in numerical magnetic modeling by finite elements. That is why, it is crucial to develop a model allowing a strong coupling in the sense of the variational formulation, of magnetic problem, external electrical circuit and finally current flow problem at the commutator interface. In this context, the use of a modeling opensource platform ONELAB has led to the development of an original model that performs projection of physical properties and sources on a fixed mesh. This method avoids remeshing during iterative processes such as geometric optimization or movement modeling. This method has led to the dynamic coupling of the magnetic problem and the external electrical circuit to the current flow formulation at the commutator interface as well as to prospects for multi-physical simulations at this interface. Finally, the choice of an optimization algorithm adapted to the numerical models (black box with a high evaluation cost) allowed the development of a tool for the design of commutator machines adapted to many topologies in order to fulfill industrial requirements

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2000.
Includes bibliographical references (p. 258-263).
Interior permanent magnet (IPM) synchronous machines are attractive because they can achieve constant-power operation over a wide speed range with limited magnet strength requirements and reduced power electronics cost. These characteristics provide the IPM machine with advantages over alternative machine types in applications such as spindle and traction drives. An important challenge for high-performance IPM machine design is to model the magnetic saturation of the core in a manner that is accurate, flexible, and computationally fast for design optimization. A magnetically-saturable lumped parameter model (LPM) is developed for the optimized design of high-performance IPM synchronous machine drives. Using equivalent magnetic circuit analyses, the dq-frame inductances and magnet flux linkage are calculated for transversely-laminated IPM machines. The lumped parameters are employed to predict machine drive system performance for both rated-torque and constant-power operation. The results of saturable model calculations and finite element analysis (FEA) match very closely for the machine inductances, magnet flux linkage, and converted torque. Further validation is presented by comparing measurements of existing experimental machines to predictions from the saturable lumped parameter model. Agreement of measurements and predictions for the highly nonlinear saturable q-axis inductance is within 5% in the saturated excitation range. The utility of the saturable LPM is then demonstrated by developing a cost-optimized design for an automotive integrated starter/generator (ISG) that is rated at 4 to 6 kW during generating operation. This ISG machine is mounted in a direct-drive mechanical configuration on the engine crankshaft. Agreement between the saturable LPM and FEA calculations for q- and d- axis inductances and PM flux linkage are all within 5% for the entire excitation range. Results of this model have been combined with structural FEA and demagnetization studies to produce a machine design that is predicted to meet all key ISG performance requirements. For this application and the chosen cost model, it is shown that optimizing the combined machine and drive system versus optimizing only the machine reduces the overall cost prediction by 12%.
by Edward Carl Francis Lovelace.
Ph.D.

Since antiquity, steganography has existed in protecting sensitive information against unauthorized unveiling attempts. Nevertheless, digital media’s evolution, reveals that steganography has been used as a tool for activities such as terrorism or child pornography. Given this background, steganalysis arises as an antidote to steganography. Steganalysis can be divided into two main approaches: universal – also called blind – and specific. Specific methods request a previous knowledge of the steganographic technique under analysis. On the other hand, universal methods which can be widely practiced in a variety of algorithms, are more adaptable to real-world applications. Thus, it is necessary to establish even more accurate steganalysis techniques capable of detecting the hidden information coming from the use of diverse steganographic methods. Considering this, a universal steganalysis method specialized in images is proposed. The method is based on the typical steganalysis process, where feature extractors and classifiers are used. The experiments were implemented on different embedding rates and for various steganographic techniques. It turns out that the proposed method succeeds for the most part, providing dignified results on color images and promising results on gray-scale images.

The Synchronous Reluctance Motor (SynRM) has been studied. A suitable machine vector modelhas been derived. The influence of the major parameters on the motor performance has beentheoretically determined.Due to the complex rotor geometry in the SynRM, a suitable and simple combined theoretical(analytical) and finite element method has been developed to overcome the high number ofinvolved parameters by identifying some classified, meaningful, macroscopic parameters.Reducing the number of parameters effectively was one of the main goals. For this purpose,attempt has been made to find and classify different parameters and variables, based on availableliteratures and studies. Thus a literature study has been conducted to find all useful ideas andconcepts regarding the SynRM. The findings have been used to develop a simple, general, finiteelement aided and fast rotor design procedure. By this method rotor design can be suitablyachieved by related and simplified finite element sensitivity analysis.The procedure have been tested and confirmed. Then it is used to optimize a special rotor for aparticular induction machine (IM) stator. This optimization is mainly focused on the torquemaximization for a certain current. Torque ripple is also minimized to a practically acceptablevalue. The procedure can also be used to optimize the rotor geometry by considering the othermachine performance parameters as constrains.Finally full geometrical parameter sensitivity analysis is also done to investigate the influence ofthe main involved design parameters on the machine performance.Some main characteristics like magnetization inductances, power factor, efficiency, overloadcapacity, iron losses, torque and torque ripple are calculated for the final designs and in differentmachine load conditions.Effects of ribs, air gap length and number of barriers have been investigated by means of suitableFEM based method sensitivity analysis.

