Model predictive control in thermal management of multiprocessor systems-on-chip

MultiProcessor Systems-on-Chip (MPSoC) are the core of nowadays and next generation computing platforms. Their relevance in the global market continuously increase, occupying an important role both in everydaylife products (e.g. smartphones, tablets, laptops, cars) and in strategical market sectors as aviation, defense, robotics, medicine. Despite of the incredible performance improvements in the recent years processors manufacturers have had to deal with issues, commonly called “Walls”, that have hindered the processors development. After the famous “Power Wall”, that limited the maximum frequency of a single core and marked the birth of the modern multiprocessors system-on-chip, the “Thermal Wall” and the “Utilization Wall” are the actual key limiter for performance improvements. The former concerns the damaging effects of the high temperature on the chip caused by the large power densities dissipation, whereas the second refers to the impossibility of fully exploiting the computing power of the processor due to the limitations on power and temperature budgets. In this thesis we faced these challenges by developing efficient and reliable solutions able to maximize performance while limiting the maximum temperature below a fixed critical threshold and saving energy. This has been possible by exploiting the Model Predictive Controller (MPC) paradigm that solves an optimization problem subject to constraints in order to find the optimal control decisions for the future interval. A fully-distributedMPC-based thermal controller with a far lower complexity respect to a centralized one has been developed. The control feasibility and interesting properties for the simplification of the control design has been proved by studying a partial differential equation thermal model. Finally, the controller has been efficiently included in more complex control schemes able to minimize energy consumption and deal with mixed-criticalities tasks

Optimal current control for sub-unity power factor correction

Switching mode power supplies (SMPS) are subject to low power factor and high harmonic distortions. Active power-factor correction (APFC) is a technique to improve the power factor and to reduce the harmonic distortion of SMPSs. However, this technique results in double frequency output voltage variation which can be reduced by using a large output capacitance. Using large capacitors increases the cost and size of the converter. Furthermore, the capacitors are subject to frequent failures mainly caused by evaporation of the electrolytic solution which reduce the converter reliability. This thesis presents an optimal control method for the input current of a boost converter to reduce the size of the output capacitor. The optimum current waveform as a function of weighing factor is found by using the Euler Lagrange equation. A set of simulations are performed to determine the ideal weighing which gives the lowest possible output voltage variation as the converter still meets the IEC-61000-3-2 class-A harmonics requirements with a power factor of 0.8 or higher. The proposed method is verified by the experimental work. A boost converter is designed and it is run for different power levels, 100 W, 200 W and 400 W. The desired output voltage ripple is 10 V peak to peak for the output voltage of 200 Vdc. This ripple value corresponds to a ± 2.5% output voltage ripple. The experimental and the simulation results are found to be quite matching. A significant reduction in capacitor size, as high as 50%, is accomplished by using the proposed method.

The inverse problem for Schwinger pair production

The production of electron–positron pairs in time-dependent electric fields (Schwinger mechanism) depends non-linearly on the applied field profile. Accordingly, the resulting momentum spectrum is extremely sensitive to small variations of the field parameters. Owing to this non-linear dependence it is so far unpredictable how to choose a field configuration such that a predetermined momentum distribution is generated. We show that quantum kinetic theory along with optimal control theory can be used to approximately solve this inverse problem for Schwinger pair production. We exemplify this by studying the superposition of a small number of harmonic components resulting in predetermined signatures in the asymptotic momentum spectrum. In the long run, our results could facilitate the observation of this yet unobserved pair production mechanism in quantum electrodynamics by providing suggestions for tailored field configurations.

Solution of optimal continuous low-thrust transfer using Lie transforms

This paper addresses the problem of optimal constant continuous low-thrust transfer in the context of the restricted two-body problem (R2BP). Using the Pontryagin’s principle, the problem is formulated as a two point boundary value problem (TPBVP) for a Hamiltonian system. Lie transforms obtained through the Deprit method allow us to obtain the canonical mapping of the phase flow as a series in terms of the order of magnitude of the thrust applied. The reachable set of states starting from a given initial condition using optimal control policy is obtained analytically. In addition, a particular optimal transfer can be computed as the solution of a non-linear algebraic equation. Se investiga el uso de series y transformadas de Lie en problemas de optimización de trayectorias de satélites impulsados por motores de bajo empuje

