Pune: At a time when global energy markets
are witnessing volatility due to geopolitical tensions and disruptions in oil
supply chains, researchers at MIT World Peace University (MIT-WPU), Pune,
have developed advanced artificial intelligence (AI) and machine learning
(ML) models that can help improve oil recovery from mature reservoirs and
accurately forecast future production. This research could play an important
role in strengthening India’s energy security by enabling more efficient
extraction of oil from existing fields and reducing dependence on crude oil
imports.
India’s rapidly expanding economy
continues to drive rising energy demand. Oil and gas account for nearly 32–37%
of the country’s total energy consumption, and India spent an estimated USD
161 billion on crude oil imports, according to government data. Increasing
domestic production from existing oil fields has therefore become a strategic
priority, particularly at a time when global oil markets are influenced by
geopolitical tensions and supply disruptions.
Researchers from the Department of
Petroleum Engineering at MIT-WPU, the only dedicated upstream oil and gas
academic department in Maharashtra, are applying artificial intelligence to
address complex challenges in petroleum reservoir management.
A research team led by Dr. Rajib
Kumar Sinharay, Professor in the Department of Petroleum Engineering, MIT
WPU, along with his PhD student Dr. Hrishikesh K. Chavan, has developed
a machine learning model capable of identifying the most suitable Enhanced
Oil Recovery (EOR) techniques for complex reservoirs. The model was trained
using data from numerous oil-producing fields worldwide and achieved an
accuracy of 91% in predicting the most effective recovery methods.
Their findings were published in the international
journal Petroleum Science and Technology. The AI-based model
significantly reduces the time required to evaluate oil recovery
strategies—from several months using conventional methods to just a few
hours.
Dr. Rajib Kumar Sinharay said, “Artificial
intelligence has the potential to transform reservoir management in the oil and
gas industry. Our research focuses on developing data-driven tools that can
help operators select the most effective recovery techniques and make more
accurate production forecasts, particularly for mature oil fields.”
In another breakthrough, Prof.
Samarth Patwardhan and his PhD student Dr. Soumitra Nande developed
a deep learning model capable of identifying carbonate reservoir rocks with
97% accuracy. These rock formations are similar to those found in Bombay
High, India’s largest offshore oil field. Their research was published in
the Arabian Journal for Science and Engineering in 2025.
The MIT-WPU research team has also
developed a machine learning model for forecasting oil production in mature
oil fields, achieving 92% accuracy (R² score) when tested using real
field data from an Indian onshore reservoir. Reliable production forecasting is
critical for the petroleum industry, as it influences investment decisions,
reservoir management strategies, and long-term supply planning. The
research was published in the internationally reputed journal Physics of
Fluids.
In addition, the team has developed an
AI-based model for optimizing oil production tubing design, which helps
determine the appropriate pipe size for efficient oil extraction. This research
was presented at the International Conference on Computational Science and
Applications and later published in Springer Nature’s Algorithms for
Intelligent Systems series. The researchers have also secured a patent
for this technology.
Currently, the MIT-WPU team is working
on identifying “sweet spots” in unconventional hydrocarbon reservoirs
and developing sustainable drilling fluids suitable for high-temperature and
high-pressure environments.
These innovations highlight the
growing role of AI-driven research in improving oil recovery, increasing
efficiency in mature fields, and supporting India’s efforts to strengthen
domestic energy production in an increasingly uncertain global energy
landscape.