A new consistent characterisation method has been developed to describe the complex composition of petroleum mixtures in terms of the molecular type and homologous series. The petroleum mixture is conceived as a matrix in which the rows represent carbon numbers, while each column represents a homologous series. The concentration of each individual component in the matrix can be measured using modern analytical tools such as gas chromatography (GC), high-performance liquid chromatography (HPLC), mass spectrometry (MS), field ionisation mass spectrometry (FIMS), sulphur chemiluminescence detection (SCD), etc. To evaluate the impacts of crude composition and refining chemistry on the composition and quality of refinery products, a novel method is proposed to predict the properties of petroleum mixtures based on the compositional information contained in the matrix. In this method, molecular structure-property correlations have been developed first to predict the boiling point and density of the molecular type homologous series in the matrix with high accuracy. Then the ASTM distillation curve and bulk density of the petroleum mixtures can be calculated with an assumed mixing rule. To predict other properties such as critical constants, freezing point, cetane number, pour point, cloud point, etc., well-tested correlations based on the distillation curve and bulk density are used along with the compositional information in the matrix. In addition, gasoline octane number can be predicted from molecular composition-based correlations. A simple but accurate method is also proposed to predict the molecular composition of a new feed through blending of fully characterised petroleum mixtures, thus expensive and time-consuming experimental analyses can be spared. The consistent molecular level characterisation of petroleum mixtures has enabled the development of refinery reaction and separation models based on the underlying process chemistry and thermodynamic principles. In addition, with the molecular information provided by the new characterisation, more efficient optimisation and integration can be conducted in the context of overall refinery.

4D seismic data bears valuable spatial information about production-related changes in the reservoir. It is a challenging task though to make simulation models honour it. Strict spatial tie of seismic data requires adequate model complexity in order to assimilate details of seismic signature. On the other hand, not all the details in the seismic signal are critical or even relevant to the flow characteristics of the simulation model so that fitting them may compromise the predictive capability of models. So, how complex should be a model to take advantage of information from seismic data and what details should be matched? This work aims to show how choices of parameterisation affect the efficiency of assimilating spatial information from the seismic data. Also, the level of details at which the seismic signal carries useful information for the simulation model is demonstrated in light of the limited detectability of events on the seismic map and modelling errors. The problem of the optimal model complexity is investigated in the context of choosing model parameterisation which allows effective assimilation of spatial information in the seismic map. In this study, a model parameterisation scheme based on deterministic objects derived from seismic interpretation creates bias for model predictions which results in poor fit of historic data. The key to rectifying the bias was found to be increasing the flexibility of parameterisation by either increasing the number of parameters or using a scheme that does not impose prior information incompatible with data such as pilot points in this case. Using the history matching experiments with a combined dataset of production and seismic data, a level of match of the seismic maps is identified which results in an optimal constraint of the simulation models. Better constrained models were identified by quality of their forecasts and closeness of the pressure and saturation state to the truth case. The results indicate that a significant amount of details in the seismic maps is not contributing to the constructive constraint by the seismic data which is caused by two factors. First is that smaller details are a specific response of the system-source of observed data, and as such are not relevant to flow characteristics of the model, and second is that the resolution of the seismic map itself is limited by the seismic bandwidth and noise. The results suggest that the notion of a good match for 4D seismic maps commonly equated to the visually close match is not universally applicable.

No description available.

Forecasting reservoir production has a large associated uncertainty, since this is the final part of a very complex process, this process is based on sparse and indirect data measurements. One the methodologies used in the oil industry to predict reservoir production is based on the Baye’s theorem. Baye’s theorem applied to reservoir forecasting, samples parameters from a prior understanding of the uncertainty to generate reservoir models and updates this prior information by comparing reservoir production data with model production response. In automatic history matching it is challenging to generate reservoir models that preserve geological realism (obtain reservoir models with geological features that have been seen in nature). One way to control the geological realism in reservoir models is by controlling the realism of the geological prior information. The aim of this thesis is to encapsulate sedimentological information in order to build prior information that can control the geological realism of the history-matched models. This “intelligent” prior information is introduced into the automatic history-matching framework rejecting geologically unrealistic reservoir models. Machine Learning Techniques (MLT) were used to build realistic sedimentological prior information models. Another goal of this thesis was to include geological parameters into the automatic history-match framework that have an impact on reservoir model performance: vertical variation of facies proportions, connectivity of geobodies, and the use of multiple training images as a source of realistic sedimentological prior information. The main outcome of this thesis is that the use of “intelligent” sedimentological prior information guarantees the realism of reservoir models and reduces computing time and uncertainty in reservoir production prediction.

The use of GIS tools in marine spatial planning has become widespread. Such tools however often prescribe sterilized zones from a developer’s perspective (e.g. protected areas) and use surrogate indicators of wave and tidal resource with these being used to suggest areas of likely commercial development. The work undertaken in this thesis follows a more dynamic approach which has developed software to model the development appraisal process for wave and tidal projects. This means the most economically feasible sites for development can be located, taking into account factors (such as cable costs) ignored where resolve parameters alone are used in marine spatial planning. Moreover the model developed enables contrasting scenarios for differing harvesting technologies, grid connection points, cable types, port facilities to be examined and for specific improvement plans for such infrastructure to be investigated.