Flow machines are very important to industry, being widely used on various processes. Performance improvements are relevant factors and can be achieved by using optimization methods, such as topology optimization. Thus, this work aims to develop a method to design radial flow machine rotors operating on laminar regime, by implementing a topology optimization formulation based on density model. The design of a rotor involves firstly modelling the fluid flow by using the Navier-Stokes equations on a rotating reference frame and using the Finite Element Method for solving the differential equations. To determine the material distribution on the domain, a porous flow model based on the Darcy equation is employed by using an inverse permeability that interpolates between fluid and solid. In the optimization phase, it is defined a multi-objective function that aims to minimize the viscous energy dissipation, vorticity and power. The optimization problem is implemented using the FEniCS environment and the libraries dolfin-adjoint and pyIpopt. The optimized topologies are verified with the ANSYS software. The resulting topologies are post-processed and a CAD model is created. The rotors are manufactured by using a 3D printer, the complete prototype is built by coupling an electric brushless motor and an experimental characterization is performed by measuring fluid flow and pressure head given by the pumps. Experimental and computational results are compared and the improvement is verified.
Máquinas de fluxo são muito importantes para a indústria, sendo utilizadas em diversos processos. Assim, melhorias de desempenho são fatores relevantes e podem ser alcançadas com a utilização de métodos de otimização, como a otimização topológica. Este trabalho visa desenvolver uma metodologia para projetar rotores de máquinas de fluxo radiais que operam em escoamento laminar implementando-se a formulação de otimização topológica baseada no modelo de densidades. O projeto de rotores envolve, primeiramente, a modelagem do escoamento utilizando-se as equações de Navier-Stokes em um referencial rotativo e a utilização do Método de Elementos Finitos para a resolução das equações diferenciais. A distribuição de material no domínio é feita empregando-se um modelo de escoamento em meio poroso baseado nas equações de Darcy, utilizando-se a permeabilidade inversa que interpola o elemento entre sólido e fluido. Na fase de otimização é definida uma função multi-objetivo, que visa minimizar dissipação de energia viscosa, a vorticidade e a potência. O problema de otimização é implementado utilizando-se o ambiente FEniCS para a resolução do sistema de elementos finitos e as bibliotecas dolfin-adjoint e pyIpopt para o algorithmo de otimização. As topologias otimizadas são verificadas com o software ANSYS. As topologias resultantes são pós-processadas para a criação de um modelo CAD dos rotores. Os rotores são construídos utilizando-se a impressão 3D, o protótipo completo é montado acoplando-se um motor elétrico sem escovas e a caracterização experimental é feita medindo-se a vazão e o ganho de pressão dados pelas bombas. Por fim, os resultados experimentais e computacionais são comparados e uma melhoria de desempenho é observada.

Includes abstract.
Includes bibliographical references (p. 117-120).
The primary focus of this thesis is in core loss measurement and modeling techniques and their impact in machine design. In practice, steel manufacturers usually supply core loss data either at 50/60Hz, 1.5T or curves (core loss vs. flux density) at 50 and/or 60Hz. There is growing need for lamination characterization at high flux densities (2T) and high frequencies (3.2 kHz) for novel electric machine designs operating at high speeds. The core loss measurement concept is reviewed first. Two core loss measurement formulae are compared using core loss results from different testing frames and materials.

