Laboratory study of hydraulic fracturing pressure data—how valid is their conventional interpretation?

Abstract

This paper presents some recent results in laboratory-controlled hydrofracturing tests in Niagara dolomite subjected to unequal principal in situ stresses. Our new computerized hydrofracturing set-up is capable of both test parameter control and data acquisition and analysis under the most general state of far-field stress. The objective of these tests was to verify the reliability of suggested techniques for determining the critical pressures involved in stress determination, such as the fracture reopening pressure P r and the shut-in pressure P s. The determination of P r was carried out by comparing digital pressure-time data of the ascending portions of the first and subsequent pressurization cycles, all conducted under the same flowrate. Our preliminary finding is that the value of P r, taken as the point of departure in the third or fourth cycle, appears to reflect the value of the least horizontal principal stress rather than the minimum tangential stress at the borehole wall. Shut-in pressures were correctly estimated (within 8%) using a statistical exponential pressure-decay model in which a non-linear regression curve fitting technique was employed. We also plotted the pressure-time derivative with respect to pressure, and found that both a bilinear and linear-quadratic regression of the recorded digital data provided good estimates of P s. Acoustic emission was detected during both fracture extension and closure.