It’s one of the most essential and critical parameters influencing drilling efficiency is the rate of penetration. To increase the drilling efficiency, lower costs, and reduce NPT, necessary to predict ROP before drilling operations. Fog computing can deliver a strong response for applications through preprocessing and filtering the data. Trimmed data can then be transferred to the cloud for further analysis. The project's goal is to build a real-time framework for selecting the optimal penetration rate during the drilling operations.
Generally, two types of ROP prediction models can be classified into (1) traditional models (Maurer,
Bourgoyne & Young’s, …) and (2) data-driven models (Neural Network, SVM, ...). In this work, to predict the penetration rate, Bourgoyne & Young's and Neural Network models are applied in terms of ROP modeling based on drilling data that has been taken from the fields of southern Iraq. The two models are then compared to see which one is the most accurate in predicting ROP. As a result, understanding the behavior of drilling data is a crucial part of developing an optimal ROP prediction model in real-time. Where the calculations were made and the results were presented using programming languages and the Spotfire software.
The Neural Network model and Bourgoyne & Young's Drilling model are applied to drilling parameters obtained from the fields of southern Iraq, which are used to predict the ROP. Then a comparison is made between the two models for the same data. The result proved the prediction of the ANN model is preferable than that of Burgoyne & Young’s drilling model. The purpose of this project is to demonstrate how data models and learning methodologies can be applied to drilling calculations. ML algorithms are being developed to predict the penetration rate across the well. This model was expanded to maximize ROP for a given section by optimizing parameters such as (weight on bit, revolution per minute, pressure standpipe, and flow rate). Therefore, this model can be used for real-time on drilling rig surface but without using the down-hole parameters because such a model is easy to implement. 
The use of fog computing can introduce Internet services such as cloud computing to advanced technology, providing control, computation, storage, communication, and service capacity. Prediction and optimization of ROP will be done in real-time data using machine learning algorithms in programming languages, then converted to a web page. So, we can monitor these results by using smart devices (watch, phone, and PC).