Pages 6-12 missing.
Thesis (MScEng)--University of Stellenbosch, 2005.
ENGLISH ABSTRACT: The transverse flux machine (TFM) offers the opportunity of high torque to volume ratios which makes it an excellent candidate for direct wheel drives and low speed generator applications. TFMs have a three dimensional flux path which eliminates iron laminates as a viable core material. Soft magnetic composites have been adopted in these machines due to their isotropic nature. There are three main variants of TFMs, namely, active rotor (with magnets on the rotor), passive rotor (with magnets on the stator), and reluctance (with no magnets). As a relatively recent development in electrical machines, the TFM still has many hurdles facing its adoption in industry. Some of these hurdles are high cogging torque, a difficult construction, and expensive materials. This thesis focuses on the design of a three phase 50 kW passive rotor machine. Finite element simulation is used to determine the optimal configuration, and the final machine is analysed in detail. The construction process and associated problems are also detailed. The completed machine did not perform to the desired specification, but much knowledge was gleaned about the TFM, the construction caveats, and future potential directions.
AFRIKAANSE OPSOMMING: Die transversale-vloed masjien (TFM) met sy hoe draaimoment tot volume verhouding, is 'n uitstekende kandidaat vir direkte wiel aandrywing en lae spoed generator toepassings. Die vioed pad van die masjiene is drie-dimensioneel, wat yster laminasies as kern materiaal elimineer. "Soft magnetic composites" kan gebruik word vir hierdie masjiene as gevolg van hulle isotropiese eienskappe. Daar bestaan drie hoof variante van die TFM, naamlik, die aktiewe rotor (met magnete op die rotor), passiewe rotor (met magnete op to stator), en reluktansie (sonder magnete). Die TFM is 'n redelike nuwe tipe masjien en daar is nog probleme wat opgelos moet word voordat die industrie sal begin om dit te gebruik. Van hierdie probleme is "cogging" draaimoment, 'n moeilike konstruksie en duur materiale. Die fokus van hierie tesis is op die ontwerp van 'n 50 kW drie-fase passiewe rotor masjien. Eindige element simulasie is gebruik om die optimale konfigurasie te kry, en 'n analise is gedoen op die finale masjien. Die konstruksie proses en die probleme wat daarmeer saam gaan is ook beskryf. Die prototipe masjien wat gebou is het nie aan sy oorspronklike spesifikasie voldoen nie, maar baie kennis is opgedoen oor die TFM, die konstruksie proses, en potensiele toekomende toepassings.

Les travaux de recherches menés dans cette thèse concernent la modélisation et l’optimisation de l’alternateur synchrone à griffes utilisé dans les véhicules mild-hybrid. Ce dernier voit son utilisation élargie à la fonction d’assistance de traction au moteur thermique et de récupération d’énergie dans les phases de freinage. La structure spéciale du rotor rend nécessaire l’utilisation de modèles éléments finis 3D pour modéliser finement les effets tridimensionnels. Ces derniers requièrent une utilisation intensive des ressources informatiques (RAM, temps CPU) qu’il est crucial de réduire pour envisager une démarche d’optimisation de dispositifs magnétiques. Dans ce contexte, l’utilisation d’une plateforme de modélisation ouverte Gmsh et GetDP, permettant l’implémentation de géométries et de formulations appropriées, a abouti au développement de modèles précis et suffisamment rapides pour être intégrés dans des algorithmes d’optimisation. L’originalité de cette démarche de modélisation consiste à hybrider une approche numérique (éléments finis 3D) et une approche semi-analytique (réseau de perméances) de sorte que l’usage des éléments finis 3D soit réduit aux zones où le champ magnétique est tridimensionnel. Enfin, le choix d’un algorithme d’optimisation adapté aux modèles numériques (boîte noire à fort coût d’évaluation) a permis de développer un outil de pré-dimensionnement et de dimensionnement des machines synchrones à griffes en vue de l’obtention d’une géométrie optimisée en réponse à un cahier des charges industriel
The research work carried out in this thesis concerns the modeling and optimization of claw-pole synchronous alternator used in mild-hybrid vehicles. The use of this machine is extended to assistit traction engine and energy recovery in the braking phases. The special structure of the rotor makes it necessary to use 3D finite element models to analyse efficiently the three-dimensional effects, which requires an intensive use of computing resources (RAM, CPU time). The latter must be reduced in order to consider an optimization process on such magnetic devices. In this context, the use of a Gmsh and GetDP, open modeling platform, led to the development of precise models, fast enough to be integrated into optimization algorithms. The originality of this modeling approach is based on the hybridization of a numerical approach (3D finite elements) and a semi-analytical approach (permeance network) so that the use of 3D finite elements is reduced to areas where The magnetic field is three-dimensional. Finally, the choice of an optimization algorithm adapted to numerical models (black box with high evaluation cost) allows to develop a pre-design and a design tools for claw-pole synchronous machines in order to obtain an optimized geometry achieving an industrial specification