Analysis of the geometric altimetry to support aircraft optimal trajectories within the future 4d trajectory management

Over the past few years, the common practice within air traffic management has been that commercial aircraft fly by following a set of predefined routes to reach their destination. Currently, aircraft operators are requesting more flexibility to fly according to their prefer- ences, in order to achieve their business objectives. Due to this reason, much research effort is being invested in developing different techniques which evaluate aircraft optimal trajectory and traffic synchronisation. Also, the inefficient use of the airspace using barometric altitude overall in the landing and takeoff phases or in Continuous Descent Approach (CDA) trajectories where currently it is necessary introduce the necessary reference setting (QNH or QFE). To solve this problem and to permit a better airspace management born the interest of this research. Where the main goals will be to evaluate the impact, weakness and strength of the use of geometrical altitude instead of the use of barometric altitude. Moreover, this dissertation propose the design a simplified trajectory simulator which is able to predict aircraft trajectories. The model is based on a three degrees of freedom aircraft point mass model that can adapt aircraft performance data from Base of Aircraft Data, and meteorological information. A feature of this trajectory simulator is to support the improvement of the strategic and pre-tactical trajectory planning in the future Air Traffic Management. To this end, the error of the tool (aircraft Trajectory Simulator) is measured by comparing its performance variables with actual flown trajectories obtained from Flight Data Recorder information. The trajectory simulator is validated by analysing the performance of different type of aircraft and considering different routes. A fuel consumption estimation error was identified and a correction is proposed for each type of aircraft model. In the future Air Traffic Management (ATM) system, the trajectory becomes the fundamental element of a new set of operating procedures collectively referred to as Trajectory-Based Operations (TBO). Thus, governmental institutions, academia, and industry have shown a renewed interest for the application of trajectory optimisation techniques in com- mercial aviation. The trajectory optimisation problem can be solved using optimal control methods. In this research we present and discuss the existing methods for solving optimal control problems focusing on direct collocation, which has received recent attention by the scientific community. In particular, two families of collocation methods are analysed, i.e., Hermite-Legendre-Gauss-Lobatto collocation and the pseudospectral collocation. They are first compared based on a benchmark case study: the minimum fuel trajectory problem with fixed arrival time. For the sake of scalability to more realistic problems, the different meth- ods are also tested based on a real Airbus 319 El Cairo-Madrid flight. Results show that pseudospectral collocation, which has shown to be numerically more accurate and computa- tionally much faster, is suitable for the type of problems arising in trajectory optimisation with application to ATM. Fast and accurate optimal trajectory can contribute properly to achieve the new challenges of the future ATM. As atmosphere uncertainties are one of the most important issues in the trajectory plan- ning, the final objective of this dissertation is to have a magnitude order of how different is the fuel consumption under different atmosphere condition. Is important to note that in the strategic phase planning the optimal trajectories are determined by meteorological predictions which differ from the moment of the flight. The optimal trajectories have shown savings of at least 500 [kg] in the majority of the atmosphere condition (different pressure, and temperature at Mean Sea Level, and different lapse rate temperature) with respect to the conventional procedure simulated at the same atmosphere condition.This results show that the implementation of optimal profiles are beneficial under the current Air traffic Management (ATM).

Probability distribution modelling to improve stability in nonlinear MIMO control

We consider the direct adaptive inverse control of nonlinear multivariable systems with different delays between every input-output pair. In direct adaptive inverse control, the inverse mapping is learned from examples of input-output pairs. This makes the obtained controller sub optimal, since the network may have to learn the response of the plant over a larger operational range than necessary. Moreover, in certain applications, the control problem can be redundant, implying that the inverse problem is ill posed. In this paper we propose a new algorithm which allows estimating and exploiting uncertainty in nonlinear multivariable control systems. This approach allows us to model strongly non-Gaussian distribution of control signals as well as processes with hysteresis. The proposed algorithm circumvents the dynamic programming problem by using the predicted neural network uncertainty to localise the possible control solutions to consider.

Efficient Control in Multistage Stochastic Optimization Problem

AMS subject classification: 90C31, 90A09, 49K15, 49L20.

