This paper describes the acquisition and interpretation of long-term pressure-buildup data in a plugged and abandoned deepwater appraisal well. To accomplish the test objectives at an acceptable cost, a novel combination of well testing, wireless-gauge technology, and material-balance techniques was used to allow the collection and interpretation of reservoir-pressure data over a planned period of 6 to 15 months following the well test. The final buildup duration was 428 days (14 months).
Three interpretation methods of increasing complexity were used to provide insights into the reservoir. First, material balance was used to produce an estimate of the minimum connected reservoir volume. The advantage of material balance is that it requires very few input assumptions and produces a high-confidence result. Second, analytical models in commercial pressure-transient-analysis software were used to investigate near-wellbore properties and distances to boundaries. Finally, finite-difference-simulation models were used to investigate reservoir properties and heterogeneity throughout the entire tested volume. With increasing model complexity came additional insights into the reservoir properties and architecture but reduced solution uniqueness.
A key complication for the interpretation of the recorded pressure data was the potential for gauge drift. This was incorporated into the uncertainty range used in all three interpretation methods.
Analysis of conventional well-test designs (with varying flow rates and buildup periods) showed that the cost of resolving the key uncertainties exceeded the value of information significantly. To justify the appraisal, a way was needed to extend either the flow period or the buildup period without a rig on station and with the well left in a permanently abandoned state. To meet this objective, the potential of wireless-gauge technology to extend the buildup length was evaluated. Two competing wireless technologies were available, acoustic and electromagnetic transmission, both occurring up the tubing/casing. The key differentiator was that acoustic transmission required that cables be run through any cement plugs, which violated the barrier standards for abandoned wells. Accordingly, electromagnetic transmission was selected for the final system. The post-abandonment well concept is shown in Fig. 1. Of note is that the wellhead was not recovered and the top of the 20- and 36-in. casings have not been severed.
One critical design feature was the use of redundant gauges (four), repeaters (four), and subsea modems (four) to ensure no single point of failure existed within the wireless system. This also resulted in a narrowing of the gauge-drift and accuracy-uncertainty range as the response of individual gauges was thought to be independently and identically distributed.
Data were recovered from the wireless gauges in three tranches (corresponding to visits by a supply vessel to the well vicinity), with the subsea modems recovered to surface as part of the final tranche.
The well-test design called for an initial cleanup flow, initial buildup, multirate test, extended high-rate-flow period, and a final buildup. For operational reasons, an additional shut-in during the extended flow period was necessary to restock methanol for hydrate inhibition on the rig. The final buildup was split into three components: The first 24 hours were planned to be recorded with conventional drillstem-test (DST) gauges run on the completion to provide high-frequency data that could be interpreted for kh (the product of formation permeability and thickness) and any near-wellbore features. This was to be followed by a break in the recorded data while the long-term wireless-gauge system was installed and commissioned. The final stage was the long-term buildup recorded by the wireless gauges, which had a much lower acquisition frequency.
Interpretation of the high-frequency-gauge data provided information on the skin and non-Darcy skin, formation permeability, heterogeneity (through indications of two-layer behavior with crossflow away from the wellbore), and the two closest boundaries.
Inclusion of the wireless-gauge data also allowed for resolution of the third boundary and provided an indication of the location of the fourth boundary, but the derivative response at the time of the last data point was still not indicative of a fully bounded system. Inclusion of the wireless-gauge data in the interpretation increased the distance investigated by the well test by a factor of ten, with the final path length to the outermost boundary estimated to be 15 km. The full test pressure match, however, indicated that the gas volume within the analytical model was a significant overestimate of the actual gas volume, which is consistent with the authors’ understanding of this environment of deposition.
Use of wireless gauges allowed the buildup duration to be extended from the 3 days used on previous DSTs in the region to 428 days, with the available rig day allocation used to produce a larger-pressure perturbance. This allowed the well test to investigate a reservoir volume similar to planned development wells (with the outermost boundary being 15 km from the well).
This paper shows how a progression of interpretation techniques of increasing complexity allowed additional insights to be drawn from the collected data but at the expense of increasing complexity, more degrees of freedom, and additional external constraints. While the more-complex techniques provide additional insights, the simplest technique provides a higher degree of confidence in its (more-limited) conclusions.
石油圈
