Optimize drilling parameters for improved performance and reduced operational costs.
Drilling comprises a significant portion of oil and gas budgets, and any downtime experienced during drilling operations can be very expensive,
leading to project cost escalation. Thus, any measure which saves time can reduce operational costs.
Since the Rate of Penetration (ROP) is a direct measure of the drilling time,
machine learning algorithms and data analytics were employed to predict and maximize it.
A random forest was built on the training data and ROP was predicted throughout the depth
of the well using RPM of the bit, WOB, UCS of rock, and flowrate as input features.
An optimization algorithm was applied to change the WOB and RPM of the test data to
find the maximum attainable ROP.
Category :Oil Well Drilling
Client : Oil & Gas
Location : India
Completed Date : 2021
15% of drilling time was saved by maximizing the ROP based on
the predictions from the ML algorithm.
The daily operating cost was significantly reduced for drilling each oil well.
The performance of the ML model was further improved by feature engineering.