Sidetrack during field development and ongoing production arises to exploit bypassed reserves (unswept areas under secondary and/or tertiary recovery), unexploited zones and unforeseen conditions likely to build due to uncertainties and heterogeneity in initially characterizing a reservoir. Whereas recompletion is prone due to sequential production of stacked reservoirs or multiple pay zones that is necessitated by regulation on comingling. The purpose of this paper is to optimize the time for sidetrack/recompletion job in multiple pay zones in view of maximizing returns on investment. A simple case to elicit the workflow is based on two pay zones and applies Experimental Design (ED), and economic analysis in the form of Net Present Value (NPV). The procedure followed for this study started with the identification of the reservoir uncertainties and their ranges; focusing basically on net pay thickness, porosity, permeability, and time to perform the sidetrack; the reservoirs are assumed to be undersaturated. The Box-Behnken response surface design in ED was used to reduce the number runs to make by generating the most effective combination of variables for the experiment. Experimental runs were conducted with a Black Oil reservoir simulator to give production profile. NPV values using stipulated Oil price, CAPEX, OPEX, tax rate, royalty and production rates were estimated, which was used afterwards to produce the Proxy model in ED and this developed correlation was used to analyse the effect of time changes. NPV computed from the Proxy Model was reasonable; however, a higher level D-Optimal design as against the 2-level design used in this study may be required for a reasonable match of NPV with respect to sidetrack (recompletion) time. In conclusion, the incorporation of time was successful and gives way to studying the impact of not only reservoir uncertainties but the uncertainty that arises from a success or failure of the sidetrack that further gives way to the application of decision analysis and the evaluation of NPV by Proxy Models for each possible outcome.

Sidetrack during field development and ongoing production arises to exploit bypassed reserves, unexploited zones and to tackle unforeseen conditions that are likely to occur due to uncertainties and heterogeneity in initially characterizing a reservoir. Whereas recompletion is due to sequential production of stacked reservoirs or multiple pay zones that is caused by regulation on commingling. The purpose of this paper is to investigate the application of experimental design (ED) in optimizing sidetrack/recompletion time of multiple pay zones in view of maximizing returns on investment by net present value (NPV) and expected monetary value (EMV). A hypothetical reservoir of two pay zones with uncertainties associated with pay thickness, porosity, permeability and time to perform the sidetrack was considered. The ED method of Box-Behnken response surface design was used to reduce the number of runs made by generating the most effective combination of variables for the experiment. Experimental runs were conducted with a Black Oil reservoir simulator to give production profile and the computed NPV was used afterwards to produce the Proxy model in ED. NPV computed from the Proxy Model was reasonable compared to that of the production profile from the simulator. However, a higher level D-Optimal or factorial design may be required for a reasonable match of EMV with respect to obtaining a realistic sidetrack/recompletion time.

Injection of carbon dioxide into deep saline aquifers is one way to reduce greenhouse gas emissions. Carbon dioxide, usually a super critical fluid at aquifer pressure and temperature conditions, is lighter than the resident brine and so forms a gas cap above the water. However, over time it dissolves in the water, creating a density inversion which induces gravitational instability. Understanding whether the dominant mixing mechanism is convective mixing rather than pure diffusion is important as this controls the timescale over which the carbon dioxide-saturated brine mixes with the unsaturated brine. This paper presents numerical simulations, using a finite difference reservoir simulator, to evaluate the predictions of analytical solutions for stability analysis and growth rate of the fingers of different wavenumbers at different Rayleigh numbers (Ra). The effects of density difference, permeability anisotropy and diffusion (both longitudinal and transverse) on fingering behaviour were investigated through the dimensionless Rayleigh number. The density difference and the vertical permeability were found to mainly control the degree of instability. At Rayleigh numbers greater than 800, fingers are present and the degree of fingering increases with Rayleigh number. Growth rate analysis showed that growth rate is directly proportional to Rayleigh number and time. The critical time (at which flow becomes unstable) varies inversely with the Rayleigh number whilst the corresponding critical wavenumber number varies linearly with the Rayleigh number. These results are consistent with previously reported linear stability analyses providing a validation of the simulator. Numerical simulation results were also validated against experiments. These validations both show that the simulator is robust and can thus be used to investigate more complex situations (heterogeneity) that cannot be analysed mathematically.