Improvement of Production Optimization Using Derivative-Free Methods

• Grant number: 25/10629-0
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: August 01, 2025
• End date: July 31, 2029
• Field of knowledge: Engineering - Mechanical Engineering
• Principal Investigator: Marcio Augusto Sampaio Pinto
• Grantee: Meysam Hasan Nezhad
• Host Institution: Escola Politécnica (EP). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Company: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Mecânica (FEM)
• Associated research grant: 17/15736-3 - Engineering Research Centre in Reservoir and Production Management, AP.PCPE

Abstract

The production optimization of an oil field can increase oil recovery, reducing water production, maximizing profits and improving the management of the field. One of the main challenges linked to the production optimization is finding an efficient optimization method. Even today, many applications are dependent on the estimations of the gradient (direct search methods), such as the Stochastic Simplex Approximate Gradient (StoSAG) (FONSECA et al., 2016), making the optimization computationally time-consuming. This has a significant impact on the oil industry, since the reservoir models can be compositional and have millions of cells, demanding even more simulation time. Moreover, there is a need to evaluate an ensemble of models instead of a deterministic one, due to geological and technical uncertainties. Heuristic and metaheuristic optimization has become the state-of-the-art in the optimization techniques, providing better performance solutions (KUMAR and YADAV, 2022). Thus, the main motivation is the search for more efficient optimization methods that demand a lower computational cost. In this way, the main objective of this research is to evaluate and develop derivative-free optimization methods to improve the production optimization process. For this purpose, optimization benchmarks will be used, allowing comparisons between the most diverse methods and proposing improvements. The algorithms will be tested under the same conditions and ranked under several criteria, including their ability to find near-global solutions. The objective function will be the net present value (NPV), adjusting the well rates of the reservoir model. Improvements may arise from mixed approaches, by proposing hybrid methods, combining benefits of different existing methods, or even creating new methods. (AU)