Optimization Problem in a Class of Linear System. Application to Multiple Drug Administration

2000 Mathematics Subject Classification: 62H15, 62P10.

Optimal Controller Designs for Rotating Machines - Penalising the Rate of Change of Control Forcing

In the context of active control of rotating machines, standard optimal controller methods enable a trade-off to be made between (weighted) mean-square vibrations and (weighted) mean-square currents injected into magnetic bearings. One shortcoming of such controllers is that no concern is devoted to the voltages required. In practice, the voltage available imposes a strict limitation on the maximum possible rate of change of control force (force slew rate). This paper removes the aforementioned existing shortcomings of traditional optimal control.

Branch and Bound algorithms in greenhouse climate control

The horticultural sector has become an increasingly important sector of food production, for which greenhouse climate control plays a vital role in improving its sustainability. One of the methods to control the greenhouse climate is Model Predictive Control, which can be optimized through a branch and bound algorithm. The application of the algorithm in literature is examined and analyzed through small examples, and later extended to greenhouse climate simulation. A comparison is made of various alternative objective functions available in literature. Subsequently, a modidified version of the B&B algorithm is presented, which reduces the number of node evaluations required for optimization. Finally, three alternative algorithms are developed and compared to consider the optimization problem from a discrete to a continuous control space.

Real-time load-side control of electric power systems

Two trends are emerging from modern electric power systems: the growth of renewable (e.g., solar and wind) generation, and the integration of information technologies and advanced power electronics. The former introduces large, rapid, and random fluctuations in power supply, demand, frequency, and voltage, which become a major challenge for real-time operation of power systems. The latter creates a tremendous number of controllable intelligent endpoints such as smart buildings and appliances, electric vehicles, energy storage devices, and power electronic devices that can sense, compute, communicate, and actuate. Most of these endpoints are distributed on the load side of power systems, in contrast to traditional control resources such as centralized bulk generators. This thesis focuses on controlling power systems in real time, using these load side resources. Specifically, it studies two problems.

(1) Distributed load-side frequency control: We establish a mathematical framework to design distributed frequency control algorithms for flexible electric loads. In this framework, we formulate a category of optimization problems, called optimal load control (OLC), to incorporate the goals of frequency control, such as balancing power supply and demand, restoring frequency to its nominal value, restoring inter-area power flows, etc., in a way that minimizes total disutility for the loads to participate in frequency control by deviating from their nominal power usage. By exploiting distributed algorithms to solve OLC and analyzing convergence of these algorithms, we design distributed load-side controllers and prove stability of closed-loop power systems governed by these controllers. This general framework is adapted and applied to different types of power systems described by different models, or to achieve different levels of control goals under different operation scenarios. We first consider a dynamically coherent power system which can be equivalently modeled with a single synchronous machine. We then extend our framework to a multi-machine power network, where we consider primary and secondary frequency controls, linear and nonlinear power flow models, and the interactions between generator dynamics and load control.

(2) Two-timescale voltage control: The voltage of a power distribution system must be maintained closely around its nominal value in real time, even in the presence of highly volatile power supply or demand. For this purpose, we jointly control two types of reactive power sources: a capacitor operating at a slow timescale, and a power electronic device, such as a smart inverter or a D-STATCOM, operating at a fast timescale. Their control actions are solved from optimal power flow problems at two timescales. Specifically, the slow-timescale problem is a chance-constrained optimization, which minimizes power loss and regulates the voltage at the current time instant while limiting the probability of future voltage violations due to stochastic changes in power supply or demand. This control framework forms the basis of an optimal sizing problem, which determines the installation capacities of the control devices by minimizing the sum of power loss and capital cost. We develop computationally efficient heuristics to solve the optimal sizing problem and implement real-time control. Numerical experiments show that the proposed sizing and control schemes significantly improve the reliability of voltage control with a moderate increase in cost.

