Machine learning and petroleum engineering

This work takes place in the context of the Volve oil field in the North Sea (Norway) and addresses the crucial problem of predicting and optimizing the productivity of oil wells. As the oil industry is strategic it is essential to develop robust models to maximize production while minimizing costs and environmental impacts. The central objective here is to develop robust prediction models using advanced Machine Learning algorithms. More specifically it aims to demonstrate the power of coupling Deep Learning with fluid flow simulation. In addition the study focuses on the development of mathematical prediction models using the response surface method while seeking to optimize well productivity parameters.