Identifying Reservoir Fluids by Wavelet Transform of Well Logs

Abstract

This article, written by Technology Editor Dennis Denney, contains highlights of paper SPE 88559, "Identifying Reservoir Fluids by Wavelet Transform of Well Logs," by Wenzheng Yue, SPE, Tsinghua U.; Guo Tao, Petroleum U. of China; and Zhengwu Liu, CNPC, prepared for the 2004 SPE Asia Pacific Oil and Gas Conference and Exhibition, Perth, Australia, 18-20 October. The wavelet-transform (WT) method has been applied to log data to extract reservoir-fluid information. In addition to time (depth)/frequency analysis generally performed with the wavelet method, energy-spectral analysis for time-/frequency-domain signals was performed by the wavelet transform. Forty-two models from an oil field in China were studied with this method, and, subsequently, these rules were applied to interpret reservoir layers. It was found that identification by use of this method was in very good agreement with the results of well tests. Introduction An important log-analysis application is determining reservoir-fluid properties. It is common to calculate water and oil saturations from electric-log data. Neutron, density, and acoustic logs identify gas-bearing formations on the basis of overlapping curves. Dielectric-constant and acoustic logs can identify reservoir-fluid types. Reservoir-fluid typing also can be made with differential-spectrum data from nuclear-magnetic-resonance (NMR) logs. However, the NMR measurements are taken mainly in the flushed and invaded zones where the pore fluids could have been replaced by mud filtrate. Primarily, this method provides information on the mud filtrate. Because all interpretations of reservoir-fluid properties from conventional logs are subject to interference from vugs, fractures, and clay content, the reliability of these interpretations is limited. The WT technique was developed with the localization idea from Gabor’s short-time Fourier analysis and further developed it. Wavelets provide the ability to perform local analysis (i.e., analyze a small portion of a larger signal). This localized analysis represents the next logical step: a windowing technique with variable-sized regions. Wavelet analysis allows the use of long time intervals, in which more-precise low-frequency information is needed, and shorter regions, where high-frequency information is needed. Wavelet analysis is capable of revealing such aspects as trends, breakdown points, discontinuities in higher derivatives, and self-similarity. In well-log-data processing, wavelet analysis has been used to identify formation boundaries and to increase vertical resolution. However, for data interpretation, identifying hydrocarbon-bearing zones by wavelet analysis is still under investigation. In this study, a wavelet-energy-spectrum analysis was developed to identify reservoir-fluid types. This technique was used for field-data interpretation and has achieved very good results. The full-length paper details the WT-calculation technique used. Wavelet-Energy-Spectrum Analysis If the information of a signal carrying specific information consists of different components, the individual contribution of any component to the total signal is different from one carrier to another. For example, everybody’s voice contains different scale characteristics of wavelet frequencies (i.e., the individual frequencies have different contributions to the total energy of the voice). Therefore, the signal can be distinguished by the characteristics of its energy spectrum. To set up the correlation between the target signal and the wavelet-energy spectrum, a model database must be built. By use of energy-spectrum analysis for model signals, the corresponding spectral characteristic for a specific model can be obtained and then stored in an identity data-base of the energy spectra. The spectral characteristics in the database then can be compared with those from the target-signal-energy spectrum. After comparison and analysis, the target signals are classified.