 |
(tel) 617.253.8027 (fax) 617.253.6385 |
|
(tel) 617.253.8027 (fax) 617.253.6385 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Petroleum Source Rock Modeling
J.D. Mendelson
Submitted
to the Department of Earth, Atmospheric, and Planetary Sciences on February
8, 1985 in partial fulfillment of the requirements for the degree of Doctor
of Philosophy
Abstract
Regional
distributions of organic content are an important aid in developing basin
evolution and hydrocarbon generation models. An approach to evaluate hydrocarbon
source rocks using resistivity, sonic, density, neutron and natural gamma
ray logs is developed. Organic matter, as a constituent in sedimentary rocks,
has a relatively low density, slow velocity, and is high in hydrogen content.
Source rocks generally have low water content, and often exhibit abnormally
high concentrations of uranium. These effects combine to make an in-situ estimation
of organic content plausible. Evolution of kerogen to bitumen, oil, and gas
systematically affects the above properties and it is possible to obtain a
qualitative assessment of the state of maturation of a known source bed.
In this thesis logs and core data from wells in two separate oil provinces
are used to test the methods of predicting total organic carbon content from
log data. Two approaches are followed. The first method treats the organic
matter as a rock constituent and calculates the log responses as a function
of organic content. Two (rock and organic matter) and three (rock matrix,
water and organic matter) component models are tested. This approach suffers
because of the uncertainties of the physical properties of the organic matter.
For each log type (i.e. sonic, gamma, resistivity, etc.) log values are correlated
with the laboratory measured total organic content. Bivariate regression helps
to illustrate the efficacy of the models. In the second method, multivariate
equations based on linear combinations of individual correlation coefficients
are obtained. The importance of combining several logs which are organic content
predictors is demonstrated. These equations can be used to predict total organic
carbon content using only log data, in different parts of an oil province.
