Proximate analysis provides the moisture, ash, volatile matter and fixed carbon content of coal samples. The estimation of proximate analysis with conventional logs has played an increasingly important role with the booming development of coalbed methane (CBM) plays in China. One method to estimate them is based on correlations between logs and core-based information, but it requires a lot of core data and cannot be transported without re-calibration. The other is the petrophysical volume model method that couldn't provide the accurate content of volatile matter and moisture, respectively. This is because the density, hydrogen index, and slowness of volatile matter and moisture are similar.
In this paper, we propose an improved volume model which divides the coal into moisture, ash, volatile matter and fixed carbon to solve the problems above. A resistivity response equation is creatively introduced into the model to distinguish the volatile matter and moisture by using the differences of their conductivities. The establishment of resistivity response equation is based on the assumption that the electric current of coal flow through four paralleled systems of conductive circuits that are ash, carbon (fixed carbon and volatile matter), water in matrix pores, and water in fracture pores, respectively. The moisture is the sum of water in matrix pores and fracture pores. The volume of water in fracture pores is first solved by Aguilera's dual lateral logging model. The other volumes are then obtained iteratively by combining three log response equations and the ‘sum of volumes’ constraint equation.
This study also presents a method that could expediently determine the log response values of coal components of coalbed with different ranks, which further support the model modified.
The improved model has been successfully applied to three CBM fields of Ordos basin and Qinshui basin in China, where the coal ranks vary from high volatile bituminous coal to anthracite. And the results estimated have also been well verified by core data. The advantages of this study are that all the input parameters can be acquired from logs information, which make the method predict the proximate analysis independent of core data accurately. Thus, it is worthy of applying to other coal basins worldwide.
