In industrial fuel gas preparation, there are several compositional properties that must be controlled within specified limits. This allows client plants to use the fuel gas mixture safely without having to adjust and control the composition themselves. The variables to be controlled are the Higher Heating Value (HHV), Wobbe Index (WI), Flame Speed Index (FSI), and Pressure (P). These variables are controlled by adjusting the volumetric flow rates of several inlet gas streams of which some are makeup streams (always available) and some are wild streams that vary in composition and availability (by-products of plants). The inlet streams need to be adjusted in the correct ratios to keep all the controlled variables (CVs) within limits while minimising the cost of the gas blend. Furthermore, the controller needs to compensate for fluctuations in inlet stream compositions and total fuel gas demand (the total discharge from the header). This dissertation describes the modelling and model validation of an industrial fuel gas header as well as a simulation study of three different Model Predictive Control (MPC) strategies for controlling the system while minimising the overall operating cost. / Dissertation (MEng)--University of Pretoria, 2011. / Electrical, Electronic and Computer Engineering / unrestricted

Inverters play key roles in connecting sustainable energy (SE) sources to the local loads and the ac grid. Although there has been a rapid expansion in the use of renewable sources in recent years, fundamental research, on the design of inverters that are specialized for use in these systems, is still needed. Recent advances in power electronics have led to proposing new topologies and switching patterns for single-stage power conversion, which are appropriate for SE sources and energy storage devices. The current source inverter (CSI) topology, along with a newly proposed switching pattern, is capable of converting the low dc voltage to the line ac in only one stage. Simple implementation and high reliability, together with the potential advantages of higher efficiency and lower cost, turns the so-called, single-stage boost inverter (SSBI), into a viable competitor to the existing SE-based power conversion technologies. The dynamic model is one of the most essential requirements for performance analysis and control design of any engineering system. Thus, in order to have satisfactory operation, it is necessary to derive a dynamic model for the SSBI system. However, because of the switching behavior and nonlinear elements involved, analysis of the SSBI is a complicated task. This research applies the state-space averaging technique to the SSBI to develop the state-space-averaged model of the SSBI under stand-alone and grid-connected modes of operation. Then, a small-signal model is derived by means of the perturbation and linearization method. An experimental hardware set-up, including a laboratory-scaled prototype SSBI, is built and the validity of the obtained models is verified through simulation and experiments. Finally, an eigenvalue sensitivity analysis is performed to investigate the stability and dynamic behavior of the SSBI system over a typical range of operation.

It is well known that the inflation linked breakeven inflation, defined as the difference between a nominal yield and an inflation linked yield, sometimes is used as an approximation of the market’s inflation expectation. D’Amico et al. (2009, [5]) show that this is a poor approximation for the US market. Based on their work, this thesis shows that the approximation also is poor for the Swedish bond market. This is done by modelling the Swedish bond market using a five-factor latent variable model, where an inflation linked bond specific premium is introduced. Latent variables and parameters are estimated using a Kalman filter and a maximum likelihood estimation. The conclusion is drawn that the modelling was successful and that the model implied outputs gave plausible results.

We examine the related nature of two identification algorithms, subspace identification (SID) and Dynamic Mode Decomposition (DMD), and their correctness properties over a broad range of problems. This investigation begins by noting the strong relationship between the two algorithms, both drawing significantly on the pseudoinverse calculation using singular value decomposition, and ultimately revealing that DMD can be viewed as a substep of SID. We then perform extensive computational studies, characterizing the performance of SID on problems of various model orders and noise levels. Specifically, we generate 10,000 random systems for each model order and noise level, calculating the average identification error for each case, and then repeat the entire experiment to ensure the results are, in fact, consistent. The results both quantify the intrinsic algorithmic error at zero-noise, monotonically increasing with model complexity, as well as demonstrate an asymptotically linear degradation to noise intensity, at least for the range under study. Finally, we close by demonstrating DMD's ability to recover system matrices, because its access to full state measurements makes them identifiable. SID, on the other hand, can't possibly hope to recover the original system matrices, due to their fundamental unidentifiability from input-output data. This is true even when SID delivers excellent performance identifying a correct set of equivalent system matrices.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第20024号 / 情博第619号 / 新制||情||108(附属図書館) / 33120 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 松山 隆司, 教授 熊田 孝恒, 准教授 川嶋 宏彰 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Our data consist of measurements of 22 sharks' movements within a 366-acre tidal basin. The measurements are made at irregular time points over a 16-month interval. Constant-length observation intervals would have been desirable, but are often infeasible in practice. We model the sharks' paths at short constant-length intervals by inferring their behavior (feeding vs transiting), interpolating their locations, and estimating parameters of motion (speed and turning angle) in environmental and ecological contexts. We are interested in inferring regional differences in the sharks' behavior, and behavioral interaction between them. Our method uses particle filters, a computational Bayesian technique designed to sequentially model a dynamic system. We discuss how resampling is used to approximate arbitrary densities, and illustrate its use in a simple example of a particle filter implementation of a state-space model. We then introduce a particular model formulation that uses conditioning to introduce unobserved parameters for the shark's behaviors. We show how the irregularly-observed shark locations can be modeled by interpolation as a set of movements at constant-length time intervals. We use a spline method for generating approximations of the ground truth at these intervals for comparison with our model. Finally, we demonstrate our model's estimates of the sharks' behavioral and ecological parameters of interest on a subset of the observed data. / Statistics

