Prediction of Brine Hydrate Formation Temperature Using ANN-Based Model

DOI: 10.14800/IOGR.1267

Abstract

This study aims to use Artificial Neural Networks (ANN) to predict Hydrate Formation Temperature of the brine which is used of completion, workover, and well intervention operations. It seeks to build an ANN model with inputs weight percentages of the brine salts components, gas specific gravity, pressure, and gas impurities percentages (N₂, H₂s, Co₂) to produce the Hydrate Formation Temperature as an output with high accuracy for both divalent and monovalent brines by using real dataset of about 200 for monovalent brines and about 300 for divalent brines with good range, optimization processes are performed to determine the best configuration to be used in ANN, so it involves training function optimization and number of hidden neurons to be used in ANN models, so The novel ANN models in this study will give a correlation that can be used to estimate the HFT directly.