StatoilHydro has implemented the pump-in/flowback test as the preferred way of performing an XLOT (extended leak-off test) with mud for minimum stress estimation. In this paper we briefly review our standard method as presented in previous publications. We provide new data supporting one of our main conclusions from previous work: traditional XLOTs tend to overestimate the stress in tight formations. Further, we show that the method is also very valuable in permeable formations, by presenting an example of how a very clear stress determination was possible from the flowback phase, while the interpretation from the shut-in phase in another test cycle was at best ambiguous. We further discuss the difference between topside and downhole data, focusing on the magnitude of friction during the various phases of a test. Next, we discuss some examples of flowback tests that did not give a credible minimum stress interpretation. Three circumstances are identified which may hinder stress estimation: 1) Unfavorable combination of well azimuth and inclination, 2) Poor integrity of the casing shoe, 3) Unfavorable fracture shape.
The extended leak-off test (XLOT) has been the industry-standard for stress determination during drilling for some time [1, 2]. In hydraulic fracturing the pump-in flowback method has been one of the standard methods for minimum stress estimation for several decades [3 ,4].
In StatoilHydro, we have for the past 10 years included a flowback phase in our standard XLOT (extended leakoff tests) with drilling mud.
Two main differences from the standard flowback procedures have been implemented: We do not attempt to keep a fixed flowback rate; rather we use a fixed choke to control the flowback. Further, we measure the volume flowed back as a function of time. The corresponding pressure and volume data allow the computation of the changes in system stiffness during the test, which has provided valuable insight into how the tests should be interpreted.
We define the system stiffness as the change is pressure per change in volume content in the system. The system compliance is the inverse of the system stiffness.  The term system underlines that both are extensive quantities.
We have documented our experience in several previous papers [5–9]. In this paper we present new field data from three different wells in different fields, all on the Norwegian continental shelf, which support our previous work.
Further, we discuss and show field examples of cases where the flowback method appears not to work.
PUMP-IN/FLOWBACK TESTS WITH MUD
Test proceduresWe normally combine the flowback test with a standard XLOT with a shut-in period. Our basic procedure calls for a minimum of two cycles [6, 7]:
Pump-in followed by a shut-in period of typically 15 minutes, followed by a flowback phase with constant choke setting.
Pump-in followed directly by a flowback phase with constant choke setting. In recent years the procedures also include a minimum 3 minutes shut-in period after the flowback phases to check for pressure rebound.