The advent of synchronized wide-area frequency measurements obtained from frequency disturbance recorders and phasor measurement units has presented the power industry with special opportunities to study power system dynamics. I propose the use of wide-area frequency measurements in identifying system disturbances based on power system post-event modal properties. In this work, power system dynamics are examined from an internal system energy viewpoint. Since an electric power system is composed of coupled rotating machines (large generators) which have air gap magnetic fields that are essentially static, or quasi-static, the power system may be modeled as a system with energy stored in quasi-static magnetic fields. The magnetic fields in the machines do change with time but may be modeled as static as far as wave propagation is concerned. The dynamic model that I develop treats this magnetic energy specifically as potential energy. Each rotating machine also contains an inertia due to the mass and motion of its rotor train and so each machine contains a rotational kinetic energy. Thus the internal system energy for a power system dynamic model may be considered to be contained in potential (magnetic) and kinetic (rotating mass) energies. This notion of internal energy lends itself to the use of a state-space model where each system state is associated with either a kinetic energy or a potential energy. An n-machine system would have a total of 2n states and would thus be a 2n-th order system. For many power system disturbances, I postulate that a linearized version of this model may be used to examine system natural response in terms of frequency and phasor measurements. The disturbances that I will investigate include generator and line outages. For any particular outage, the power system exhibits a very specific natural response in terms of its kinetic and potential energies. Kinetic energy in the system is directly related to each specific machine's rotational speed. I propose that the kinetic energy corresponds directly with bus frequencies through a linear transformation. Likewise magnetic field energy in each machine corresponds directly with a torque angle. The potential energy in the system thus corresponds directly with bus angles through a linear transformation. The primary focus of this work is on frequency deviation modal characteristics – specifically damped oscillation frequencies, mode shapes, and damping ratios. This work presents how specific disturbances on a power system will lead to specific oscillation frequencies in the deviation quantities and that these oscillation frequencies may be used to identify the disturbance. The idea of disturbance identification stems out of previous work done in locating disturbances by using a distributed parameter (DP) model of an electric power system. This DP model, which assumes a wave-like motion of frequency and phase quantities, was used to locate disturbances via a triangulation method. This present work, instead of using a DP model of the power system, assumes lumped parameters and focuses on disturbance identification strictly via modal characteristics – particularly oscillation frequency in the frequency deviations. This model is not concerned with geographic location but focuses on system topology, loading, and machine mass as lumped parameters. Advantages of disturbance identification include mainly reliability enhancements but can also be used in marketing applications. The state-space model used to realize this theory is verified via simulation using small, "academic" systems which should prove useful in classroom settings. Additionally the model is verified on a larger test system in order prove its validity and potential usefulness on large power systems. / Ph. D.

