Improvement of Reservoir Management Efficiency Using Stochastic Capacitance Resistance Model

Abstract

Lost Hills is a dual permeability Diatomite reservoir that is distinct from conventional reservoirs. Application of numerical simulation has been limited throughout field life due to the complex nature of the diatomite including low permeability (∼ 1 md) but high porosity (∼ 50%) and weak rock strength (∼ 100,000 psi of Young's modulus). Thus, many reservoir management practices are based on trial and error methods which are sub-optimal. This work aims to enhance the efficiency of reservoir management activities. The usage of Capacitance Resistance Model (CRM) has been increasing due to its simple but effective capabilities of analyzing waterflooding performances. However, CRM has innate limitations as well. The core calculation of CRM is solving nonlinear regression. Solution of nonlinear regression is not unique since it is not guaranteed to find the global minimum and is affected by solver algorithms and initial guesses. In addition to the innate limitations, due to the lack of bottomhole pressure data in the Lost Hills and its high oil viscosity (∼ 20 °API), the accuracy of Lost Hills CRM solution is not enough to be used in daily operations. Stochastic CRM (SCRM) was developed to mitigate these limitations by combining bootstrap with CRM and provides stochastic answers. SCRM estimates probabilities of an initial solution using bootstrap, substitutes low probable parameter values with P50 values, and updates injector-producer connection pairs and its interwell connectivity. SCRM was developed for analyzing waterflooding operations such as identification of ineffective injector-producer connection pairs and estimation of reservoir pressure. SCRM analysis results were benchmarked against the Lost Hills tracer test data and demonstrated that SCRM provided a better solution than CRM. Compared to the fact that the deterministic solution from CRM found only 50% of connection pairs which tracer identified, SCRM solution identified 10 out of the 12 tracer identified connections. After the verification, SCRM was applied to find out connected injectors which cause Fluid Over Pump (FOP) wells. The existing workflow for identifying connected injectors was a trial and error method and hard to find connected injectors if connected injectors are located farther than 300 ft from FOP wells (chronic FOP wells). The novel workflow has been deployed so that FOP wells can be mitigated systematically and enable the optimization team to improve its reservoir management efficiency. In 2017, 10 chronic FOP wells were mitigated by identifying connected injectors with the novel workflow.