Identifier

etd-12152003-154103

Degree

Master of Science in Petroleum Engineering (MSPE)

Department

Petroleum Engineering

Document Type

Thesis

Abstract

Drilling operations represent the major cost in finding and developing new petroleum reserves. Poor drilling performance when drilling deep shale and strong rocks, as evidenced by slow rate of penetration (ROP), has a significant detrimental impact on drilling costs. Also, it has been concluded that bit balling is the main cause of low ROP when drilling deep, clay-rich shale with water-based mud. In addition, it is estimated that a potential saving in drilling cost of hundreds of millions of dollars a year can be obtained if bit balling is mitigated and ROP is improved. Several methods have been developed in order to improve bit performance. Recently, Arash Aghassi 1 and John R. Smith 37 proposed one of them, which uses simple drilling data to identify bit balling and lithology change as two separate effects through the calculation of five diagnostic parameters and comparing these values to a “baseline” zone. The objective of this research is to apply, evaluate, and improve the method proposed by Aghassi and Smith. A set of down-hole well data and several sets of surface well data were used to evaluate the method. The diagnostic parameters of Aghassi’s method were calculated, first using the down-hole data, and then with surface data. All of the results were correlated with and compared to wire-line logs. As a result, the utility of using surface data was confirmed. The overall utility of the method and its diagnostic parameters for detecting the occurrence of, and increases in, the severity of bit balling as distinguished from drilling into a stronger rock were evaluated. The results were very sensitive to the selection of the baseline; also, when drilling strong rock, the interpretation of the diagnostic parameters is sometimes that the bit is balled. Statistical “Logistic Regression” models were developed and evaluated as a means to solve these problems. Those models were applied using several sets of well data. As a result, it was determined that the logistic regression can potentially provide a basis for distinguish between bit balling and strong rock. It can be used independently, but it is more effective as a complement to Aghassi & Smith’s method.

Date

2004

Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

John R. Smith

DOI

10.31390/gradschool_theses.578

Share

COinS