The concept of piezohydraulic actuation is to transfer the reciprocal small stroke displacement of piezoceramics into unidirectional motion by frequency rectification through a hydraulic fluid. It takes advantage of the high force capabilities that piezoelectric materials have and couples it with very stiff media such as hydraulic fluid to amplify and create this unidirectional motion. Inlet and outlet valves are connected to a pumping chamber where pressure is built by the displacement of the piezoelectric material and released by the opening of the outlet valve, thus achieving a variable flow rate that is used to push a hydraulic cylinder. Loads may be connected to this hydraulic cylinder for measuring/achieving mechanical power. As part of this research, a benchtop piezohydraulic actuator with active piezohydraulic valves has been developed and the concept of piezohydraulic actuation has been demonstrated. Displacement of a hydraulic cylinder by driving a piezoelectric stack has been achieved while the cylinder was loaded or unloaded. Lumped-parameter state-space models have been developed in order to simulate the dynamics of the active valves and entire actuator system. The model simulates the chamber pressure, displacement of the hydraulic cylinder, and power of the piezohydraulic unit. A four-stage cycle simulation was used to model the pumping operation and dynamic response of the system. Experimental results demonstrate the importance of fluid compressibility, valve timing, and fluid circuit components in the optimization of the output power of the actuation system. An array of different timing tests run on the inlet and outlet valves shows that their timing is crucial to the performance of the system. Also shown is that the optimal timing conditions change slightly while under different loads. When operating at higher frequencies (above 140 Hz), it is shown that the hydraulic fluid circuit does not respond quickly enough for the piston to fully extend against the fluid and loaded cylinder. There is not sufficient time when operating at higher frequencies to push all the fluid from the chamber into the hydraulic cylinder, operation is too fast for the dynamics of the fluid circuit. The four stage lumped-parameter model achieves good approximations of the experimental results when the load inertia was neglected while operating at frequencies below 120 Hz and under loads at or below 12.825 kg. Memory limitations caused the number of elements included in the lumped-parameter model to be limited, and are believed to be the source of the errors for the higher operation frequencies and loads. The model never converged due to the lack of elements, and the simulated system did not respond quickly enough to accurately model the fluid exiting the chamber. When operating at frequencies above the 120 Hz value, this error in modeling the fluid exiting the valves becomes very important. The simulation predicts higher values than the experiment and fails to correlate to the actual results at the higher frequencies and while under the higher loads. The errors at higher loads may also be attributed to the neglected inertia. The most recent tests on the benchtop set-up were all run with a pre-pressure value of 190 psi, a piston duty cycle of 50%, valve duty cycles of 40% for each, and a 5% outlet valve offset. Slightly better operation performance might be achieved at frequencies higher than 140 Hz by increasing the piston duty cycle and varying the valve parameters. Also, increaing the pre-pressure of the fluid may help by stiffening the system to create a faster response, however this will have an adverse effect also by creating more force against piston motion. Lastly, the hydraulic cylinder was built for high pressures and had considerable friction associated with it. Obtaining a different cylinder with less friction may also help the response time of the fluid circuit. / Master of Science

State space modeling represents a statistical framework for exponential smoo- thing methods and it is often used in time series modeling. This thesis descri- bes seasonal innovations state space models and focuses on recently suggested TBATS model. This model includes Box-Cox transformation, ARMA model for residuals and trigonometric representation of seasonality and it was designed to handle a broad spectrum of time series with complex types of seasonality inclu- ding multiple seasonality, high frequency of data, non-integer periods of seasonal components, and dual-calendar effects. The estimation of the parameters based on maximum likelihood and trigonometric representation of seasonality greatly reduce computational burden in this model. The universatility of TBATS model is demonstrated by four real data time series.

Improving agricultural water management is important for conserving water during dry seasons, using limited water resources in the most efficient way, and minimizing environmental risks (e.g., leaching, surface runoff). The understanding of water movement in different zones of agricultural production fields is crucial to developing an effective irrigation strategy. This work centered on optimizing field water management by characterizing the spatial patterns of soil hydraulic properties. Soil hydraulic conductivity was measured across different zones in a farmer’s field, and its spatial variability was investigated by using geostatistical techniques. Since direct measurement of hydraulic conductivity is time-consuming and arduous, pedo-transfer functions (PTFs) have been developed to estimate hydraulic conductivity indirectly through more easily measurable soil properties. Due to ignoring soil structural information and spatial covariance between soil variables, PTFs often perform unsatisfactorily when field-scale estimations of hydraulic conductivity are needed. The performance of PTFs in estimating hydraulic conductivity in the field was therefore critically evaluated. Due to the presence of structural macro-pores, saturated hydraulic conductivity (Ks) showed high spatial heterogeneity, and this variability was not captured by texture-dominated PTF estimates. However, the general spatial pattern of near-saturated hydraulic conductivity can still be reasonably generated by PTF estimates. Therefore, the hydraulic conductivity maps based on PTF estimates should be evaluated carefully and handled with caution. Recognizing the significant contribution of macro-pores to saturated water flow, PTFs were further improved by including soil macro-porosity and were proven to perform much better in estimating Ks compared with established PTFs tested in this study. Additionally, the spatial relationship between hydraulic conductivity and its potential influencing factors were further quantified by the state-space approach. State-space models outperformed current PTFs and effectively described the spatial characteristics of hydraulic conductivity in the studied field. These findings provided a basis for modeling water/solute transport in the vadose zone, and sitespecific water management.