Use of Drilling Parameters To Predict In-Situ Stress Bounds

Abstract

Abstract This paper reports theory, procedure and results on the use of drilling parameters collected during typical drilling operations to predict bounds on minimum principal in-situ stress of rock. These predictions are desired in order that hydraulic fracturability of reservoir rock can be better determined and fracturing programs designed without the need for expensive fracturing stress tests, guesswork, or empiricism. A high fidelity tri-cone roller bit drilling model is used in an "inverted" mode to predict in-situ ultimate compressive rock strength. This compressive rock strength is a function of effective confining pressure available from published laboratory data for various rock types. Knowledge of the compressive rock strength failure as a function of confining pressure can be used to obtain the Mohr failure envelope at a given depth for rock. The angle of internal friction is determined from the Mohr failure envelope, which can be used to calculate a "coefficient for earth at rest." This coefficient, together with known overburden and pore pressure, can be used to calculate an upper bound on the minimum horizontal stress for each foot drilled. The calculated in-situ stress bound profiles are compared, with good results, with experimental field closure stress data obtained in 46 tests in four GRI (Gas Research Institute) wells, SFE (Staged Field Experiment) wells #I, 2, 3, and 4. Introduction In order to properly design and complete effective hydraulic fracturing operations, Voneiff and Holditch [1] and Holditch, et al [2] have shown that apriori knowledge of in-situ stress profiles is an extremely important ingredient. Typical methods for estimating such profiles have assumed elastic formation properties, and have used Poisson's ratio together with measurements taken from electric logs, with mixed results [3–4]. These procedures require expensive logging operations, which in shallow wells might constitute a large fraction of the total well cost. During the drilling of a well, drilling operating parameters, mud properties, and mudlogger samples are usually collected. By using such data in the approach proposed here, a potentially less expensive method is available for in-situ stress determination. Instrumentation for collection of drilling data is becoming more routine with most drilling operations, and will therefore add little extra cost. In what follows, we present a new method, using data collected during normal drilling operations, to estimate bounds on in-situ stress. The theoretical background is presented, a step-by-step procedure is listed, and results are compared with field data from four GRI SFE wells. DRILLING MODELS The use of drilling data to predict drilling rock strength has developed over a number of years as drilling models for various types of bits have steadily improved. Although penetration rate models have been proposed for polycrystalline diamond compact bits and natural diamond bits, the more traditional tricone roller bit has received the most attention [5–7] because of its widespread use. Consequently, penetration rate models of this bit are the most highly-developed, and a recent article by Winters, et al [8] has demonstrated high fidelity in predicting penetration rates. In this model, penetration rate of the drill bit is calculated as a function of known operating conditions, bit coefficients, mud properties and hydraulics, and ultimate compressive rock strength and ductility. P. 457^