Tezislər / Theses

 
 
APPLICATION OF MACHINE LEARNING TECHNIQUES FOR 
RATE OF PENETRATION PREDICTION
Asad Safarov 
Azerbaijan State Oil and Industry University 
Baku, Azerbaijan 
[e-mail] 
Supervisor: David Solomonov 
Keywords:
machine learning, rate of penetration, drilling
Over the last decade, many techniques have been introduced to the 
drilling industry to estimate a rate of penetration. Overall, these methods are 
divided into two groups: deterministic and data-driven. In the deterministic 
models, the correlation equations are derived from a given data set whereas 
the data-driven models are generated from machine learning algorithms. 

---

THE 3
rd
 INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS 
dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
70
Machine learning is one of the newest approaches for accurate estimations 
and it gives a minimum value of human-related error [Asad, 2021, 1].
Rate of penetrations indicates a drilling rate which is a key factor for 
better drilling operations. A greater value of penetration rate leads to faster 
drilling which in turn contributes to the overall drilling performance. For years, 
many deterministic correlations have been used in the industry for 
penetration rate predictions. These models are based on laboratory 
experiments or poor correlation between real drilling data. The technological 
advancements in computational science and machine learning allow us to 
model new data-driven prediction models. These models use different 
algorithms to estimate a rate of penetration [Chiranth, 2018, 2]. 
This paper is written with the purpose of the development of a machine 
learning model generated from a drilling data set. For this, data is classified 
into two categories: train and test data. The train data is used to generate a 
model, while the test data helps to increase the accuracy of the model. 
Several algorithms will be used for the machine learning architectures and 
their results will be visualised in order to realize the performance accuracy of 
each approach. 
References 
[1] Asad M. E., "Application of Machine Learning Techniques for Real Time Rate of 
Penetration Optimization," Abu Dhabi, 2021 
[2] Chiranth H., "Rate of Penetration (ROP) Modelling Using Hybrid Models: Deterministic 
and Machine Learning", Houston, 2018.