A 2.5-D velocity-stress finite-difference code is described that models acoustic propagation in a borehole penetrating a generally anisotropic formation. The excitation may be a dilatation (monopole), or a point force (dipole) in an arbitrary direction. The anisotropic formation is homogeneous along the axis of the borehole but may be inhomogeneous in the transverse plane. The borehole cross section and location of the source in the borehole are arbitrary. Synthetic time-domain waveforms are displayed for arrays of monopole and dipole receivers deployed along the borehole axis in both fast and slow anisotropic formations. The specific anisotropy model employed for the numerical results is a transversely isotropic (TI) formation with its axis of symmetry inclined with respect to the borehole axis by an arbitrary angle.Flexural/shear and Stoneley wave slowness and attenuation estimates are extracted from the synthetic waveforms using a variant of Prony's method for a range of borehole inclinations relative to the formation axis of symmetry. In a fast formation, with borehole and formation symmetry axes perpendicular, flexural mode dispersion curves for quasi-SV and $H polarizations separate only at low frequency where moderate attenuation is also observed. In a slow formation, distinct dispersion curves are obtained for quasi-$V and $H polarizations over the entire frequency range. Moderate attenuation is again observed at low frequency. Scaled laboratory experiments confirm the numerical procedure. Experimental and numerical waveforms for monopole and several polarizations of dipole excitation in a transversely isotropic model formation overlay with excellent agreement.

Optimal control of oil-field hydraulic fracturing operations will benefit from reliable, automated, real-time microseismic monitoring. Existing microseismic processing techniques are often either unreliable, or impractical for real time use. In this paper we present the method of Coalescence Microseismic Mapping (CMMapping) for detection and localization of microseismic events. The technique is both automatic and robust, and explicitly allows for the inclusion of velocity model uncertainty into its formulation. We present the general basis for the method and illustrate its use on data acquired during a hydraulic fracture well stimulation. Introduction One of the clear benefits of being able to generate maps of the fracture network in real time as the stimulation progresses is that the operator can potentially adjust pumping and stimulation parameters, such as pump rate, fluid properties and volumes, to better keep the stimulation within its original design. Current techniques of microseismic event location rely on either fully-automated or interactive, semi-automated picking of discrete seismic arrivals at each of one or more seismic detectors. However, reliable automated arrival time picking remains a challenge and interactive picking of potentially many seismic arrivals is impractical for real-time monitoring. Tarantola and Valette[1] detail forward and inverse problems applied for seismic event location. In their example on hypocenter location, there is a reliance on the picking of discrete seismic arrivals. The inverse problem is then the simultaneous inversion for location and origin time for each discrete seismic event. In this paper we formulate the inverse problem in terms of both the probability of occurrence of a microseismic event and its spatio-temporal coordinates. We have implemented a method to detect and locate events, which does not require the identification and picking of discrete arrivals at each sensor.We do so by continuously updating and applying event detection to a spatial map of the probability of microseismicity occurrence. This map is generated by transforming and then forward mapping the signals from multiple seismic detectors, thereby allowing the joint detection and location of microseismic events, without requiring discrete arrival detection and time picking at each of the detectors individually. The method enhances the detectability of microseismic events and is a robust and fully automated method of microseismic event detection and location. By making use of today's faster computers and incorporating careful formulation and optimization, the method can be applied to generate reliable maps of the fracture network with near zero latency. In Parts 1–3, we discuss the theory, simplifications, and assumptions made in the formulation. In the process, we look to the general theory of Tarantola and Valette to understand how to account for forward model uncertainty in our formulation.[2] We then discuss implementation, and show the comparative result for one implementation, and for the result obtained using a traditional method of microseismic event detection and location, using arrival times picked on individual sensors.

A frequency-wavenumber (co -k) representation is given for the response of eccentric multipole sensors in a fluid-filled borehole. Specializing to eccentric dipole sensors, synthetic time-domain waveforms from an array of receivers deployed along the axis of the borehole are displayed for a range of eccentricity magnitude and orientation, formation slowness, and borehole radius. Flexural wave slowness estimates within several frequency bands are extracted from the synthetic waveforms using a semblance technique. For eccentricity small relative to the borehole radius, the character of dipole waveforms and the estimated flexural slowness are little changed. When the eccentricity is a substantial fraction of the borehole radius, flexural wave amplitude is greatly increased, and strong nonmonotonic trends are introduced due to interference with other modes. If eccentricity is not parallel to the transmitting dipole axis, substantial signals are detected on cross-dipole receivers, perpendicular to that of the transmitting dipole. Eccentricity effects decrease with frequency. Even for large eccentricity, flexural slowness estimates are affected only slightly. Scaled laboratory experiments confirm the numerical results. Comparison of the experimental and numerical waveforms is made for both centered and eccentric dipoles in several model formations. The predicted increase and nonmonotonic variation of flexural wave amplitude versus receiver offset for eccentric sensors is observed. Experimental and numerical waveforms overlay with excellent agreement.PACS numbers: 43.20.Rz, 43.20.Mv

Simulations of cement bond logging (CBL) have shown that wellbore fluid effects can be segregated from sonic-signal response to changing cement strengths. Traditionally, the effects have been considered negligible and the CBL's have been interpreted as if water were in the wellbore. However, large variations in CBL's have become apparent with the increasing number of logs run in completion fluids, such as CaC1 2 , ZnBr2' and CaBr2'To study wellbore fluid effects, physical and numerical models were developed that simulated the wellbore geometry. Measurements were conducted in 5-, 7-, and 9%-in. casings for a range of wellbore fluid types and for both densities and viscosities. Parallel numerical modeling used similar parameters.Results show that bond-log amplitudes varied dramatically with the wellbore fluid acoustic impedance-i.e., there was a 70% increase in signal amplitudes for 11.5 Ibm/gal (1370-kg/m3) CaC1 2 over the signal amplitude in water. This led to the development of a fluid-compensated bond log that corrects the amplitude for acoustic impedance of various wellbore fluids, thereby making the measurements more directly related to the cement qUality.

Transmitted energy levels for wide-band $OFAR transmission are calculated for arbitrary sound-speed profries using ray-mode analysis [R. P. Porter, J. A½oust. Soc. Am. 54, 1081-1091 (1973)]. The analysis of the received field has been generalized to include arbitrary normal modes and range-varying, soundspeed profiles. It is shown that computations of wide-band propagation based on ray-mode analysis require an order of magnitude fewer calculations than do calculations by harmonic analysis. A computer code has been developed and used to estimate the wide-band loss for profiles from the Mediterranean and the North Atlantic. Comparisons with continuous tone, normal-mode calculations yield agreement to within 3 dB. Subject Classification: 30.20, 30.25' 20.20, 20.40. LIST OF SYMBOLS c(z), Co kO q R•, R} t To sound speed at depth z and at source depth z0 frequency dependence of source ray cycle length in the horizontal plane = w/c(z), wavenumber where c(z) is the sound speed wavenumber at depth z0 mode number, number of modes eigenray order = k(z)/ko, index of refraction referred to source depth slope for N •' -linear sound-speed profile cylindrical coordinates reflection coefficients looking up and down time arrival time of energy traveling at axial sound speed Co U W Zu, Zt Y o 7' r group velocity bandwidth of transmitted pulse upper and lower turning points at which • = k(z) scale height of phase integral vertical propagation constant phase integral of vertical propagation constant normalization factor horizontal propagation constant dispersive delay referred to T o = (t-To)/T o , relative dispersive delay pressure field ray exit angle from the source z-dependent eigenfunction for mode m pressure field spectrum circular frequency