Heavy duty powertrains are complex systems with components from various domains, different response times during transient operations and different efficient operating ranges. To ensure efficient transient operation of a powertrain, e.g. with low fuel consumption or short transient duration, it is important to come up with proper control strategies. In this dissertation, optimal control theory is used to calculate and analyze efficient heavy duty powertrain controls during transient operations in different applications. This is enabled by first developing control ready models, usable for multi-phase optimal control problem formulations, and then using numerical optimal control methods to calculate the optimal transients. Optimal control analysis of a wheel loader operating in a repetitive loading cycle is the first studied application. Increasing fuel efficiency or reducing the operation time in such repetitive loading cycles sums up to large savings over longer periods of time. Load lifting and vehicle traction consume almost all of the power produced by a diesel engine during wheel loader operation. Physical models are developed for these subsystems where the dynamics are described by differential equations. The model parameters are tuned and fuel consumption estimation is validated against measured values from real wheel loader operation. The sensitivity of wheel loader trajectory with respect to constrains such as the angle at which the wheel loader reaches the unloading position is also analyzed. A time and fuel optimal trajectory map is calculated for various unloading positions. Moreover, the importance of simultaneous optimization of wheel loader trajectory and the component transients is shown via a side to side comparison between measured fuel consumption and trajectories versus optimal control results. In another application, optimal control is used to calculate efficient gear shift controls for a heavy duty Automatic Transmission system. A modeling and optimal control framework is developed for a nine speed automatic transmission. Solving optimal control problems using the developed model, time and jerk efficient transient for simultaneous disengagement of off-going and engagement of in-coming shift actuators are obtained and the results are analyzed. Optimal controls of a diesel-electric powertrain during a gear shift in an Automated Manual Transmission system are calculated and analyzed in another application of optimal control. The powertrain model is extended by including driveline backlash angle as an extra state in the system. This is enabled by implementation of smoothing techniques in order to describe backlash dynamics as a single continuous function during all gear shift phases. Optimal controls are also calculated for a diesel-electric powertrain corresponding to a hybrid bus during a tip-in maneuver. It is shown that for optimal control analysis of complex powertrain systems, minimizing only one property such as time pushes the system transients into extreme operating conditions far from what is achievable in real applications. Multi-objective optimal control problem formulations are suggested in order to obtain a compromise between various objectives when analyzing such complex powertrain systems.

Most smart actuators exhibit rate-dependant hysteresis when the working frequency is higher than 5Hz. Although the Preisach model has been a very powerful tool to model the static hysteresis, it cannot be directly used to model the dynamic hysteresis. Some researchers have proposed various generalizations of the Preisach operator to model the rate-dependant hysteresis, however, most of them are application-dependant and only valid for low frequency range. In this thesis, a first-order dynamic relay operator is proposed. It is then used to build a novel dynamic Preisach model. It can be used to model general dynamic hysteresis and is valid for a large frequency range. Real experiment data of magnetostrictive actuator is used to test the proposed model. Experiments have shown that the proposed model can predict all the static major and minor loops very well and at the same time give an accurate prediction for the dynamic hysteresis loops. The controller design using the proposed model is also studied. An inversion algorithm is developed and a PID controller with inverse hysteresis compensation is proposed and tested through simulations. The results show that the PID controller with inverse compensation is good at regulating control; its tracking performance is really limited (average error is 10 micron), especially for high frequency signals. Hence, a simplified predictive control scheme is developed to improve the tracking performance. It is proved through experiments that the proposed predictive controller can reduce the average tracking error to 2 micron while preserve a good regulating performance.

This thesis proposes a complete modeling and control design methodology for a multifunctional single-phase bidirectional PWM converter in renewable energy systems. There is a generic current loop for different modes of operation to ease the transition between different modes, including stand-alone inverter mode, grid-tied inverter mode, grid-tied rectifier mode and grid-tied charger/discharger mode. Under stand-alone mode operation, ac voltage regulation is of importance because of the sensitive loads. In this thesis, different multi-loop-based control schemes are investigated and compared, especially between the load current feedback control, PR control and capacitor current loop control. It shows that PR controller reduces the steady-state error, while load current feedback controller improves the transient response. However, the load current feedback controller and capacitor current loop controller presents unstable outputs under some filter load condition. Single-phase d-q frame control is also studied. In order to ease the implementation effort, an unbalanced d-q frame control is proposed to achieve zero steady-state error voltage regulation without generating β-axis component. Based on the same principle, a d-q frame-based single-phase PLL is also proposed to achieve the fast dynamic response with the zero steady-state error phase tracking. The entire control system is verified on a modified 7 kW single-phase PWM converter prototype with a simple DSP-based digital implementation. The load step response test is presented under different modes of operation. The controllers for stand-alone mode are also done under no load, 1 kW resistive load, 1kVar capacitive load, and non-linear load conditions verifying that the single-phase d-q achieves 70% steady-state error improvement if taking the normal PID controller as the baseline design. In the end, the proposed PLL is compared with the standard PLL by experiments showing that the steady-state error can be reduced by 80%. / Master of Science

