Abstract
The aim of the study was to develop an analytical approach to minimising lost production in sand-prone oil wells based on operational data from the Caspian region, including facilities in Azerbaijan. The research methodology was based on an empirical analysis of monthly production observations from 48 producing wells for the period 2024-2025 and included a comparison of stable and unstable production intervals, a quantitative assessment of the contribution of operational factors to the formation of lost production, predictive modelling of unstable states and production losses, as well as an engineering verification of the identified patterns under real operational constraints. It was established that the transition to an unstable operational state was accompanied by a 23.5% decrease in oil flow rate, a 47.9% increase in water cut, a 69.8% increase in pressure drop across the filter elements, a 3.8-fold increase in downtime duration, and a 154.5% increase in the frequency of operational interventions. A statistical comparison showed that the most pronounced differences between stable and unstable intervals were associated with the rate of decline in flow rate, an increase in water cut, the cumulative duration of downtime over a six-month window, and a reduction in the proportion of stable production without signs of sand production. Regression analysis revealed that the greatest contribution to the change in lost production was made by the pressure drop across the filter elements, the proportion of stable operation, cumulative downtime and the frequency of interventions, whilst in the predictive modelling block, the XGBoost model demonstrated the best results, with an average absolute error of 2.18 m³ per month and a coefficient of determination of 0.79. The practical significance of the results lies in their potential application in production monitoring and operational management systems for sand-prone wells, enabling the early detection of unstable conditions, the justification of operational interventions, and the adjustment of production rates to minimise lost production