Assessment of “Lost Gas” Estimation Technique by Replicated Canister Desorption Testing

Satya Harpalani and Abhijit Mitra
Southern Illinois University Carbondale, Mining and Mineral Resources Engineering, IL 62901, [email protected]

Determination of “lost gas” is critical in order to accurately estimate the gas content of a coalbed. It is generally estimated by plotting the volume of gas desorbed from drill cuttings against desorption time and extrapolating the initial portion of the plot for “lost time”. However, the procedure assumes the initial part of the plot to be linear, which might not be true given the extremely rapid desorption during initial desorption period. The rate slows down with continued desorption, and eventually, the plot looks more like a sorption isotherm, in which case, the amount of “lost gas” may be underestimated significantly.

This paper presents the results of an experimental study aimed at verifying this hypothesis by carrying out replicated canister tests in a laboratory environment, that is, with zero “lost time” and “lost gas”. The results showed that the error in estimating “lost gas” is less than 12% for Illinois coals and almost 25% for San Juan coals. Furthermore, the results were also used to estimate the sorption time, which was almost eleven days for Illinois sample and one day for San Juan sample. Finally, reduction in lost time for Illinois coal reduced the error to 10%, indicating the error reduction to be trivial for coals with low gas content. However, the findings have valuable implication for the deeper coals with higher gas content and might provide an explanation for gas production rates encountered in the San Juan Basin for over twenty years, which has surpassed all expectations.

AAPG Search and Discovery Article #90095©2009 AAPG Eastern Section Meeting, Evansville, Indiana, September 20-22, 2009