#TITLE_ALTERNATIVE#

• Main Author: RAMADHAN (NIM : 12208085), REZA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/30373
• Institution: Institut Teknologi Bandung

Accurate prediction of pressure drop in pipeline under multiphase flow condition is needed to obtain proper design of surface facilities such as pumps, compressors, and separators. This study presents Artificial Neural Network (ANN) method to predict the pressure drop in pipeline under multiphase flow conditions. Artificial Neural Network is one form of artificial intelligence that mimics the thought processes of the human brain. Artificial Neural Network does not need formula and rules, it learns from experience and have the ability to recognize pattern from data.A total of 500 hypothetical data sets were used in this study. . The hypothetical data was simulated using Schlumberger-Pipesim 2008. Data were divided into three portions: training, validation, and testing sets. The testing data sets, which were not used by the ANN model during the training phase, was used to test the prediction accuracy of the model. The result of this final project show that the accuracy of the Artificial Neural Network prediction depends on ratio of data sets (training, validation, and testing) and exact number of hidden neurons.