This thesis has evaluated a wide range of techniques to militate one of the big challenges of petroleum industry, water production. The techniques discussed waterflooding management, all aim to reduce excessive water production. The injection and production history at a well and field level are most common available data in any oil field, especially when nowadays we can have these data in real time with the implementation of the digital oil field and the intelligent well completion. This research aims to understand the strength and weaknesses of the existing techniques and “repackage” them to provide an optimum combination for more effective waterflood management by analysing injection and production data history. The first part of this research reviewed, tested and compared the analytical techniques that have been previously used for analysing the injection and production. The methods studied fell in to two distinct classes: those that monitor the waterflood performance secondly, methods for determining the inter-well connectivity. The second part of this thesis showed that an improved workflow used the captured information from the phase one methods could be combined to give more effective waterflood management via combination of reservoir voidage management (RVM), water allocation management (WAM) and production allocation management (PAM). Finally, a semi-analytical method was introduced in this thesis for performing RVM. Two approaches were defined for WAM and new techniques developed for PAM, all of which employed only the production and injection history. The results from these techniques were compared with the more advanced reservoir simulation methodologies such as gradient free optimisation. This comparison showed the reliability of the proposed techniques.

A novel polynomial chaos proxy-based history matching and uncertainty quantification method is presented that can be employed for complex geological structures in inverse problems. For complex geological structures, when there are many unknown geological parameters with highly nonlinear correlations, typically more than 106 full reservoir simulation runs might be required to accurately probe the posterior probability space given the production history of reservoir. This is not practical for high-resolution geological models. One solution is to use a "proxy model" that replicates the simulation model for selected input parameters. The main advantage of the polynomial chaos proxy compared to other proxy models and response surfaces is that it is generally applicable and converges systematically as the order of the expansion increases. The Cameron and Martin theorem 2.24 states that the convergence rate of the standard polynomial chaos expansions is exponential for Gaussian random variables. To improve the convergence rate for non-Gaussian random variables, the generalized polynomial chaos is implemented that uses an Askey-scheme to choose the optimal basis for polynomial chaos expansions [199]. Additionally, for the non-Gaussian distributions that can be effectively approximated by a mixture of Gaussian distributions, we use the mixture-modeling based clustering approach where under each cluster the polynomial chaos proxy converges exponentially fast and the overall posterior distribution can be estimated more efficiently using different polynomial chaos proxies. The main disadvantage of the polynomial chaos proxy is that for high-dimensional problems, the number of the polynomial chaos terms increases drastically as the order of the polynomial chaos expansions increases. Although different non-intrusive methods have been developed in the literature to address this issue, still a large number of simulation runs is required to compute high-order terms of the polynomial chaos expansions. This work resolves this issue by proposing the reduced-terms polynomial chaos expansion which preserves only the relevant terms in the polynomial chaos representation. We demonstrated that the sparsity pattern in the polynomial chaos expansion, when used with the Karhunen-Loéve decomposition method or kernel PCA, can be systematically captured. A probabilistic framework based on the polynomial chaos proxy is also suggested in the context of the Bayesian model selection to study the plausibility of different geological interpretations of the sedimentary environments. The proposed surrogate-accelerated Bayesian inverse analysis can be coherently used in practical reservoir optimization workflows and uncertainty assessments.

No description available.

Naturally fractured reservoirs (NFR) contain a significant amount of remaining petroleum reserves and are now considered for Enhanced Oil Recovery (EOR) schemes that involve three-phase flow such as water-alternating-gas (WAG) injection. Accurate numerical simulation of flow in NFR is essential for sound reservoir management decisions to maximise oil recovery and minimise the cost of field development. In this thesis, two important issues related to flow simulation in NFR are investigated. First, a step-wise upscaling approach is developed to evaluate the accuracy of dual porosity models in estimating matrix-fracture transfer duringWAG injection. It was found that the classical dual porosity models generally overestimate recovery from matrix blocks. Hence, a double block model was developed and extended to a multi-rate dual porosity (MRDP). The multi-rate double block model showed significant improvements in matching detailed fine grid simulations of three-phase matrix-fracture transfer. Second, the accuracy of upscaling discrete fracture networks (DFN) is assessed and its impact on history matching was investigated on a real fractured reservoir. A new method to upscale the shape factors needed for MRDP models from DFN is presented. This method is a notable step towards more accurate but still efficient reservoir simulation in NFR.

This study uses an extensive data‐set from a sand‐rich contourite depositional system located in the Gulf of Cadiz. Newly acquired data, targeting a modern contourite sandy depositional system in the eastern Gulf of Cadiz, has been integrated with historical data, including a seismic survey over a Pliocene mixed system in the northern Gulf of Cadiz. Seismic, core and sediment analysis has been used to assess the controls on the system, and characterise the sediments. The presented study gives a new complete view of past circulation of the Mediterranean Outflow Water and contourite deposition in the eastern Gulf of Cadiz over the upper, mid and lower slope. It also identifies a mixed system in the Pliocene section of the northern Gulf of Cadiz. This allows us to make the following contributions to the field; 1. A new model from Quaternary evolution of the eastern sandy contourites, and Pliocene evolution of a sandy mixed system the northern Gulf of Cadiz. 2. A detailed examination of the sand facies and sediment of the Gulf of Cadiz and a new set of contourite sand facies models proposed. 3. A thorough examination of the controls on contourite formation, and their integration in a sequence stratigraphic framework. 4. A review of the hazards bottom currents pose to deep water exploration, and the reservoir potential of contourite sands and other facies. The Gulf of Cadiz provides a good modern day analogue for contourite sands. They show various facies and sedimentological characteristics that are tied to depositional processes. They are also highly cyclic in nature, on a variety of time scales. If they can be positively identified in the subsurface, they have the capacity to make potential hydrocarbon reservoirs. The facies and sedimentological models presented here, and their integration into a sequence stratigraphic model will aid the positive identification of these deposits in the future.