Oilfield sustainable development: An operation optimization study for the water-flooding pipeline network system based on data fusion

Abstract

Water flooding is a widely employed technique for enhancing oilfield recovery, and the oilfield water-flooding pipeline network system (OWPNS) is a major energy consumer in the oilfield. However, the difficulty in describing the injection-production relationship and the complex network structure present challenges for optimizing the operation of OWPNS in current studies. A multi-objective mixed-integer nonlinear programming model was developed, incorporating the reservoir characteristics described by the injection-production relationship and considering various constraints, such as injection well operation, pressure balance, and wastewater treatment process. Due to the nonlinearity and multi-objective nature of the proposed model, the piecewise linearization method and augmented epsilon-constraint method were employed to convert the optimization model into a mixed-linear programming problem. In addition, the reservoir system was conceptualized as a signal system, and the injection-production relationship was quantitatively evaluated using the Extended Kalman Filter, thereby providing boundary parameters of the injection-production relationship for the optimization model. Numerical experiments demonstrated that Extended Kalman Filter not only effectively quantified the injection-production relationship but also tracked its variations over time. Furthermore, case studies showed that the optimization model exhibited good applicability to complex network structures. For example, stop valves effectively regulated the flow distribution, and storage tanks played a crucial buffering role, enhancing operational flexibility. This study achieved the coupling of the OWPNS with reservoir characteristics and offers valuable insights for developing a digital twin-driven intelligent water-flooding system.