Analysis of produced water data to model and identify geochemical reactions occurring in the reservoir
Water injection has been commonly used to maintain reservoir pressure and improve oil recovery in many oil fields. The equilibrium between reservoir rock and formation brine at initial reservoir conditions would be disturbed by the introduction of a non-native water, which mixes and reacts with the formation brine and interacts with minerals present in the formation. A series of brine/brine and/or brine/rock interactions would then take place in the flow paths from the injection well to the production well, potentially leading to dissolution and/or precipitation of minerals; knowledge of these interactions is crucial for the evaluation and the management of oilfield scale problems. In this thesis, thermodynamic models, reactive transport models and reservoir simulation models are used to identify the geochemical reactions occurring in the reservoir and investigate how the reservoir interactions affect the produced water composition. Brine composition data have been collected from 26 fields, and examples from four selective fields provide the basis for the analysis in this thesis. For Field X, a typical sandstone reservoir located in the North Sea region, a thermodynamic prediction model was used to calculate the risk of scale precipitation based on a series of individual produced water samples, thus providing an evaluation of the actual scaling risk in these samples, which is then compared with the usual theoretical estimate based on endpoint formation and injection brine compositions, and the erroneous assumption that no reactions in the reservoir impact the produced water composition. The occurrence of barium sulphate precipitation and calcium magnesium ion exchange reaction are identified by the modelling results. The Cation Exchange Capacity was identified as a modest 0.05 mol/kgw (50 meq/L) for this field. Since ion exchange capacity is an important parameter for some chemical EOR method, this a promising technique for EOR evaluation. An available history matched streamline reservoir simulation model of the Miller Field was then integrated with produced water chemical data. Streamline simulation is applied to better model brine mixing through reducing the numerical dispersion which cannot be effectively controlled in finite difference simulation. A simplified model of barite scale precipitation was included in the streamline simulation, and the calculation results with and without considering barite precipitation were compared with the observed produced water chemical data. The streamline simulation model assumes scale deposition is possible everywhere in the formation, whereas in reality the near production well zones were generally protected by squeezed scale inhibitor, and thus the discrepancies between modelled and observed barium concentrations at these two given wells diagnose the effectiveness of the chemical treatments to prevent formation. 1D and 2D reactive transport models were developed to identify the geochemical reactions occurring in the Gyda Field where there is a high reservoir temperature and formation water is high salinity. Anhydrite and barite precipitation are identified as the two dominant mineral reactions taking place deep within the reservoir. Anhydrite is deposited due to mixing of formation and injection waters in the area before this zone is cooled, and the precipitated anhydrite is gradually dissolved as the local reservoir temperature is lowered by cool injection water. The dissolved anhydrite then re-precipitates downstream in the at high temperature zones since the propagation of the temperature front is much slower than the brine mixing front. This creates a risk of late life anhydrite deposition in the producer. Finally, a carbonate reservoir study was performed for Ekofisk field where seawater flooding has been implemented. The 1D reactive transport model provides a good match with observed produced water chemistry data when the primary calcite mineral phase, calcium magnesium carbonate precipitation, temperature change and initial source of CO2 were modelled. In Ekofisk Field, calcite dissolution drives anhydrite and calcium magnesium carbonate precipitation. The modelled combination of calcium magnesium carbonate precipitation and ion exchange remove magnesium from the brine, also as observed from the produced water data. Simulation results also demonstrate that calcite dissolves quickly at first due to CO2 partitioning from the hydrocarbon phase into the brine. It was also shown that calcite dissolution is promoted by an increase in sulphate concentration in the injection water due to the coupled anhydrite precipitation. This body of work develop a methodology for systematically storing and analyzing produced brine data, and using modelling tools to identify what geochemical reactions have taken place. The methodology is then applied to various reservoir scenarios, leading to insights that impact scale management in these systems, and may also have a bearing on chemical EOR methods.