Integrating Diverse Measurements to Predict Pore Pressure With Uncertainties While Drilling

Abstract

Abstract An accurate estimate of the formation pore pressure is a key element for safe and cost-effective drilling. We describe here a methodology that integrates information from diverse measurements to predict pore pressure in a well before and during drilling. The data we can incorporate include surface seismics, sonic and density well logs, checkshots, mud weights, and measured pore pressures. The distinctive feature of our method is that uncertainties in the predicted velocities and pore pressures are quantified and updated while drilling. Quantifying uncertainties allows us to properly integrate all data and to show how measurements acquired while drilling reduce the uncertainty in the predicted pore pressures. In practice, we use a Monte Carlo method to sample the posterior probability distribution (given all the data) of the compressional-wave velocity profile and of coefficients in the relationships that relate velocity to pore pressures. For each sampled value of velocities and other coefficients, we compute the corresponding pore-pressure profile along the well trajectory; the range of all sampled pore pressures describes the pore-pressure uncertainty. We demonstrate the methodology in two examples where the uncertainty in pore pressure predicted before drilling (from seismic velocities obtained by reflection tomography) is reduced by additional data acquired while drilling (checkshot, sonic log, mud weights, and pore-pressure data). Introduction To safely drill a deep well for hydrocarbon exploration or production, it is necessary to prevent formation fluids from flowing into the well. This is typically done by adjusting the density of the drilling mud so that the wellbore pressure throughout the open hole is above the pressure of formation fluids (the pore pressure). On the other hand, the mud density cannot be so great as to cause hydraulic fracturing of the formation (it cannot exceed the fracture pressure). The pore-pressure and fracture-pressure gradients thus provide minimum and maximum values that define a mud-weight window, and this window is used to decide the depth of casing points.1 For safe and cost-effective drilling, it is therefore important to employ a method of estimating pore and fracture pressures before drilling and to update these estimates as the well is drilled and new information is acquired. In this paper, we first describe a way to predict pore pressure from a combination of various measurements and simple empirical relationships. We then describe how to deal with uncertainties in the fundamental inputs to the empirical relationships used to predict pore pressure; the resulting prediction has uncertainties determined by the available data. Finally, we show sample applications of this method to two wells and demonstrate how the uncertainty in predicted pore pressure is related to the information provided by measurements. While this paper deals with pore-pressure prediction, uncertainties in predicted fracture pressure can also be evaluated with the same approach. Deterministic Pore-Pressure Prediction An approach widely used to estimate pore pressure is based on measurements of compressional-wave velocities, formation resistivities, or drilling penetration rates. The fundamental assumption is that departures from a normal trend of these measurements are related to corresponding anomalies in pore pressure (for a general treatment and additional references on this subject, see Chapter 6 of Ref. 1). For example, if elevated pore pressures are a result of undercompaction of shales, the sediment porosities will be anomalously high and the velocities anomalously low. In this paper we concentrate on the use of compressional-wave velocity to estimate pore pressure, because velocity estimates are typically available before drilling from the processing of surface seismic data2 and can then be refined during drilling from measurements such as well logs and borehole seismic data.3,4