Show simple item record

dc.contributor.advisorMacBeth, Professor Colin
dc.contributor.advisorChassagne, Doctor Romain
dc.contributor.authorGeng, Chong
dc.date.accessioned2019-09-09T09:37:44Z
dc.date.available2019-09-09T09:37:44Z
dc.date.issued2018-06
dc.identifier.urihttp://hdl.handle.net/10399/4017
dc.description.abstractGenerally, reservoir simulation is used to predict field performance and analyse uncertainties for assistance in decision making, while history matching is a key step in reservoir simulation, which is a process of model adjustment and simulation runs with different reservoir parameter settings until the differences between simulated data and historical data reach minima. An efficient reservoir simulation model must be the one that can predict reservoir performance and update history matching results continuously by modifying the reservoir model as long as new data become available. However, reservoir simulation can be very time consuming, depending on the complexity of the reservoir model, and history matching is even more computationally expensive, since it requires many simulation runs. Recently, intelligent technology advances in the oil and gas industry, have initiated a new era of big data. As different varieties of data have been integrated to make better decisions, together with the generation of high frequency data streams, a major concern for petroleum engineers is how reservoir simulation should be calibrated in line with the real time data without compromising the simulation time. In the seismic history matching (SHM) workflow this may be a more obvious issue than in the conventional well production history matching. In order to address this problem, many studies have been undertaken. Besides increasing computational power, various types of research have focused on speeding up the reservoir simulation process, especially history matching, by either implementing optimisation algorithms or generating efficient proxy models. Nevertheless, there has not yet been a standard method recognized in reservoir simulation. In this study, a novel method has been proposed as an attempt to investigate the possibility of achieving efficient seismic history matching by data-driven proxy models. This thesis essentially involves detailing background motivations, proxy model building, followed by its testing and application in SHM. Comparisons of proxy models with conventional simulators have been made from different aspects. The objective is mainly focused on examining the capability of the proxy models as a simplification of the conventional physics-based simulators in SHM. According to the simulation results, the feasibility of the combination of proxy models has been proven to be successful and efficient. Importantly, huge amounts of time and effort have been saved in the reservoir simulation process. In addition, it is suggested that other challenges of SHM, such as multi-domain comparison and multi-field communication, could be tackled by using the proxy method.en
dc.language.isoenen
dc.publisherHeriot-Watt Universityen
dc.publisherEnergy, Geoscience, Infrastructure and Societyen
dc.rightsAll items in ROS are protected by the Creative Commons copyright license (http://creativecommons.org/licenses/by-nc-nd/2.5/scotland/), with some rights reserved.
dc.titleSeismic history matching using proxy modelsen
dc.typeThesisen


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record