Research on calculation model of bottom of the well pressure based on machine learning

摘要

Based on the research on the calculation model of the bottom of the well pressure of the Managed Pressure Drilling, a set of accurate and effective monitoring methods for the bottom of the well pressure of the Managed Pressure Drilling is proposed.Through the analysis of the hydraulic model, the feasibility of the artificial intelligence algorithm in the bottom of the well pressure calculation and monitoring method is studied. Based on the analysis of the wellbore flow hydraulic model, the simulated annealing (SA) algorithm is combined with the support vector regression (SVR) machine to establish a set based on the simulated annealing algorithm to improving the supported vector regression machine (SA-SVR) is used to optimize the monitoring method of bottom of the well pressure of Managed Pressure Drilling (MPD).Combining the hydrostatic column pressure, Annulus pressure loss and surface back pressure data, Managed Pressure Drilling bottom of the well pressure monitoring model based machine learning is constructed to realize bottom of the well pressure data monitoring without PWD (Pressure While Drilling) instruments. The bottom of the well pressure monitoring model based on machine learning is used to calculate and analyze the bottom of the well pressure, provide theoretical support for Bottom of the well pressure monitoring of drilling operations, and guide safe drilling at the construction site.