Autonomous Viscosity/Density Sensing System for Drilling Edge-Computing System

Abstract Current mud monitoring practices are limited due to their reliance on manual measurements such as funnel viscometers, weight balances, or basic field rheometers. These manual practices impose restraints on the quantity and quality of the available data that are essential to ensure optimal and safe drilling operations. In this study, we introduce a new autonomous mud viscosity/density system based on an electromechanical tuning fork resonator. The system was integrated into an edge-computing system for improved data collection and deployment of machine learning models. The system was tested during a live drilling campaign. The viscosity/density sensor is based on an electromechanical tuning fork resonator. The sensor was integrated into a submergible housing for in-tank measurements. Two systems were developed for simultaneous measurements at inflow (possum belly) and outflow (suction pit). The data from the two systems were broadcast wirelessly to the central computer room at the rig for real-time display and data aggregation by the edge-computing system for the development of time-series analysis models using machine learning. During initial field testing, data from a single sensor were collected for various hours at a rate less than a sample per second. The test allowed for continuous monitoring of the mud consistency not accessible by current measurement practices. The data demonstrated the potential to perform real-time calculation and display of drilling parameters and to detect anomalies in the fluid that might be indicative of developing operational problems, which would enable the instrument to be used as an early-warning system and real-time calculation of drilling parameters. The system detailed here provides an essential building block to enable drilling automation. The robustness and compactness of the instrument allow it to be installed at various points in the mud circulation system for the generation of large data sets that can be processed using modern analytics algorithms in an edge-computing framework.