Dynamic Stress Analysis of Critical and Cyclic Loads for Production Casing in Horizontal Shale Wells

Abstract

Recent development of a new dynamic model for tubular stress analysis is now extended to the design challenges and failure modes characteristic of long production casing strings in extended horizontal shale wells. In particular, the issue of cyclic loading due to repeated sequences of multi-stage fracturing has not been addressed until now. The new model provides the ideal means of analysis of cyclic thermal loads as well as critical impact of compression due to initial running friction. The new dynamic model of tubular stress solves the one-dimensional momentum equation over a time step sequence initiated from the original running of the string into the wellbore. Friction is modeled in a fully history dependent manner, with damping provided naturally by the wellbore fluid viscosity. Local pipe velocity as well as magnitude and orientation of sliding friction is solved at each node with friction aggregated at the connection upset and joint mid-point. Unconventional shale wells pose critical design challenges especially in regard to the long production casing strings run in extended horizontal or lateral sections. Compressive frictional loads accumulated during running are trapped in the string by cement, packers and the wellhead. Thus the initial load state must fully account for the initial frictional state in order to be realistic and conservative. Hydraulic fracturing at high flow rates and significant pump pressures, including the possibility of screen-out, represents a critical design load on the casing which can also significantly alter the orientation and magnitude of tubular/wellbore frictional contact. The particular phenomenon of repeated fracturing treatements in a multi-stage stimulation compounds the design challenge. Cycles of cold stimulation followed by renewed hot production can lead to unexpected migration of axial loads and localization of critical stresses. The cyclic nature of loading due to repeated sequences of multi-stage re-fractures and renewed production has not received industry attention due to the unavailability of appropriate models. Lack of adequate models has perhaps resulted in the problem being overlooked. A dynamic model is ideally suited to the analysis of cyclic loads because of its inherent ability to account for a full history of friction loads. The dynamics of loading and unloading are also critical to this new ability to address the design problem. Previous static-based stress models have been unable to provide a comprehensive basis of design.