Colleges and Schools
http://hdl.handle.net/1969.1/2815
2019-01-23T18:32:32ZStreamline Tracing and Sensitivity Calculation in Fractured Reservoir with Complex Geometry: Field Application to History Matching and Flood Optimization
http://hdl.handle.net/1969.1/174386
Streamline Tracing and Sensitivity Calculation in Fractured Reservoir with Complex Geometry: Field Application to History Matching and Flood Optimization
The popularity of streamline application mainly depends on two aspects: efficient tracing algorithm to generate streamline, and effective flow and transport analysis along streamline. Previous studies proved its applicability for conventional resources such as waterflood in single and dual porosity models. Streamline technology has limited success in extension to fractured reservoir with discrete fracture networks due to lack of efficient tracing method in the complex porous media geometry. Streamline based application such as history matching and rate optimization also has limitation to gas reservoir depletion or fractured reservoir waterflood due to lack of effective streamline-based flow and transport analysis for highly compressible fluid and highly contrasted porous media.
In this study, we first develop streamline tracing method in complex geometry such as faults and discrete fractures. The discrete fractures here are depicted by embedded discrete fracture model (EDFM). We are going to propose novel methods to construct boundary layers for fault non-neighbor connections and EDFM non-neighbor connections. The novel methods reduce the treatment of complex grid geometry to a minimum level and honor the flux of each connection. The utility and validity of this proposed approach is demonstrated using both 2D and 3D examples.
Second, we propose an amended streamline-based travel time sensitivity formulation. This novel sensitivity formulation has improved accuracy than the legacy one when compared to numerical perturbed sensitivity, thus results in faster data misfit reduction. We also develop general streamline-based bottom hole pressure sensitivity
calculation method suitable for highly compressible fluids or complex geometry caused by non-neighbor connections. The bottom hole pressure sensitivity calculation is validated by a successful history matching application to a high pressure high temperature gas reservoir.
Finally, we develop a rate allocation optimization method based on fast estimation of oil recovery, which also applies to fractured reservoirs. The oil recovery is estimated along streamline within the drainage volume by the end of optimization period. The injection/production rates are updated to maximize the field oil recovery. The novel optimization method results in better performance than equalizing well pair injection efficiency or equalizing well pair time of flight when applying to a waterflood case in fractured reservoir. Its validation is further established by the waterflood optimization application to a field scale EDFM reservoir.
We concluded that our proposed approach of streamline tracing, inversion and optimization algorithm extends streamline technology application to fractured media represented by discrete fracture networks and highly compressible fluid, leading to a highly effective reservoir management tool.
2018-11-19T00:00:00ZRobust Signal Processing Techniques for Wearable Inertial Measurement Unit (IMU) Sensors
http://hdl.handle.net/1969.1/174385
Robust Signal Processing Techniques for Wearable Inertial Measurement Unit (IMU) Sensors
Activity and gesture recognition using wearable motion sensors, also known as inertial measurement units (IMUs), provides important context for many ubiquitous sensing applications including healthcare monitoring, human computer interface and context-aware smart homes and offices. Such systems are gaining popularity due to their minimal cost and ability to provide sensing functionality at any time and place. However, several factors can affect the system performance such as sensor location and orientation displacement, activity and gesture inconsistency, movement speed variation and lack of tiny motion information.
This research is focused on developing signal processing solutions to ensure the system robustness with respect to these factors. Firstly, for existing systems which have already been designed to work with certain sensor orientation/location, this research proposes opportunistic calibration algorithms leveraging camera information from the environment to ensure the system performs correctly despite location or orientation displacement of the sensors. The calibration algorithms do not require extra effort from the users and the calibration is done seamlessly when the users present in front of an environmental camera and perform arbitrary movements. Secondly, an orientation independent and speed independent approach is proposed and studied by exploring a novel orientation independent feature set and by intelligently selecting only the relevant and consistent portions of various activities and gestures. Thirdly, in order to address the challenge that the IMU is not able capture tiny motion which is important to some applications, a sensor fusion framework is proposed to fuse the complementary sensor modality in order to enhance the system performance and robustness. For example, American Sign Language has a large vocabulary of signs and a recognition system solely based on IMU sensors would not perform very well. In order to demonstrate the feasibility of sensor fusion techniques, a robust real-time American Sign Language recognition approach is developed using wrist worn IMU and surface electromyography (EMG) sensors.
2018-11-20T00:00:00ZAlgorithms for Computing Edge-Connected Subgraphs
http://hdl.handle.net/1969.1/174384
Algorithms for Computing Edge-Connected Subgraphs
This thesis concentrates on algorithms for finding all the maximal k-edge-connected components
in a given graph G = (V, E) where V and E represent the set of vertices and the set of edges,
respectively, which are further used to develop a scale reduction procedure for the maximum clique
problem. The proposed scale-reduction approach is based on the observation that a subset C of
k + 1 vertices is a clique if and only if one needs to remove at least k edges in order to disconnect
the corresponding induced subgraph G[C] (that is, G[C] is k-edge-connected). Thus, any clique
consisting of k + 1 or more vertices must be a subset of a single k-edge connected component of
the graph. This motivates us to look for subgraphs with edge connectivity at least k in a given
graph G, for an appropriately selected k value.
We employ the method based on the concept of the auxiliary graph, previously proposed in
the literature, for finding all maximal k-edge-connected subgraphs. This method processes the
input graph G to construct a tree-like graphic structure A, which stores the information of the edge
connectivity between each pair of vertices of the graph G. Moreover, this method could provide
us the maximal k-edge-connected components for all possible k and it shares the same vertex set
V with the graph G.
With the information from the auxiliary graph, we implement the scale reduction procedure
for the maximum clique problem on sparse graphs based on the k-edge-connected subgraphs with
appropriately selected values of k. Furthermore, we performed computational experiments to evaluate
the performance of the proposed scale reduction and compare it to the previously used k-core
method. The comparison results present the advancement of the scale reduction with k-edge-connected
subgraphs. Even though our scale reduction algorithm based has higher time complexity,
it is still of interest and deserves further investigation.
2018-11-20T00:00:00ZVINS-mono Optimized: A Monocular Visual-inertial State Estimator with Improved Initialization
http://hdl.handle.net/1969.1/174383
VINS-mono Optimized: A Monocular Visual-inertial State Estimator with Improved Initialization
State estimation is one of the key areas in robotics. It touches a variety of applications in practice such as, aerial vehicle navigation, autonomous driving, augmented reality, and virtual reality. A monocular visual-inertial system (VINS) is one of the popular trends in solving state estimation. By fusing a monocular camera and IMU properly, the system is capable of providing the position and orientation of a vehicle and recovering the scale.
One of the challenges for a monocular VINS is estimator initialization due to the inadequacy of direct distance measurement. Based on the work of Hong Kong University of Technology on monocular VINS, a checkerboard pattern is introduced to improve the original initialization process. The checkerboard parameters are used along with the calculated 3D coordinates to replace the original initialization process, leading to higher accuracy. The results demonstrated lowered cross track error and final drift, compared with the original approach.
2018-12-10T00:00:00Z