Using Bayes Theorem to Quantify and Reduce Uncertainties when Monitoring Varying Marine Environments for Indications of a Leak
Using Bayes Theorem to Quantify and Reduce Uncertainties when Monitoring Varying Marine Environments for Indications of a Leak
复制标题
在监测变化的海洋环境中是否存在泄漏迹象时，使用贝叶斯定理来量化和减少不确定性
DOI：
10.1016/j.egypro.2017.03.1492
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
Alendal G
中科院分区：
文献类型：
--
作者：
Alendal G
由于环境的多变性以及二氧化碳迁移可能渗入海底的区域范围，监测海洋环境中地质封存项目的泄漏是一项挑战。由于存在环境风险，不应允许相关泄漏继续存在而未被发现。由于海上作业成本高昂，因此还存在成本问题，因此应避免误报。那么主要的问题是：监测数据的偏差有多大应该引起确认和本地化程序的动员？这里建议将贝叶斯定理和贝叶斯决策理论作为量化确定性并在决策过程中实施误报（误报）和漏报（未检测到的泄漏）成本的工具。使用模拟的自然 CO2 含量变化和来自模拟泄漏的预测 CO2 信号来举例说明该过程。
Monitoring the marine environment for leaks from geological storage projects is a challenge due to the variability of the environment and the extent of the area that migrating CO2 might seep through the seafloor. Due to the environmental risk associated leaks should not be allowed to continue undetected. There is also a cost issue since marine operations are expensive, so false alarms should be avoided. The main question is then: how large a deviation in the monitoring data should cause mobilization of confirmation and localization procedures? Here Baye's theorem and Bayesian decision theory is suggested as a tool for quantifying certainties and to implement costs for false positives (false alarms) and false negatives (undetected leaks) in the decision procedure. The procedure is exemplified using modeled natural CO2 content variability and the predicted CO2 signal from a simulated leak.