Frontiers in Determination of Solution Gas-Oil Ratio using Artificial Neural Networks with Multi-Layers Perceptron

For solving complex problems, it’s needed to go beyond standard mathematical techniques. Instead, it’s necessary to complement the conventional analysis methods with a number emerging methodologies and soft computing techniques such as expert system, artificial neural network, fuzzy logic, genetic algorithm, probabilistic reasoning, and parallel processing techniques. Soft computing differs from conventional (hard computing) in that, unlike hard computing, it is tolerant of imprecision, uncertainty, and partial truth. Soft computing is also tractable, robust, efficient and inexpensive. This paper presents a technique to model the behavior of crude oil systems. The proposed technique is using Multi-Layers Perceptron neural network. The model predicts solution gas-oil ratio. Input data to the Multi-Layers Perceptron are reservoir pressure, temperature, stock tank oil gravity, and separator gas gravity. The proposed Multi-Layers Perceptron is tested using PVT properties of other samples that have not been used during the training process. Result show good accuracy between the Multi-Layers Perceptron predicted data and actual data.

Keywords:

Multi-Layers Perceptron , Neural Networks , Neural Network Regression Techniques , Solution Gas-Oil Ratio