A Fast and Reliable Network Solver Using Numerical Proxy for Reservoir Simulation

Abstract

Abstract Reliable and efficient modeling of subsea production systems is critical for predicting performance of offshore reservoirs. Iterative numerical procedures are typically required if the production system includes flow networks where wells may compete and interact with each other. The network problem can become challenging since production constraints, hydraulic flow stability, and other issues need also to be considered when solving for numerical values of pressure and rates. In this work, a new technique, called stability proxy, for solving flow network problems in reservoir simulation has been developed by generalizing the concept of inflow performance relationship (IPR). IPR is traditionally used for representing the quantitative relationship between flow rate and bottom-hole pressure (BHP) or tubing head pressure (WHP) of a well. With stability proxy, IPR-like relationships are constructed for the entire production network, including wells and facilities downstream to wellheads. This approach handles production constraints and hydraulic stability and employs approximation techniques for computational efficiency. Stability proxy offers several advantages compared to traditional techniques for solving flow network problems. First, stability proxy method determines pressure/rate operating ranges for wells and locations downstream in a flow network. Given a pressure or rate boundary condition, the computational procedure identifies wells that are open to production and wells which cannot flow due to lack of pressure support, flow instability, and other reasons. After well state is determined, rates and pressure anywhere in the flow network can be easily determined from the stability proxy dataset. In contrast, traditional methods for solving the facility network problem start computations assuming wells are open and often require good initial guess to converge successfully. The stability proxy method does not require initial guess, eliminating numerical artifacts from the solution process. Second, performance of traditional iterative approach often degrades when wells are operating near the stability limit, and multiple network calls are often required to ensure that well production allocation is consistent with the network model downstream to wells. With the proxy algorithm, computational efficiency is significantly improved since proxies can be reused in many situations without compromising accuracy. Using stability proxy results, a semi-analytical method is derived for obtaining derivatives of production rates with respect to flow control variables, which provides a fast and convenient tool for quantifying well interference. Effective gas-oil ratio (GOR) and water-cut (WCUT) accounting for well interference are defined for flow networks based on the derivative results. Tests show that effective GOR and WCUT can be used as a more realistic basis for production allocation in field management.