Cybersees Type 2: Cyber-Enabled Water and Energy Systems Sustainability Utilizing Climate Information

不断增加的水需求(由于人口增长)和燃料成本威胁到水和能源系统的可靠性，也增加了运营成本。此外，气候多变性和长期变化增加了这两个系统的脆弱性。例如，水库系统主要依赖于月到季的降水量；而电力系统的需求取决于季节中日温度的变化。目前，这两个系统都考虑短期(0-3个月)管理的气候平均值，这忽略了气候的不确定性，导致水库水电减少(即泄漏增加)，以及过度储存燃料增加了电力系统的运营成本。尽管季节性气候预报在美国部分地区冬季和夏季都包含相当高的技能水平，但由于缺乏将概率预报吸收、可视化并传达到管理模型中的框架，这些预报在水电系统共同优化方面的应用一直受到限制。该项目将开发一个原型可视化分析模拟-优化框架，以便更好地交流和可视化决策空间上的概率信息，以利用每月到季度的多模式气候预报来改善供水和电力系统的可持续性。该项目的主要目标是开发一种综合的基于网络的方法，通过以下方式提高水和能源的可持续性：a.使用Optimus-Prime开发一个网络基础设施框架，通过纳入降水、电力需求和风力发电组合，共同优化水和电力系统的配置。B.通过利用视觉分析的最新发展，开发用于交互探索多目标决策空间和与优化引擎交互的视觉分析框架。C.开发细粒度和粗粒度的资源利用策略，以针对基于异质多核/加速器的架构上的计算的不同部分。D.分析风险管理战略，包括改进燃料储备、定期工厂维护和水库运行计划，以利用试点流域现有的多模式电力需求、风力发电和径流预测将运营成本降至最低。F.进行情景分析，考虑发电组合中可再生能源--水电和风能--的增加，在当前情况下减少二氧化碳排放，并根据多模式气候预测增加水需求潜力。F.综合目标(A-F)的结果，考虑在全国具有代表性的各种虚拟水和电力系统配置，以便推广和更广泛地应用。该提案的根本贡献在于开发了一个原型网络基础设施，用于利用气候预报改善水和电力系统的管理。利用该框架，研究报告建议，通过考虑利用多模式气候(降水、温度和风)预报最大限度地发挥可再生能源潜力的各种方案，最大限度地减少这两个相互依赖的系统的业务成本。通过利用径流和电力需求预测，水和电力系统管理的模式转变是有针对性的，以促进各种积极主动的战略，如燃料远期采购、减少排放和满足目标水库储存，以确保两个系统的可靠性。为了克服概率预测带来的计算挑战，支持HPC的网络基础设施将跨集合并行计算，并在水电随机优化模型之间提供无缝交互。为了互动探索水利和电力部门的多目标决策空间，将采用视觉分析的新发展，以改进试点流域利益攸关方推动的解决方案，并通过虚拟系统促进学习和理解。开发的研究模块和成果将被纳入NCSU水资源和环境工程、计算机辅助工程和电力系统工程课程的各种高级研究生课程。开发的基于气候信息的风险管理工具也将用于NCSU的气候变化和社会专业科学硕士项目，该项目专门吸引具有跨学科背景的专业人员。调查人员还将与北卡罗来纳州国家气候办公室和联邦机构PNNL和ORNL合作，为更广泛的应用实施该项目中开发的算法。研究人员还将与NCSU的工程部门合作，开发定制的播客和模块，讨论气候-水-能源的关系及其对高中生可持续发展的意义。

Continually increasing water demand (due to population growth) and fuel costs threaten the reliability of water and energy systems and also increase operational costs. In addition, climatic variability and long-term change increase the vulnerability of these two systems. For instance, reservoir systems primarily depend on monthly to seasonal precipitation; whereas power systems demand depend on variations in diurnal temperature over the season. Currently, both systems consider climatological averages for their short-term (0-3 months) management, which ignores uncertainty in the climate resulting in reduced hydropower (i.e., increased spill) from reservoirs and increased operational costs for power systems from excessive fuel stockpiling. While seasonal climate forecasts contain appreciable levels of skill over parts of the US in both winter and summer, the uptake of these forecasts for co-optimization of water and power systems has been limited due to lack of a framework to assimilate, visualize and communicate probabilistic forecasts into management models. The project will develop a prototype visual analytic simulation-optimization framework for better communication and visualization of probabilistic information on the decision space for improving water and power systems sustainability utilizing monthly to seasonal multimodel climate forecasts. The primary goal of this poject is to develop an integrated cyber-enabled approach for improved water and energy sustainability utilizing the monthly to seasonal climate forecasts by: A. Developing a cyberinfrastructure framework using Optimus-PRIME that co-optimizes water and power system allocations by incorporating precipitation, power demand and wind power ensembles. B. Developing a visual analytic framework for interactive exploration of multi-objective decision space and interaction with the optimization engine by leveraging recent developments in visual analytics. C. Developing fine-grained and coarse-grained resource utilization strategies for targeting different parts of the computation on heterogeneous many-core/accelerator based architectures. D. Analyzing risk management strategies that include improved fuel stockpiling, scheduled plant maintenance and reservoir operational schedules to minimize operational costs utilizing the multimodel electricity demand, wind power, and streamflow forecasts available for the pilot basin. F. Performing scenario analyses that consider increased renewable energy availability -- hydropower and wind -- in the power generation mix, reduced CO2 emission scenarios under current and increased water demand potentials contingent on multimodel climate forecasts. F. Synthesizing the findings from objectives (A-F) by considering various virtual water and power system configurations that are typical across the country for generalization and broader application. The fundamental contribution of the proposal lies in developing a prototype cyberinfrastructure for improving water and power systems management utilizing climate forecasts. Using the framework, the study proposes to minimize the operational costs of these two interdependent systems by considering various scenarios for maximizing the renewable energy potential utilizing multimodel climate (precipitation, temperature and wind) forecasts. By utilizing the streamflow and power demand forecasts, a paradigm shift in water and power systems management is targeted that promotes various proactive strategies such as forward purchasing of fuels, reducing emissions and meeting target reservoir storage to ensure both systems reliability. To overcome the computational challenges arising from probabilistic forecasts, an HPC-enabled cyberinfrastructure will parallelize the computations across ensembles and provide a seamless interaction between water and power stochastic optimization models. For interactive exploration of multi-objective decision spaces in the water and power sectors, new developments in visual analytics will be employed that can lead to improved stakeholder facilitated solutions for the pilot basin and promote learning and understanding through virtual systems.Developed research modules and findings will be incorporated into various advanced graduate courses at NCSU in the water resources and environmental engineering, computer-aided engineering and power system engineering curricula. The developed climate information based risk-management tools will also be used in the NCSU's Climate Change and Society Professional Science Master's program, which exclusively draws professionals with interdisciplinary background. The investigators will also work with the State Climate Office of NC and federal agencies, PNNL and ORNL, to implement the algorithms developed in the project for broader applications. The investigators will also develop customized podcasts and modules on climate-water-energy nexus and their significance to sustainability for high school students by collaborating with the engineering place at NCSU.