Coupling Seismic Data With Simulated Annealing Method Improves Reservoir Characterization

Abstract

Abstract Limitation in quantity as well as distribution of reservoir data leads to unsatisfied realization of inter-well reservoir characterization. Recent development in seismic data processing and analysis, that is seismic inversion, has been able to furnish a footstep to define such physical reservoir data in inter-well vertical section. A stochastic "cloud transform" method instead of deterministic ones has been applied to change the acoustic impedance to porosity. The present work integrated this stochastic procedure with Simulated Annealing Method (SAM) to optimize reservoir description. Due to strong relationship between impedance and porosity, semivariogram in lateral direction is derived from acoustic impedance data, whereas that in vertical direction is defined from log data in reference wells. A case study on one of Indonesian limestone reservoirs has been conducted. Results showed that seismic impedance-derived porosity is sensitive to lithologic changes. Coupling seismic data with a conventional simulated annealing method gave better simulation result. This work leads to improved understanding of the importance of seismic data in characterizing reservoirs. Detailed results will be presented and discussed. Introduction Reservoir characterization or description may be defined in general as a process of quantifying reservoir attributes based on the available data. Usually, these data sampled are limited in both quantity and sampling points. An attempt, however, should be made to characterize the entire reservoir laterally and vertically as well. Based on a limited data set, statistical method may be employed to estimate reservoir parameters at unsampled points. Engineers usually use geostatistical methods to study reservoir heterogenity. One of the methods commonly used is conditional geostatistics that can create stochastic modelling to estimate reservoir properties at unsampled points based on soft data, such as seismic data, without modifying the sampled data values. Simulated annealing method has been developed recently to generate a random variable distribution that holds statistical boundaries given, such as conditioning data (sample data), cumulative distribution function (CDF), and the model semivariogram. Geophysics technology has been advancing rapidly, enabling one to translate surface seismic cross section into acoustic impedance by employing inversion technique. To estimate physical properties of reservoir, porosity for instance, from acoustic impedance data, a stocastic or simulated annealing method may be used. The advantage of this method is its capability to minimize the ambiguity of seismic data by implementing a cloud transform technique. The objective of this work was to conduct a case study integrating well logs and seismic data in an effort to obtain better information about the reservoir under study. The Reservoir and Data Preparation The Lembak gas field located geographically in South Sumatra, Indonesia, has been chosen in the present study based on the complete data availability. The field is producing gas and a small amount of oil and condensate from a limestone reservoir of Baturaja formation. The depth interval studied was 1259.8 -1368.3 meters. The seismic line 801- C7357 crossing wells No. 4, No. 7, and No. 10 at trace 162,96 and 126, respectively, was used (see Fig. 1). Acoustic impedance data were obtained by applying a band limited inversion method of the Charisma Reservoir modelling modul to the seismic data. We assumed here that the surface seismic data was corrected to zero phase, no multiple exists after deconvolution, and ELAN-interpreted porosity from the sonic and density well logs were accurate. P. 87