Additive Manufacturing (AM) has demonstrated great potential to advance product design and manufacturing, and has showed higher flexibility than conventional manufacturing techniques for the production of small volume, complex and customised components. In an economy focused on the need to develop customised and hi-tech products, there is increasing interest in establishing AM technologies as a more efficient production approach for high value products such as aerospace and biomedical products. Nevertheless, the use of AM processes, for even small to medium volume production faces a number of issues in the current state of the technology. AM production is normally used for making parts with complex geometry which implicates the assessment of numerous processing options or choices; the wrong choice of process parameters can result in poor surface quality, onerous manufacturing time and energy waste, and thus increased production costs and resources. A few commonly used AM processes require the presence of cellular support structures for the production of overhanging parts. Depending on the object complexity their removal can be impossible or very time (and resources) consuming. Currently, there is a lack of tools to advise the AM operator on the optimal choice of process parameters. This prevents the diffusion of AM as an efficient production process for enterprises, and as affordable access to democratic product development for individual users. Research in literature has focused mainly on the optimisation of single criteria for AM production. An integrated predictive modelling and optimisation technique has not yet been well established for identifying an efficient process set up for complicated products which often involve critical building requirements. For instance, there are no robust methods for the optimal design of complex cellular support structures, and most of the software commercially available today does not provide adequate guidance on how to optimally orientate the part into the machine bed, or which particular combination of cellular structures need to be used as support. The choice of wrong support and orientation can degenerate into structure collapse during an AM process such as Selective Laser Melting (SLM), due to the high thermal stress in the junctions between fillets of different cells. Another issue of AM production is the limited parts’ surface quality typically generated by the discrete deposition and fusion of material. This research has focused on the formation of surface morphology of AM parts. Analysis of SLM parts showed that roughness measured was different from that predicted through a classic model based on pure geometrical consideration on the stair step profile. Experiments also revealed the presence of partially bonded particles on the surface; an explanation of this phenomenon has been proposed. Results have been integrated into a novel mathematical model for the prediction of surface roughness of SLM parts. The model formulated correctly describes the observed trend of the experimental data, and thus provides an accurate prediction of surface roughness. This thesis aims to deliver an effective computational methodology for the multi- objective optimisation of the main building conditions that affect process efficiency of AM production. For this purpose, mathematical models have been formulated for the determination of parts’ surface quality, manufacturing time and energy consumption, and for the design of optimal cellular support structures. All the predictive models have been used to evaluate multiple performance and costs objectives; all the objectives are typically contrasting; and all greatly affected by the part’s build orientation. A multi-objective optimisation technique has been developed to visualise and identify optimal trade-offs between all the contrastive objectives for the most efficient AM production. Hence, this thesis has delivered a decision support system to assist the operator in the "process planning" stage, in order to achieve optimal efficiency and sustainability in AM production through maximum material, time and energy savings.

In the thesis, automatic flight control system is designed for Tilt Duct Unmanned Aerial Vehicle (UAV). The vehicle is a Vertical Take-Off Landing (VTOL) type with two symmetric rotors on the wings, one aft rotor on the aft body. It behaves like a helicopter but with higher speeds in forward flight. Transition flight of the aircraft from hover to cruise or take-off to forward flight is the primary concern of the thesis study with the nonlinearities and instabilities encountered, together with the over-actuated controls in this mode. A nonlinear simulation code is developed including nonlinear equations of motion together with the nonlinear aerodynamics, environmental eects, and rotor dynamics. Trim and linearization codes are also developed. Trim conditions for the transition flight phase are calculated for two different transition scenarios. Linear controllers are developed and nonlinear controller is designed for the transition mode. Nonlinear controller uses the state dependent Ricatti equation SDRE approach by using extended linearization. Two loop approach is used in order to increase controllability. In the inner loop, attitude rates are fed back and SDRE approach is used to calculate the feedback gain matrix online. In the outer loop, vehicle attitude is controlled using the eigenvalue assignment. Blended inverse algorithm based control allocation method is used in control of the over-actuated transition phase. This algorithm is shown to be quite effective among different methods in not only generating necessary forces needed for the control, but also allocating with more control authority on the desired actuator.

