Abstract
In many cases, the intensity of tide signal present in a pressure transient test is so strong as to render its interpretation unfeasible. This noise (from tide) to signal (from reservoir) ratio is a function of the transfer of tidal stresses to porous pressure as well as of some parameters from a downhole pressure test, specially, the flow rate before shut-in, formation transmissibility, shut-in elapsed time and indirectly, the wellbore skin factor.
The purpose of this work is to disseminate a predictive method of tidal effect simulation based on TMD, Tide Model Driver, with the necessary procedures and adjustments, which allows the extraction of the tidal signal components from the pressure signal recorded in wells under test through downhole pressure gauges, thus allowing analysis of test data under intense tide signal interference.
Furthermore, this work establishes a relationship between Bourdet derivative smoothing parameter L and the Nyquist Sampling Theorem. This relationship acts as a criterion on the choice of maximum L, limited by Nyquist Theorem and is designed to preserve derivative tide signal shape and amplitude for the Bourdet well test analysis method, not allowing its misinterpretation as reservoir behavior. In addition, this work present simple relationships or even a full tide simulator to predict tidal effects on pressure buidup or falloff and logarithmic derivative, considering the actual or estimated time of test, contributing thus to well test design as well.
The installation of PDGs, Permanent Downhole Gauges (pressure and temperature piezo-electric sensors), in wells completed with ICVs, Intelligent Completion Valves and designed to perform extended well tests, or EWTs, has made it possible to recover characteristic tide pressure signals from relatively long periods of pressure record. In some cases, it has become possible to detach the tide signal in order to characterize it, through the Discrete Fourier Transform, DFT, determining the main tide frequency components present therein. The comparison of those frequency constituents with the ones present in oceanic and earth tide signals led to the use of a tide simulator, TMD MatLab Toolbox, as a tidal effect filter with excellent results, without the need of time lag adjustments.
The tidal effect present in pressure data from PDGs installed at oilfield wells always comes together with reservoir signal. A simple method to recover an almost pure tidal signal was developed, consisting of a polynomial fit to represent the reservoir signal. The recovered tide signal was then submitted to Discrete Fourier analysis, DFT, and the resulting frequency components were compared with oceanic and coastal tidal data available through tide stations, and later on, compared also with simulated tidal predictions, leading to the use of a tide predictor as a tidal filter. With this result from DFT, a tide signal generator, the TMD software, was selected with excellent results. A non-linear regression is run over the recovered tide data, to adjust the amplitude level of tide signal transmited to the reservoir pore pressure by rock stress. The procedure for using the TMD and necessary adjustments to conform and extract the tide signal is illustrated by a field case with strong tidal effect, to the point of totally spoiling the characteristic transmissibility levels of the pressure derivative on pressure shut-ins. On the other hand, after the proper treatment through the proposed filter, the reservoir "tide-free" signal is recovered, turning possible the analysis of the data with increased sensibility. Once the filter is adjusted to pressure data from one well, it is ready for use on other wells operating at the same field and reservoir.
