Formation Drilling RoP Prediction via Deep Neural Networks with TensorFlow

Isam, Sherrif and Zhang, Qichun (2022) Formation Drilling RoP Prediction via Deep Neural Networks with TensorFlow. In: 2022 27th International Conference on Automation and Computing, 01-09-2022 Through 03-09-2022.

Full text not available from this repository. (Request a copy)

Abstract

Predicting the speed for the drill penetrates the formation (RoP, Rate of Penetration) has been considered as an essential factor for the drilling operation and efficiency. Complexity in predicting the RoP arises from its dependency on many factors, for example, drilling fluid properties, drilling parameters, and the characteristics of the drilled formation. Based on the mentioned factors, the objective of this paper is to use Deep Neural Networks (DNN) with TensorFlow to predict RoP based on the 8 distinct parameters, which are obtained in the Final Well Report-Well 15/9-F-15 of the Volve oil field data provided by Equinor. The obtained results confirmed that the three DNN Models with TensorFlow techniques are applied to estimate the complex lithologies RoP with varying accuracies. As a comparative study, the presented Mode13 in this paper which includes an input layer with 32 neurons, 5 deep layers with 64 neurons, and a single neuron output layer has outperformed the previous 2 smaller models and it has been recorded how all the performance parameters has recorded smaller error values as the models adopted new deep layer and neurons.

Item Type:	Conference or Workshop Item (Paper)
Additional Information:	Publisher Copyright: © 2022 IEEE.; 27th International Conference on Automation and Computing
Keywords:	deep neural networks, formation drilling, Rate of penetration, TensorFlow
Depositing User:	RED Unit Admin
Date Deposited:	27 May 2025 14:03
Last Modified:	27 May 2025 14:03
URI:	https://bnu.repository.guildhe.ac.uk/id/eprint/20264

Actions (login required)

Edit Item