Advantages of electrochemical wastewater treatment over conventional wastewater treatment include its smaller footprint, modularity, and ability to meet increasingly stringent government regulations. A power supply that can be packaged with an electrochemical stack could make electrochemical wastewater treatment more cost-effective and scalable. For this application, the series and series-parallel resonant converters are suitable power converter candidates. With an output current specification of 100A, the series-parallel resonant converter (SPRC) is superior due to its simpler output stage. The thesis presents the design of a 500W SPRC for a wastewater treatment cell stack. A rudimentary cell model is derived experimentally. The closed loop analysis, controller design and simulation results are presented. The output voltage and current are estimated using sensed quantities extracted from the high voltage, low current primary side. Low voltage experimental results verify the operation of the power stage and voltage estimation circuitry in open loop pulsed operation.

Advantages of electrochemical wastewater treatment over conventional wastewater treatment include its smaller footprint, modularity, and ability to meet increasingly stringent government regulations. A power supply that can be packaged with an electrochemical stack could make electrochemical wastewater treatment more cost-effective and scalable. For this application, the series and series-parallel resonant converters are suitable power converter candidates. With an output current specification of 100A, the series-parallel resonant converter (SPRC) is superior due to its simpler output stage. The thesis presents the design of a 500W SPRC for a wastewater treatment cell stack. A rudimentary cell model is derived experimentally. The closed loop analysis, controller design and simulation results are presented. The output voltage and current are estimated using sensed quantities extracted from the high voltage, low current primary side. Low voltage experimental results verify the operation of the power stage and voltage estimation circuitry in open loop pulsed operation.

