Recent years advances within well deployment and instrumentation technology offers huge potentials for increased oil recovery from reservoir production. Wells can now be equipped with controllable valves at reservoir depth, which may possibly alter the production profitability of the field completely, if the devices are used in an intelligent manner. This thesis investigates this potential by using model predictive control to maximize reservoir production performance and total oil production. The report describes an algorithm for nonlinear model predictive control, using a single shooting, multistep, quasi-Newton method, and implements it on an existing industrial MPC platform - Statoil's in-house MPC tool SEPTIC. The method is an iterative method, solving a series of quadratic problems analogous to sequential quadratic programming, to find the optimal control settings. An interface between SEPTIC and a commercial reservoir simulator, ECLIPSE, is developed for process modelling and predictions. ECLIPSE provides highly realistic and detailed reservoir behaviour and is used by SEPTIC to obtain numerical gradients for optimization. The method is applied to two reservoir examples, Case 1 and Case 2, and develops optimal control strategies for each of these. The two examples have conceptually different model structures. Case 1 is a simple introduction model. Case 2 is a benchmark model previously used in Yeten, Durlofsky and Aziz (2002) and models a North Sea type channelized reservoir. It is described by a set of different realizations, to capture a notion of model uncertainty. The report addresses each of the available realizations and shows how the value of an optimal production strategy can vary for equally probable realizations. Improvements in reservoir production performance using the model predictive control method are shown for all cases, compared to basic controlled references cases. For the benchmark example improvements range up to as much as 68% increase in one realization, and 30% on average for all realizations. This is an increase from the results previously published for the benchmark, with a 3% average. However, the level of improvement shows significant variation, and is only marginal for example Case 1. A thorough field analysis should therefore be performed before deciding to take the extra cost of well equipment and optimal control.

The project includes a numerical model of heat flow in robotic welding. The modelling is done in finite element method, and afterwards exported to Matlab where the model will be linearized and model reduced. The ultimate goal is to use the model in an automated robotic welding system. In order to decide if IR-thermography may be used as a decision parameter in automated system, experimental work on temperature measurements of welding has been performed.

Drilling into mature, depleted fields is often difficult because of tight pressure margins. Increasing the pressure control will enable wells that previously were considered undrillable, to be drilled. Enabling drilling and increased oil recovery from depleted fields would most likely lead to a substantial increase in profit margains. A better pressure control will also increase the safety of the drilling crew, because the risk of unwanted situations such as a kick or a blow-out is decreased, also reducing the risk of unwanted environmental influence, e.g. oil spill. To compensate for the lack of a continuous measurement of the bottomhole pressure during drilling operations, an adpative observer of the bottomhole pressure is implemented. The observer implemented is tested, and shows promising results in estimating both the bottomhole pressure and the friction coefficient in the well during a pipe connection procedure. To control the pressure in the well, a low-order nonlinear model predicitve controller is developed, and it has been tested to perform well during the pipe connection procedure, where it maintains the pressure within the predefined boundaries. In this thesis both the obsever and the controller will be tested against an artificial well; simulated by a commercial software.

In coupled reservoir simulations a model of a reservoir and the production system are joined together to obtain realistic responses. One way of performing production optimization is to make use of a gas oil ratio (GOR) estimate when calculating the optimal flow rates. Today this estimate is either not accurate enough or calculating it is a time consuming process. The thesis tries on an alternative online approach for estimating the GOR. It makes use of basis functions in the form of polynomials and normalized radial basis functions together with a recursive least squares (RLS) algorithm. This yields a simple and effective optimization strategy. When using a second order polynomial together with the fast convergent recursive least squares algorithm one achieves a suitable fit to the estimated production data. The algorithm has not yet been tested together with a production optimization tool and it has not been compared to the existing methods for estimating the GOR. Still it seems to have a lot of potential and the RLS is fast, convergent and proper for the objective of this thesis.

Prostheses are artificial body parts that can be used by amputees. Myoelectric prostheses are controlled by so-called surface electromyograms (sEMG) that are acquired on the skin surface of the residual limb. A well-known problem in myoelectric prostheses is motion artifacts, these artifacts cause unwanted behaviour of the prosthesis. The purpose of this study is to try to cancel the effect motion artifacts have on myoelectric prosthesis control, in order to avoid unsolicited prosthesis behaviour. The subject of myoelectric prostheses and motion artifacts are outlined in this study, together with the development and characterisation of a sensor that can do simultaneous measurements of sEMG and contact forces between a surface electrode and the skin. A protocol has been developed for the recording of the different signals in a laboratory. Suitable data sets have been recorded from one test subject, and signal processing and pattern recognition methods have been applied on these data sets to generate muscle force estimates. The pattern recognition methods were linear and quadratic mapping functions, and multi-layer perceptron network. To achieve better force estimates when motion artifacts are presence, signals from FSRs are taken into consideration together with sEMG signals. A qualitative comparison reveals obvious improvements for the sEMG sensor when FSR measurement is included. The system presently undergoes quantitative assessment of static and dynamic performance. The final step will hopefully be to integrate FSRs in a real prosthesis.

