摘要

In cased boreholes, acoustic logging waveforms are often contaminated by casing waves, making it difficult to extract formation acoustic velocity from the acoustic data, especially when the casing is poorly bonded with formation. Various methods have been tested to solve the problem with only limited success, due to the weak signal amplitude and poor coherence compared to the dominant casing waves. The new technique of this paper utilizes the modulation of wave power spectrum caused by the interference of casing and formation waves. The advantage of the technique is its ability to suppress the coherence of the casing wave and significantly enhance the coherence of the formation wave, allowing for determining the formation acoustic velocity from the latter wave signal. For an acoustic logging instrument composed of an array of equally spaced receivers, the waveform data containing casing and formation waves are windowed and transformed into frequency domain to obtain the power spectrum for each receiver in array. The average power spectrum of the array is calculated and subtracted from the power spectrum of each receiver, yielding a residue power spectrum. The residue spectrum and its Hilbert transform form the real and imaginary parts of an analytical signal, which, after transformation into time domain, results in a new waveform array data whose phase moveout across array is entirely controlled by the velocity of the formation wave. Processing the new array data using an array velocity analysis method, e.g., the slowness-time-coherence (STC) method, yields the formation slowness behind the steel casing. @@@ The new technique was tested on synthetic waveform data calculated for a poorly bonded casing. The synthetic data processing example shows that the technique effectively cancels the coherence of the casing signal and significantly enhances the coherence of the desired formation signal. The technique has also been applied to processing field acoustic logging data acquired in cased boreholes. In depth intervals with poor bonding conditions, the STC result of the new technique shows that the casing wave coherence almost disappears and the formation wave coherence is significantly enhanced. Further, the validity of the velocity curve is verified by comparing it with its counterpart from the raw data. The much improved result in the poorly bonded section indicates the ability of the new technique to extract and enhance weak formation signals in the presence of strong casing wave interference.