This thesis describes the work done during the several stages of the design, analyse, manufacture and test of a high speed synchronous reluctance machine capable of delivering 5 kW at 80 krpm. In order to meet such demanding speed requirement, several multi-disciplinary design exercises have been carried out having different aims. First the influences of the speed-dependent limiting factors on the machine performance have been investigated by analytical methodologies. After the preliminary analytical design, the main challenges related with the structural and electromagnetic FE-based design refinements have been identified. A comparative study has been then presented with the purpose of identify the most effective rotor design approach in terms of performance of the final design and computational effort related for its achievement. Once the design strategy has been chosen, the advantages of considering rotor parametrization of increasing complexity are evaluated via a comparative study showing the results of several structural optimizations. This study indicated the optimal geometry to manufacture. Prior to the machine prototyping, the influence of the rotor manufacturing tolerance as well as the thermal limitations on the machine performance have been deeply analysed. Tests carried out on the prototype have essentially validated the proposed design approach. In addition, an investigative study aimed at identifying and understanding the reasons of the found discrepancy between the measured and expected performance is also reported.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2008.
Includes bibliographical references (leaves 235-244).
The search for the global minimum energy conformation (GMEC) of protein side chains is an important computational challenge in protein structure prediction and design. Using rotamer models, the problem is formulated as a NP-hard optimization problem. Dead-end elimination (DEE) methods combined with systematic A* search (DEE/A*) have proven useful, but may not be strong enough as we attempt to solve protein design problems where a large number of similar rotamers is eligible and the network of interactions between residues is dense. In this thesis, we present an exact solution method, named BroMAP (branch-and-bound rotamer optimization using MAP estimation), for such protein design problems. The design goal of BroMAP is to be able to expand smaller search trees than conventional branch-and-bound methods while performing only a moderate amount of computation in each node, thereby reducing the total running time. To achieve that, BroMAP attempts reduction of the problem size within each node through DEE and elimination by energy lower bounds from approximate maximurn-a-posteriori (MAP) estimation. The lower bounds are also exploited in branching and subproblem selection for fast discovery of strong upper bounds. Our computational results show that BroMAP tends to be faster than DEE/A* for large protein design cases. BroMAP also solved cases that were not solvable by DEE/A* within the maximum allowed time, and did not incur significant disadvantage for cases where DEE/A* performed well. In the second part of the thesis, we explore several ways of improving the energy lower bounds by using Lagrangian relaxation. Through computational experiments, solving the dual problem derived from cyclic subgraphs, such as triplets, is shown to produce stronger lower bounds than using the tree-reweighted max-product algorithm.
(cont.) In the second approach, the Lagrangian relaxation is tightened through addition of violated valid inequalities. Finally, we suggest a way of computing individual lower bounds using the dual method. The preliminary results from evaluating BroMAP employing the dual bounds suggest that the use of the strengthened bounds does not in general improve the running time of BroMAP due to the longer running time of the dual method.
by Eun-Jong Hong.
Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 91-93).
This thesis addresses the problem of logical topology design for optical backbone networks subject to traffic following a Gaussian distribution. The network design problem is broken into three tasks: traffic routing, capacity allocation, and link placement. The routing and capacity allocation problems are formulated as a convex mathematical program. To extend this formulation to discrete optimization problems, such as the link placement sub-problem, it is reformulated as a mixed integer linear program (MILP) by extending tools from robust optimization to Gaussian variables. Bounds are presented to relate capacity allocation to the probability of traffic overflow on a link. Lastly, the link placement subproblem is formulated as an MILP and network topologies for deterministic traffic are compared with those for stochastic traffic. Additionally, this thesis presents a scheme in which a dedicated backup network is designed to provide protection from random link failures. Upon a link failure in the primary network, traffic is rerouted through a preplanned path in the backup network. We introduce a novel approach for dealing with random link failures, in which probabilistic survivability guarantees are provided to limit capacity over-provisioning. We show that the optimal backup routing strategy in this respect depends on the reliability of the primary network. Specifically, as primary links become less likely to fail, the optimal backup networks employ more resource sharing amongst backup paths. We apply results from the field of robust optimization to formulate an ILP for the design and capacity provisioning of these backup networks. We then propose a simulated annealing heuristic to solve this problem for large-scale networks, and we present simulation results to verify our analysis on optimal backup networks.
by Matthew R. Johnston.
S.M.

