Methods and models for analyzing and controlling the safety in operations of autonomous marine systems

Abstract

Autonomous marine systems (AMS), such as autonomous underwater vehicles (AUVs), and unmanned surface vehicles (USVs) have evolved over the past decades. Maritime autonomous surface ships (MASS) are gradually being developed and commissioned. AMS are applied in different types of industries and research. Examples include the application of USVs or AUVs for ocean monitoring, and the development of MASS for future cargo and personnel transportation. In these operations, AMS can help reduce the risk of personnel exposure to harsh environments, reduce the operational costs, and improve the efficiency and performance of the human operators. However, compared to conventional marine systems, new types of failure might be introduced to AMS operations due to unforeseen interdependencies in the system design, dynamic operating environments, maintenance challenges, insufficient situation awareness and decision-making from human operators, etc. Also, AMS functions are constantly being improved, and the operations of AMS are becoming more complex and advanced. The safety issues of these systems have become even more critical. Techniques for analyzing and controlling the safety of AMS operations are therefore required. The overall aim of this PhD project is to develop methods and models for analyzing and controlling safety in operations of AMS. It is refined into the following three research objectives that are addressed in five research articles: • Identify and analyze hazards and hazardous events in the operation of autonomous marine systems and evaluate the applicability of relevant methods as a basis for online risk modeling of autonomous marine systems. • Analyze the dynamic changes in the operating environment and system status, and model their impacts on the safe operation. • Propose a general method for developing online risk models for autonomous marine systems and operations, supporting risk-based control. The work presented here reviews the existing methods and models and identifies the main research challenges and gaps with respect to the above research objectives. The research presented in the thesis addressed some of these issues. The main contributions of this thesis are summarized as follows: • Investigation of the potential hazards/ hazardous events during the operation with multiple AMS and how these hazards/ hazardous events may affect the safe and reliable operations of AMS. The results highlight the importance of considering unsafe interactions in hazard identification or risk assessment in AMS operations. • Comprehensive hazard identification works with a number of potential hazards/ hazardous events that may affect the safe operation of an under-ice AUV operation through various methods. The results contribute research and practical implications for improved engineering design and operational procedures to enhance the safety and robustness of future AMS operations in the Arctic. • Identification of a list of evaluation criteria for online risk models for AMS and comprehensive evaluation of the applicability of several existing methods for online risk modeling of AMS. The evaluation results contribute to an appropriate first step towards a general framework for online risk modeling for AMS. • Proposal for a dynamic risk analysis method to determine the dynamic changes in the operating environment and assess the environmental impact on the safe operation of AMS. • Proposal for a novel dynamic maintenance planning method for AMS that addresses challenges in maintenance planning, including the high consequence of system shutdown, limited and irregular maintenance opportunities, and various dependencies among components. • Proposal for a general framework for the online risk modeling of AMS to enhance the intelligence of the AMS, its situation awareness, and decision-making. The proposed framework addresses several challenges in developing online risk modeling, e.g., evidence uncertainty. • Proposal for a two-level strategy to develop a supervisory risk control (SRC) system for AMS operations based on the developed online risk model. The SRC system can improve the intelligence of AMS by enabling its risk-based control. In conclusion, the research and findings presented in the thesis provide researchers and practitioners in the field with a comprehensive overview of safety issues in AMS operations, and novel methods and models for analyzing and handling these. The proposed methods and models are expected to improve the safety of future AMS operations.