Estimation of the annular pressure at critical locations in the well is crucial for high-precision pressure control. Certain parameters which are important in order to determine the pressure profile of the well (in particular the friction factor, bulk modulus and density in the annulus), are encumbered with high uncertainty and are besides, continuously, but slowly changing. The objective of the thesis is to employ various Kalman filter designs estimation of the bottomhole pressure and certain important parameters/slowly varying variables, during drilling. Topics that should be addressed are: 1) Literature review of the theory of Kalman filters design 2) Design and implement an Extended Kalman filter (EKF) and possibly the Unscented Kalman filter (UKF) for a) Estimation of the annulus bottom-hole pressure b) Esimation of the friction coefficient, bulk modulus and density in the annulus 3) Analyse the performance/robustness of the observer in important cases, in particular: a) Pipe connection b) Pump ramp up/down 4) Examine observer performance to unmodeled dynamics by testing the Kalman filter against data sets from Wemod 5) Evaluate performance of the observer against experimental data from Grane.

This thesis develops a Dynamic Positioning (DP) system for small marine craft by using the LQR controller approach. Development has been done with a 'Viknes 830' vessel in mind, which is operated by the company 'Maritime Robotics AS' and will be equipped for DP operation the during summer of 2008. A Matlab-based simulator designed for DP simulations has been developed, and is used throughout the thesis. Furthermore, a Hardware-In-the-Loop (HIL) simulator has been used in order to localize and resolve as many implementation issues as possible prior to full-scale installation. A discussion on the general use of a HIL simulator for DP is included. Three variations of a feedback LQR station-keeping controller have been implemented and compared; a simple LQR controller, an LQR controller with modeled actuator dynamics, and nally an LQR controller with actuator dynamics and integral action. A feedforward controller has been added in order to provide enhanced station-keeping performance, as well as bumpless transfer from station keeping to low-speed maneuvering. A reference model has been created for smooth transfer in-between station-keeping reference points, and as input for the feedforward controller. A passive Luenberger DP observer has been applied in order to lter out high-frequency wave loads. Simulation results reveal that the LQR controller with actuator dynamics and integral action is most likely to perform well in real-life application. The largest performance enhancement is gained from the inclusion of actuator dynamics in the controller. It is discovered that the performance turns out better if the actuator dynamics is modeled faster in the controller due to unmodeled actuator saturation limits. V

This report considers the real-time implementation approach of an integration between an Inertial Navigation System (INS) and a Global Positioning System (GPS). The integration has been performed, using a GlobalSat EM--411 GPS receiver and a Microstrain 3DM--GX1 Inertial Measurement Unit (IMU). This has been performed by incorporating a Kalman filter, and aiding the INS estimates through GPS measurements. The goal of this thesis is to create an integrated application able to achieve performance of existing solutions three times the cost. The implementation has been made in real-time in c++, and off-line in Matlab. However the c++ code has not been sufficiently tested due to computer processing problems. Also the code has not been tested on an actual unmanned surface vehicle. The integrated solution worked sufficently when the GPS was online. However, during GPS droupout, the result is subject to high position drift, resulting in position errors of up to 400 meters after 20 seconds. Although it is unknown quite how large the position deviation of other, existing solutions are. However, high drift during GPS dropouts renders the IMU estimates quite useless for navigation. Thus this experiment has been unsuccessful.

In this thesis, we investige the applicability of the certain numerical methods for the solution of certain systems of partial differential equations. Numerical methods for this purpose are studied, in particular the finite element method. These systems arise from the study of extending performance analysis to general convergent nonlinear systems. It is argued that, for simplex meshes, the finite element method is not applicable to the systems. Further, it is argued that the finite element method should not be pursued as an alternative to the solution of these systems, as far better alternatives have recently been developed.

The focus of the work shall be on DVB-T PBR issues of range and Doppler migration of targets (of opportunity) in resolution cells due to high signal bandwidth and long integration time.