Quantifying the Effect of Stress Hysteresis on the Drilling Window: How Mud Weight Variations Can Affect Wellbore Strength

Abstract

The estimation of the drilling window limits ensures that lost circulation and wellbore instability events are minimized. These limits are conventionally defined during the pre-drilling phase based on offset wells data. As drilling commences, mud weights are selected to fit within these limits and they can be adjusted to react to different drilling scenarios as long as they don't violate the defined limits. This process fails to consider the effect of the initial mud weight and its subsequent adjustments on the strength of the wellbore. The concept of stress hysteresis dictates that when a body is subjected to a certain load, such as the one exerted by the hydrostatic pressure of the mud, its state will be altered in a manner that can shift its strength limits. This work presents a model that quantifies the changes in the drilling window due to variations in mud weight. The objective is to ensure that any subsequent mud weight changes will fall within the updated drilling window limits. The analysis is carried out using a novel process of a 3D poro-elasto-plastic finite element model (FEM) that is integrated with a machine learning (ML) algorithm. The integrated FEM-ML model uses offset wells data along with the best fitting failure criterion to estimate the initial limits of the drilling window. The offset wells data used consist of wireline logs, drilling reports, and mechanical testing lab results belonging to the formation of interest. The integrated model uses this data to estimate the stress distribution and learn the failure patterns. The model is then used to run different scenarios of mud weight variations while drilling a specific hole section to quantify their effect on the drilling window. The end result of each scenario is an update of the drilling window, which reflects the effect of stress hysteresis. When examining the initial estimations of the drilling window against those reflecting the stress path effect, a significant discrepancy in the window size is quantified. This examination is carried out for an offset well, which experienced multiple mud weight changes as a response to various drilling events. Subsequently, the changes in the drilling window and the actual mud weights used are analyzed in view of the drilling difficulties experienced in that specific offset well for the purpose of providing a form of validation. The model results show that the drilling window had shrunk significantly enough for the mud weight to violate the wellbore stability limit. Failure to consider the stress hysteresis effect in this well led major wellbore instability, tight hole, and overpull. The modelling effort presented in this work allows for a new aspect of dynamic responses to drilling events as they occur.