Presenting a Graphical Tool to Predict the Drilling Rate of Penetration through Intelligent Approaches

The prediction of drilling rate is one of the important issue because of its role in minimizing drilling costs to optimize the drilling process. Field data analysis is a key element in reducing costs and improving drilling operations. Furthermore, developing field information analysis tools and providing prediction modelsare two alternatives to improve drilling operations. When a drilling system is deployed, there are only a few limited parameters which can be controlled and changed. In general, the main purpose of this research is to apply intelligent techniques and provide graphical tools for predicting drilling performance. For this purpose, a database of field data such as well depth, drill weight, drill speed, drill chuck, weight on the hook and the torque was established from one of the southern fields of Iran. In this research, two different types of graphical tools were proposed to predict the drilling rate of penetration as well as to calculate the cost per foot, using a fuzzy neural network and Neuro-fuzzy approaches. The goals of the economic evaluation are the drill performance and the cost-per-foot calculation. The results showed that a good correlation coefficient (R2=0.94) was obtained to predict the penetration rate using the neural network. In order to improve the findings, the fuzzy neural network method was applied. The results demonstrated that a very good relationship with high precision having a coefficient of determination (R2=0.99) was obtained and thereby it depicted a significant improvement in the accuracy of the prediction models.