Differential sticking is known to be influenced by drilling fluid properties and other parameters, such as the characteristics of rock formations. In the past, multivariate statistical analysis techniques and simulated sticking testes using different drilling fluids have been performed to identify and modify parameters that lead to differential pipe sticking in order to minimize or prevent sticking. Recently, an application of neural network methodology to predict differential pipe sticking incidents in Gulf of Mexico has been published by Halliburton1 . This paper presents two different types of artificial neural network that can provide solutions for problems associated with differential pipe sticking.
A stuck pipe database was developed with data from 64 side tracked and horizontal wells drilled in reservoir section using oil based and synthetic drilling fluid from different fields in the Persian Gulf. Two three-layers feed forward networks; Multi Layer Perceptron (MLP) and Radial Basis Functions (RBF) with back propagation training algorithm were used to develop stuck pipe predictive models for oil base and synthetic drilling fluid together. Using these models, prediction of the probability of stuck pipe may be undertaken to monitor drilling operations for stuck pipe avoidance. A sensitivity analysis was also done by applying data of different fields separately to identify the parameters that had more effect on tendency to differential pipe sticking. The proposed methodology can be used for optimum drilling fluid design during well development in Persian Gulf, Offshore Iran.