AI Model Trained to Measure Subsurface Pressure During Drilling
The system is expected to help reduce resource use in oil and gas operations.
Researchers at Perm Polytechnic, working with colleagues in China, have developed a hybrid intelligent system designed to optimize drilling processes. The algorithm analyzes formation pressure and calculates the precise volume of drilling fluid required for injection into a well.
The neural network can simultaneously process nine key parameters. Based on this data, the system determines both minimum and maximum stress levels within the rock formation.
To train the model, researchers used data from 10,000 measurements collected across three wells in the Junggar Basin in northwestern China. The field is considered geologically complex, meaning the system is designed to handle a wide range of conditions.
Accurate data on horizontal rock stress is critical for hydraulic fracturing. The system’s output helps guide fractures toward resource-rich zones rather than into non-productive rock.
