Performance analysis of large numbers of bit runs is often anecdotal and uses historical cost data. To this end, there are numerous problems with this approach. There is no uniform approach to identifying good performance. At best, the analysis provides an imprecise picture of overall performance. Large datasets need to be condensed into runs of interest. Difficulties arise when comparing multiple runs through long intervals with variable thicknesses of hard stringers. Since BHA, rig, and other costs change over time, it is problematic using historical cost per foot (CPF) data for the current target well. Finally, how does one determine if long slow runs or short fast ones are better since both could have the same CPF?
In this paper, the authors discuss a structured benchmarking method that can be applied regardless of the application or area studied. The basic process is simple and can be tailored to the requirements of different applications. The goal is to deliver a statistical benchmarking process that helps filter large sets of data and facilitates a consistent approach to bit performance analysis that is independent of historical cost data. A process flow chart is developed to guide engineers step-by-step through the benchmarking method. Good offsets are identified and included in the benchmarking population. Eligible bit runs are then ranked by a new key performance indicator (KPI): ROP*Distance Drilled. No historical cost data is included in the analysis. A detailed engineering study is then carried out on the identified best runs to develop recommendations for future applications. As the last step of the process, a financial analysis is carried out using cost data for the current well.
The paper will describe the use of this process to analyze bit performance in the operator's gas drilling operation and show how it allowed the identification of ‘true' unbiased top performance. The benchmarking process standardizes performance analysis and ensures sound engineering principles are applied resulting in a better understanding of past performance and better recommendations for future applications.