The Mangala Field is the largest discovered oil field in the Barmer Basin of Rajasthan, India with a STOIIP of over 1 billion barrels contained in stacked clastic fluvial reservoirs of the Fatehgarh Formation (Compton, 2009). The Fatehgarh Formation is divided into two broad reservoir units. The upper Fatehgarh (average net to gross ~40%), consists of low sinuosity channel system sands, and the lower Fatehgarh, consists of 30m to 40m of stacked braided fluvial sands (average net to gross ~90%) having resistivity of more than 5000 ohm-m.

The flow simulation model demonstrated that horizontal wells are the best way to develop the lower Fatehgarh reservoirs. The key challenges in successful drilling of horizontal wells are the uncertainty in formation top, the position of field bounding fault, presence of gravitational collapse faults, and navigating within the ultra-high resistive reservoirs.

Offset well correlation, high resolution sampling of drill cuttings, close monitoring of grain size variations and real-time density image log analyses were used to overcome the challenge of landing the well in target reservoir zone at desired borehole inclination. After landing the well, the strategy was to build borehole angle to ~90° and geosteer within the reservoir zone. The directional resistivity tool used is sensitive to resistivity contrasts providing an early indication of non-reservoir ahead of bit. However the ultra-high resistive reservoirs limited the boundary detection capability of this tool thus reducing the reaction time to take corrective measures. The true stratigraphic position (TSP) of the wellbore was also continuously monitored using gamma ray correlation.

Eleven horizontal wells have been drilled in the Mangala field using various data integration approach. Accurate placement of wells in the best part of the reservoir, avoiding interlayer mudstones, has resulted in more than 90% reservoir contact and an average production of ~10000 bopd/well.

You can access this article if you purchase or spend a download.