Application of Autoregressive Integrated Moving Average for Modeling and Forecasting of Crude Oil Production

Abstract

Most of the crude oil production forecasting and modeling studies have focused on the traditional model of decline curve analysis as techniques used for generating forecast values of future oil and gas in place in conventional and unconventional wells. To many production reservoir economists, the importance of forecast estimations is critical because it allows forecast users to have a profound interest in monitoring and improving forecast performance. It also provides clear indications for the directions, strategies, and bottom line of both the National Oil Companies (NOCs) and International Oil Corporations (IOCs). While the traditional DCS model techniques have well-grounded mathematical underlining. This statistical design does necessarily assure that predicated function regardless of the values of the predictive variables. Moreover, the power traditional oil reservoir production forecasting technique is inefficient. This current research attempts to provide appropriate modeling and forecasting techniques for reservoirs utilizing a time series approach. It reveals how the historical oil production data can be used to project future oil reservoir production and how these projections influence future oil reservoir production decisions. Hence, the main objective of this research study is to practically explore the possibility of the autoregressive integrated-moving average model as a feasible function preference for predicting crude oil production. The historical oil reservoir production time series were used to establish respective autoregressive integrated moving average models through the time series technique by Box–Jenkins and the suitable models were designated with four performance criteria: maximum likelihood, standard error, Schwarz Bayesian criterion, and Akaike criterion for seven elected regions oil reservoir production and the established models conformed to the ARIMA (p, d, q). Once the process is identified, the adequacy of the forecasts will be determined and compared with the traditional decline curve analysis. Thus, as an accurate and effective oil production prediction for stretching a reservoir life cycle and enhancing reservoir productivity and recovery factors. These results provide production economists and reservoir engineers with quick, reliable, consistent, and real-time guidelines in budgeting, planning, and making decisions regarding field development. Finally, ARIMA forecasting models are more precise and deliver operational efficiency for dynamic forecasting of oil reservoir production.

Keywords: Oil and gas production forecasting, decline curve analysis, time series, ARIMA modeling,

DOI: 10.7176/JETP/13-1-05

Publication date: February 28^th 2023