Identify Fracture With Wavelet Multi-scale Analysis Method

Fracture prediction and identification is the key to the exploration of fractured reservoirs. The keynote of this article is identifying fracture with well logging multi-scale analysis. Based on systematic study on the multi-scale character of conventional well logging signals and full wave logging signals, extract the sensitive information that reflects development of fracture from the conventional well logging signals to determine the fracture developing intervals more quickly and exactly; extract different waveform from full waveform and analyse the character variation of different waveform to determine the fracture intervals and the developing effectivity of fracture.According to wavelet multi-scale analysis theory, we select Dual-tree Complex wavelet which has better characters. We use multi-scale analysis for conventional well logging signals, and build multi-scale analysis of conventional well logging signals; Analyse conventional logging respond to fracture, like that fracture development makes the acoustic logging curve create high frequency variations. By the multi-scale analysis, we reconstruct well logging signals which can reflect the development of fracture at the fitting scale. Based on this, we identify intervals of fracture with multi-scale reconstructed information of several conventional well logging signals.We analyse the time-frequency characteristic of reflected P wave, compressional wave, shearing wave and stoneley wave in acoustic full waveform. According to wavelet multi-scale analysis theories, we build the relationship of scale and frequency. We separate those waveforms which have a great different frequencies. Fully using the characteristic that wavelet multi-scale analysis remain the time domain, we combine wavelet multi-scale analysis with slowness time coherence (STC) method which is usually used to compute the slowness of acoustic full wave for separating reflected P wave, compressional wave and shearing wave which have near frequencies and smaller amplitude. Based on the waveform separation, we compute the energy curves of compressional wave, shearing wave and stoneley wave, and analysis the characteristic of every waveform in acoustic full waveform while fracture developing. According to amplitude and energy curve of the separated waveform, we identify fracture intervals and the effectivity of fracture development with the slowness curves of compressional wave and shearing wave in packed bed.All these researches provide a completely new method for fully using well logging information to identify fractures.