Abstract
Given the potentially catastrophic human and environmental results of an unplanned well collision, a meaningful estimate of the probability of wellbore intersections provides a valuable risk management tool. A number of estimation techniques are currently in use in the industry; however, all of these have limitations which may not be clearly understood. For example, many computations currently in use fail when the two wells are parallel, they ignore the fact that the drilled well may have reached its current location without collision, or they return an identical result regardless of whether the drilling well is thought to be approaching or departing from the existing well.
A rational approach to the problem begins by clearly defining the probability to be estimated. This paper presents a method to ascertain the probability that a particular interval of the well being drilled might intersect an existing adjacent wellbore. An overall collision probability can then be obtained by combining the probabilities associated with successive intervals along the reference well. The estimates of probability are based on knowledge of the surveyed wellpaths and their survey uncertainties, expressed in the form of position covariance matrices.
A body of relevant research exists in the field of astrodynamics, concerned with estimating the risk of collision between orbiting satellites. This paper draws on prior work from both the oilfield and aerospace in developing a more rigorous assessment of collision probability, overcoming the limitation of previous methods. Examples demonstrate that the new method agrees with analytic results for both parallel and non-parallel wellpaths in simple cases where both wells are straight in the region of interest.
