An accurate bottom hole pressure estimation is critical for the correct evaluation of any unconventional shale reservoir as it generates the necessary input for production diagnostic plots (such as the Linear Flow plot and the Log-Log plot), Rate Transient Analysis and Numerical Simulation models. Down hole permanent gauges may not be available in all wells (especially when the exploratory phase of the field has been left behind or when overall well costs want to be reduced) and so, the reservoir engineer must apply different methods to estimate the bottom hole pressure using production data and well head information. Dynamic gradients should be run periodically in the wellbore to build vertical lift performance curves and generate pressure loss estimations along the tubing. For wells with artificial lift systems (such as gas lift or rod pumping systems), this calculation may not be so straightforward and additional information about the pumping system may be needed to estimate the BHP.
The goal of this work is to build a Machine Learning data-driven model that can predict the BHP for multi-fractured horizontal wells of the Vaca Muerta Formation in Argentina. Input variables for the model training phase process include well and field location, landing zone, production data, artificial lift system parameters, completion design and reservoir and fluid properties.
The bottom hole flowing pressure (BHP) for any oil well can be obtained either from a direct measurement, or from an indirect calculation from production rates and well head pressure (WHP). For the former case, a down hole permanent gauge must be attached to the tubing to measure the bottom hole pressure versus time in a continuous trend for the given depth where the pressure sensor has been set. For the latter, a pressure drop equation has to be solved in order to relate well head and bottom hole pressures using surface production rates, fluid properties and tubular geometry.