Perpustakaan Digital ITB

Multi-stage hydraulic fracturing is standard stimulation technique for improving productivity of horizontal wells completed in tight gas reservoir, yet the number of fracture stage that maximizes gas deliverability is not self-evident: under a fixed proppant mass constraint, increasing stage count reduces the propped half- length available per stage. This study evaluates the effect of fracture stage number and fracture geometry on the productivity of a horizontal well in the tight gas Y Field (Natuna Sea) using numerical simulation in CMG-IMEX, in place of the analytical Unified Fracture Design approach applied to the same field data in prior work. A no-fracture base case was benchmarked against the analytical horizontal-well solution across a bottomhole pressure (BHP) range of 4000–750 psi, yielding a maximum deviation of 4.70% at 4000 psi and averages below 2% at lower BHP, consistent with the transition from early transient flow toward the pseudosteady-state assumption embedded in the analytical model. Six fracture-stage scenarios (n = 8, 11, 14, 17, 20, and 21) were simulated and cross-checked against the multi-stage fracturing model, with deviations remaining below 5% (maximum 4.89% for n = 21). The resulting inflow performance relationship (IPR) curves intersected across the tested BHP range: n = 11 delivered the highest gas rate at BHP ? 2000 psi, while n = 14 became dominant at BHP ? 1000 psi, producing 25,038.1 MSCF/day at BHP = 1000 psi, 2.7% higher than n = 11. Because low-BHP operation represents the majority of the well's producing life under reservoir depletion, n = 14 was selected as the best case, a choice independently corroborated by the prior analytical study, which identified n = 14 as the only configuration achieving full fracture penetration at the maximum permissible half-length of 750 ft. Geometric optimization of the n = 14 configuration was then performed in CMOST using the Design Exploration Controlled Evolution (DECE) algorithm across four parameters — fracture half-length, height, width, and permeability — yielding 22 converged experiments. Sobol sensitivity analysis showed fracture half-length accounting for approximately 99% of the variance in gas rate, reflecting the dominance of matrix-fracture contact area over fracture conductivity in this very-low-permeability system. The resulting proxy model achieved R2- training of 0.951 and R2-adjusted of 0.946, with verification-point deviations averaging 0.54%. The optimum configuration increased gas rate by 5.12% relative to the n = 14 base case, with the optimal half- length located at an interior point of the tested range. These results indicate that, under fixed proppant constraints, gas deliverability in this formation is governed primarily by achievable fracture penetration rather than stage count or conductivity, and that relying on a single high-BHP criterion can lead to a suboptimal fracture-stage selection for long-term well performance. These findings provide a quantitative basis for prioritizing fracture penetration over stage count and conductivity when proppant volume is constrained in similarly tight, low-permeability horizontal wells.