An Innovative System Architecture for Real-Time Monitoring and Alarming for Cutting Transport in Oil Well Drilling

Abstract

A new system architecture is developed to provide decision aids on the prediction and prevention of downhole problems related to inadequate hole-cleaning and wellbore stability. The developed adoptive algorithm includes model calibration, real-time monitoring and alarm generation module when an anomaly is detected. An innovative approach is proposed to develop an unsteady state one-dimensional wellbore model, and model is capable do the real-time calculation of equivalent circulation density (ECD), and standpipe pressure drop (SPP). The one-dimensional wellbore model is developed by integrating different sections of the mudflow. The unsteady one-dimensional wellbore model is integrated with Hershel-Bulkley model to predict both equivalent circulation density, and standpipe pressure drop (SPP), wherein the model parameter of the empirical equations are tuned to adapt to different types of rigs, mud systems, formations, and drilling scenarios. The mathematical model is first tuned with available historical data of the same well. Henceforth, the tuned model is used for monitoring the SPP and ECD profile across different sections of the wellbore. The developed model is successfully tested in the oil field for real-time monitoring of ECD and SPP. The tuned model found to be capable of predicting the SPP below 5% error. The monitoring procedure of drilling activity was improved with a calibrated mathematical model. The system was able to detect the downhole problems related to hole-cleaning and hydraulic management, namely, excessive ECD, cutting accumulation in wellbore annulus and the possibility of stuck and kick in real-time. The false alarm generation due to sensor fault is found to be one of the challenging issues to resolve. Further, we observed that the data reconciliation and preprocessing of real-time sensor data could reduce false alarm for downhole complications. Further model accuracy can be improved by improving the accuracy of the sensors used for mud density, mud loss, and cutting size. Unlike previous research works, in this work the annulus section of wellbore is divided into many small Continuous Stirred Tanks (CST) (i.e. Dynamic lumped parameter model) and they are connected in series to improve the accuracy of cutting transport model (i.e. to consider spatial variation of cutting concentration along the depth of the wellbore). Further simplified one-dimensional unsteady state wellbore model can be used for real-time calculation