Wibowo, Rahmat Catur and Sarkowi, Muh and Dewanto, Ordas and Mulyatno, Bagus Sapto and Dani, Ilham (2024) Multiple Linear Regression Method for Thermal Maturity Prediction Based On Well Logs. JOURNAL OF GEOSCIENCE, ENGINEERING, ENVIRONMENT, AND TECHNALOGY, 9 (2). pp. 103-108. ISSN 2541-5794

Text JGEET 9(2) - Juni 2024 - Multiple Linear Regression Method for Thermal Maturity Prediction Based On Well Logs.pdf Download (1MB)

Abstract

Accurate estimation of thermal maturity is essential in characterizing a source rock, especially using vitrinite reflectance (Ro). The limitations of laboratory data related to the high cost of analysis require a special reliable method to measure the Ro value indirectly in the source rock layer. The proposed method is a continuous prediction of the value of Ro from well logs data using the Multiple Linear Regression (MLR) technique in the Palembang Sub-Basin, South Sumatra Basin. A total of 25 Ro data from 2 wells (RCW-01 and RCW-02) are available from the laboratory's core data analysis results. The Ro data varies from 0.39% to 0.76%, with an average of 0.54%. Prediction of the value of Ro is carried out using the MLR method, which is then carried out training and validation for continuous Ro. The training was carried out using one well (RCW-01) at 2287-3027 m and testing at other intervals (1848-2286 m). The results of the training show an estimation accuracy of R2 0.99, while the test results produce R2 0.81. The MLR formula in the RCW-01 well was then applied to the RCW-02 well for the validation test phase. The well RCW-02 produces a good correlation estimate equal to R2 0.85. Prediction of the value of Ro using the MLR method can be used to evaluate the source rock layer of a sedimentary basin in the form of a continuous interval.

Item Type:	Article
Subjects:	Q Science > QE Geology
Divisions:	Fakultas Teknik (FT) > Prodi Teknik Geofisika
Depositing User:	M.T Ilham Dani
Date Deposited:	01 Jul 2024 01:11
Last Modified:	01 Jul 2024 01:11
URI:	http://repository.lppm.unila.ac.id/id/eprint/53960

Actions (login required)

View Item