Benchmarking Recovery Factors of Individual Wells Using a Probabilistic Model of Original Gas in Place to Pinpoint the Good, Bad and Ugly Producers

摘要

Estimation of original gas in place (OGIP) in a rock volume provides an upper limit for the expected ultimate recovery. Calculation of OGIP for a drainage region may introduce significant errors when based on discrete values; the deterministic OGIP may be either overestimated (positively biased) or underestimated (negatively biased). For example, parameters like porosity, water saturation and adsorbed gas density may vary spatially, which must be accounted for to obtain realistic OGIP estimations. Our objective was to create a more accurate OGIP model and use it to probabilistically asses OGIP, estimated ultimate recovery (EUR) and recovery factor (RF) for shale gas reservoirs like the Marcellus shale. The conventional OGIP model was updated to include recent developments in shale geology and gas adsorption. Corrections to traditional free gas calculations are made by subtracting adsorbed gas reservoir volume from free gas volume in order to obtain improved OGIP estimations. This change was assessed in the context of Langmuir and BET isotherm adsorption models. A 25-year EUR response surface model was created using a semi-analytical model from our previous work. Both OGIP and EUR models were coupled during Monte Carlo simulation to produce a probability distribution for RF. When adsorbed gas was included in the pore space available for free gas, OGIP for the Marcellus was reduced 14% from previous estimates. With this model, changing from the traditional Langmuir isotherm to a BET isotherm resulted in a marginal gains in OGIP. Using the limited tuning parameters available for the BET isotherm, a 14-24% reduction in Marcellus OGIP was observed. The coupled OGIP-EUR simulation produced a P50 OGIP estimate of 1,320 Tcf, P50 EUR of 492 Tcf, and P50 RF of 38%.