Optimization of Gas Injection Method Using New Approach: Integrated Production Modeling

Nowadays, according to the exponential growth of oil production, the subjects related to the enhanced oil recovery such as gas injection have become more important. In this regard, selection the appropriate scenario with meeting the operational and injection constraints is an essential task. These main constraints in gas injection process are: the volume of injected gas, the amount of allocated gas between the injection wells and the proper production rate of producing wells. In this paper, the impact of all of these factors (as input factors) was examined using the Net Present Value (NPV) as an objective function (output factor). To reach the maximum value of the NPV, the developed set of petroleum engineering software tools of the PETEX Company were employed to model both the production and injection systems as an integrated system.Integrated system has the ability to increase the reliability and comprehensibility of the results, since several components of the injection and production model can be considered simultaneously. To reduce calculations run time, and also to find out the relationship between inputs and output, proxy model was utilized. The proxy model was built by composition of Box- Behnken design and constrained mixture design. Using an adaptive network-based fuzzy inference system (ANFIS) and by training, testing and validation of the neuro-fuzzy network, the impacts of the inputs on the objective function were understood and the optimal value of each factor was acquired.Results showed that despite the rising of injection costs, increasing the gas injection volume contributes to the more revenue as the enhanced produced oil greatly overcomes the injection costs. Moreover, based on the optimization, the first injection well requires 0.35 the second well needs 0.65 of gas injection volume, respectively. This matter displayed the importance of wells position effect on the objective function. In addition to this, the best production rate was determined as 6050 bpd, in order to prevent sudden reduction of reservoir pressure during oil production period. This study revealed that the error value using the fuzzy neural network is really negligible, and hence, the integrated production modeling and the fuzzy neural network has correctly helped determine the examined function variables. Furthermore, the NPV was improved percent compared to the base case (equal gas injection between wells).

Optimization of Gas Injection Method Using New Approach: Integrated Production Modeling Keywords:

Gas Injection , Integrated System , PETEX Software , Proxy model , ANFIS