This paper focuses on the development of numerical models and simulation software to predict the integrity of wells, from their construction to abandonment. The manuscript presents different numerical modelling equations and techniques to capture the construction of a well, the curing process of well cements, the accumulation of post-curing initial state of stress, debonding between the different components of the well and their fracturing. Key results are presented for standard well conditions and materials, they highlights the complexity of well integrity simulations; particularly regarding the curing of the cement. Cement curing is captured by leveraging existing chemical and volumetric models and by developing a new stress model. The limitations of such models are discussed and it is argued that they should not be used without extensive validation and calibration work and that the large, impactful uncertainties of well conditions should always be considered when drawing conclusions from their outputs. 1 INTRODUCTION Ensuring the integrity of active and abandoned wells is paramount to minimize the risk of leaks to the atmosphere and prevent shallow aquifer contamination. Structurally sound wells are particularly important for subsurface applications which involve greenhouse gases and/or valuable resources, such as oil & gas, gas fuel storage (CH4 and H2) and CO2 storage which is the focus of this research paper. The well's permanent subsurface structure is composed of imbricated cylindrical sections made of one or more cement annulus and steel casings; surrounded by the rock formation. The integrity of the well can be lost upon fracturing of the casing, cement or formation, the debonding between these different components and microscopic changes inside materials resulting from THMC (thermo-hydro-mechanical and chemical) processes. To make an assessment on the integrity of wells, there are three categories of methods. First field measurements, by collecting data downhole they provide the most direct assessment of integrity. Their drawback is cost, availability at scale and technical limitations on the data that can be acquired. For instance, well logging tools can difficulty acquire data relating to the cement sheath and formation when measurements need to be made through the thickness of one or more casings. Second, experimental methods which provide extremely valuable insights and data on the mechanisms of well integrity loss. Experiments reproduce the well behaviour in a simplified environment where the quality and the possibilities for different types of measurements is much greater than in the field. Their drawback is the complexity of the experimental endeavour, the simplifications made with respect to the field conditions and the limitations on the number of experimental runs that are practically possible to perform. The third type of assessment is numerical and mathematical methods which enable predictions of the THMC behaviour of the well. Provided that they are correctly validated and/or calibrated with field and experimental data, they make fast, cheap and reliable predictions. The reliability of such predictions is however tightly linked to the uncertainty of the model input parameters. When the input is uncertain, numerical methods can still provide a range of possible behaviours based on a reasonable uncertainty range on the input data.
No takes yet. Share an insight, caveat, or question.
Joulin et al. (2024) studied this question.