Cost Effective Risk Based Corrective Action Design for Contaminated Groundwater

本研究的目的是发展一个耦合的优化和模拟风险管理模型，以调查在不确定性条件下人类健康风险和纠正措施设计之间的关系。 该方法将确定最佳的基于风险的纠正行动（RBCA）设计方案，从自然衰减到工程羽流控制和补救，以及这些方法的组合。将探讨提高模型计算效率的理论进步，以便可以考虑大规模的现场。 噪声遗传算法将与称为Modflow和RT3D的综合流量和命运以及运输模拟模型相结合。 噪声遗传算法类似于遗传算法，但每个候选解的适应度（价值）通过从概率分布中采样来评估。 对于这种应用，概率分布表示空间可变的水力传导率测量值的不确定性和用于估计人类健康风险的参数的可变性。 提高模型计算效率的方法包括开发高效和有效的采样策略，测试采样容差对数值模型中网格近似的影响，以及研究注入岛遗传算法的性能，该算法使用多个分辨率的多个种群的解决方案来有效地识别和微调强解。 预计拟议的工作将为监管机构和受监管社区提供一个宝贵的风险管理工具，确定既具有成本效益又能充分保护人类健康的风险减少战略。

The objective of this research is to develop a coupled optimization and simulation risk management model for investigating the relationships between human health risk and corrective action design under conditions of uncertainty. The methodology will identify optimal risk-based corrective action (RBCA) design options ranging from natural attenuation to engineered plume control and remediation, as well as combinations of these approaches. Theoretical advancements for improving computational efficiency of the model will be explored so that large-scale field sites can be considered. A noisy genetic algorithm will be coupled with comprehensive flow and fate and transport simulation models called Modflow and RT3D. Noisy genetic algorithms are similar to genetic algorithms, but the fitness (worth) of each candidate solution is evaluated through sampling from probability distributions. For this application, probability distributions represent uncertainty in spatially-variable hydraulic conductivity measurements and variability in parameters used to estimate human health risks. Methods for improving computational efficiency of the model to be investigated include developing efficient and effective sampling strategies, testing the implications of sampling tolerance on grid approximations in the numerical models and investigating the performance of an injection island genetic algorithm that uses multiple populations of solutions at multiple resolutions to efficiently identify and fine-tune strong solutions. The proposed work is expected to result in a valuable risk management tool for both regulators and the regulated community, identifying risk reduction strategies that are both cost-effective and sufficiently protective of human health.