Search

In coalbed-methane (CBM) projects, conventional logs, such as density were used heavily for coal identification, coal-property derivation, and also potentially for coal ranking. This paper presents a novel work flow of density-log normalization for CBM wells. It addresses two major challenges to current normalization methods: absence of a pervasive reference layer and removal of geological effects after normalization. To overcome these disadvantages, the proposed work flow undertakes six steps, which are data quality control (QC), geological- area segmentation, subarea type-well selection, density-control points, variable linear transformation and normalized-log result QC with box plots, matrix crossplots, and several histograms. This work flow removes the systematic errors caused by acquiring logs at different times with different logging instruments provided by different vendors. Furthermore, it possesses the advantages of the automated QC work flow, type-well control, and reference-layer methods without incurring their respective disadvantages. It was applied successfully to approximately 600 wells covering a large area in the Surat basin of Australia. Proper log normalization will reduce the uncertainties involved in 3D geological modeling and will increase the accuracy of reservoir evaluation. This paper should therefore be of interest to petrophysicists and geologists involved in CBM-field characterization. [ABSTRACT FROM AUTHOR]

The development of a coalbed-methane (CBM) field in its early stage is often plagued by the lack of well control and the scarcity of geological data across a large geographical area. Therefore, constructing a representative static model to estimate the in-place volume presents a formidable challenge. In this paper, we propose a work flow to overcome this challenge and apply it to a CBM field in the northern Bowen basin of Australia. One may consider this work flow as a best practice for the following reasons. First, it makes use of data from various sources including cores, well logs, seismic interpretation, and topography. Second, it performs rigorous quality control on these data, such as depth shift and log normalization. Third, coal-ply division and correlation and subsequent structural modeling are based on three types of correlation: well-to-well, well-to-seismic, and well-seismic- geographic information system. Fourth, it establishes the low, base, and high trends for the most-important reservoir properties. Fifth, it constructs a base-case static model by combining the aforementioned structural and reservoir-property models. Sixth, it uses sensitivity analysis, which varies one reservoir parameter at a time, to rank the impact of reservoir parameters on in-place volume. Seventh, it uses uncertainty analysis that varies all reservoir parameters simultaneously to arrive at the P10, P50, and P90 in-place volumes and their corresponding static models that one can use for reservoir simulations to estimate the recoverable volumes. [ABSTRACT FROM AUTHOR]