Traditionally the machine learning community has viewed the No Free Lunch (NFL) theorems for search and optimization as a limitation. I review, analyze, and unify the NFL theorem with the many frameworks to arrive at necessary conditions for improving black-box optimization, model selection, and machine learning in general. I review meta-learning literature to determine when and how meta-learning can benefit machine learning. We generalize meta-learning, in context of the NFL theorems, to arrive at a novel technique called Anti-Training with Sacrificial Data (ATSD). My technique applies at the meta level to arrive at domain specific algorithms and models. I also show how to generate sacrificial data. An extensive case study is presented along with simulated annealing results to demonstrate the efficacy of the ATSD method.

Mechatronic systems are widely used in modern manufacturing. The key machinery of a manufacturing system should be reliable, flexible, intelligent, less complex, and cost effective, which indeed are distinguishing features of a mechatronic system. To achieve these goals, continuous or on-demand design improvements should be incorporated rapidly and effectively, which will address new design requirements or resolve existing weaknesses of the original design. With the advances in sensor technologies, wireless communication, data storage, and data mining, machine health monitoring (MHM) has achieved significant capabilities to monitor the performance of an operating machine. The extensive data from the MHM system can be employed in design improvement of the monitored system. In that context, the present dissertation addresses several challenges in applying MHM in design optimization of a mechatronic system. First, this dissertation develops a systematic framework for continuous design evolution of a mechatronic system with MHM. Possible design weaknesses of the monitored system are detected using the information from MHM. The proposed method incorporates an index to identify a possible design weakness by evaluating the performance, detecting failures and estimating the health status of the system. Second, improved approaches of intelligent machine fault diagnosis (IMFD) that can be applied to more general machinery and faults, are presented. This dissertation develops an IMFD approach based on deep neural networks (DNN). It uses the massive unlabeled MHM data to learn representative features. Using very few items of labeled data, this approach can achieve superior diagnosis performance. The dissertation presents another IMFD approach, which uses the convolutional neural networks (CNN) and sensor fusion and has increased diagnosis accuracy and reliability. The end-to-end learning capability of the two approaches enables diagnosis of fault types or machines for which limited prior knowledge is available. Third, a hierarchical DNN-based method of remaining useful life (RUL) prediction is developed. It achieves high accuracy of RUL prediction by modeling the system degradation on different health stages. This method generates a better estimate of the system RUL, which provides accurate information for the evaluation of system design.
Applied Science, Faculty of
Mechanical Engineering, Department of
Graduate

