Drilling

Integrated Approach for Overpressure Prediction in an HP/HT Well Offshore Malaysia

Pore-pressure prediction plays an important role in well planning as exploration targets shift to deeper, overpressured reservoirs.

jpt-2017-04-hphtchallengesfocushero.jpg
Source: Getty Images.

Pore-pressure prediction plays an important role in well planning as exploration targets shift to deeper, overpressured reservoirs. Pressure-related problems in such zones are mainly associated with narrow operating windows, resulting in severe well-control incidents, sometimes even leading to early abandonment. Uncertainties in prediction models arise from input data, assumptions used in the work flow, and the complexity of the geological or structural conditions. It is important to analyze these uncertainties and develop an understanding of them before drilling.

Introduction

Central Luconia is a geological province of the Sarawak Basin characterized by extensive development of Miocene to Holocene carbonate buildups. The buildups reveal complex seismic geomorphologies, facies, and structural setting. There have been several discoveries in the shallow-water (approximately 90-m depth) area of central Luconia province offshore Sarawak. A predrill pore-pressure- and fracture-gradient-prediction study was initiated for a planned high-pressure/high-­temperature (HP/HT) well with three offset wells (Offsets 1, 2, and 3) drilled in the vicinity of the study area. The objective of the study was to help in well design to reach desired deeper target reservoirs. The prospect is situated in a structure characterized by a northwest/southeast normal fault. It was formed on the foot wall block for the shallow (post-carbonate clastic) reservoirs and crossing the fault to the hanging wall block to hit the main targets in deeper (precarbonate clastic) cycles. The lithology for the shallow cycles is interbedded fine-grained siliciclastic shale and sand with limestone streaks. The deeper cycles consist of predominantly sandstones with interbedded shales. Two offset wells (Offsets 2 and 3) penetrated a thick carbonate sequence, and the other offset well (Offset 1) encountered a thick sand/shale sequence with interbedded carbonate stringers, which is similar to the prognosed-well result. Therefore, in the predrill modeling, there was only one relevant offset well (Offset 1) available in the study area that penetrated similar lithostratigraphic units.

Well Issues

A detailed analysis of all the drilling problems encountered in the offset wells was catalogued. This was achieved by reviewing the daily drilling reports, final well reports, and mud-logging reports and documenting all the events on a daily basis. Overall, all the offset wells analyzed during the study indicated several drilling issues related to pore pressure, including internal blowout with violent gas flow from the bell nipple, influxes, gas-cut mud, high gases, connection gases, pumps-off gas, and cavings. In addition, wells showed significant challenging drilling conditions such as equivalent static- and circulating-density contrast because of high temperature, wellbore breathing caused by a narrow mud window, and several hole-stability issues such as tight hole, lost circulation, and stuck pipe, eventually leading to multiple sidetracks.

Apart from the offset-well-drilling challenges, several other challenges, such as lack of calibration for deeper undrilled cycles, a well penetrating reservoirs in different fault blocks, lithological variation across the field, overpressure top and magnitude, reservoirs with centroid effect, and the quality of 2D-seismic velocities, were also considered during the predrill modeling.

Process

Considering uncertainties from several sources, several scenarios of predrill pore-pressure and fracture-­gradient models were considered. It was also recommended to update the predrill model as soon as the new data became available during drilling. This actually helped to reduce the uncertainty window during drilling.

Offset-well pore-pressure analysis was carried out using petrophysical logs calibrated with direct pressure measurements and reported drilling events. Overburden estimation at the offset wells was performed using the available density logs. A velocity-to-density transform using the velocity was carried out in all the offset wells, and a calibration was established from all the offset wells. This transform was used to calculate pseudo­densities for the sections where density logs were ­either absent or not reliable. A semi­regional density trend using all ­offset-well density logs was also established and used for the shallow section from the mudline to the top of the first available actual log data. The estimated profiles showed a well-constrained overburden across the offset wells because there is negligible difference in water depth. For pore-pressure estimation, Eaton’s method was used for resistivity and acoustic, while the ­equivalent-depth method was used for density. Similar compaction profiles were created for all the offset wells. An in-depth analysis to understand the prevalent pore-­pressure regimes in terms of stratigraphical and structural settings helped to generate different scenarios of possible shale-pressure evolution as well as sand/shale pressure relationships that could be expected with respect to the regional geological understanding. The top of overpressure is significantly shallow in the relevant offset (Well 1) as compared with the other two offsets (Wells 2 and 3), where the top of the overpressure is much deeper. The occurrence of overpressure at shallow depth in Well 1 could possibly be attributed to a thick dominantly shale/sand sequence that is replaced by carbonate in the other two offset wells that show a lesser degree of overpressure. There was a significant challenge in modeling the top of the overpressure and its magnitude in the shallow section at the relevant offset (Well 1) because of the nonavailability of any pressure measurements and the poor quality of log data. This well was drilled in the 1970s, with very limited information available. Drilling events were used as a guide to model pore pressure in the shallower part of the well. The petrophysical log data appear to be useful for predicting overpressure in shale/claystones but are of very limited use because of dominantly nonshaly lithology such as sands and carbonates.

