An adaptive moment estimation framework for well placement optimization

Abstract

In this study, we propose the use of a first-order gradient framework, the adaptive moment estimation (Adam), in conjunction with a stochastic gradient approximation, to well location and trajectory optimization problems. The Adam framework allows the incorporation of additional information from previous gradients to calculate variable-specific progression steps. As a result, this assists the search progression to be adjusted further for each variable and allows a convergence speed-up in problems where the gradients need to be approximated. We argue that under computational budget constraints, local optimization algorithms provide suitable solutions from a heuristic initial guess. Nonlinear constraints are taken into account to ensure the proposed solutions are not in violation of practical field considerations. The performance of the proposed algorithm is compared against steepest descent and generalized pattern search, using two case studies — the placement of four vertical wells and placement of 20 nonconventional (deviated, horizontal and/or slanted) wells. The results indicate that the proposed algorithm consistently outperforms the tested methods in terms computational efficiency and final optimum value. Additional discussions regarding nonconventional parameterization provide insights into simultaneous perturbation gradient approximations.

Performance Evaluation on Resolution Time Prediction Using Machine Learning Techniques

“…Hijry and Olawoyin [27] proposed a solution that studies deep learning techniques for historical queueing variables that will be used in addition to, or instead of, queueing theory to anticipate patient waiting times in a system (QT). They employed four optimization strategies, including Stochastic Gradient Descent (SGD) [28], adaptive moment estimation (Adam) [29], [30] Root Mean Square Propagation (RMSprop), and Adaptive Gradient (AdaGrad) [31], The model is evaluated using the mean absolute error (MAE) method. A conventional mathematical simulation is adopted for further analysis.…”

Accelerated Ensemble Optimization using Momentum Methods

“…Applying some form of momentum to reservoir management optimization is not entirely untested. Arouri and Sayyafzadeh 14 applied the Adam method together with the SPSA (Simultaneous Perturbation Stochastic Approximation) gradient to form Adam-SPSA, which was then applied to the problem of placing production wells, and later to well control optimization 15 . Both of the references report significant improvements in terms of stability and final objective function value.…”

Deep Learning Pada Detektor Jerawat: Model YOLOv5

“…Adam [33], [34] merupakan algoritma gabungan dari algoritma RMSProp dan Momentum dengan menyimpan learning rate RMSProp dan weight dari rerata momentum [35], [36].…”