Search

Relative permeability is a substantial parameter for estimating multi-phase fluid flow in porous rocks. It is a complex physical property that is influenced by the behavior and interactions between the fluid and rock phases. Relative permeability measurement of rock samples in laboratory can be carried out using steady-state or non-steady-state techniques. Permeability measurement is relatively difficult and time consuming. Because of the difficulty in measurement, empirical models are often used to estimate relative permeability or extrapolate to limited laboratory data. Artificial neural network (ANN) is a method that can be applied to obtain complex correlations of parameters that influence each other. In this study, ANN is used to predict the relative permeability of oil and water. The proposed model evaluates the relative permeability of a phase as a function of rock absolute permeability, porosity, depth, permeability of other phases and water saturation. A total of 159 relative permeability data from Talang Akar Formation were used for the training and testing processes. Based on the comparison between measured and calculated data, the correlation coefficients for relative permeability to water and oil using ANN method are 0.77 and 0.94 respectively. While those using regression analysis are 0.88 and 0.73 respectively.

Designing drilling mud rheology is a complex task, particularly when it comes to preventing filter cakes from obstructing formation pores and making sure they can be easily decomposed using breakers. Incorporating both multiphysics and data-driven numerical simulations into the design of mud rheology experiments creates an additional challenge due to their symmetrical integration. In this computational intelligence study, we introduced numerical validation techniques using 498 available datasets from mud rheology and images from filter cakes. The goal was to symmetrically predict flow, maximize filtration volume, monitor void spaces, and evaluate formation damage occurrences. A neural-objective and image optimization approach to drilling mud rheology automation was employed using an artificial neural network feedforward (ANN-FF) function, a non-ANN-FF function, an image processing tool, and an objective optimization tool. These methods utilized the Google TensorFlow Sequential API-DNN architecture, MATLAB-nftool, the MATLAB-image processing tool, and a single-objective optimization algorithm. However, the analysis emanating from the ANN-FF and non-ANN-FF (with neurons of 10, 12, and 18) indicated that, unlike non-ANN-FF, ANN-FF obtained the highest correlation coefficient of 0.96–0.99. Also, the analysis of SBM and OBM image processing revealed a total void area of 1790 M µm2 and 1739 M µm2, respectively. Both SBM and OBM exhibited notable porosity and permeability that contributed to the enhancement of the flow index. Nonetheless, this study did reveal that the experimental-informed single objective analysis impeded the filtration volume; hence, it demonstrated potential formation damage. It is, therefore, consistent to note that automating flow predictions from mud rheology and filter cakes present an alternative intelligence method for non-programmers to optimize drilling productive time. [ABSTRACT FROM AUTHOR]

Locating petroleum-productive wells using informed geological data, a conventional means, has proven to be tedious and undesirable by reservoir engineers. The former numerical simulator required a lengthy trial-and-error process to manipulate the variables and uncertainties that lie on the reservoir to determine the best placement of the well. Hence, this paper examines the use of a global genetic algorithm (GA) to optimize the placement of wells in complex reservoirs, rather than relying on gradient-based (GB) methods. This is because GB approaches are influenced by the solution's surface gradient and may only reach local optima, as opposed to global optima. Complex reservoirs have rough surfaces with high uncertainties, which hinders the traditional gradient-based method from converging to global optima. The explicit focus of this study was to examine the impact of various initial well placement distributions, the number of random solution sizes and the crossover rate on cumulative oil production, the optimization of the synthetic reservoir model created by CMG Builder, CMOST, and IMEX indicated that using a greater number of random solutions led to an increase in cumulative oil production. Despite the successful optimization, more generations are required to reach the optimal solution, while the application of GA on our synthetic model has proven efficient for well placement; however, different optimization algorithms such as the improved particle swarm (PSO) and grey wolf optimization (GWO) algorithms could be used to redefine well-placement optimization in CMG. [ABSTRACT FROM AUTHOR]