Abstract

Rate of penetration is defined as the rate at which a drill bit cuts through the rock beneath it during the drilling process. the drilling and petroleum industry could benefit greatly from the ability to predict the Rate of Penetration as it could serve as a basis for decision making, operation planning, drilling process optimization, and for the evaluation of time, cost and efficiency of the drilling process. Non-linear regression is a means of modelling data and relationships that cannot be accurately represented on a straight-line graph by representing them on a curved line instead. Because multiple factors -such as rock strength, brittleness, weathering rate and density as well as drill bit type, operating conditions and hydraulics- affect the ROP, multiple regression analysis is used to make an accurate prediction. From literature we can see that non-linear regression is a valid means of predicting the ROP in a well-bore using information from previous wells in a field. This endeavour yields results with varying degrees of accuracy based on the parameters used to create the models. Modelling and predicting the ROP involves gathering the data for both the rock and the drill bit through methods such as core tests and other geophysical methods, then running them through regression software such as SPSS and MATLAB. Multiple models resulted from this process and the accuracy of the resulting models was ascertained using metrics such as the determination coefficient, correlation coefficient and standard error. The most accurate models are models that included both rock and drill bit characteristics. Conclusively it can be shown that the ROP can be predicted to a reasonable degree of accuracy using non-linear regression.

Key words: Rate of Penetration, Non-linear Regression, Regression model