A Theoretical Framework for Optimising Gas Processing Train Configurations Under Stochastic Feed Composition Variability in the Nigerian Gas Gathering and Processing System

Gas processing plant train configuration optimisation under conditions of stochastic feed gas composition variability is a challenging operational research problem that cannot be adequately addressed by deterministic design methods calibrated to design-point composition alone. Nigerian gas gathering systems aggregate feed gas streams from multiple reservoir sources with varying compositional properties, creating composition variability that propagates as performance uncertainty through the processing train. This study develops a theoretical framework for optimising gas processing train configuration under stochastic feed composition variability applicable to the Nigerian gas gathering and processing system. The framework integrates stochastic process simulation, multi-objective optimisation, and robust design theory. Feed composition variability is characterised from ten years of chromatographic data from Nigerian Gas Company field operations using copula-based multivariate statistical models that preserve inter-component correlations. Aspen HYSYS steady-state process simulations are embedded within a Monte Carlo-based stochastic optimisation framework using differential evolution algorithm to optimise train configuration variables including refrigeration cycle pressure levels, number of theoretical stages in fractionation columns, and lean oil recirculation rates. A novel robustness index combining expected value and variance of product specification compliance is derived as the primary optimisation objective. Application to a six-train gas processing facility delivering pipeline quality sales gas demonstrates that the robust optimal configuration maintains product specification compliance in 94 percent of simulated feed composition scenarios compared to 76 percent compliance for the deterministic optimal design. Keywords: gas processing optimisation, stochastic feed variability, robust design, Nigerian gas system, process simulation.