Development of Optimal Energy Management Strategies for a Hybrid Boat

The present work proposes different approaches to extend the mathematical methods of supervisory energy management used in terrestrial environments to the maritime sector, that diverges in constraints, variables and disturbances. The aim is to find the optimal real-time solution that includes the minimization of a defined track time, while maintaining the classical energetic approach. Starting from analyzing and modelling the powertrain and boat dynamics, the energy economy problem formulation is done, following the mathematical principles behind the optimal control theory. Then, an adaptation aimed in finding a winning strategy for the Monaco Energy Boat Challenge endurance trial is performed via ECMS and A-ECMS control strategies, which lead to a more accurate knowledge of energy sources and boat’s behaviour. The simulations show that the algorithm accomplishes fuel economy and time optimization targets, but the latter adds huge tuning and calculation complexity. In order to assess a practical implementation on real hardware, the knowledge of the previous approaches has been translated into a rule-based algorithm, that let it be run on an embedded CPU. Finally, the algorithm has been tuned and tested in a real-world race scenario, showing promising results.

Modeling and Model Predictive Control of an Industrial Hydraulic System

This thesis aims to illustrate the construction of a mathematical model of a hydraulic system, oriented to the design of a model predictive control (MPC) algorithm. The modeling procedure starts with the basic formulation of a piston-servovalve system. The latter is a complex non linear system with some unknown and not measurable effects that constitute a challenging problem for the modeling procedure. The first level of approximation for system parameters is obtained basing on datasheet informations, provided workbench tests and other data from the company. Then, to validate and refine the model, open-loop simulations have been made for data matching with the characteristics obtained from real acquisitions. The final developed set of ODEs captures all the main peculiarities of the system despite some characteristics due to highly varying and unknown hydraulic effects, like the unmodeled resistive elements of the pipes. After an accurate analysis, since the model presents many internal complexities, a simplified version is presented. The latter is used to linearize and discretize correctly the non linear model. Basing on that, a MPC algorithm for reference tracking with linear constraints is implemented. The results obtained show the potential of MPC in this kind of industrial applications, thus a high quality tracking performances while satisfying state and input constraints. The increased robustness and flexibility are evident with respect to the standard control techniques, such as PID controllers, adopted for these systems. The simulations for model validation and the controlled system have been carried out in a Python code environment.

Design and implementation of an optimal motion cueing algorithm for the Stuttgart Driving Simulator

Driving simulators emulate a real vehicle drive in a virtual environment. One of the most challenging problems in this field is to create a simulated drive as real as possible to deceive the driver's senses and cause the believing to be in a real vehicle. This thesis first provides an overview of the Stuttgart driving simulator with a description of the overall system, followed by a theoretical presentation of the commonly used motion cueing algorithms. The second and predominant part of the work presents the implementation of the classical and optimal washout algorithms in a Simulink environment. The project aims to create a new optimal washout algorithm and compare the obtained results with the results of the classical washout. The classical washout algorithm, already implemented in the Stuttgart driving simulator, is the most used in the motion control of the simulator. This classical algorithm is based on a sequence of filters in which each parameter has a clear physical meaning and a unique assignment to a single degree of freedom. However, the effects on human perception are not exploited, and each parameter must be tuned online by an engineer in the control room, depending on the driver's feeling. To overcome this problem and also consider the driver's sensations, the optimal washout motion cueing algorithm was implemented. This optimal control-base algorithm treats motion cueing as a tracking problem, forcing the accelerations perceived in the simulator to track the accelerations that would have been perceived in a real vehicle, by minimizing the perception error within the constraints of the motion platform. The last chapter presents a comparison between the two algorithms, based on the driver's feelings after the test drive. Firstly it was implemented an off-line test with a step signal as an input acceleration to verify the behaviour of the simulator. Secondly, the algorithms were executed in the simulator during a test drive on several tracks.

Neural network-based H(infinity) control for fully actuated and underactuated cooperative manipulators

This paper develops H(infinity) control designs based on neural networks for fully actuated and underactuated cooperative manipulators. The neural networks proposed in this paper only adapt the uncertain dynamics of the robot manipulators. They work as a complement of the nominal model. The H(infinity) performance index includes the position errors as well the squeeze force errors between the manipulator end-effectors and the object, which represents a complete disturbance rejection scenario. For the underactuated case, the squeeze force control problem is more difficult to solve due to the loss of some degrees of manipulator actuation. Results obtained from an actual cooperative manipulator, which is able to work as a fully actuated and an underactuated manipulator, are presented. (C) 2008 Elsevier Ltd. All rights reserved.