Integrated Production Facilities Clustering and Time-Series Forecasting Derived from Large Dataset of Multiple Hydrocarbon Flow Measurement

Authors

  • Adityapati Rangga Politeknik Caltex Riau
  • Yohana Dewi Lulu Widyasari Politeknik Caltex Riau
  • Dadang Syarif Sihabudin Sahid Politeknik Caltex Riau

DOI:

https://doi.org/10.59190/stc.v2i2.207

Keywords:

Oilfields, Forecast, Clustering, VAR, LSTM, K-Means, Principal Component Analysis

Abstract

In the complex, mature, and large oilfields, there is a need for Integrated solution in order to have a helicopter view of entire facilities throughput. The real time metering information provides an on-demand daily data and trend; However, it is rarely being connected to analytics solution for business intelligence such as, prediction, optimization, decision support and forecast. This paper cover about exploratory data analysis of large dataset of multiple hydrocarbon facilities metering within integrated network, performing multi-feature data clustering and making a time-series forecasting techniques. K-Means and PCA are combined to make cluster of production facilities which resulted with gas processing cluster, high oil producer, high water processing station, and the lowest performer in term in hydrocarbon processing. Furthermore, VAR and LSTM is compared as forecasting tools for day-to-day fluid prediction, to maintain normal operational scenario.

Downloads

Published

2022-02-28