In the next phase, a seismic-­velocity-based analysis was carried out at the relevant offset (Well 1) with the existing 2D-seismic velocities. The key idea of selecting Well 1 was to understand the ­seismic-velocity predictive quality in view of the depth of the overpressure top and overpressures in the deeper undrilled sections possibly encountering similar lithofacies. A comparative analysis of the velocities showed that the seismic is similar to acoustic velocities in the shallower section. In the deeper section, the seismic velocity is faster compared with acoustic and checkshot velocities acquired in Well 1. This necessitated using a different compaction trend than that from acoustic velocities, with certain adjustments to match with actual pore pressure for seismic-based predictions. Keeping in mind the results from Well 1, it was expected that use of existing 2D-surface-seismic velocity for predrill prediction might provide reasonable results. The data quality was good in the shallow section, where it was characterized by a continuous-frequency reflector. In the middle and deeper part, where it was dominated by lower-frequency discontinuous events, the data quality can be classified as moderate to challenging to interpret. The well was expected to cross a fault that would penetrate the shallow and deeper reservoirs across different fault blocks, which added a considerable amount of uncertainty to seismic-based predictions. To improve on the quality of prediction and reduce the uncertainty, seismic velocities were extracted from the upthrown and downthrown side of the fault block to determine the lateral and vertical variations. In addition, very-coarse seismic-velocity points were available for analysis, which was challenging to compare with the high-frequency acoustic velocities of the offset well. Also, the velocity model was generated with a single time/depth-function method obtained from offset wells. A best-fit velocity model was developed on the basis of these data and used for the depth conversion to generate a depth-structure map and well-depth prognosis. The offset wells did not penetrate the deeper cycles; hence, time-to-depth conversion performed on the basis of the offset-well function carried a considerable uncertainty with depth. On the basis of the seismic-velocity variations around the planned locations and offset-well experiences, three possible pore-pressure scenarios were predicted for each prospect. The fracture gradient was estimated on the basis of the effective-stress-ratio approach by use of the leakoff-test (LOT) results available from Well 1. A similar three-case scenario for fracture pressure was proposed for different cases of pore pressure.

The well was designed to drill to a depth of 3500 m with an overpressured Miocene/Oligocene sediment column of 2500 m (water depth of 90 m, maximum prognosed pressure of 10,000 psi, and temperature of approximately 190°C). An integrated approach was adopted through a robust predrill analysis with real-time monitoring of the well and updating of the predrill model wherever required by use of the new information. On the basis of post-drill analysis (Fig. 1), the actual pore pressure was within the uncertainty model, which helped in making decisions on well design in real time. The predrill model predicted a manageable mud window, with a very narrow mud window for deeper cycles. The well was terminated within the last cycle because of the nondrillable window, which was already prognosed by the predrill high-case scenario.

jpt-2017-04-18909fig1.jpg
Fig. 1—Comparative analysis of post-drill and predrill analyses. PP=pore pressure. MW=mud weight. ECD=equivalent circulating density. 

 


This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper IPTC 18909, “An Integrated Approach for Overpressure Prediction in a Wildcat High-Pressure, High-Temperature (HP/HT) Exploration Well Offshore Sarawak, Malaysia,” by Avirup Chatterjee, Amitava Ghosh, and Sanjeev Bordoloi, Baker Hughes, and Anifadora Bt. Mustapha, Petronas, prepared for the 2016 International Petroleum Technology Conference, Bangkok, Thailand, 14–16 November. The paper has not been peer reviewed. Copyright 2016 International Petroleum Technology Conference. Reproduced by permission.