The minifrac procedure used in the oilfield services industry is a one-step procedure. The most important information one tries to obtain from minifrac testing is an effective fluid loss coefficient, C, which is the average of the early time fluid loss coefficient, C, and the late time fluid loss coefficient, C. Since C could be much larger than C, this procedure works reasonably well only if the early fluid loss period, commonly referred to as spurt time, T, is short compared to the overall job time. However, since the spurt time is sometimes on the order of several minutes, which is in the same order as the minifrac pumping time, this method tends to overestimate the magnitude of C, resulting in the overdesign of a job treatment when spurt time is large.
A mathematical algorithm has been developed in this paper to determine both the C and C, T and spurt volume, V, from the conventional minifrac treatment such that they can be adequately used to design the full-scale treatment.
There are four parameters that are needed to determine the fluid loss characteristics of a fracturing treatment. They are spurt volume, V, early time fluid loss coefficient, C, spurt time, T, and late time fluid loss coefficient, C. In most fracture design programs, two inputs are required; they usually are V and C. The other two, T and C, are calculated based on formation characteristics. The values of V and C are determined in the laboratory.
If a minifrac treatment is performed before a full-scale treatment, only an average effective fluid loss coefficient, C, can be determined, which is the average of the early and late time fluid loss coefficients, C and C. Since C is usually more than twice as large as C, replacing C and C by an average C is tolerable only if the spurt time is relatively short with respect to overall job time. However, the minifrac treatment does not yield any information regarding the magnitudes of any of the three parameters, C, T and V ; thus, whether C determined from a treatment of short duration is applicable to full-scale treatment of much longer duration is questionable. The contribution to overall fluid loss by spurt is much more significant in a short treatment than along one. Frequently neglecting the spurt leads to excessive requirement for pad volume.
Since the fluid loss rate is proportional to the square root of time, a quick estimate of the effective fluid loss coefficient, C, can be derived from the expression:
(1)
where Tp is the job time assumed to be larger than the spurt time, Ts.
P. 355^
