A risk analysis model for underground gas storage well integrity failure

Abstract

Loss of underground gas storage (UGS) well integrity can result in serious consequences. Risk analysis is critical to maintaining the integrity of a UGS well and reducing potential accidents. This study proposes a risk analysis model for UGS well integrity failure. First, we employ the bow-tie (BT) approach to establish a systematic well integrity failure model, qualitatively depicting the causal relationship between well integrity failure and potential consequences. Subsequently, hierarchical Bayes analysis (HBA) is used to model source-to-source uncertainty among the aggregated reliability data and provide reasonable estimates for the reliability parameters of components (i.e., well components and safety barriers). This process allows us to consider UGS well attributes using collected data. Furthermore, to address the issues of the BT approach in risk quantification, the developed BT model is mapped to a corresponding Bayesian network (BN). Predictive analysis is performed to estimate the occurrence probabilities of well integrity failure and potential consequences. Probability updating is performed to identify the key contributing factors given the observation of an undesired event. Finally, to obtain a more case-specific result, risk updating is conducted by using probability adapting of the BN and the new information accumulated over time. The results show that the proposed model can effectively handle uncertainty in risk calculation and provide more accurate results. This study has practical value for improving UGS safety.