Intelligent decision support for marine oil spill response in harsh environments

海洋溢油（如2010年英国石油公司溢油）不仅对环境和社会经济造成重大负面影响，而且对海洋生物和人类健康构成长期危害。对泄漏的响应是一个时间敏感的、动态的、多方面的和复杂的过程，取决于各种因素，例如油的数量/性质、位置、环境条件、响应资源和人的判断/行为。成功在很大程度上取决于信息和资源的利用效率和有效性，以及决策和行动的最佳和快速程度。它还面临着加拿大近海地区（如北极/北大西洋）普遍存在的恶劣海洋环境（低温、波涛汹涌的海面、海冰等）的进一步挑战。由于技术/后勤困难、安全和快速的石油运输，响应机会窗口非常短。最近在纽芬兰近海发生的三起漏油事件就证明了这一点。因此，动态、高效的响应决策支持是迫切需要的，但目前还缺乏研究和实践。此外，人为因素/错误占溢油事故的60-80%，但没有很好地低估其对溢油反应的影响。因此，该研究旨在通过开发适用于恶劣环境的智能海洋溢油响应决策支持系统（iMOS-DSS）来帮助解决这些挑战，重点是：1)全面审查和开发海洋溢油i-Base（溢油/溢油数据库和响应知识库）；2)智能体和人为因素分析辅助溢出模拟；3)深度学习辅助响应过程模拟-优化耦合，支持滗析实验（脱脂、分离、处理、处置）；4)人工智能辅助海上溢油应急决策支持；5)虚拟现实技术为海上溢油应急决策提供可视化支持。最常用的响应选项（仅在加拿大合法允许），滗析过程，将主要针对实验室实验，以支持过程模拟和优化工作。系统验证和评估将通过当地的溢油应急人员课堂和NL的海上培训演习进一步进行。拟议的项目将1)产生新的/先进的知识和方法，以解决提高加拿大及其他地区应对能力的迫切需求；2)全面丰富和完善环境应急管理和决策的理论和技术；3)通过跨学科、公平、多元和包容（EDI）的研究/培训环境培养下一代专业人员和研究人员；4)帮助我的团队和大学在全球海洋溢油研究中定位，扩大国际合作/影响，并在全球范围内应用成果；5)通过支持负责任的工业发展和保护我们的海洋环境，带来显著的短期/长期效益。

Marine oil spills (e.g., 2010 BP spill) can not only cause significantly negative impacts on the environment and socio-economy but constitute a long-term hazard to marine life and human health. The response to a spill is a time-sensitive, dynamic, multifaceted and complex process, depending on various factors e.g., oil quantity/properties, location, environmental conditions, response resources, and human judgement/behaviours. The success much relies on how efficiently and effectively the information and resources can be utilized and how optimally and quickly the decision and actions can be made. It is further challenged by harsh marine environments (low temperature, rough sea, sea ice, etc.) prevailing in Canadian offshore areas such as the Arctic/Northern Atlantic oceans. The response window of opportunity is very short due to technical/logistic difficulties, safety, and rapid oil movement. This has been evidenced by three recent spills in Newfoundland offshore. Thus, dynamic and efficient response decision support is much desired but still lacking of research and practice. Furthermore, human factors/errors contribute 60-80% in spill accidents but without good understating of their effects on spill response. Therefore, the research is to help address these challenges by developing an intelligent Marine Oil Spill response Decision Support System (iMOS-DSS) for harsh environments with focuses on: 1) comprehensive review and development of marine oil spill i-Base (oil/spill databases and a response knowledge base); 2) intelligent agent and human factor analysis aided spill simulation; 3) deep learning aided response process simulation-optimization coupling supported by decanting experiments (skimming, separation, treatment, and disposal); 4) artificial intelligence aided marine oil spill response decision support; and 5) virtual reality aid visualization for marine oil spill response decision making. The most commonly used response option (and only legally allowed in Canada), decanting process, will be mainly targeted for lab experiment to support process simulation and optimization work. System validation and evaluation will be further conducted via local spill responders' classroom and at-sea training exercises in NL. The proposed program will 1) produce new/advanced knowledge and methodologies to address urgent needs in improving response capabilities in Canada and beyond; 2) enrich and improve the theories and technologies of environmental emergency management and decision making in general; 3) train next-generation professionals and researchers through a cross-disciplinary, equitable, diverse and inclusive (EDI) research/training environment; 4) help position my group and university globally in marine oil spill research to expand international collaborations/impacts and apply the outcomes worldwide; and 5) bring significant short-/long-term benefits by supporting responsible development of industry and protecting our marine environments.