Extended Probabilistic Risk Assessment of Autonomous Underwater Vehicle Docking Scenarios Considering Battery Consumption

Abstract

Autonomous underwater vehicles (AUVs) play a crucial role in marine environments, such as in inspecting marine structures and monitoring the condition of subsea pipelines. After completing their mission, AUVs dock with recovery systems at designated locations. However, underwater docking carries a significant risk of failure due to unpredictable maritime conditions. Considering the limitations in communication during the mission, docking failure can lead to the loss of collected data and failure of the entire AUV mission. In this study, a hypothetical AUV docking scenario was defined based on expert knowledge and without actual operational data. A Markov chain-based probabilistic model was employed to quantitatively assess the risk of the system during the mission. Environmental factors were excluded from the evaluation, and the simulation results were classified into five categories: success, timeout, internal component failure, exceeding a predefined sequence repetition limit, and spending the electrical energy under the battery SOC threshold. By analyzing the failure points of each category, strategies to improve the scenario success rate were discussed. This study quantitatively identified the interactions between constraints and risk factors that should be considered when establishing AUV docking plans through a virtual scenario-based failure analysis, thereby providing an evaluation framework that can be utilized in actual design.