Gas Reservoir Management: How to Improve Gas Production Allocation Per Reservoir in Commingled Wells Using Geochemistry Technology?

Abstract

Abstract In a commingled gas reservoir, it is crucial to determine the respective contribution of each produced reservoir for an efficient and reliable Reservoir Management, and to match the reservoir models. The Production Logging Tool (PLT) is commonly used to allocate the production to the individual layers. The limit of this process is that PLT results and therefore the layer-based equation change with time due to the evolution of reservoir characteristics such as reservoir pressure, mechanical skin, non-Darcy coefficient and sometimes PVT properties. PLT acquisition must be repeated to quantify this evolution and to determine time-dependent layer allocated equations. The repetition of PLT acquisition must be optimized taking into account the needs and the impact of time on results, but is as well highly constrained by cost and logistics, which prevent is some cases to get the desired frequency. Geochemistry technology can provide additional and reliable data on the hydrocarbon fluid provenance in order to improve accuracy of the layer-allocated gas rate equations. Indeed sampling and geochemical analysis can be done at a higher frequency and much lower cost between each PLT. Fingerprinting is one of well-known techniques, which was tested recently in Qatargas in cooperation with Total and results appeared to be encouraging. All the analyses and data processing of the condensate samples were achieved in Qatar and a specific methodology was developed to get the best layer-based allocation process. This geochemical approach is based on a statistical interpretation of High Resolution Gas Chromatography (HRGC) data of the condensate fraction from the produced fluids. There are several prerequisites for such a method to be applicable: identical separator conditions, excellent sample storage and handling, highest GC standards, appropriate statistical tools and availability of representative reservoir fluids. The individual reservoir fluids are considered as end-members which must be known to determine quantitatively the contribution of each reservoir in the produced well fluid. This paper describes the details of the fingerprinting technique used, including how the samples are analyzed and how data are processed in an innovative way to get the most reliable results. In our case, the compositions of the different condensates show very subtil differences. As a result, we had sometimes to consider more than 1000 chromatograms peak ratios to differentiate the fluids coming from different layers. We concluded that geochemical production allocation was possible in some (but not all) areas of the field and that results were consistent with other types of dynamic and geological data. These positive conclusions allowed us to envisage an operational application of this research feasibility study.