Performance prediction of a reservoir under gas injection, using output error model

The pursuit of knowledge on the future performance of reservoirs is a longstanding subject in the downstream oil and gas industry. Current methods employed are time-consuming and sometimes less reliable. In this study, we propose output error model structure for developing a forecasting model of a reservoir under gas injection. A case study which involves a synthetic reservoir model under gas injection has been utilized to prove the hypothesis that system identification using output error model structure is useful for performance prediction of a reservoir under gas injection. The case study was a medium sized reservoir model with 18017 active cells. It consists of five gas injectors and seven producers. Output error model structure was utilized to develop forecasting models for oil production rate, gas-oil ratio, and water-cut. The models were cross validated using dataset which were not used during modelling. The result of the study confirms that system identification by output error model structure has a very good potential for use in reservoir performance prediction under gas injection. © 2016 Berihun M. Negash et al.