The effect of inaccurate directional surveying on lateral wellbore spacing is commonly overlooked. The purpose of this paper is to demonstrate how inaccuracy in standard directional-surveying methods affects wellbore position and to recommend practices to improve surveying accuracy for greater confidence in lateral spacing.

Introduction

Determining optimal wellbore spacing for a given field requires consideration of several factors such as hydraulic-fracture geometry and reservoir quality. These variables are simulated using advanced reservoir models to identify ideal parameters, such as lateral spacing distance, which will lead to the greatest production value. Models are then validated against field spacing tests to confirm results. However, an important consideration that is usually ignored is wellbore positional accuracy. 

Most reservoir models assume accurate wellbore placement. When reservoir models do account for wellbore positional uncertainty, the approximation is only a fraction of the actual positional error expected from standard wellbore-surveying techniques. This creates a problem when positional errors occur during spacing tests, because the actual lateral spacing distance could vary significantly from the values used to create the reservoir model. If production and reservoir simulations are misinterpreted because of wellbore positional errors, then economic consequences could be significant. 

Horizontal wells are directionally drilled using measurement-while-drilling (MWD) surveying to characterize the well path. Standard MWD surveying is subject to many error sources that lead to significant positional uncertainty and can result in inaccurate well placement. However, enhanced surveying methods such as use of multistation analysis (MSA) in field referencing can be an effective strategy for reducing a majority of errors. Positional uncertainty can be modeled by 3D ellipsoids of uncertainty at each survey point in the well path, which represent a statistical distribution of where the actual survey might exist.