Prediction of Threshold Sand Rates from Acoustic Monitors using Mechanistic Models and Artificial Intelligence

A framework is being developed and optimized to utilize mechanistic modeling and Artificial Intelligence-Machine Leaning (ML) Algorithms to correlate the TSR from Acoustic Monitors with pipe size, sand size, inclination, superficial liquid/gas velocities, mixture density/viscosity, flow regime, and erosion rate. The current AI model agreement with experimental data is encouraging and suggests that this methodology can be extended for a variety of flow conditions and pipe sizes not tested before. After the development of the AI models, these models are implemented in an Excel spreadsheet using a code that is written in a statistical software that runs in the backend. The spreadsheet is built in a way to add data being collected currently and in the future and can be trained as the data bank is updated.