Robot manipulators are getting more and more attention nowadays. This is due to their high precision and the speed they provide while executing their tasks. The desires for such high standards are increasing exponentially due to the extended workspace that manipulators provide. Therefore, a safe controller is needed to make it possible for the robot to work alongside people considering the safety precautions. These safety preconditions are widely spread, even when the needs for better human-friendly robots are rising. This thesis will introduce and explain a way to model a 6-axis robot by using its dynamical properties as well as the development of a joint space inverse dynamic controller. The controller will be tested in various different ways. Firstly by adding noise to the measured data. Then testing the robustness of the control model, while the simulated model includes properties different from those used for the controller itself. The different properties would for example be payloads and the inertia of the links. Thereafter, evaluating the precision of a followed path that is given by an operational space trajectory. The outcome of these experiments show promising results. The results show that the controller is able to manage a noise in both the joint angle and joint velocity. It also shows that an error in the payload data will give a small error in the joint angles, sequentially that gives an acceptable error for the end-effector in the operational space. Furthermore, the controller manages to keep the maximum errorin the joint angle low, while it is following a trajectory in the operational space.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第23504号 / 工博第4916号 / 新制||工||1768(附属図書館) / 京都大学大学院工学研究科機械理工学専攻 / (主査)教授 松野 文俊, 教授 泉田 啓, 教授 藤本 健治 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Récemment, les sources d’énergie renouvelables telles que l’énergie éolienne et solaire, sont devenues des éléments essentiels des réseaux électriques en tant qu’alternatives d’énergie propre aux combustibles fossiles. Cependant, la qualité de la production de telles ressources énergétiques dépend de différents facteurs incertains, tels que les conditions météorologiques. Par conséquent, la gestion intermittente des sources d’énergie renouvelables est l’un des principaux défis à relever pour une utilisation à plus grande échelle.Une solution possible pour réduire les effets de l'intermittence des ressources énergétiques sur la production d'énergie et la stabilité du réseau consiste à utiliser les technologies de stockage d'énergie. Les stations de transfert d’énergie par pompage (STEP) semblent être la méthode de stockage propre unique qui peut être utilisée pour lutter contre la nature intermittente de l’énergie éolienne et solaire. Les STEP utilisent des pompes-turbines réversibles pouvant fonctionner comme des pompes pour stocker l'excès d'énergie électrique dans le réseau et comme des turbines pour générer de l'énergie électrique, lorsque davantage d'énergie électrique est nécessaire. Ainsi, les STEP aident à stabiliser le réseau en présence de ressources en énergies renouvelables intermittentes.Ce travail met l’accent sur les conditions de fonctionnement de la turbine pour le démarrage des STEP. Dans les STEP, la condition de fonctionnement de démarrage est généralement visitée plusieurs fois, à la suite d'un basculement entre les modes de pompage et de turbine. Ainsi, l'amélioration des performances des régulateurs de vitesse utilisés pour le démarrage devient plus importante lorsque l'on traite avec des STEP afin de permettre une récupération rapide de la tension.Cette thèse s'inscrit dans le cadre du projet pluridisciplinaire INNOVHYDRO, qui regroupe différents laboratoires et entreprises tels que, GIPSA-lab où cette thèse a été préparée, G2Elab, GE et EDF.Dans cette thèse, une architecture de contrôleur prenant en compte les limitations informatiques des microcontrôleurs existants utilisés chez GE est proposée. Elle apporte une solution au problème du démarrage rapide de la turbine tout en évitant l'excitation de fortes oscillations de pression. De plus, les contraintes de couple s'intègrent facilement pour permettre un démarrage en douceur, ce qui réduit la fatigue des composants mécaniques, résultant du démarrage répétitif des turbines.Des solutions sont proposées pour ajuster les gains du contrôleur, tout en tenant compte de la dynamique non linéaire de l'actionneur utilisé chez GE. Pour commencer, une méthodologie de réglage est décrite pour garantir la stabilité asymptotique et les performances en boucle fermée, tout en minimisant la limite supérieure de l'erreur de suivi en sortie. De plus, une approche d'optimisation systématique est développée pour sélectionner les gains du contrôleur afin de minimiser le temps nécessaire pour obtenir une connexion stable au réseau, tout en respectant les contraintes de couple maximales. De plus, des algorithmes sont utilisés pour choisir les paramètres du contrôleur de sorte que des certificats de robustesse soient obtenus sur le contrôleur résultant.De plus, un simulateur a été développé pour les centrales hydrauliques et utilisé pour tester le contrôleur proposé. Le simulateur est constitué d’un système d’équations différentielles continues qui modélisent systématiquement le comportement des différents composants de la centrale hydraulique, tels que les conduites forcées, les tunnels, les réservoirs et les cheminées d’équilibre. De plus, le comportement non linéaire et les caractéristiques en S des régions instables des turbines hydrauliques, généralement modélisées par des diagrammes de Hill, sont pris en compte avec succès. De plus, la dynamique non linéaire de l'actionneur est incluse dans le modèle mathématique complet. / Recently, renewable energy resources such as, wind and solar energy, have become integral parts of electric grids as clean energy alternatives to fossil fuels. However, the quality of production of such resources of energy depends on different uncertain factors, for instance, weather conditions. Therefore, dealing with the intermittent nature of renewable energy resources is one of the main challenges when using them on a larger scale.A possible solution to reduce the effects of energy resources intermittency on energy production and grid's stability, is to use energy storage technologies. Pumped storage power plants (PSPs) seem to be the unique clean storage method that can be used to counteract the intermittent nature of wind and solar energy. PSPs make use of pumps-turbines which are capable of working as pumps to store excess electric energy in the grid, and as turbines to generate electric energy, when more electric energy is needed. Thus, PSPs help in stabilizing the grid in the presence of intermittent renewable energy resources.The emphasis in this work is on turbine start-up operating mode for PSPs. In PSPs, the start-up operating mode is usually visited multiple times, as a result of switching back and forth between pumping and turbine modes. Thus, enhancing the performance of the speed governors used for starting-up becomes more important when dealing with PSPs to enable a rapid voltage recovery.This PhD thesis is part of the multidisciplinary INNOVHYDRO project that includes different laboratories and enterprises such as, GIPSA-lab where this thesis was prepared, G2Elab, GE and EDF.In this thesis, a controller architecture that takes into account the computational limitations of existing microcontrollers in use at GE, is proposed. It provides a solution to the problem of fast turbine start-up, while avoiding the excitation of sharp pressure oscillations. In addition, torque constraints are easily integrated to achieve smoother start-up, which reduces the fatigue of the mechanical components, resulting from repetitive start-up of turbines.Different approaches are proposed to tune the controller gains, while taking into account the nonlinear dynamics of the actuator used at GE. To begin with, a tuning methodology is outlined to guarantee the asymptotic stability and the closed-loop performance, while minimizing the guaranteed upper bound on the output tracking error. In addition, a systematic optimization approach is developed to select the controller gains to minimize time needed to get a stable start-up, while respecting maximum torque constraints. Moreover, randomized algorithms are used to choose the controller parameters such that robustness certificates are obtained on the resulting controller.Furthermore, a simulator has been developed for hydraulic power plants and used to test the proposed controller. The simulator constitutes of a system of continuous differential equations, which systematically model the behavior of the different components of the hydraulic power plant such as, penstocks, tunnels, reservoirs and surge tanks. In addition, the nonlinear behavior and unstable regions 'S-characteristics' of hydraulic turbines, usually modeled by Hill charts, are successfully taken into consideration. Moreover, the actuator's nonlinear dynamics are included in the overall mathematical model.