Abstract
Completion optimization and development optimization are two terms that can mean many different things to people in our industry. What are we optimizing? Minimizing completion cycle time, maximizing a production parameter (90 day cumulative, IP or EUR), minimizing well cost, maximizing perf cluster efficiency, or an economic focus, maximizing NPV, ROI, ROR or minimizing time to payout and so on. This paper will focus on methods that have been used to maximize NPV and ROI for both individual well completions and field development in unconventional multi-stage wells. The methods used in this paper allow us to define the key reservoir and completion characteristics that govern the production behavior of these wells and reservoirs. These three key reservoir and completion parameters are; the formation permeability; the number of producing fractures in the completion; and the average effective fracture half-length of those producing hydraulic fractures. There are many other parameters that must be known to solve the very complex problem that is completion and development optimization, such as net pay, porosity, water saturation, reservoir pressure and fluid properties. We are going to rely on other geological and engineering team members to provide how these values may change across a development area.
Many operators are attempting to solve this problem through the analysis of "Big Data". Large amounts of data are a good thing and should be analyzed statistically, through machine learning and through artificial neural networks, however most of these processes lack the ability to extrapolate outside of their range of existing data without incorporating some type of physical model. "Big Data" costs a significant amount of capital to develop. Wells must be drilled and completed and produced and while these processes rely on data from producing wells, meaningful results can be achieved with only a small fraction of the data required for machine learning. After evaluating countless horizontal, muti-stage well production curves, over 80% behave in the manner described in this paper. There are exceptions, most of which have mechanical problems. Where implemented, the recommendations derived from the processes described in this paper have resulted in the reduction of significant capital cost and increases in individual well hydrocarbon recoveries.
