A case study on improving waterflood surveillance aided by a better understanding of the correlation between various water injectors and oil producers completed in the shallowest sub-layer of a giant multi-layered matured carbonate reservoir in Mumbai Offshore Basin is presented here. This understanding is then used to gauge effectiveness of the prolonged waterflood programme and to identify ‘target wells’ for optimizing water injection rate. The inferences of this analysis were tested using a simulation model.
Production, injection and pressure data of all wells completed in this sub-layer were extracted. The reservoir injection and withdrawal rates were computed using PVT data which were subsequently fed into an in-house developed streamline simulation program that generates a matrix of flow-based well rate allocation factors (WAF) correlating injection to withdrawal for each individual well as a part of its output. The analysis of injection efficiency per well was carried out in two scenarios viz. with current rates for effective waterflood surveillance and at a cumulative level with averaged rates to identify areas of deficiencies and optimize future injection rates.
Flow-based allocation factors provided a better picture than traditionally employed distance weighted technique owing to the underlying physics involved in describing streamline distribution in the reservoir. Results of analysis at the cumulative level indicated wells where injection efficiency, as measured by the ratio of injection rate to sum of streamlines-weighted withdrawal rates from connected producers, substantially deviates from 1. Few wells had an injector efficiency significantly higher than 1 which defined over-injection and potential recycling while a large number of injector wells had ratios of less than 1, highlighting the need to step-up injection rates and devise strategies for rigorous surveillance. To achieve the latter objective, injection-centric WAF's were regenerated at current situation with current rates and the dynamic nature of these factors could be observed by noting their slight difference with respect to previously estimated factors. This is attributed to averaged-out flow rates limiting the influence of newer high-rate producers and injectors. Nonetheless, wells in areas demanding attention are identified and requisite injection rates are assigned. These changes are included in the history-matched simulation model used for redevelopment activities and results were compared with a do-nothing case. The significant incremental recovery proves as a validation of the methodology adopted.
Waterflood surveillance on a well-to-well basis is always difficult in a matured field where water injectors are deployed in a ubiquitous fashion. This approach has rarely been employed in a reservoir of the size of Mumbai High and can be extended to other sub-layers subject to positive results from field implementation. Thus it is an endeavour to monitor waterflood effectiveness at a large field scale and could be beneficial for similarly developed fields.