Multi-Lateral Well Design Advisory System; an Inverse-Looking Solution

Abstract

The shifting trend towards drilling multilateral wells has motivated engineers and researchers to investigate ways to improve their decision-making when designing multilaterals. The problem becomes more challenging in unconventional reservoirs, where tighter and highly heterogeneous and discontinuous reservoirs are targeted. Achieving the right design of a multilateral well configuration is a complex problem due to the vast possibilities of well forms (i.e., number, location, length and orientation of laterals) that need to be evaluated. Accordingly, arises the need for an inverse-looking approach to come up with a sound decision. Conventional ways in designing multilaterals involves investigation of different possible scenarios using a reservoir simulator model and comparing results against an objective function. However this addresses the problem only from a forward-looking point of view. This procedure is time and labor intensive and tremendously increases the computational cost. This paper illustrates the utilization of the artificial neural network (ANN) concept to develop an advisory system for designing multi-lateral well configurations for a given target reservoir performance in fraction of the time, cost and effort required by a numerical simulator. From that point onwards an engineer can improve on the recommended design or modify it to adhere to certain field, cost or operational constraints.