Petroleum production flow prediction

This repository contains the necessary scripts for oil production flow prediction models, these scripts are submittable in a spark (with Hadoop based infrastructure) cluster in order to make use of scalable learning. The submission of these scripts and model creation is triggered by the web component of a full information system.

Physical model

This classic model calculates oil flow using Gilbert's equation.

Artificial intelligence based models

It has become imperative to make use of AI models due to their success in the applied sciences and the industry. We had to make tests on our computing services using multiple model types. The scripts are categorised into:

ANN model

Static model

Dynamic model

LSTM model

Datasets

We worked on the local oil production flow. the features were changing at each learning job submission, so we didn't see the need to include it.