Evaluasi Forecasting Produksi Minyak Menggunakan Algoritma Machine Learning TCN: Studi Kasus pada Lapangan Produksi Volve

Abstract

Temporal Convolutional Networks have been demonstrated to be effective for processing time series data, particularly in the context of oil and gas production forecasting, due to their ability to capture long-range dependencies and temporal patterns while maintaining a stable network architecture. This study utilized the Volve field as a case study, which was in operation for eight years from 2008 to 2016, to evaluate the potential for continued production before reaching an assumed production limit of 100 barrels of oil per day. Decline Curve Analysis was employed to forecast future production trends and to assess the remaining production life of the field. Through careful tuning of hyperparameters, the Temporal Convolutional Network was able to generate a reliable forecast model. The results of this study indicated that, assuming the Volve field continues to operate until the production limit is reached and without factoring in operational costs, the field could potentially sustain production for an extended period. This research contributes to the understanding of the application of Temporal Convolutional Networks in forecasting production life in mature oil fields and demonstrates the viability of using advanced machine learning techniques in the oil and gas industry for predictive analysis.