--> --> Seismic Detection of Mud Losses – An Application to the Colombian Foothills

2018 AAPG International Conference and Exhibition

Datapages, Inc.Print this page

Seismic Detection of Mud Losses – An Application to the Colombian Foothills


Mud losses, particularly large losses, can negatively impact the success of a drilling program. Detecting the depth intervals where potential mud losses may occur during drilling involves both geomechanical analysis and optimization of operational parameters. One method of detection is one dimensional analysis, geomechanical and operational, which provides a linear representation of a three dimensional space. Representing just one dimension limits the outcomes of the analysis and increases uncertainty. Incorporating 3D seismic data highlights faults and seismically visible fractured intervals, reducing uncertainty and improving the quality of the results. In this context, integrating seismic interpretation, well data and geomechanical analysis allows the interpreter to predict with greater certainty potential well behaviour while drilling with a specific mud weight, thus preventing mud losses and wellbore stability problems. In this presentation, we show how incorporating 3D seismic data in the geomechanical analysis, and, more specifically, using a seismic attribute called Fault Likelihood, made it possible to find a relationship between reflector discontinuities associated to tensile faults parallel to the maximum horizontal stress in the area and mud losses that occurred while drilling in the correlation wells. Once this relationship was found, it was easy to explain why some of the correlation wells with the lowest mud density reported presented massive mud losses, while others with higher mud density in the same lithostratigraphic interval did not present any mud losses while drilling. This methodology was applied to an oil field located in the Colombian foothills and used to predict depth intervals where potential mud losses might materialize while drilling in one of the field’s prospect wells. The result was a mud losses prediction effectiveness higher than 80%.