SPE-KSA T&PP’S Distinguished Lecturer Program Hosts Dr. Pallav Sarma
The SPE-KSA Technical & Professional Programs successfully hosted its second event of the Distinguished Lecturer Program (DLP) series titled “Physics Embedded Machine Learning for Modeling and Optimization of Mature Fields.” Held on May 19th, 2024, the insightful lecture was delivered by Dr. Pallav Sarma, Co-founder and Chief Scientist at Tachyus.
The event highlighted a novel approach known as Data Physics, which combines the strengths of physics-based modeling and machine learning to expedite predictive modeling processes. Dr. Sarma introduced Data Physics as an innovative method that enhances the accuracy and speed of modeling and optimization tasks. This approach is particularly beneficial for mature fields, where extensive historical data is available, allowing for more precise and quicker predictions. During his presentation, Dr. Sarma showcased a range of machine learning models, offering a comprehensive view of how artificial intelligence (AI) and machine learning (ML) can revolutionize reservoir simulation. He emphasized the future potential of AI and ML applications in the energy sector, illustrating how these technologies can optimize production and improve decision-making processes.
Among the modeling techniques discussed were Fast Forward Modeling and Closed Loop Optimization Modeling. Both of which proved to be successful when employed in predicting the optimal injection rates for reservoirs. Notably, Dr. Sarma demonstrated an application of the Data Physics approach to a complex waterflood, where a 15% increase in cumulative production was achieved since implementation. This was followed by addressing the limitations of the Data Physics approach. While it is highly effective under the right conditions, its success depends on the availability of high-quality historical data. Despite these limitations, he argued that Data Physics and conventional methods can complement each other effectively, providing a more robust framework for advancing reservoir modeling and optimization.
The event underscored the importance of integrating reservoir physics, modeling, and simulation with cutting-edge AI and ML techniques. By combining these disciplines, the industry can achieve significant advancements in the management and optimization of mature fields. Dr. Sarma’s lecture was a testament to the progress being made in this field, demonstrating the potential of Data Physics to transform traditional modeling practices.
The DLP session saw excellent attendance, with participants actively engaging in discussions, asking insightful questions, and contributing enriching thoughts andadditions. This made the event a great success, highlighting the community’s interest and commitment to leveraging advanced technologies in reservoir management. Overall, the DLP event provided valuable insights into the future of reservoir simulation, showcasing how the integration of physics and machine learning can drive innovation and efficiency in the energy sector