Next-generation cellular systems will rely on millimeter wave (mmWave) bands to meet the increasing demand for wireless connectivity from end user equipment. Given large available bandwidth and small-sized antenna elements, mmWave frequencies can support high communication rates and facilitate the use of multiple-input-multiple-output (MIMO) techniques to increase the wireless capacity. However, the small wavelength of mmWave yields severe path loss and high channel uncertainty. Meanwhile, using a large number of antenna elements requires a high energy consumption and heavy communication overhead for MIMO transmissions and channel measurement. To facilitate efficient mmWave communications, in this dissertation, the challenges of energy efficiency and communication overhead are addressed. First, the use of unmanned aerial vehicle (UAV), intelligent signal reflector, and device-to-device (D2D) communications are investigated to improve the reliability and energy efficiency of mmWave communications in face of blockage. Next, to reduce the communication overhead, new channel modeling and user localization approaches are developed to facilitate MIMO channel estimation by providing prior knowledge of mmWave links. Using advance mathematical tools from machine learning (ML), game theory, and communication theory, this dissertation develops a suite of novel frameworks using which mmWave communication networks can be reliably deployed and operated in wireless cellular systems, UAV networks, and wearable device networks. For UAV-based wireless communications, a learning framework is developed to predict the cellular data traffic during congestion events, and a new framework for the on-demand deployment of UAVs is proposed to offload the excessive traffic from the ground base stations (BSs) to the UAVs. The results show that the proposed approach enables a dynamical and optimal deployment of UAVs that alleviates the cellular traffic congestion. Subsequently, a novel energy-efficient framework is developed to reflect mmWave signals from a BS towards mobile users using a UAV-carried intelligent reflector (IR). To optimize the location and reflection coefficient of the UAV-carried IR, a deep reinforcement learning (RL) approach is proposed to maximize the downlink transmission capacity. The results show that the RL-based approach significantly improves the downlink line-of-sight probability and increases the achievable data rate. Moreover, the channel estimation challenge for MIMO communications is addressed using a distributional RL approach, while optimizing an IR-aided downlink multi-user communication. The results show that the proposed method captures the statistic feature of MIMO channels, and significantly increases the downlink sum-rate. Moreover, in order to capture the characteristics of air-to-ground channels, a data-driven approach is developed, based on a distributed framework of generative adversarial networks, so that each UAV collects and shares mmWave channel state information (CSI) for cooperative channel modeling. The results show that the proposed algorithm enables an accurate channel modeling for mmWave MIMO communications over a large temporal-spatial domain. Furthermore, the CSI pattern is analyzed via semi-supervised ML tools to localize the wireless devices in the mmWave networks. Finally, to support D2D communications, a novel framework for mmWave multi-hop transmissions is investigated to improve the performance of the high-rate low-latency transmissions between wearable devices. In a nutshell, this dissertation provides analytical foundations on the ML-based performance optimization of mmWave communication systems, and the anticipated results provide rigorous guidelines for effective deployment of mmWave frequency bands into next-generation wireless systems (e.g., 6G).
Doctor of Philosophy
Different kinds of new smart devices are invented and deployed every year. Emerging smart city applications, including autonomous vehicles, virtual reality, drones, and Internet-of-things, will require the wireless communication system to support more data transmissions and connectivity. However, existing wireless network (e.g., 5G and Wi-Fi) operates at congested microwave frequency bands and cannot satisfy needs of these applications due to limited resources. Therefore, a different, very high frequency band at the millimeter wave (mmWave) spectrum becomes an inevitable choice to manage the exponential growth in wireless traffic for next-generation communication systems. With abundant bandwidth resources, mmWave frequencies can provide the high transmission rate and support the wireless connectivity for the massive number of devices in a smart city. Despite the advantages of communications at the mmWave bands, it is necessary to address the challenges related to high-frequency transmissions, such as low energy efficiency and unpredictable link states. To this end, this dissertation develops a set of novel network frameworks to facilitate the service deployment, performance analysis, and network optimization for mmWave communications. In particular, the proposed frameworks and efficient algorithms are tailored to the characteristics of mmWave propagation and satisfy the communication requirements of emerging smart city applications. Using advanced mathematical tools from machine learning, game theory, and wireless communications, this dissertation provides a comprehensive understanding of the communication performance over mmWave frequencies in the cellular systems, wireless local area networks, and drone networks. The anticipated results will promote the deployment of mmWave frequencies in next-generation communication systems.

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2019
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 213-230).
Due to the limited size of training data available, machine learning models for biology have remained rudimentary and inaccurate despite the significant advance in machine learning research. With the recent advent of high-throughput sequencing technology, an exponentially growing number of genomic and proteomic datasets have been generated. These large-scale datasets admit the training of high-capacity machine learning models to characterize sophisticated features and produce accurate predictions on unseen examples. In this thesis, we attempt to develop advanced machine learning models for functional genomics and therapeutics design, two areas with ample data deposited in public databases and tremendous clinical implications. The shared theme of these models is to learn how the composition of a biological sequence encodes a functional phenotype and then leverage such knowledge to provide insight for target discovery and therapeutic design.
First, we design three machine learning models that predict transcription factor binding and DNA methylation, two fundamental epigenetic phenotypes closely tied to gene regulation, from DNA sequence alone. We show that these epigenetic phenotypes can be well predicted from the sequence context. Moreover, the predicted change in phenotype between the reference and alternate allele of a genetic variant accurately reflect its functional impact and improves the identification of regulatory variants causal for complex diseases. Second, we devise two machine learning models that improve the prediction of peptides displayed by the major histocompatibility complex (MHC) on the cell surface. Computational modeling of peptide-display by MHC is central in the design of peptide-based therapeutics.
Our first machine learning model introduces the capacity to quantify uncertainty in the computational prediction and proposes a new metric for peptide prioritization that reduces false positives in high-affinity peptide design. The second model improves the state-of-the-art performance in MHC-ligand prediction by employing a deep language model to learn the sequence determinants for auxiliary processes in MHC-ligand selection, such as proteasome cleavage, that are omitted by existing methods due to the lack of labeled data. Third, we develop machine learning frameworks to model the enrichment of an antibody sequence in phage-panning experiments against a target antigen. We show that antibodies with low specificity can be reduced by a computational procedure using machine learning models trained for multiple targets. Moreover, machine learning can help to design novel antibody sequences with improved affinity.
by Haoyang Zeng
Ph. D.
Ph.D. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1996.
Includes bibliographical references (leaves 45-46).
by Jonathan E. MIchelson.
M.Eng.