This paper presents the analysis of measured bottom hole pressure data for 45 hydraulic fracture treatments pumped in Alaska and California in the past three years. On 30 of these wells, bottom hole data was obtained from memory gauges deployed on the tubing or in side-pocket mandrels, while the other 15 wells had reflected pressure up an open annulus. The fluid systems were either delayed borate guars, or linear gel HECs.
The paper starts with a comparison of these measurements with the Keck et al4 correlation. The results show that there are large systematic errors both in the magnitude and trends of the correlation. We illustrate the effects these errors have on net pressure analysis and on the success of executing these tip screen out treatments. The data also shows an unexpected transient effect in almost all the wells, whereby changes in surface pressure of 100's to 1000's of psi are found to change the bottom hole pressure by 10's or 100's of psi, respectively. We speculate on possible causes for this phenomenon, but conclude that additional work is needed for accurate predictions of bottom hole pressures during fracturing treatments.
Introduction
Accurate methods to predict bottom hole treating pressures of hydraulic fracturing treatments have long been hoped for in our industry, but still do not exist. Bottom hole pressures drive mini-frac analysis, net pressure analysis, history-matching with models, and so the stakes for accurate bottom hole values are quite high. Given that measuring bottom hole pressures in real time is quite costly, and therefore is rarely routinely performed, we rely on calculated values from correlations.
Since the first use of polymer based fluids, the industry has recognized that accurate predictions are complicated by two fundamental problems. First, even non-crosslinked linear gels are highly viscoelastic, showing drag reducing properties in turbulent flow that we take advantage of to lower friction losses. No theory, however, exists that allows scale-up of small sized laboratory test results to field sized tubulars, and so we are forced to use correlations that try to extrapolate from limited data. The second problem is that the addition of proppant to these fluids further distances us from theoretical calculations and moves us more into reliance on correlations.
A comprehensive review of the literature in this area is beyond the scope of this paper, and the reader interested in works published before 1990 might start with the SPE monograph by Gidley et al.1 or the review text by Economides and Nolte2. Since then, there have been several studies on friction pressures of water-based fracturing fluids, including, for example, linear gel friction pressures (Shah3, Keck et al.4), delayed borate fluid systems (Tan5), laboratory studies on proppant effects (Shah and Lee6, Keck et al.4), and field studies (Jennnings7, Bilden8).