Estimate oil production in Iraq using an artificial neural network

Estimating petroleum production is crucial in strategic planning
and decision-making within the petroleum industry. In this study,
we propose the use of artificial neural network (ANN) methods,
implemented through MATLAB software, to forecast petroleum
production in Iraq, a major oil-producing nation with vast reserves
and complex geological characteristics. By using 50 years of
historical production data (1973–2023) and a range of relevant
input factors, including (1) normal conditions, (2) wartime
conditions, (3) economic blockade conditions, and (4) epidemic
conditions, we used and train ANN models to estimate future
petroleum production levels in Iraq over the next five years (2024–
2028). The study compares the ANN estimation results with the
Iraqi Oil Ministry’s official production plan for the same period.
According to the Ministry’s plan, production is expected to reach
7 million barrels per day (Mb/d) by the end of 2028. The ANN
model forecasts indicate that, under normal conditions, the
Ministry can achieve its planned production target with a
maximum margin of error of 2.7%, corresponding to a shortfall of
approximately 0.189 Mb/d. However, in scenarios affected by
war, economic blockades, or epidemics, the ANN estimations
show that these factors could negatively impact production,
hindering the achievement of the planned target by up to 80% in
the worst-case scenario.