Compensating for the Compensation Effect Using Simulated and Experimental Kinetics From the Bakken and Red River Formations, Williston Basin, North Dakota

Abstract

The application of the Arrhenius equation to the problem of petroleum generation promises to predict hydrocarbon generation rates given experimentally defined kinetic parameters and temperature. There are several ways to determine kinetic parameters. In most, if not all of these methods, small variations in experimental conditions result in a linear covariance between activation energy (Ea) and the natural log of the frequency factor (A). This covariance is usually referred to as the compensation effect, the underlying cause of which is debated. However, experimental compensation effects frequently include frequency factors that are thermodynamically impossible. Based on thermodynamic arguments, the frequency factor should be largely unaffected by temperature and as a consequence be nearly constant. The covariance in A and Ea is explained from a statistical standpoint by noting that random experimental errors, particularly in temperature, result in “best fit” solutions to some experimental variable, usually temperature. Solutions for Ea and A that incorporate these errors lie within an extremely elongated ellipse that coincides with the compensation effect. An analytical solution exists that relates the harmonic mean of experimental temperatures containing random errors to the slope of the resulting compensation effect. “Compensation effects” from numerical simulations of programmed pyrolysis experiments containing random temperature errors are consistent with the analytical solution. Furthermore, calculated compensation effects using experimental temperatures are very close to the compensation effect obtained from non-isothermal kinetic analyses of samples from the Bakken (Miss-Dev) and Red River (Ord.) formations. The compensation effect is largely due to temperature errors that may be corrected to a constant value of A using the harmonic mean of the peak reaction temperatures. Assuming a constant frequency factor allows “correct” activation energies to be obtained. When done, corrected values of Ea from the Bakken (Type II) change more rapidly with depth than do samples from the Red River (Type I). A map of “corrected” activation energies from the Bakken Formation is consistent with current notions of the formation's thermal maturity within the Williston Basin of North Dakota.

AAPG Datapages/Search and Discovery Article #90216 ©2015 AAPG Annual Convention and Exhibition, Denver, CO., May 31 - June 3, 2015